A loaded reefer trailer at a cold dock at dusk, where every door-open event is a temperature excursion clock that real-time KPIs either catch before threshold or miss until the product record tells the story.
Temperature is either controlled or it isn’t. In cold chain and refrigerated logistics, there is no graceful degradation: a reefer unit that drifts two degrees above setpoint for four hours doesn’t slow things down, it destroys product, triggers regulatory reporting, and can pull a customer contract the same week.
Most cold chain failures don’t announce themselves. They accumulate quietly across a fleet of trailers, a warehouse with 40 temperature zones, and a network of third-party cross-docks where your visibility stops at the gate. By the time a driver notices a problem or a warehouse scan flags a pallet, the excursion is already hours old. The 10 KPIs below give you the real-time signal you need to catch problems while they’re still recoverable, not after the product is already compromised.
Temperature Excursion Rate
Why it Matters: Every excursion carries financial and regulatory exposure. High rates signal systemic equipment or process failures across the fleet or facility.
What it Measures: Percentage of shipments or storage zones that exceed defined temperature thresholds during a given period.
What Happens if Missed: Product spoilage, regulatory noncompliance, and customer chargebacks accumulate before anyone in operations has a clear picture.
Formula: (Number of Excursion Events / Total Shipments or Zones) x 100
Indicator Type: Current. Reflects the live state of thermal integrity across active shipments and storage assets right now.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking trailers or warehouse zones by excursion frequency.
Frequency: Real-time, continuous polling from temperature sensors.
Data Required: Setpoint temperature, actual recorded temperature, shipment or zone count, excursion event log.
Pro Tip: Segment excursion rate by lane, carrier, and asset type. Distribution center exits and last-mile handoffs tend to be the highest-risk transition points.
Red Flag: Excursion rate rising on a specific asset class or lane while everything else holds steady almost always points to a maintenance backlog or a carrier compliance gap.
Reefer Unit Runtime and Fuel Consumption
Why it Matters: Abnormal runtime or fuel burn is an early mechanical indicator before a unit fails mid-haul with a full load on board.
What it Measures: Hours of continuous engine operation and fuel volume consumed per unit over a defined operating period.
What Happens if Missed: Compressor failures, refrigerant leaks, and fuel system faults go undetected until the unit fails in transit or during a live delivery window.
Formula: Fuel Consumed (liters or gallons) / Operating Hours
Indicator Type: Leading. Elevated consumption-per-hour precedes most reefer mechanical failures by hours or days.
Unit of Measure: Liters per hour (L/h) or gallons per hour (gal/h)
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking units by highest fuel consumption deviation from fleet baseline.
Frequency: Real-time, from telematics and onboard diagnostics.
Data Required: Fuel consumption readings, engine runtime hours, unit identifier, ambient temperature at time of measurement.
Pro Tip: Normalize fuel consumption against ambient temperature. A unit burning more fuel on a hot day isn’t the same problem as one burning more fuel on a mild night.
Red Flag: A unit trending 15% or more above its fleet-peer baseline for three or more consecutive trips usually signals a refrigerant charge issue or condenser fouling.
Pre-Trip Inspection Pass Rate
Why it Matters: Reefer units that enter service with unresolved faults are a liability on every lane they run. Pass rate tells you whether your pre-dispatch process is actually catching problems.
What it Measures: Percentage of reefer units that pass pre-trip inspection without defects on the first check, before departure.
What Happens if Missed: Units with latent faults depart loaded, excursions occur mid-route, and the root cause traces directly back to a failed or skipped inspection step.
Formula: (Units Passing First Inspection / Total Units Inspected) x 100
Indicator Type: Leading. Declining pass rate predicts excursions and roadside failures before they happen.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI trend with real-time alerts; bullet chart showing current pass rate against target threshold.
Frequency: Per dispatch cycle; aggregated daily.
Data Required: Inspection outcome per unit, inspection timestamp, unit identifier, defect category if failed.
Pro Tip: Track first-pass rate by depot and by inspector shift. Consistent failure patterns at one location often reflect training gaps or maintenance backlog, not random equipment issues.
Red Flag: Pass rate dropping on a single equipment cohort (same model year or manufacturer) points to a fleet-wide mechanical issue that needs a systematic response, not individual unit repairs.
Cold Storage Zone Compliance Rate
Why it Matters: Warehouse temperature zones that drift outside spec put entire product categories at risk, and regulators will ask for the logs.
What it Measures: Percentage of monitored storage zones maintaining temperature within the defined operating range over a rolling time window.
What Happens if Missed: A zone running warm overnight can result in thousands of pallets of product requiring disposition review, with all the labor, write-off, and documentation that follows.
Formula: (Time in Compliance / Total Monitoring Period) x 100, per zone
Indicator Type: Current. Reflects the live thermal state of each warehouse zone against its specification.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI Map View showing all zones with live status by color; Pareto chart when ranking zones by total out-of-spec time.
Frequency: Real-time, from fixed sensor arrays.
Data Required: Zone setpoint, actual zone temperature, sensor read timestamps, zone identifier, product category stored.
Pro Tip: Set tighter alert bands than your regulatory requirement. A zone drifting toward the limit gives you time to intervene. A zone that’s already out of spec gives you paperwork.
Red Flag: A zone repeatedly falling out of compliance during the same time window each day is almost always a loading dock door event or an HVAC cycling issue. It’s a process problem, not a sensor problem.
Delivery Temperature Compliance Rate
Why it Matters: This is the proof-of-delivery equivalent for cold chain. It tells you whether a product actually arrived within spec, regardless of what happened in transit.
What it Measures: Percentage of deliveries where product temperature at the point of receipt falls within the contracted or regulatory range.
What Happens if Missed: Customer disputes, product rejection, and food safety authority involvement are the operational and commercial consequences of chronic delivery noncompliance.
Formula: (Compliant Deliveries / Total Deliveries) x 100
Indicator Type: Lagging. Confirms whether the full chain from origin to receiver actually maintained temperature.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking lanes, carriers, or customers by noncompliant delivery count.
Frequency: Per delivery event; aggregated daily and by lane.
Data Required: Receiver-recorded delivery temperature, product specification limit, shipment identifier, lane, carrier, and timestamp.
Pro Tip: Compare delivery compliance rate against your own pre-departure temperature readings. A gap between the two isolates where in the chain the excursion is actually happening.
Red Flag: A carrier with high delivery compliance overall but persistent noncompliance on a specific lane or product type suggests a specific handling or equipment issue on that route.
Mean Time to Detect Temperature Excursion
Why it Matters: It doesn’t matter that you have sensors if the alert takes 45 minutes to reach the person who can act on it. Detection speed determines whether an excursion is recoverable.
What it Measures: Average elapsed time between the start of a temperature excursion event and the first alert or notification reaching operations.
What Happens if Missed: Slow detection converts a recoverable equipment issue into a product loss event. Time is the variable that determines the outcome.
Formula: Average (Alert Timestamp – Excursion Start Timestamp), across all events
Indicator Type: Leading. Improving detection speed reduces the severity and cost of every future excursion.
Unit of Measure: Minutes
Ideal Visualization(s): KPI trend with real-time alerts; histogram showing distribution of detection times across all events.
Frequency: Per excursion event; trended continuously.
Data Required: Excursion start time from sensor data, alert generation timestamp, notification delivery timestamp, event identifier.
Pro Tip: Measure detection time separately for in-transit and in-facility excursions. They have different root causes and different response paths.
Red Flag: Detection time increasing over a rolling 30-day window usually means sensor polling frequency has been reduced, an alert rule has been changed, or someone has been silencing notifications.
Reefer Unit Availability Rate
Why it Matters: You cannot run a cold chain without available assets. Low availability means loading delays, lane substitutions, and contracted shipments moving on equipment that wasn’t assigned for a reason.
What it Measures: Percentage of reefer units in the fleet available for dispatch at any given time, excluding units down for maintenance, repair, or inspection hold.
What Happens if Missed: Fleet availability shortfalls create cascading scheduling problems. The downstream impact on lane coverage and customer commitments is usually larger than the maintenance event itself.
Formula: (Available Units / Total Fleet Units) x 100
Indicator Type: Current. Reflects the live operational capacity of the refrigerated fleet right now.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI block with live status; bullet chart showing current availability against minimum threshold; Pareto chart when ranking depots by availability rate.
Frequency: Real-time, updated on maintenance status change events.
Data Required: Unit status per asset (available, in service, in maintenance, inspection hold), unit identifier, depot location.
Pro Tip: Track availability by depot, not just fleet-wide. A high aggregate rate can mask a single depot with a chronic maintenance backlog that’s driving disproportionate lane risk.
Red Flag: Availability dropping below 85% fleet-wide for more than 48 consecutive hours is a scheduling and maintenance prioritization problem that needs executive visibility immediately.
Cross-Dock Dwell Time
Why it Matters: Every minute a temperature-sensitive pallet sits in an ambient or inadequately controlled cross-dock environment is thermal risk accumulation. Dwell time is where the cold chain breaks down most often without being captured.
What it Measures: Average elapsed time between pallet arrival and pallet departure at cross-dock and transshipment facilities.
What Happens if Missed: Excessive dwell at an ambient dock drives excursions that show up in delivery compliance metrics with no obvious in-transit cause, making root cause analysis harder and slower.
Formula: Average (Outbound Scan Timestamp – Inbound Scan Timestamp) per pallet or shipment unit
Indicator Type: Current. Indicates live throughput performance at each cross-dock location relative to product hold time limits.
Unit of Measure: Minutes or hours
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking cross-dock locations by average dwell time.
Frequency: Per scan event; aggregated in real time.
Data Required: Inbound scan timestamp, outbound scan timestamp, shipment identifier, cross-dock location identifier, product temperature specification.
Pro Tip: Set dwell time alerts relative to the product’s thermal tolerance, not a fixed limit. Frozen goods can tolerate a different dwell window than chilled produce or pharmaceuticals.
Red Flag: Dwell time spiking at one location during the same shift window consistently points to a staffing or equipment constraint at that facility, not a systemic network issue.
Refrigerant Charge Health Index
Why it Matters: Refrigerant loss is the leading mechanical cause of temperature excursions in reefer fleets. By the time a unit fails to hold temperature, it’s already been underperforming for some time.
What it Measures: Composite indicator of refrigerant system health per unit, derived from suction pressure, discharge pressure, and temperature differential across the evaporator coil.
What Happens if Missed: Units with degraded refrigerant charge continue to operate while temperature-holding capacity declines, typically without any visible alarm until the threshold is crossed mid-haul.
Formula: N/A
Indicator Type: Leading. Declining index scores precede temperature excursions by hours to days.
Unit of Measure: Index (0-100)
Ideal Visualization(s): Bullet chart showing current index score against health bands; KPI trend with real-time alerts; Pareto chart when ranking units by lowest health index score.
Frequency: Real-time, from onboard refrigeration telemetry.
Data Required: Suction pressure, discharge pressure, evaporator inlet and outlet temperature, ambient temperature, unit identifier.
Pro Tip: Use this index to prioritize which units get pulled for preventive refrigerant service before they fail a load. Units below 70 on the index should not be assigned high-value or pharmaceutical lanes.
Red Flag: A unit with an index score dropping more than 10 points in a single operating day has a refrigerant system issue that needs workshop attention before the next dispatch.
On-Time Cold Chain Handoff Rate
Why it Matters: Temperature integrity depends on every link in the chain handing off on schedule. Late handoffs create dwell time, break cold room access windows, and force product to sit in conditions that weren’t planned for.
What it Measures: Percentage of scheduled handoffs (loading, transfer, delivery, or cross-dock arrival) completed within the contracted or planned time window.
What Happens if Missed: Late handoffs are the most common untracked driver of cold chain failures. They compound downstream, and the thermal cost accumulates at every subsequent node in the network.
Formula: (On-Time Handoffs / Total Scheduled Handoffs) x 100
Indicator Type: Current. Tracks live adherence to the operational schedule across every handoff node in the network.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking lanes or partners by handoff delay frequency.
Frequency: Per event; aggregated by shift, daily, and by lane.
Data Required: Planned handoff time, actual handoff time, handoff type (loading, transfer, delivery), location identifier, carrier or partner identifier.
Pro Tip: Correlate handoff compliance with your delivery temperature compliance rate. The nodes with the lowest handoff rates are almost always the same nodes generating your excursions.
Red Flag: A third-party partner with declining on-time handoff rates and no corresponding excursion reports probably isn’t reporting excursions accurately. Audit that node.
Why Real-Time Visibility Matters
Cold chain operations run on extremely narrow tolerances, and the gaps between those tolerances and real conditions are often invisible until the damage is done. A warehouse that looks fine on a morning walkthrough can have three zones in creeping excursion by mid-afternoon. A reefer fleet with 95% delivery compliance can have a dozen units quietly accumulating refrigerant system risk that will surface as a lane failure next week.
The difference between an operations leader who is reacting and one who is in front of the problem is almost always a matter of detection speed and visibility depth. Real-time KPIs that alert before thresholds are breached, that surface patterns across a fleet or network, and that make the current state obvious without requiring someone to go looking, are what keep cold chain compliance from being a permanent fire drill.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
An aerial view of a high-pressure pipeline and valve manifold system at a commercial-scale hydrogen or gas processing facility. At this level of complexity, manual monitoring falls short. Real-time KPI visibility across flow, pressure, and system state is what keeps operations running safely and on target.
Green hydrogen electrolysis is a precision chemistry process masquerading as a utility operation, and the margin between optimal performance and expensive waste is measured in millivolts and milliliters per minute. When stack health degrades undetected, efficiency curves slide quietly in the wrong direction. Faradaic losses compound. Water quality drifts outside tolerance, and by the time the shift report surfaces the issue, you’ve already burned the energy and lost the production. In a process where power costs dominate your cost of goods, flying blind is not a recoverable position.
The 10 KPIs below give you real-time command of what’s actually happening inside your electrolyzers, your balance of plant, and your hydrogen output quality before problems escalate into unplanned shutdowns or off-spec product.
Stack Cell Voltage Deviation
Why it Matters: Cell voltage spread across a stack is the earliest indicator of degradation, membrane fouling, or flow distribution problems. Catching it early protects stack life.
What it Measures: The variance between individual cell voltages within an electrolyzer stack at a given current density. Tight spread means a healthy, balanced stack.
What Happens if Missed: Uneven cell loading accelerates membrane wear on high-voltage cells. A stack replacement program that should run 80,000 hours starts looking like 40,000.
Formula: Max Cell Voltage – Min Cell Voltage (per stack, per operating point)
Indicator Type: Leading. Rising deviation warns of stack imbalance weeks before overall efficiency or output quality degrades measurably.
Unit of Measure: mV
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking stacks by cell voltage deviation to prioritize inspection.
Frequency: Real-time (per scan cycle)
Data Required: Individual cell voltage readings, stack current, operating temperature, stack identifier
Pro Tip: Set deviation thresholds relative to baseline at commissioning, not absolute values. Stack behavior shifts with age, and a fixed threshold either alarms too early or too late.
Red Flag: Any single cell trending consistently 20 mV or more above its stack average is a membrane failure in progress, not a sensor artifact.
Specific Energy Consumption
Why it Matters: This is your primary cost-of-production metric. Every deviation from the design curve translates directly into dollars per kilogram of hydrogen produced.
What it Measures: Electrical energy consumed per unit mass of hydrogen produced, typically expressed at standard conditions. It benchmarks actual efficiency against the thermodynamic and design baseline.
What Happens if Missed: Efficiency losses hide inside acceptable-looking production totals. You’re making your tonnage target but paying a premium that erodes your economics quietly over months.
Formula: Total kWh Consumed / kg H2 Produced
Indicator Type: Current. It reflects the real-time efficiency state of the system and responds immediately to stack health, operating pressure, and temperature conditions.
Unit of Measure: kWh/kg H2
Ideal Visualization(s): KPI trend with real-time alerts; bullet chart comparing actual specific energy against design target and best-demonstrated performance.
Frequency: Real-time, updated every minute
Data Required: Active power draw (kW), hydrogen mass flow rate, operating temperature, operating pressure, stack count in service
Pro Tip: Segment this KPI by stack cluster. A system-level average masks a degraded cluster being carried by healthier ones.
Red Flag: Specific energy rising consistently above the design curve at rated current density, with no change in water quality or temperature, points directly at stack degradation rather than a balance-of-plant issue.
Feed Water Conductivity
Why it Matters: Electrolyte contamination is a stack killer. Water purity is not a peripheral utility concern; it is a core process variable that governs membrane life and cell performance.
What it Measures: Electrical conductivity of the deionized water feed to the electrolyzers, as a proxy for total dissolved solids and ionic contamination levels.
What Happens if Missed: Ionic contaminants poison catalysts, accelerate membrane degradation, and introduce corrosion pathways into the stack. The damage is cumulative and largely irreversible.
Formula: N/A
Indicator Type: Leading. Water quality problems appear in the feed line before they manifest as stack performance changes, giving you a response window if you’re watching.
Unit of Measure: µS/cm
Ideal Visualization(s): KPI trend with real-time alerts; status history trend to correlate conductivity excursions with upstream water treatment events.
Frequency: Real-time (continuous sensor)
Data Required: Feed water conductivity, polisher outlet conductivity, resin bed pressure differential, flow rate
Pro Tip: Set two alert thresholds: a caution band that triggers a review of your water treatment loop, and a hard limit that initiates an automatic electrolyzer load reduction before the water reaches the stacks.
Red Flag: A sudden conductivity spike that resolves quickly often means a resin bed is channeling. The average looks fine; the peak does the damage.
Hydrogen Purity
Why it Matters: Off-spec hydrogen is unsaleable and can create safety hazards downstream. Purity is both a product quality requirement and a process health indicator.
What it Measures: Hydrogen gas purity at the electrolyzer outlet, with oxygen and water vapor as the primary monitored impurities. Purity requirements vary by application but commonly target 99.9% or higher.
What Happens if Missed: Oxygen crossover above the lower flammability limit creates an explosive atmosphere in the hydrogen circuit. Even sub-critical crossover accumulation degrades downstream purification efficiency and increases operating cost.
Formula: (1 – Impurity Concentration) x 100
Indicator Type: Current. Purity is a real-time output quality signal and a simultaneous indicator of membrane integrity and differential pressure control.
Unit of Measure: % H2 (volume)
Ideal Visualization(s): KPI trend with real-time alerts; KPI block showing live purity against spec limit for immediate operator awareness.
Frequency: Real-time (continuous analyzer) or every 1-5 minutes depending on analyzer type
Data Required: H2 purity analyzer output, O2 concentration in H2 stream, dew point, differential pressure across membrane, stack current
Pro Tip: Correlate purity drops with differential pressure excursions. A membrane pinhole shows up in the oxygen-in-hydrogen trace before any other signal.
Red Flag: O2 content in the hydrogen stream rising above 0.4% demands immediate action. At that point, you’re approaching the safety threshold, not just a quality threshold.
Differential Pressure Across the Membrane
Why it Matters: Membrane differential pressure (delta-P) governs gas crossover rates. Mismanaged delta-P is the fastest path to a safety incident or a destroyed stack.
What it Measures: The pressure difference between the hydrogen side and the oxygen side of the electrolyzer membrane. Proper control keeps gas crossover within safe limits.
What Happens if Missed: Excessive delta-P drives hydrogen into the oxygen stream or oxygen into the hydrogen stream. Both scenarios are dangerous. Either one also accelerates membrane mechanical degradation.
Formula: P(H2 side) – P(O2 side)
Indicator Type: Current. Delta-P is an active control variable with immediate safety implications if it moves outside the operating envelope.
Unit of Measure: mbar
Ideal Visualization(s): KPI trend with real-time alerts; KPI block with status coloring (green / yellow / red) visible at a glance on the operations display.
Frequency: Real-time (per scan cycle)
Data Required: Hydrogen circuit pressure, oxygen circuit pressure, stack operating current, liquid level in gas-liquid separators
Pro Tip: Delta-P instability during load transients is normal within limits. It’s sustained delta-P drift at a steady state that requires investigation.
Red Flag: Delta-P swings that correlate with separator liquid level fluctuations point to a level control issue, not a membrane issue. Treat the right root cause.
Stack Operating Temperature
Why it Matters: Temperature directly controls electrolyzer efficiency and membrane durability. Running too hot accelerates degradation; running too cold wastes energy and risks condensation damage.
What it Measures: Average and distribution of operating temperature across the electrolyzer stack, monitored at key thermal points in the stack and electrolyte circuit.
What Happens if Missed: Thermal runaway in a poorly cooled stack can destroy membranes and electrodes within a single shift. Chronic overcooling suppresses efficiency and inflates energy costs.
Formula: N/A
Indicator Type: Current. Temperature is an active operating parameter that requires continuous regulation relative to load, ambient conditions, and coolant performance.
Unit of Measure: °C
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking stacks by temperature deviation from setpoint to identify cooling circuit problems.
Frequency: Real-time (per scan cycle)
Data Required: Stack inlet and outlet temperatures, coolant flow rate, coolant supply temperature, ambient temperature, stack current density
Pro Tip: Temperature gradients within a stack are more informative than average temperature. A stack averaging 75°C with a 15°C internal gradient has a cooling flow distribution problem.
Red Flag: Stack temperature rising faster than the coolant response can explain during a stable load period means you have an internal resistance increase. Pull the stack voltage profile immediately.
Faradaic Efficiency
Why it Matters: Faradaic efficiency tells you how much of the electricity you consume is actually converting water into hydrogen, versus being lost to parasitic reactions and gas crossover.
What it Measures: The ratio of actual hydrogen produced to the theoretical maximum hydrogen producible from the charge passed through the cell. It declines as crossover and side reactions increase.
What Happens if Missed: A 2-3% drop in Faradaic efficiency at scale means thousands of kilograms of hydrogen production capacity lost annually with no visible alarm. The energy bill rises; the output tonnage quietly falls short.
Formula: (Actual H2 Production Rate / Theoretical H2 Production Rate) x 100
Indicator Type: Lagging. Faradaic efficiency reflects the cumulative effect of membrane condition, water purity, and operating discipline over time.
Unit of Measure: %
Ideal Visualization(s): KPI trend with real-time alerts; SPC trend (control chart) to distinguish normal operating variation from statistically significant degradation trends.
Frequency: Updated every 15 minutes (requires stable operating conditions for accuracy)
Data Required: Total charge passed (Ampere-hours), actual hydrogen mass flow at outlet, stack current, operating temperature and pressure
Pro Tip: Faradaic efficiency calculated at off-design current densities is misleading. Compare only at consistent operating points to avoid misinterpreting load-following artifacts as degradation.
Red Flag: Faradaic efficiency declining faster than the expected stack aging curve, without a water quality event to explain it, is a membrane integrity issue that warrants physical inspection.
Water Consumption Rate vs. Theoretical
Why it Matters: The ratio of actual water consumed to theoretically required water is a mass balance integrity check for the entire system. Deviations reveal leaks, metering errors, or unaccounted losses.
What it Measures: Actual deionized water consumption per unit of hydrogen produced, compared to the stoichiometric requirement. Excess consumption points to losses; deficits point to measurement errors.
What Happens if Missed: Undetected water losses mean undetected electrolyte inventory changes. In alkaline electrolyzers, lye concentration drifts. In PEM systems, membrane hydration may be compromised. Both degrade performance.
Formula: Actual Water Consumed (L) / H2 Produced (kg) vs. Stoichiometric Ratio (8.9 L/kg)
Indicator Type: Current. Water balance deviations appear in near real-time and often indicate problems faster than gas quality analyzers respond.
Unit of Measure: L H2O / kg H2
Ideal Visualization(s): KPI trend with real-time alerts; bar chart comparing actual vs. theoretical consumption across operating periods.
Frequency: Updated every 15 minutes
Data Required: Deionized water feed flow, hydrogen mass production, oxygen side water recovery flow, system drain and blowdown volumes
Pro Tip: Establish a baseline water-to-hydrogen ratio at commissioning for each stack configuration. A 5% deviation from baseline warrants a leak survey; a 15% deviation warrants a shutdown and inspection.
Red Flag: Water consumption rising while hydrogen production holds flat means you have a leak path somewhere in the system. It’s not an instrumentation error until you’ve proved otherwise.
Electrolyzer Availability
Why it Matters: For any hydrogen project tied to a power purchase agreement or a downstream offtake contract, availability directly determines whether you meet your commercial obligations.
What it Measures: The percentage of scheduled production time during which the electrolyzer system is in a state capable of producing hydrogen at rated conditions.
What Happens if Missed: Availability gaps compound. A 92% availability target missed by 3 points translates directly to a production shortfall and, depending on your contract structure, penalty exposure.
Formula: (Total Available Hours / Total Scheduled Hours) x 100
Indicator Type: Lagging. Availability is a summary metric that reflects the aggregate effect of all reliability, maintenance, and operational decisions made during the period.
Unit of Measure: %
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking downtime causes by duration to guide reliability improvement priorities.
Data Required: Electrolyzer operating state (run, standby, fault, planned maintenance), fault codes, planned outage schedule, actual outage start and end times
Pro Tip: Track availability separately by stack cluster and by system (including balance of plant). BOP failures account for a higher fraction of downtime than most teams expect, especially in the first two years of operation.
Red Flag: Availability trending below target in the first year without a clear root cause identified means your preventive maintenance intervals are wrong, not that the equipment is inherently unreliable.
Hydrogen Production Rate vs. Setpoint
Why it Matters: Production rate is your primary output KPI and your first-line indicator that the system is responding correctly to dispatch signals or load setpoints.
What it Measures: Actual hydrogen mass flow at the system outlet compared to the commanded setpoint, expressed as a percentage of target. It captures both steady-state performance and transient response fidelity.
What Happens if Missed: Chronic underproduction against setpoint masks itself inside daily totals. You discover the shortfall at the end of the month, not at the moment when corrective action is still possible.
Formula: (Actual H2 Production Rate / Setpoint H2 Production Rate) x 100
Indicator Type: Current. Production rate deviation is an immediate signal that something in the system, whether electrical, mechanical, or process-related, is not performing as commanded.
Unit of Measure: % of setpoint
Ideal Visualization(s): KPI trend with real-time alerts; sparklines showing production rate across active stacks to identify which units are underperforming relative to their contribution.
Frequency: Real-time, updated every minute
Data Required: Hydrogen mass flow at system outlet, active power setpoint, current setpoint per stack, number of stacks in service, stack operating state
Pro Tip: Separate setpoint deviation caused by commanded load curtailment from deviation caused by process limitations. Curtailment is planned; process-driven shortfall is a problem.
Red Flag: Production rate consistently lagging setpoint by more than 5% at steady state, with all stacks showing normal individual cell voltages, points to a flow measurement issue or a significant hydrogen inventory loss between the stack outlets and the system meter.
Why Real-Time Visibility Matters
Green hydrogen electrolysis operates at the intersection of electrochemistry, fluid dynamics, and high-power electrical systems. Each of those domains has its own failure modes, and they interact in ways that only show up clearly when you have continuous visibility across all of them simultaneously. A water quality excursion that lasts 20 minutes can degrade a membrane that was otherwise good for years. A cooling imbalance that never triggers a temperature alarm can shorten a stack replacement interval by half.
These aren’t edge cases; they’re the normal operating reality of a system running at commercial scale, and they only become manageable when you can see them as they happen. The KPIs above don’t represent a complete monitoring program on their own, but they represent the signals that separate reactive fire-fighting from proactive performance management. When these metrics are visible in real time, with alerts that reach the right people before problems escalate, your team spends its time on optimization instead of incident response.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
Precision-machined titanium components staged for dimensional inspection, where surface finish and tolerance conformance determine whether the part advances to assembly or returns to the crushing and sizing loop.
Titanium processing is a business where the gap between in-spec and scrap is measured in degrees, seconds, and parts per million. The Kroll process doesn’t forgive late corrections. Chlorination temperatures drift, reduction reactor conditions shift, and distillation cycles run long before a batch report tells you anything went wrong. By then, you’re already counting the loss. Every delay in detection ripples forward into yield penalties, contaminated product, and downstream delivery failures.
The ten KPIs below give you the real-time grip you need to run titanium operations with precision instead of hindsight.
Chlorination Reactor Temperature Deviation
Why it Matters: Chlorination temperature directly controls TiCl4 yield and upstream impurity profile. Deviations drive off-spec material that cascades through the entire process.
What it Measures: The real-time difference between actual reactor temperature and the setpoint across chlorination beds.
What Happens if Missed: Persistent deviation produces iron chloride and vanadium contamination that loads downstream purification and increases scrap rates.
Formula: (Actual Reactor Temperature – Setpoint Temperature) / Setpoint Temperature × 100
Indicator Type: Current. Temperature deviation is an immediate signal of reactor instability requiring fast operator response.
Unit of Measure: % deviation
Ideal Visualization(s): KPI trend with real-time alerts; bullet chart showing current deviation against control limits; Pareto chart when ranking reactor zones by deviation magnitude.
Frequency: Real-time (continuous)
Data Required: Reactor bed temperature (multiple zones), temperature setpoints, feed rate
Pro Tip: Set asymmetric alert bands. High-side deviation in chlorination is typically more damaging than low-side because of vanadium volatilization windows.
Red Flag: Repeated small deviations that self-correct without operator intervention usually signal a feed consistency problem, not a temperature control problem.
Kroll Reduction Cycle Time vs. Standard
Why it Matters: Cycle time variance directly signals whether magnesium reduction is completing cleanly or leaving residual TiCl4 in the reactor.
What it Measures: The elapsed time of each reduction cycle compared to the established standard for that reactor and batch size.
What Happens if Missed: Extended cycles tie up reactor capacity and increase magnesium consumption. Short cycles risk incomplete reduction and contaminated sponge.
Formula: (Actual Cycle Time – Standard Cycle Time) / Standard Cycle Time × 100
Indicator Type: Current. Cycle time deviation reflects process conditions right now and signals whether the current batch is on track.
Unit of Measure: % deviation
Ideal Visualization(s): KPI trend with real-time alerts; status history trend showing cycle completion patterns across reactors; Pareto chart when ranking reactors by cycle time variance.
Frequency: Per cycle (updated continuously during active reduction)
Data Required: Cycle start timestamp, current elapsed time, standard cycle time by reactor and batch size, reactor ID
Pro Tip: Track cycle time variance by reactor age. Older reactors trend longer before failure modes become visible in temperature or pressure data.
Red Flag: A cycle running more than 8% long with stable temperature and pressure often indicates feed purity issues rather than process control failure.
Vanadium and Iron Removal Efficiency
Why it Matters: TiCl4 purity entering the Kroll reactor sets the ceiling on final sponge grade. Vanadium contamination above threshold makes the batch unacceptable for aerospace-grade product.
What it Measures: Vanadium and iron content in purified TiCl4 feed expressed against the specification limit for the target product grade.
What Happens if Missed: Contaminated feed produces off-grade sponge that can’t be reclassified without full reprocessing. The cost is the full reduction cycle plus material write-off.
Indicator Type: Current. Purity at the purification stage is a live gate on what enters the most expensive part of the process.
Unit of Measure: % of spec limit (ppm basis)
Ideal Visualization(s): Bullet chart showing current purity against specification thresholds; KPI trend with real-time alerts; SPC trend (control chart) for long-run purity stability.
Frequency: Per batch or continuous if inline analyzers are available
Data Required: Vanadium concentration (ppm), iron concentration (ppm), specification limits by product grade, batch ID
Pro Tip: Correlation of vanadium spikes with chlorination bed temperature excursions is one of the fastest root-cause paths available. Track both in the same view.
Red Flag: Iron contamination spiking independently of vanadium suggests a vessel corrosion event, not a feed chemistry problem.
Magnesium Consumption Rate
Why it Matters: Excess magnesium consumption raises input costs and loads the MgCl2 recovery system. Insufficient consumption leaves TiCl4 unreduced in the reactor.
What it Measures: Actual magnesium consumed per kilogram of titanium sponge produced, compared to the theoretical consumption ratio.
What Happens if Missed: Consumption trending above target often indicates incomplete previous distillation or reactor condition degradation, both of which are expensive to address reactively.
Formula: Actual Mg Consumed (kg) / Ti Sponge Produced (kg)
Indicator Type: Current. This ratio reflects in-progress batch efficiency and flags process imbalances before cycle completion.
Unit of Measure: kg Mg / kg Ti
Ideal Visualization(s): KPI trend with real-time alerts; bullet chart showing current ratio against target and control limits.
Frequency: Real-time (continuous during reduction cycle)
Data Required: Magnesium feed mass, titanium sponge yield (estimated from reactor conditions), reactor ID, batch ID
Pro Tip: Magnesium consumption ratio combined with cycle time variance gives you a two-signal early warning for reactor condition that neither KPI delivers alone.
Vacuum Distillation Residual Pressure
Why it Matters: Distillation removes residual magnesium and MgCl2 from sponge. Insufficient vacuum means incomplete removal, which contaminates the sponge and fails chemistry specs.
What it Measures: Chamber pressure during the distillation phase measured against the required operating envelope.
What Happens if Missed: Residual magnesium in sponge causes chemistry failures at the melting stage, typically discovered after ingot production when reversal is impossible.
Formula: Actual Chamber Pressure vs. Required Pressure Range (binary in-spec / out-of-spec or continuous mbar value)
Indicator Type: Current. Vacuum integrity is a live process condition that determines whether the current distillation cycle will achieve spec.
Unit of Measure: mbar
Ideal Visualization(s): KPI trend with real-time alerts; bullet chart showing current pressure against minimum vacuum requirement.
Frequency: Real-time (continuous during distillation)
Data Required: Chamber pressure (mbar), distillation phase timestamp, required pressure envelope by product grade
Pro Tip: A vacuum that achieves target pressure but degrades over time during the hold phase usually indicates a seal or gasket failure, not pump capacity.
Red Flag: Repeated pressure excursions in the same distillation unit on consecutive cycles signal equipment degradation before a formal maintenance event surfaces it.
Sponge Crushing and Sizing Yield
Why it Matters: Crushing yield determines how much of the sponge mass converts to usable product fractions. Poor yield means either over-crushing (fines loss) or under-crushing (oversized rejection).
What it Measures: The percentage of sponge mass that falls within the target particle size distribution after crushing.
What Happens if Missed: Yield losses at crushing are unrecoverable in most cases. Fines are typically reprocessed at a discount or blended away, which erodes margin on already-tight product grades.
Formula: (On-spec Sponge Mass / Total Crushed Sponge Mass) × 100
Indicator Type: Current. Yield tracks batch-by-batch performance and captures equipment wear trends before they become rejection events.
Unit of Measure: %
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking crushing lines by yield loss; box plot for yield distribution across shifts.
Frequency: Per batch
Data Required: Pre-crush sponge mass, on-spec output mass, off-spec fraction mass (oversized, fines), crusher ID
Pro Tip: Yield degradation that correlates with shift changes is almost always a setup or calibration issue, not an equipment issue.
Red Flag: A steady drift in fines percentage over two to three weeks is a leading indicator of crusher wear, not random variation.
Ingot Melt Electrode Consumption Rate
Why it Matters: Electrode consumption rate in VAR (vacuum arc remelting) directly affects energy cost per kilogram and cycle time for ingot production.
What it Measures: The rate at which the consumable electrode melts per unit of arc energy applied, compared to the standard for that electrode specification.
What Happens if Missed: Abnormal consumption rate signals arc instability, electrode contamination, or power supply issues. Any of these can produce segregation in the ingot that isn’t detectable until testing.
Formula: Electrode Mass Loss (kg) / Arc Energy Applied (kWh)
Indicator Type: Current. Consumption rate during an active melt cycle reflects real-time arc conditions and electrode quality.
Unit of Measure: kg / kWh
Ideal Visualization(s): KPI trend with real-time alerts; SPC trend (control chart) for melt-to-melt consistency; Pareto chart when ranking VAR furnaces by consumption deviation.
Frequency: Real-time (continuous during melt cycle)
Data Required: Electrode mass (start and real-time), arc current, arc voltage, melt cycle timestamp, furnace ID
Pro Tip: Electrode consumption rate combined with arc voltage variance is the earliest non-destructive signal for cold-shut formation risk in the ingot.
Chlorine Recovery and Recycle Rate
Why it Matters: Chlorine recovery directly drives raw material cost and site safety risk. Unrecovered chlorine becomes an emissions and compliance problem, not just an efficiency one.
What it Measures: The percentage of chlorine generated in the process that is captured and recycled back into chlorination feed.
What Happens if Missed: Low recovery rates increase chlorine purchase cost and trigger regulatory reporting thresholds that carry operational and reputational consequences.
Indicator Type: Leading. Recovery rate trends ahead of emissions and compliance events and reflects the health of the recovery loop before a formal exceedance occurs.
Unit of Measure: %
Ideal Visualization(s): KPI trend with real-time alerts; bullet chart showing current recovery against compliance threshold; KPI block for shift-level summary.
Frequency: Real-time (continuous)
Data Required: Chlorine generation rate, chlorine recycled volume, vent stack flow rate, recovery system status
Pro Tip: Chlorine recovery rate dropping without a corresponding change in generation rate almost always points to a condenser or absorption column fouling event.
Red Flag: A recovery rate that fluctuates by more than 3 to 4 points within a single shift suggests a control valve or pressure regulation issue in the recovery loop.
Unplanned Reactor Downtime Rate
Why it Matters: Titanium’s long cycle times mean unplanned downtime in chlorination or reduction carries a cost multiplier that batch industries with short cycles don’t face.
What it Measures: The percentage of scheduled operating hours lost to unplanned stoppages across chlorination and reduction reactors.
What Happens if Missed: Unplanned downtime that isn’t tracked in real time tends to get absorbed into shift reports rather than driving maintenance priority changes. The pattern repeats until a major failure forces action.
Indicator Type: Lagging. This KPI measures what has already happened, but drives predictive maintenance prioritization for the next period.
Unit of Measure: %
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking reactors or equipment classes by downtime contribution; status history trend for visualizing stoppage patterns by shift.
Frequency: Real-time (event-driven, updated on each stoppage event)
Data Required: Reactor run/stop status, stoppage timestamps, stoppage cause code, scheduled hours, reactor ID
Pro Tip: Pareto the downtime by failure mode, not just by equipment. Seal failures on multiple reactors rank differently than a single catastrophic event, and the response is completely different.
Red Flag: If unplanned downtime concentrates on a single shift pattern, the problem is almost never the equipment.
Product Grade Achievement Rate
Why it Matters: Grade achievement is the ultimate measure of whether the entire upstream process is operating within spec. It’s the score that customers see.
What it Measures: The percentage of finished product batches that achieve the target grade on first assessment, without rework or downgrade.
What Happens if Missed: Grade failures require either reprocessing (time and energy cost) or downgrade sales (margin loss). Either outcome erodes the economics of a process that already runs on tight margins.
Formula: (On-Grade Batches / Total Batches Assessed) × 100
Indicator Type: Lagging. Grade achievement reflects the cumulative performance of all upstream steps and is the lagging confirmation of process health.
Unit of Measure: %
Ideal Visualization(s): KPI trend with real-time alerts; Pareto chart when ranking failure modes by grade impact; bar chart comparing grade achievement by product specification.
Frequency: Per batch (updated on assessment completion)
Data Required: Batch ID, grade specification, assessment result, downgrade classification, rework flag
Pro Tip: Tracking grade achievement alongside TiCl4 purity and VAR consumption rate in a single view gives you the fastest correlation path from upstream excursion to final product impact.
Red Flag: Grade achievement declining on a single product specification while others hold steady points to a process step specific to that grade’s requirements, not a site-wide drift.
Why Real-Time Visibility Matters
Titanium processing compresses risk into long, interdependent cycles where a problem in chlorination becomes a quality failure in ingots three to five steps later. Operating without real-time KPI visibility means you’re managing consequences rather than conditions. By the time a batch report confirms the yield loss or the grade failure, the decision window has already closed.
The KPIs above aren’t a dashboard exercise. They’re the minimum real-time signal set you need to keep a titanium operation running at grade, on yield, and within the process envelope that makes the economics work. Each one connects to a decision point where acting on current data costs less than acting on historical data. That gap, between knowing now and knowing later, is where margin is made or lost in this business.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
Dawn at a CO2 conditioning and injection terminal, where the columns hold compressed CO2 inside a narrow phase envelope. Operators walking the unit work to the same purity, pressure, and fugitive readings the control room is tracking in real time, because any drift here lands at the wellhead within minutes.
A Carbon Capture and Storage (CCS) operation is judged by what stays underground. Everything upstream of that, from the absorber column to the wellhead, exists to make sure the molecule you captured doesn’t end up back in the atmosphere or pooling somewhere it shouldn’t.
The chain is long: capture, conditioning, compression, transport, injection, storage. A problem at any point eventually shows up at every other point. Run blind for a shift and you risk an off-spec product entering a pipeline, missed tonnage commitments, or an induced seismic event that nobody flagged until it ended up on a regulator’s desk. Here are the ten real-time KPIs every CCS Operations Leader should have on the wall.
CO2 Capture Rate
Why it Matters: Capture rate drives every downstream metric. It defines whether the facility meets emissions targets, tonnage commitments, and credit obligations.
What it Measures: The percentage of CO2 removed from the inlet flue gas versus what slips through the absorber overhead vent.
What Happens if Missed: Slipping below target releases unaccounted CO2, jeopardizes 45Q credits, and forces costly solvent and steam corrections downstream.
Formula: (CO2 in – CO2 out) / CO2 in × 100
Indicator Type: Current. It reflects live capture performance and reacts within minutes to load swings or solvent degradation.
Data Required: Inlet flue gas CO2 concentration, outlet vent CO2 concentration, inlet flow rate, outlet flow rate
Pro Tip: Run capture rate alongside reboiler duty on the same screen. The two move together, and the relationship exposes solvent health early.
Red Flag: Capture rate drifting down while reboiler duty climbs signals amine degradation, foaming, or fouling building inside the absorber.
Specific Reboiler Duty
Why it Matters: Reboiler duty is the largest energy cost in amine-based capture. Small drifts compound into millions of dollars across a year.
What it Measures: Thermal energy delivered to the reboiler per tonne of CO2 captured, expressed in gigajoules per tonne.
What Happens if Missed: Rising duty erodes margins, increases parasitic load on the host plant, and signals solvent or column problems forming underneath.
Formula: Reboiler heat input (GJ) / CO2 captured (tonnes)
Indicator Type: Current. It tracks energy efficiency continuously and surfaces creeping degradation before it shows up on the operating bill.
Unit of Measure: GJ per tonne CO2
Ideal Visualization(s): KPI blocks, KPI trend with real-time alerts, Histogram
Frequency: Real-time
Data Required: Steam flow to reboiler, steam enthalpy, condensate return rate, CO2 production rate
Pro Tip: Benchmark against commissioning baseline weekly. The first 0.1 GJ per tonne of drift is the cheapest one to investigate.
Red Flag: Duty rising with stable capture rate usually means heat exchanger fouling or solvent contamination, not a real load change.
CO2 Product Stream Purity
Why it Matters: Off-spec CO2 corrodes pipelines, damages compressors, and threatens injection well integrity. Specifications exist for hard physical reasons.
What it Measures: Concentration of CO2 in the compressed product stream, plus impurity levels including water, oxygen, H2S, SOx, and NOx.
What Happens if Missed: Wet or oxygenated CO2 accelerates pipeline corrosion. Shutting down transport for cleanup costs more than continuous inline monitoring.
Formula: N/A
Indicator Type: Current. Product analyzers run continuously, and any deviation cascades into transport and storage systems within minutes.
Unit of Measure: mol % (CO2), ppm (impurities)
Ideal Visualization(s): KPI blocks, KPI trend with real-time alerts, Tables
Frequency: Real-time
Data Required: Product stream CO2 concentration, H2O content, O2 content, H2S content, SOx and NOx levels
Pro Tip: Track moisture and oxygen on the same trend. Most pipeline operators treat these two as the gating impurities.
Red Flag: A sudden rise in moisture downstream of dehydration almost always means glycol carryover or a desiccant bed breakthrough.
CO2 Injection Rate
Why it Matters: Injection rate is the throughput number that ties storage commitments, capture-side production, and credit volumes together in one place.
What it Measures: Mass flow of CO2 entering the injection well at the wellhead, reconciled against allocation across the storage complex.
What Happens if Missed: Under-injection stalls upstream production and burns inventory. Over-injection risks exceeding permitted pressures and triggering regulatory review.
Formula: Cumulative mass flow / time period
Indicator Type: Current. It reflects active injection conditions and ties directly to pipeline throughput and capture-side production.
Unit of Measure: Tonnes per day (or tonnes per hour)
Ideal Visualization(s): KPI blocks, KPI trend with real-time alerts, Pareto chart when ranking wells by deviation from target, Group rollup bars
Frequency: Real-time
Data Required: Wellhead mass flow, wellhead pressure, wellhead temperature, injection valve position
Pro Tip: Plot actual rate against permit limits and target curve on the same chart. Operators stop second-guessing each other when both are visible.
Red Flag: Injection rate stable but wellhead pressure climbing fast usually points to near-wellbore plugging, not reservoir filling.
Bottomhole Pressure
Why it Matters: Bottomhole pressure defines the operating envelope for the entire storage complex. Exceed it, and you risk fracturing caprock or inducing seismicity.
What it Measures: Pressure measured downhole at injection depth, indicating reservoir response to cumulative CO2 placement over time.
What Happens if Missed: Pressure approaching fracture gradient invites induced seismicity, caprock breach, or regulatory suspension of the injection license.
Formula: N/A
Indicator Type: Leading. Rising bottomhole pressure signals approaching limits well before any surface metric catches up.
Pro Tip: Plot bottomhole pressure against cumulative tonnes injected. The slope tells you whether storage capacity behaves the way the model predicted.
Red Flag: A sudden non-linear jump in bottomhole pressure with no change in injection rate signals near-wellbore damage or reservoir compartmentalization.
Wellhead Annulus Pressure
Why it Matters: Annulus pressure is the earliest indicator of compromised well integrity. A casing leak in a CO2 well is not a minor event.
What it Measures: Pressure inside the annular space between casing strings, normally stable and tracked continuously against the established baseline.
What Happens if Missed: Undetected annular communication can vent CO2 into shallow zones or surface, triggering containment failure and mandatory remediation work.
Formula: N/A
Indicator Type: Leading. Annulus deviations precede visible leaks and offer the earliest possible warning of mechanical integrity loss.
Unit of Measure: PSI or kPa
Ideal Visualization(s): KPI blocks, KPI trend with real-time alerts, Pareto chart when ranking wells by deviation from baseline
Frequency: Real-time
Data Required: A-annulus pressure, B-annulus pressure, tubing pressure, ambient temperature
Pro Tip: Set narrow deviation alerts during early well life. A 50 psi drift in a new well is more informative than a wide static limit.
Red Flag: Annulus pressure tracking tubing pressure in real time means direct communication, which is active well integrity loss.
Microseismic Event Rate and Magnitude
Why it Matters: Induced seismicity is the regulatory and public-perception risk that ends CCS projects. Catching small events early prevents larger ones.
What it Measures: Rate, location, and magnitude of microseismic events detected within and around the storage complex by the monitoring array.
What Happens if Missed: Unmonitored event clustering can escalate to felt seismic activity, triggering injection halts, lawsuits, and erosion of the operating license.
Formula: N/A
Indicator Type: Leading. Microseismic patterns shift before macroseismic events occur, giving operators time to throttle injection or pause.
Unit of Measure: Event count and Moment Magnitude (Mw)
Ideal Visualization(s): GeoMap, KPI trend with real-time alerts, Histogram, Bar chart
Frequency: Real-time
Data Required: Event timestamp, hypocenter coordinates, magnitude, depth, event count by zone
Pro Tip: Pair event clustering with injection rate on one time axis. Causality conversations move faster when the data sits side by side.
Red Flag: A migrating cluster of events toward known faults or upward through the overburden warrants immediate review of injection strategy.
CO2 Plume Conformance
Why it Matters: Conformance is how you prove storage is going where the model said it would. Regulators, insurers, and partners all read this number first.
What it Measures: Observed CO2 plume extent and shape compared against the predicted footprint generated by the reservoir simulation.
What Happens if Missed: Plume migration outside the licensed area threatens the storage permit and pore-space leases. Recovery options are limited and expensive.
Formula: N/A
Indicator Type: Lagging. Conformance is observed after injection through monitoring wells, 4D seismic surveys, and modeled comparisons.
Unit of Measure: Percent deviation from predicted footprint, or square kilometers
Ideal Visualization(s): GeoMap, KPI Map, KPI trend with real-time alerts
Frequency: Updated with each monitoring campaign
Data Required: Time-lapse 4D seismic data, monitoring well pressure, monitoring well saturation, modeled plume boundaries
Pro Tip: Overlay each monitoring snapshot on the original simulation. Patterns of misfit show which subsurface assumptions need to be updated.
Red Flag: Plume movement faster than predicted, especially toward fault zones or property boundaries, signals a reservoir model that no longer holds.
Fugitive CO2 Emissions
Why it Matters: Every fugitive tonne cancels a captured tonne and undermines the credibility of the entire operation. Quantification is non-negotiable.
What it Measures: Detected CO2 concentrations from continuous point monitors, optical gas imaging, and ambient sensors across capture, transport, and injection sites.
What Happens if Missed: Undetected leaks compound emissions liabilities, invalidate credit claims, and put personnel and nearby communities at real risk.
Formula: Sum of leak rates across detection points × time
Indicator Type: Current. Sensors and imaging run continuously, though quantification depends on calibration and confirmation workflows.
Unit of Measure: Tonnes per day, ppm at detection point
Ideal Visualization(s): GeoMap, KPI blocks, KPI trend with real-time alerts, Pareto chart when ranking sites by detected emissions
Frequency: Real-time
Data Required: Point sensor concentrations, optical gas imaging detections, wind speed, wind direction, ambient temperature
Pro Tip: Tie wind data into every alert. A high reading downwind of a known vent is meteorology, not necessarily a leak.
Red Flag: A persistent low-level reading at a single sensor across multiple shifts is almost always a real leak, even when the magnitude looks modest.
Pipeline Flow Assurance
Why it Matters: CO2 pipelines operate in a narrow phase envelope. Drop out of dense phase and you get slugging, vibration, and unsafe shutdowns.
What it Measures: Pressure, temperature, and flow along the pipeline, evaluated against the supercritical or dense phase operating window.
What Happens if Missed: Two-phase flow damages compressors, pumps, and metering systems. Restarting from an unstable state is slow, costly, and risk-heavy.
Formula: N/A
Indicator Type: Current. Pipeline conditions update continuously and require active control to stay inside the dense phase envelope.
Unit of Measure: PSI, degrees C (or F), tonnes per hour
Ideal Visualization(s): KPI trend with real-time alerts, XY/scatter plot, Status history trends
Frequency: Real-time
Data Required: Inlet pressure, outlet pressure, temperature along the line, mass flow rate, ambient temperature
Pro Tip: Plot every operating point on a pressure-temperature scatter against the phase envelope. Operators spot drift visually long before alarms fire.
Red Flag: A widening gap between upstream and downstream pressure with stable flow signals phase change or hydrate formation along the line.
Why Real-Time Visibility Matters
CCS operates under permanent scrutiny: from regulators, from financiers, from the host plant whose emissions the facility exists to manage. Every captured tonne carries a paper trail, and the paper trail only works when the underlying data is current, accurate, and visible to the people who can act on it. Operating blind for a single shift can mean an off-spec product entering a pipeline, a missed injection target, or a microseismic cluster nobody flagged until it showed up in the news.
The chain runs from absorber to caprock, and a problem anywhere in it eventually shows up everywhere else. Real-time visibility connects capture efficiency to compression health to wellhead pressure to plume behavior, so the operations leader sees one system instead of five disconnected ones. That single picture is the difference between a project that earns credits and a project that earns headlines.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
A loaded Komatsu 960E haul truck climbs the ramp out of the pit with run-of-mine ore destined for the beneficiation plant, where recovery, grade, and throughput will decide how much of that payload becomes payable metal and how much ends up in the tailings dam.
Ore beneficiation is the narrow corridor between what the geology promised and what the smelter actually pays for. Every tonne of feed gets one chance to be liberated, separated, and concentrated, and every decision in that corridor either captures value or sends it to the tailings dam.
When a beneficiation plant runs blind, recovery slides before anyone notices, reagent dosing drifts, and grade variability shows up days later in the lab. By then the metal is gone and the only conversation left is who lost it. These ten KPIs are the ones a beneficiation operations leader should be able to see, trend, and act on in real time.
Metallurgical Recovery
Why it Matters: Recovery is the largest single lever on revenue per tonne milled. A one-point swing on a concentrator can mean millions per year.
What it Measures: The percentage of valuable minerals in the feed that reports to final concentrate, calculated from grade and tonnage on every stream.
What Happens if Missed: Metal leaves with the tailings, the loss compounds shift after shift, and the lab confirms it days after the money is gone.
Formula: Recovery (%) = (Concentrate Mass × Concentrate Grade) / (Feed Mass × Feed Grade) × 100
Indicator Type: Lagging, when calculated from assay. Current, when modelled from online analyzers. Treat any analyzer-driven recovery as a live operating signal.
Unit of Measure: Percent
Ideal Visualization(s): KPI block for current shift recovery, KPI trend with real-time alerts on downward drift, Pareto chart when ranking flotation lines by recovery deviation
Frequency: Real-time when online analyzers are available, hourly to shift-based on lab assay
Pro Tip: Pair the live analyzer recovery with the shift assay back-calculation. Persistent gaps usually point to a sampling or mass balance issue, not the process.
Red Flag: Recovery holding steady while concentrate grade falls. The plant is trading quality for the recovery number, and the penalty will follow.
Concentrate Grade
Why it Matters: Grade drives smelter payable value and penalty terms. Too low costs revenue; too high usually costs recovery.
What it Measures: The mass percentage of the target mineral in the final concentrate stream, measured by online analyzer or composite assay.
What Happens if Missed: Off-spec concentrate triggers smelter penalties, contract renegotiation, or rejection at the port if the deviation is severe.
Formula: Concentrate Grade (%) = (Mass of Target Mineral in Concentrate) / (Total Concentrate Mass) × 100
Indicator Type: Current. Online XRF and similar analyzers make this actionable within minutes rather than the next morning’s lab report.
Unit of Measure: Percent
Ideal Visualization(s): KPI block with live target band, KPI trend with real-time alerts on upper and lower limits, bullet chart against contract spec
Frequency: Real-time on analyzer, every 1-2 hours when sampled
Pro Tip: Plot grade against recovery on an XY scatter for the last seven days. The cloud shows you the operating envelope and the outliers tell you where the upset windows are.
Red Flag: A stable grade average masking widening shift-to-shift variability. Sensor drift, sampler plugging, or a feed mineralogy shift is usually starting underneath it.
Tailings Grade
Why it Matters: Every gram in the tails is lost revenue. Tailings grade is the cleanest signal that recovery is slipping before the daily reconciliation is done.
What it Measures: The residual grade of the target mineral in the final tailings stream, before discharge to the tailings storage facility.
What Happens if Missed: Persistent losses go unnoticed until month-end reconciliation, by which time the metal is unrecoverable and budget has eroded.
Formula: Tailings Grade (%) = (Mass of Target Mineral in Tails) / (Total Tails Mass) × 100
Indicator Type: Leading for recovery. A creeping tails grade almost always shows up before headline recovery moves.
Unit of Measure: Percent, or grams per tonne for precious metals
Ideal Visualization(s): KPI trend with real-time alerts on threshold breach, SPC trend (control chart), Pareto chart when ranking final tails streams across multiple banks
Frequency: Real-time on analyzer, hourly when sampled
Data Required: Tailings analyzer reading, tails sample assay, head grade for reconciliation
Pro Tip: Set the alert threshold off historical variability, not a fixed number. A 0.05% creep in copper tails over a week is worth more than a single spike that resets in an hour.
Mill Throughput
Why it Matters: Throughput sets the upper ceiling on revenue every shift. Every hour below nameplate is unrecoverable production.
What it Measures: Dry tonnes of ore processed through the comminution circuit per unit time, typically measured at SAG or ball mill feed.
What Happens if Missed: The plant misses monthly production targets, fixed costs spread over fewer tonnes, and downstream stockpiles run thin.
Indicator Type: Current. Throughput is the live heartbeat of the plant and the constraint on almost every other beneficiation KPI.
Unit of Measure: Tonnes per hour (t/h)
Ideal Visualization(s): KPI block for instantaneous t/h, KPI trend with real-time alerts on dips below target, bullet chart against shift target, Pareto chart when comparing parallel grinding lines
Frequency: Real-time
Data Required: Belt weigh scale reading, moisture content, mill power draw, mill operating status
Pro Tip: Trend throughput alongside mill power and bearing pressure on a single time axis. The three together tell you whether you are feed-limited, power-limited, or grinding-limited.
Red Flag: Throughput holding while power and pressure climb. The mill is overfilling and you are minutes from a grindout or overload trip.
Grinding Circuit P80
Why it Matters: P80 sets ore liberation, which caps downstream recovery. Get the grind wrong and nothing in flotation can recover the loss.
What it Measures: The particle size at which 80% of the cyclone overflow stream passes, indicating the fineness of the grind feeding flotation.
What Happens if Missed: Coarse particles lock up valuable minerals and report straight to tailings; over-grinding wastes energy and creates slimes that hurt flotation.
Formula: N/A
Indicator Type: Leading. P80 deviations show up in flotation recovery within one to two residence times.
Unit of Measure: Microns (µm)
Ideal Visualization(s): KPI block for current P80, KPI trend with real-time alerts on band breach, histogram of size distribution across the day
Frequency: Real-time with online particle size monitor, hourly with manual screening
Data Required: Online particle size measurement, cyclone overflow flow, cyclone feed pressure, feed hardness index
Pro Tip: Correlate P80 against recovery on a rolling 24-hour window. The sweet spot is usually narrower than the operating procedure assumes.
Specific Energy Consumption (SEC)
Why it Matters: Comminution is the largest energy user on a beneficiation plant. SEC shows how efficiently power converts into liberation.
What it Measures: Energy consumed per tonne of ore milled, expressed in kilowatt-hours per dry tonne across the grinding circuit.
What Happens if Missed: Power costs creep upward, carbon reporting suffers, and you lose the early signal of a worn liner or an inefficient grind.
Formula: SEC (kWh/t) = Total Mill Power (kWh) / Throughput (t)
Indicator Type: Current. SEC is calculated continuously and trended against feed hardness and product size for context.
Unit of Measure: kWh per dry tonne (kWh/t)
Ideal Visualization(s): KPI trend with real-time alerts on upward drift, bullet chart against design SEC, Pareto chart when ranking mills by efficiency
Frequency: Real-time
Data Required: Mill motor power, throughput, ore hardness index, liner age
Pro Tip: SEC rising with stable throughput and P80 almost always means liners are worn or the media charge is off. One wasted month of high-SEC operation usually pays for the reline.
Flotation Reagent Consumption Rate
Why it Matters: Reagents are the second-largest variable cost in a flotation plant. Overdosing burns cash; underdosing burns recovery.
What it Measures: Mass of collector, frother, depressant, and modifier consumed per tonne of feed ore, by reagent type and by circuit.
What Happens if Missed: Dosing pumps drift, operators compensate by overfeeding, and the unit cost per tonne quietly inflates without anyone owning the number.
Formula: Reagent Consumption (g/t) = Total Reagent Used (g) / Tonnes of Feed Processed (t)
Indicator Type: Current. Real-time dosing combined with feed tonnage gives a live cost per tonne by reagent.
Unit of Measure: Grams per tonne (g/t)
Ideal Visualization(s): KPI block per reagent, KPI trend with real-time alerts on consumption above target, Pareto chart when ranking reagents by spend, bullet chart against target dosage
Frequency: Real-time
Data Required: Reagent flow meter readings, tank levels, feed tonnage, reagent unit cost
Pro Tip: Always trend reagent consumption alongside recovery and head grade. Reagent decisions made in isolation usually cost more than they save.
Red Flag: Collector dose rising while recovery stays flat. Either mineralogy has shifted, the reagent is degrading in storage, or someone has manually overridden the dose loop.
Flotation Cell Froth Depth and Pulp Level
Why it Matters: Froth depth controls the trade-off between grade and recovery in every flotation cell. A wandering level is a wandering profit margin.
What it Measures: Distance from the cell lip to the pulp froth interface in each flotation cell, measured continuously by float, ultrasonic, or radar sensor.
What Happens if Missed: Cells run shallow and grade tanks; cells run deep and recovery falls; cleaner circuits then chase the upset for hours.
Indicator Type: Current. Level control is the most-tuned and most-disturbed control loop in flotation.
Unit of Measure: Millimetres (mm) or centimetres (cm)
Ideal Visualization(s): KPI trend with real-time alerts on level deviation, status history trends for control loop mode, bullet chart against setpoint, Pareto chart when comparing cells by deviation
Frequency: Real-time
Data Required: Level sensor reading, control valve position, controller setpoint, control loop mode (auto/manual)
Pro Tip: A cell sitting in manual for more than an hour is a silent recovery loss. Track time-in-auto by cell and by bank as a daily operations metric, not just an engineering one.
Cyclone Feed Slurry Density
Why it Matters: Slurry density at the cyclone feed controls classification cut size, which controls grinding load and downstream P80. Wrong density wastes both.
What it Measures: Mass of solids per unit volume of slurry entering the cyclone cluster, typically expressed as percent solids by weight.
What Happens if Missed: Cyclones bypass coarse particles to overflow, ball mill recirculating load swings, and P80 oscillates outside the recovery window.
Formula: Percent Solids (%) = (Mass of Solids) / (Mass of Solids + Mass of Water) × 100
Indicator Type: Leading. Density swings show up in P80, mill load, and ultimately recovery within minutes.
Unit of Measure: Percent solids by weight, or specific gravity (g/mL)
Ideal Visualization(s): KPI trend with real-time alerts on density excursion, SPC trend (control chart), bullet chart against target band
Frequency: Real-time
Data Required: Density gauge reading (nuclear or Coriolis), cyclone feed flow, sump level, water addition flow
Pro Tip: When sump pumps start cavitating or cycling, density is almost always the upstream cause. Trend pump speed and density together to catch it before a trip.
Plant Availability (Beneficiation Circuit)
Why it Matters: Every hour the plant is down is an hour of fixed costs with zero revenue. Availability sets the realistic ceiling on monthly throughput.
What it Measures: Percentage of scheduled production time the beneficiation circuit is actually available to process ore, by major equipment and overall plant.
What Happens if Missed: Unplanned stoppages cascade, maintenance windows compress, and the monthly plan quietly drifts out of reach by mid-month.
Indicator Type: Lagging, at plant level. Current, when measured continuously by asset. Use both views together.
Unit of Measure: Percent
Ideal Visualization(s): KPI block per major asset, KPI trend with real-time alerts on availability dip, Pareto chart when ranking equipment by downtime hours, status history trends for run/idle/fault states
Frequency: Real-time
Data Required: Equipment run/idle/fault status, downtime reason codes, scheduled operating hours, planned maintenance windows
Pro Tip: Pareto the last 30 days of downtime by reason code, not by duration. The repeating short stoppages almost always cost more than the one big event everyone remembers.
Red Flag: Availability holding flat while short-stop frequency climbs. The plant is being held together by operator intervention, and the next big trip is already on the calendar.
Why Real-Time Visibility Matters
Beneficiation plants make money in minutes and lose it in hours. A cyclone density excursion at 2am, a froth level drifting in cleaner cells, a collector pump quietly under-dosing on the rougher bank: none of these wait for the morning meeting. The plants that hit plan are the ones where the operations leader, the metallurgist, and the control room are all looking at the same live numbers and acting on them inside the same shift.
Running blind is expensive in this corridor of the operation. Recovery losses do not arrive as alarms; they show up in month-end reconciliations, in smelter penalties, and in stretched maintenance budgets. Real-time KPIs, trended against limits and pushed to the right person on the right device, turn a beneficiation plant from a reactive cost centre into a predictable revenue engine.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
Hydrogen storage vessels and interconnecting headers stage product for offtake. This is where purity excursions and small vent losses quietly compound into lost margin before monthly reconciliation catches them.
Hydrogen production is mostly a fight between conversion efficiency and the cost of every electron or molecule of methane walking in the door. The margin is thin, the spec is tight, and the regulatory ground is moving.
When a hydrogen plant runs blind, the losses don’t announce themselves. They show up as a quarter point of SEC drift, a creeping impurity at the product header, or a vent valve cycling 40 seconds too often per shift. None of it triggers an alarm. All of it compounds. The KPIs below are the ones a hydrogen operations leader needs in real time, alerted, and trended, before drift turns into lost product.
Specific Energy Consumption (SEC)
Why it Matters: SEC is the single biggest driver of hydrogen cost. A 5% drift translates directly into millions per year on any commercial-scale plant.
What it Measures: Total electrical or thermal energy consumed per kilogram of hydrogen produced, including auxiliaries and parasitic loads.
What Happens if Missed: Margin erosion hides inside normal operations. By the time monthly accounting flags it, you’ve already lost the quarter.
Formula: SEC = Total Energy Input (kWh) / H2 Produced (kg)
Indicator Type: Current. Reflects the live efficiency of the conversion process and plant balance.
Unit of Measure: kWh/kg H2
Ideal Visualization(s): KPI block with target band, KPI trend with real-time alerts, bullet chart against design SEC
Frequency: Real-time, with rolling hourly and daily averages
Data Required: Total power draw, thermal energy input, hydrogen mass flow, ambient temperature, load setpoint
Red Flag: SEC creeping up at constant load and constant feedstock quality. That’s degradation or fouling, not weather.
Hydrogen Production Rate vs. Nameplate
Why it Matters: Production rate determines whether you meet offtake commitments and recover fixed costs. Below 85% of nameplate puts contracts and economics at risk.
What it Measures: Real-time hydrogen output expressed as a percentage of design capacity at current operating conditions and ambient.
What Happens if Missed: Offtake shortfalls trigger penalty clauses. Repeated underproduction forces the operator to renegotiate or buy hydrogen on the spot market to cover.
Indicator Type: Current. Shows whether the plant is delivering against design at this moment.
Unit of Measure: kg/hr and % of nameplate
Ideal Visualization(s): KPI block, KPI trend with real-time alerts, Pareto chart when ranking trains or modules by production deviation
Frequency: Real-time
Data Required: Hydrogen mass flow, design capacity, load setpoint, feedstock availability
Pro Tip: Always compare actual rate to load setpoint, not just nameplate. A plant running at 60% setpoint isn’t underperforming, but a plant at 100% setpoint hitting only 92% output is.
Hydrogen Purity
Why it Matters: Purity defines salability. Fuel-cell grade requires 99.97% H2 with strict limits on CO, H2O, and sulfur, and off-spec product gets vented or downgraded.
What it Measures: Hydrogen concentration at the product header, plus trace impurity levels for the key contaminants of your chosen sales spec.
What Happens if Missed: A purity excursion contaminates downstream storage, damages customer equipment, and can force a full plant blowdown to recover.
Formula: Purity % = (H2 vol / Total Product Gas vol) × 100, with separate ppm tracking for O2, H2O, CO, N2, S
Indicator Type: Current. Quality measurement at the point of sale or compression.
Unit of Measure: vol% H2 and ppm per impurity
Ideal Visualization(s): KPI block per impurity, KPI trend with real-time alerts, SPC trend (control chart) for each impurity stream
Frequency: Real-time from online analyzers, with periodic lab validation
Data Required: Product hydrogen analyzer reading, dryer dew point, deoxo bed temperature, PSA tail gas composition
Red Flag: A slow rise in any single impurity at constant production. That’s a bed, membrane, or analyzer drift, not noise.
Cell or Tube Outlier Spread
Why it Matters: Hydrogen plants run hundreds of cells or tubes in parallel. One bad unit drags efficiency and accelerates the failure of its neighbors.
What it Measures: Statistical spread of cell voltage across an electrolyzer stack, or tube skin temperature across an SMR furnace, against the fleet median.
What Happens if Missed: A failing cell or hot tube can cascade into stack damage, tube rupture, or an unplanned shutdown that costs days of production.
Formula: Spread = |Individual Reading – Fleet Median| / Fleet Median, with outliers flagged above defined sigma thresholds
Indicator Type: Leading. A widening spread shows up well before a hard alarm fires.
Unit of Measure: Volts (electrolysis) or °C (SMR), and standard deviations from median
Ideal Visualization(s): Box plot of all cells or tubes, bullet chart per outlier, Pareto chart when ranking units by deviation
Frequency: Real-time
Data Required: Per-cell voltage or per-tube skin temperature, fleet median, sigma thresholds, age and last replacement date of each unit
Pro Tip: Rank outliers by deviation and by rate of change together. A unit drifting fast is more urgent than one that’s been bad and stable for months.
Stack or Catalyst Degradation Rate
Why it Matters: Degradation sets stack replacement timing, catalyst change-out schedules, and the slope of your future SEC. It defines lifetime economics.
What it Measures: Rate of irreversible performance loss in the conversion unit, normalized to constant load and ambient conditions.
What Happens if Missed: You over-run a catalyst bed or stack into severe degradation and incur an unplanned outage, often at the worst possible moment.
Formula: Degradation Rate = ΔPerformance Metric / Operating Hours (e.g., μV/hr for stacks, ΔApproach to Equilibrium per khr for catalyst)
Indicator Type: Lagging. Real performance loss is only visible across hundreds of hours of operation.
Unit of Measure: μV/hr, %/khr, or ΔT approach per thousand hours
Ideal Visualization(s): KPI trend with real-time alerts, SPC trend (control chart) on normalized performance, bar chart of degradation rate per stack or bed
Frequency: Calculated daily, reviewed weekly
Data Required: Normalized voltage or catalyst activity, hours in service, load history, feedstock quality, ambient conditions
Red Flag: A step change in degradation rate after a startup, trip, or feedstock change. That’s damage, not normal wear.
Compressor Specific Power
Why it Matters: Compression can consume 10 to 15% of total plant energy. Drifting compressor efficiency adds directly to delivered hydrogen cost.
What it Measures: Power consumed per kilogram of hydrogen compressed, segmented by stage and across the discharge pressure profile.
What Happens if Missed: Worn seals, fouled coolers, and recycle drift quietly eat margin until trip alarms force attention. By then, repairs are expensive.
Formula: Compressor Specific Power = Compressor Power Draw (kW) / H2 Mass Flow Compressed (kg/hr)
Indicator Type: Current. Reflects the live mechanical and thermal health of the compression train.
Unit of Measure: kWh/kg H2 compressed
Ideal Visualization(s): KPI block per stage, KPI trend with real-time alerts, Pareto chart when ranking stages or trains by deviation from design
Frequency: Real-time
Data Required: Compressor power draw per stage, suction and discharge pressure, suction and discharge temperature, hydrogen mass flow, recycle flow
Pro Tip: Plot specific power against discharge pressure. A vertical drift at constant pressure is internal wear. A horizontal drift is a system or cooling problem.
Feedstock-to-Hydrogen Conversion Efficiency
Why it Matters: Conversion efficiency shows how much of your methane or water actually leaves as product hydrogen. It’s the second-biggest cost lever after energy.
What it Measures: Hydrogen produced per unit of feedstock consumed, expressed as a percentage of the thermodynamic maximum for the process.
What Happens if Missed: Feedstock slip, side reactions, and methane breakthrough drive up consumption without showing up in production rate until invoices arrive.
Indicator Type: Current. Live view of how well the conversion chemistry is performing.
Unit of Measure: % of theoretical maximum
Ideal Visualization(s): KPI block, KPI trend with real-time alerts, bullet chart against design conversion
Frequency: Real-time
Data Required: Hydrogen mass flow, feedstock mass flow, feedstock composition, reformer or stack temperature, steam-to-carbon ratio (SMR), current density (electrolysis)
Red Flag: Conversion dropping while feedstock flow rises. The plant is hiding lost throughput inside higher feed consumption.
Hydrogen Loss Rate (Vent and Flare)
Why it Matters: Every kilogram vented or flared is hydrogen you paid to produce and didn’t sell. Losses above 2% wipe out efficiency gains elsewhere.
What it Measures: Hydrogen mass routed to vent stacks, flares, or fuel headers, expressed as a percentage of total production.
What Happens if Missed: Sustained losses get normalized as part of operations. They never appear on operator screens, but they show up clearly in monthly reconciliation.
Formula: Loss Rate % = (H2 Vented + H2 Flared + H2 to Fuel) / Total H2 Produced × 100
Indicator Type: Current. Live view of where product is escaping the system.
Unit of Measure: % of production, and kg/hr lost
Ideal Visualization(s): KPI block per loss path, KPI trend with real-time alerts, Pareto chart when ranking loss sources by mass
Frequency: Real-time
Data Required: Vent stack flow, flare flow with H2 fraction, fuel header H2 content, PSA tail gas mass flow, total H2 production
Pro Tip: Tag every vent and flare event with its upstream cause. Most “unexplained” losses are repeat patterns from startups, trips, or pressure swings.
Carbon Intensity per kg H2
Why it Matters: Carbon intensity defines hydrogen color and unlocks subsidies, tax credits, and offtake premiums. A 1 kg CO2e shift can move a plant across regulatory thresholds.
What it Measures: Total scope 1 and scope 2 greenhouse gas emissions per kilogram of hydrogen produced, calculated in real time.
What Happens if Missed: Plants exceed certification thresholds without knowing it, lose access to incentive markets, and rebuild offtake terms after the fact.
Indicator Type: Current. Tracks regulatory and commercial standing as it happens.
Unit of Measure: kg CO2e per kg H2
Ideal Visualization(s): KPI block, KPI trend with real-time alerts, bullet chart against certification threshold, SPC trend (control chart) for variance
Frequency: Real-time, with daily and monthly rollups for reporting
Data Required: Fuel gas consumption, grid electricity draw, on-site renewable generation, carbon capture rate, fugitive emissions, hydrogen production
Red Flag: A sustained jump after a feedstock or grid mix change. Most plants only catch this in monthly disclosures, far too late.
Plant On-Stream Factor
Why it Matters: On-stream factor determines annual production, contract delivery, and capital recovery. A 1% drop on a 30,000 tpa plant costs roughly 300 tonnes of hydrogen per year.
What it Measures: Hours of hydrogen production at or above minimum spec, divided by total available hours in the period.
What Happens if Missed: Trip patterns, slow restarts, and short outages compound. Annual availability misses the budget by a margin that’s almost impossible to recover.
Formula: On-Stream Factor = (Hours at or above Spec / Total Period Hours) × 100
Indicator Type: Lagging. Reflects accumulated uptime over a defined period.
Unit of Measure: % of period
Ideal Visualization(s): Status history trend, KPI block with monthly and yearly rollup, Pareto chart when ranking outage causes by lost hours
Frequency: Hourly updates, daily and monthly rollups
Data Required: Production state per minute, spec compliance flags, trip events, planned maintenance windows
Pro Tip: Separate planned outage time from unplanned. Unplanned hours are where operations leaders actually have leverage.
Why Real-Time Visibility Matters
Hydrogen plants don’t fail loudly. They fail slowly, through a quarter point of SEC drift, a creeping impurity, a flare event that’s become routine. The cost of any single moment looks small. The cost across a quarter does not. By the time monthly accounting closes the books, the operations leader is reading history, not running the plant.
The hydrogen plants that hit their certification thresholds, meet their offtake contracts, and stay profitable across the next round of energy-price moves are the ones that catch drift in hours, not weeks. Every KPI on this list is a leading or current indicator that, properly alerted and trended, gives the team enough time to act before the loss is locked in.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
Crushed product moves from screens to stockpiles across a web of conveyors and transfer points, where every belt is a live readout of throughput, particle size, and circuit balance. When these feeds drift out of spec, downstream recovery and grade pay the price within the next shift.
A mineral processing plant has one job: separate value from waste every second of every shift. When the circuit drifts and nobody sees it in real time, recovery slips, reagent bills balloon, and concentrate grade walks out the door with the tailings. Most plants don’t lose money in one big event.
They lose it in 2% recovery drops, 30-minute mill stalls, and grind size creeping coarser while everyone watches lagging shift reports. By the time the daily metallurgical balance lands, the metal is already gone. These ten KPIs are the ones a mineral processing operations leader should see updating live, with alerts firing on the deviations that actually matter.
Mill Throughput
Why it Matters: Throughput drives revenue. Every tonne not milled is a tonne of unrecovered metal sitting in the stockpile or the pit.
What it Measures: Tonnes of dry ore processed per hour through the grinding circuit, measured at the mill feed weightometer.
What Happens if Missed: Hidden bottlenecks compound. Operators chase phantom issues while the real constraint, often a single cyclone or pump, stays invisible.
Indicator Type: Current. It reflects live grinding circuit performance against design capacity.
Unit of Measure: Tonnes per hour (t/h)
Ideal Visualization(s): KPI blocks for current throughput, KPI trend with real-time alerts on sustained dips, Pareto chart when ranking mill lines by throughput deviation
Frequency: Real-time, 1-minute updates
Data Required: Mill feed mass flow, ore moisture content, mill operating status
Metallurgical Recovery
Why it Matters: Recovery is the difference between a profitable plant and a tailings dam full of money. Even 1% lost across a year is millions.
What it Measures: Percentage of valuable mineral in the feed that reports to final concentrate, calculated by the two-product or three-product formula.
What Happens if Missed: Recovery losses compound silently across shifts. Without live visibility, root cause hunts start days after the metal is already in the tailings.
Formula: Recovery (%) = (C × c) / (F × f) × 100
Indicator Type: Current. Modeled from online analyzers it updates live; based on shift assays it lags by hours.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI blocks for live recovery, KPI trend with real-time alerts on recovery drops, SPC trend for shift-over-shift variability
Frequency: Real-time when modeled from online analyzers, otherwise every shift
Data Required: Feed grade, concentrate grade, tailings grade, feed mass flow, concentrate mass flow
Concentrate Grade
Why it Matters: Grade determines smelter payable terms and freight efficiency. Low-grade concentrate gets penalized at the smelter and costs more to ship per tonne of metal.
What it Measures: Percentage by mass of the valuable mineral or metal in the final concentrate stream.
What Happens if Missed: Operators chase recovery and let grade slip, or chase grade and bleed recovery. Without live grade visibility, the trade-off is guesswork.
Formula: Grade (%) = mass valuable mineral / total concentrate mass × 100
Indicator Type: Current. Online analyzers deliver live readings; shift composites lag the operating decision by hours.
Unit of Measure: Percent (%) or grams per tonne (g/t) depending on commodity
Ideal Visualization(s): KPI blocks for current grade, KPI trend with real-time alerts on grade slippage, XY scatter plot for grade versus recovery trade-off
Frequency: Real-time with online analyzer, every 30 to 60 minutes with composite sampling
Data Required: Online analyzer readings, concentrate mass flow, feed grade for reconciliation
Pro Tip: The grade-recovery curve isn’t static. Watch how it shifts with feed mineralogy, reagent dose, and froth depth. The operating point should move with the ore.
Grind Size (P80)
Why it Matters: Grind size sets the ceiling on recovery. Too coarse and minerals stay locked, too fine and energy and reagents are wasted on overgrind.
What it Measures: The particle size at which 80% of the cyclone overflow passes, measured by online particle size analyzer.
What Happens if Missed: P80 drifts coarser as liners wear or feed hardens. Recovery quietly drops for hours before assay results catch up.
Formula: N/A
Indicator Type: Leading. Grind size changes show up in recovery several minutes to hours later.
Unit of Measure: Microns (µm)
Ideal Visualization(s): KPI blocks for live P80, KPI trend with real-time alerts on coarsening, histogram of the full size distribution
Frequency: Real-time, 1 to 5 minute updates from the particle size analyzer
Indicator Type: Current. It moves with load, liner condition, and ore competence in real time.
Unit of Measure: Kilowatt-hours per tonne (kWh/t)
Ideal Visualization(s): KPI blocks for live SEC, KPI trend with real-time alerts on sustained rises, Pareto chart when ranking mills by SEC deviation from baseline
Frequency: Real-time, 1-minute updates
Data Required: Mill motor power draw, mill feed tonnage, mill operating status
Pro Tip: Benchmark SEC against ore Bond Work Index, not against a static target. The ratio of actual to theoretical SEC tells you how efficiently the mill is running today.
Reagent Consumption per Tonne
Why it Matters: Flotation reagents are one of the largest variable operating costs. Overdosing is wasted money, underdosing is lost metal.
What it Measures: Mass of each reagent (collector, frother, depressant, modifier) consumed per tonne of ore processed.
What Happens if Missed: Operators dose to a fixed setpoint regardless of ore type. Recovery drifts and reagent costs creep with nobody connecting the two.
Indicator Type: Current. Live dosing rates show whether circuits are over-fed or starved against present ore conditions.
Unit of Measure: Grams or kilograms per tonne (g/t or kg/t)
Ideal Visualization(s): KPI blocks per reagent, KPI trend with real-time alerts on dose deviation, Pareto chart when ranking reagents by cost contribution
Frequency: Real-time, every 1 to 5 minutes
Data Required: Reagent flow rates, reagent density, feed tonnage, ore type indicator
Red Flag: Reagent dose climbing while recovery stays flat. That signals the circuit isn’t responding and there’s a deeper chemistry or feed mineralogy issue.
Cyclone Overflow Percent Solids
Why it Matters: Classification sets up flotation. Wrong overflow density skews the particle size reporting to flotation and degrades both grade and recovery downstream.
What it Measures: Mass percentage of solids in the hydrocyclone overflow stream feeding the flotation circuit.
What Happens if Missed: A drifting nuclear density gauge or a worn apex masks a classification problem for hours. Flotation gets blamed for grinding’s mistakes.
Formula: % Solids = mass solids / total slurry mass × 100
Indicator Type: Current. It reflects the live operating state of the classification circuit.
Unit of Measure: Percent solids by mass (%)
Ideal Visualization(s): KPI blocks for current % solids, KPI trend with real-time alerts on deviation, Pareto chart when ranking cyclones by deviation from target
Frequency: Real-time, 1-minute updates
Data Required: Cyclone overflow density, slurry flow rate, water addition rate, cyclone feed pressure
Pro Tip: Pair this with cyclone feed pressure. Pressure drifting up with stable flow usually means apex wear or roping, both of which kill classification.
Plant Availability
Why it Matters: Availability is a direct multiplier on annual production. An hour lost on the SAG mill is rarely recovered, since downstream circuits can’t run dry.
What it Measures: Percentage of scheduled time the plant or a critical asset is actually available to process ore.
What Happens if Missed: Small repeating stoppages get coded as “minor” and never escalate. A 20-minute trip three times a shift is an hour of lost production nobody is accounting for.
Formula: Availability (%) = Operating Time / Scheduled Time × 100
Indicator Type: Lagging. It summarizes recent runtime against schedule.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI blocks per circuit, status history trends for outage causes, Pareto chart when ranking downtime events by cause and duration
Frequency: Real-time status, with hourly and shift rollups
Data Required: Equipment run status, stoppage reason codes, scheduled operating hours
Red Flag: Availability above 95% with throughput below target. The plant’s running but throttled, which is a hidden bottleneck, not an uptime win.
Tailings Slurry Density and Water Recovery
Why it Matters: Water is a regulated, cost-laden resource. Thicker tailings mean more recycled water, lower pumping costs, and safer dam operation.
What it Measures: Density of thickener underflow and the percentage of process water recovered for reuse back into the plant.
What Happens if Missed: Thickener upsets push thin tailings to the dam, raise freshwater intake, and quietly inflate water balance reports until regulators or auditors notice.
Formula: Water Recovery (%) = recycled water volume / total process water consumed × 100
Indicator Type: Current. Density and recovery move with flocculant dose, bed level, and feed solids in real time.
Unit of Measure: Percent solids (%) for density, percent (%) for water recovery
Ideal Visualization(s): KPI blocks for live density and recovery, KPI trend with real-time alerts on thickener bed level, SPC trend for stability
Frequency: Real-time, 1-minute updates
Data Required: Thickener underflow density, overflow turbidity, flocculant dose rate, bed level, recycled water flow
Pro Tip: Watch flocculant consumption per tonne of solids. A rising flocculant dose with stable feed is the first sign of mineralogy or pH changes hitting settling rates.
Mass Pull
Why it Matters: Mass pull links grinding, flotation, and the grade-recovery curve. It’s a fast proxy for circuit response when assays lag.
What it Measures: Percentage of feed mass reporting to concentrate, measured by mass flow before assay results are available.
What Happens if Missed: Mass pull drifts up and operators read it as recovery success when it’s really dilution. Concentrate grade quietly drops behind it.
Formula: Mass Pull (%) = concentrate mass flow / feed mass flow × 100
Indicator Type: Leading. Mass pull moves minutes ahead of any grade or recovery confirmation.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI blocks for current mass pull, KPI trend with real-time alerts on rapid shifts, XY scatter plot for mass pull versus grade
Frequency: Real-time, 1-minute updates
Data Required: Concentrate mass flow, feed mass flow, slurry densities
Red Flag: Mass pull climbing while concentrate grade falls. That’s froth carrying gangue, often from over-aerated cells or too-shallow froth depth.
Why Real-Time Visibility Matters
Mineral processing operates on minute-by-minute trade-offs. A grind size drift, a reagent overdose, a thickener bed creeping up: none of these show up cleanly in a shift report, but each one moves recovery, grade, and operating cost in real time. When the visibility window is 12 hours wide, every shift starts behind, reacting to yesterday’s metal balance instead of today’s circuit.
The plants that pull ahead aren’t the ones with the newest equipment. They’re the ones where the control room, the metallurgist, and the maintenance lead all see the same live numbers, with alerts firing on the deviations that actually matter. Real-time KPIs turn a reactive plant into one that catches problems while the ore is still in the circuit.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
A mixed-aperture ground station at altitude, with each dish carrying its own pointing budget, link margin, and AOS clock. Real-time visibility is what keeps them aligned during the pass, not in the postmortem after.
A ground station network earns its keep in the ten minutes a satellite is overhead. Everything else is preparation or postmortem. When a pass starts going sideways, the warning signs are short. Pointing drifts before lock is lost. Link margin erodes before bit errors spike. An HPA twitches off setpoint before it trips. Teams that wait for the postmortem are already working from data that’s a customer escalation.
These are the ten KPIs that decide whether a ground station network runs ahead of trouble or behind it.
Pass Success Rate
Why it Matters: Every missed pass is unrecoverable revenue and a customer escalation. This number is what your largest accounts measure you against.
What it Measures: The percentage of scheduled passes that delivered all telemetry, command, and mission data within contracted quality thresholds.
What Happens if Missed: SLA penalties stack, customers re-evaluate provider mix, and your network reputation takes hits that take quarters to repair.
Formula: (Successful Passes / Total Scheduled Passes) × 100
Indicator Type: Lagging. Reflects what already happened, but trended over rolling windows it exposes systemic degradation early.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI trend with real-time alerts, Pareto chart when ranking ground stations by failed-pass count
Frequency: Every pass, rolled up hourly and daily
Data Required: Pass schedule, pass completion status, data quality flags, customer acceptance status
Pro Tip: Slice by satellite operator, antenna, and time of day. The averages always look fine; the slices tell the truth.
Red Flag: A station hitting target on rate but failing a single customer disproportionately. Aggregate KPIs hide concentrated SLA risk.
Link Margin (Eb/No)
Why it Matters: Link margin is your headroom against rain, interference, and pointing drift. When it collapses, the data goes with it.
What it Measures: The difference between received energy-per-bit-to-noise ratio and the threshold required for the modulation and coding scheme in use.
What Happens if Missed: BER climbs, retransmits multiply, and high-rate downlinks degrade or drop entirely. Mission data arrives late or corrupted.
Formula: Eb/No_measured – Eb/No_required (in dB)
Indicator Type: Current. A real-time view of how much link you have to spare before quality collapses.
Unit of Measure: Decibels (dB)
Ideal Visualization(s): KPI trend with real-time alerts, Bullet chart against required threshold
Frequency: Real-time, sub-second sampling
Data Required: Received Eb/No, modulation scheme, coding rate, required Eb/No for current configuration
Pro Tip: Plot margin against elevation angle for each pass. Sudden margin dips at high elevation usually mean the antenna, not the weather.
Red Flag: Margin trending down across multiple satellites and stations simultaneously. That’s not weather, that’s a calibration or reference drift.
Antenna Pointing Error
Why it Matters: A few millidegrees of pointing error at S-band is annoying. At Ka-band it costs you the pass.
What it Measures: The angular distance between commanded antenna position and actual position during active tracking, in both azimuth and elevation.
What Happens if Missed: Link margin erodes, autotrack falls off the satellite, and high-frequency missions become unworkable in marginal conditions.
Indicator Type: Lagging. Reports actual delivered capacity, and the breakdown by cause is what drives the maintenance plan.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI trend with real-time alerts, Pareto chart when ranking antennas by downtime cause, Status history trends
Frequency: Continuous, rolled up daily and weekly
Data Required: Scheduled antenna time, actual ready time, fault state, maintenance windows
Pro Tip: Track planned versus unplanned downtime separately. The blended number always looks acceptable; only the unplanned slice exposes real reliability.
Red Flag: Same antennas appearing repeatedly in unplanned downtime. Concentration in one or two assets points to a systemic fault no one is owning.
Schedule Utilization Rate
Why it Matters: Antennas are capital that depreciates whether they’re tracking or sitting idle. Empty minutes are money you’ve already spent.
What it Measures: The percentage of available antenna minutes booked against demonstrated capacity, after maintenance and reserve windows are excluded.
What Happens if Missed: Capacity planning loses credibility, new customer onboarding stalls, and the network grows in the wrong places at the wrong time.
Formula: (Booked Minutes / Available Minutes) × 100
Indicator Type: Current. Tells you in near real-time how full the network is and where there’s headroom to sell.
Unit of Measure: Percent (%)
Ideal Visualization(s): KPI trend with real-time alerts, Bullet chart, Pareto chart when ranking stations by utilization
Frequency: Hourly, with daily and weekly rollups
Data Required: Schedule database, antenna availability, maintenance calendar, contractual reservations
Pro Tip: Watch the gap between booked and executed minutes. Cancellations and reschedules quietly burn capacity that the booking number says is full.
Red Flag: High utilization with rising last-minute cancellations. The schedule looks full, but executed minutes are quietly draining and capacity planning is wrong.
Doppler Tracking Error
Why it Matters: LEO satellites move at orbital velocity. Bad Doppler tracking means lost lock and a useless pass.
What it Measures: The frequency offset between the predicted Doppler-compensated carrier and the actual received carrier during a tracked pass.
What Happens if Missed: Modems lose lock mid-pass, telemetry frames drop, and high-rate downlinks fail to close even when geometry is favorable.
Formula: |f_received – f_predicted|
Indicator Type: Leading. Tracking error climbs before lock is lost, giving you a short window to react.
Unit of Measure: Hertz (Hz)
Ideal Visualization(s): KPI trend with real-time alerts, XY scatter plot of error versus elevation, SPC trend
Frequency: Real-time, every second during pass
Data Required: Predicted Doppler, measured received frequency, reference oscillator status, modem lock status
Pro Tip: Compare Doppler error across passes for the same satellite. Persistent bias points to your reference, not the satellite.
Red Flag: Doppler error spikes only on specific satellites. Often your TLE source for that constellation is stale or low-priority in the pipeline.
HPA Forward Power Stability
Why it Matters: Unstable HPA output kills uplink quality and risks the amplifier itself. Both outcomes are expensive and visible.
What it Measures: The deviation of high-power amplifier forward output power from setpoint during active uplink, over a defined window.
What Happens if Missed: Uplink BER degrades, command rejections rise, and at the extreme you damage the HPA or the feed.
Pro Tip: Pair forward power with reflected power on one view. Swings without VSWR change usually mean the drive chain, not the antenna.
Red Flag: Forward power steady but reflected power climbing. That’s the antenna or waveguide telling you something cracked, iced, or shifted.
Why Real-Time Visibility Matters
Ground station operations live or die on what you can see during the pass, not after. A late AOS, a millidegree of pointing drift, an HPA twitching off setpoint: each one has minutes of warning before it costs you a customer minute. The teams that catch them in time are the ones with eyes on every link, every antenna, and every schedule slot, at the rate the satellites actually move.
The cost of operating blind isn’t theoretical. It’s a missed weather observation, a delayed Earth-observation tasking, a constellation operator looking at downtime credits and wondering whether your network is still the right one. The KPIs above are what turn a ground station from a reactive bottleneck into a predictable, sellable piece of capacity.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
The gates and buttresses that gate every dispatch decision. In pumped hydro, a few centimeters of reservoir level is the difference between a profitable cycle and a missed market window.
Pumped hydro storage earns its margin in the gap between a market signal and a synchronized machine, and that gap closes fast when a unit refuses to swing into generate mode. Every second of delay is arbitrage revenue or ancillary services revenue walking out the door.
The expensive failures here are quiet ones. A drifting bearing temperature, an unnoticed cavitation pattern, a pressure transient nobody trended. Run a plant like this blind and you don’t lose dollars, you lose units for months while a stator gets rewound or a runner gets pulled. These ten KPIs are the ones every operations leader should have in front of them every shift.
Round-Trip Efficiency (RTE)
Why it Matters: RTE drives the core economics of the asset. A two-point drop on a 1 GWh cycle wipes out the day’s arbitrage spread.
What it Measures: Energy delivered during generation divided by energy consumed during pumping over a complete charge and discharge cycle.
What Happens if Missed: Silent efficiency loss erodes margin, and degraded runners or worn seals stay invisible until the next overhaul reveals them.
Formula: RTE = (MWh Generated / MWh Pumped) × 100
Indicator Type: Lagging. RTE confirms what already happened across a full charge and discharge cycle, but trended unit by unit it surfaces degradation early.
Unit of Measure: Percent
Ideal Visualization(s): KPI blocks, KPI trend with real-time alerts, Pareto chart when ranking units by RTE deviation, SPC trend (control chart)
Frequency: Per cycle (typically once daily, more often for fast-cycling units)
Data Required: Energy generated per cycle, energy pumped per cycle, cycle start and end timestamps
Pro Tip: Track RTE against head and load condition. A unit at part-load looks inefficient but may be operating exactly as designed.
Red Flag: A steady multi-week decline in RTE on a single unit while peers hold flat. Look at seal wear, runner cavitation, or instrument drift.
Upper Reservoir State of Charge
Why it Matters: State of charge defines how much energy you can dispatch and how much you can absorb. Get it wrong and you forfeit market position.
What it Measures: Usable storage between minimum and maximum operating levels, expressed as a percentage of total capacity.
What Happens if Missed: Bidders commit to dispatch that can’t physically be delivered, triggering market penalties and forcing emergency curtailment.
Formula: SoC = ((Current Level − Min Operating Level) / (Max Operating Level − Min Operating Level)) × 100
Indicator Type: Current. State of charge tells you exactly where the plant stands right now and what’s available for the next dispatch window.
Unit of Measure: Percent (and MWh equivalent)
Ideal Visualization(s): Bullet chart, KPI blocks, KPI trend with real-time alerts, Status history trends
Frequency: Real-time
Data Required: Upper reservoir level, lower reservoir level, reservoir geometry, current head, conversion factor to MWh
Pro Tip: Pair SoC with rolling 24-hour and 72-hour forecasts of price and inflow. Storage value depends entirely on when you plan to use it.
Red Flag: SoC drifts down between cycles without a matching pump schedule. Leakage, evaporation underestimation, or a metering error is the usual cause.
Unit Availability Factor
Why it Matters: A pump-turbine that can’t be called is worthless. Availability is what the market actually pays for through capacity and ancillary contracts.
What it Measures: Hours each unit was available for dispatch (pumping or generating) divided by total hours in the period.
What Happens if Missed: Forced derates and surprise outages erode capacity revenue and damage the operator’s standing in ancillary services markets.
Formula: Availability = (Available Hours / Period Hours) × 100
Indicator Type: Lagging. Availability is calculated after the fact, but daily and weekly views reveal patterns that predict the next forced outage.
Unit of Measure: Percent
Ideal Visualization(s): Bullet chart, KPI trend with real-time alerts, Pareto chart when ranking units by lost availability hours, Status history trends
Frequency: Hourly rollup, monthly summary
Data Required: Unit operating status, forced outage hours, planned outage hours, derate hours
Pro Tip: Track availability separately for pump mode and generate mode. Some failures only show up in one direction and a single number hides them.
Red Flag: A unit available in generate mode but repeatedly failing in pump mode. Motor starter, exciter, or dewatering system is usually the culprit.
Mode Transition Time
Why it Matters: Fast transitions unlock high-value ancillary services revenue. Every extra minute from standstill to synchronized is paid for by someone else’s faster plant.
What it Measures: Elapsed time from a dispatch signal to the unit reaching commanded MW output in the requested mode.
What Happens if Missed: Slow transitions disqualify the asset from frequency response markets and reduce its value in fast-ramping merit orders.
Formula: Transition Time = Time at Commanded Output − Time of Dispatch Signal
Indicator Type: Current. Each transition is its own event, and trending the population of transitions surfaces slow-developing problems.
Unit of Measure: Seconds (or minutes for slower units)
Ideal Visualization(s): Histogram, Box plot, KPI trend with real-time alerts, Pareto chart when ranking units by average transition time
Frequency: Per transition event
Data Required: Dispatch signal timestamp, sync timestamp, MW setpoint, actual MW output, mode flag
Pro Tip: Separate generate-to-pump and pump-to-generate populations. They share equipment but stress it in different ways and they degrade independently.
Red Flag: Mean transition time creeping up by even a few seconds per week. Guide vane servos, dewatering pumps, or breaker timing usually cause it.
Dispatch Compliance
Why it Matters: Market operators pay for the MW you bid, not the MW you deliver. Persistent under or over-delivery brings penalties and bid rejections.
What it Measures: Actual MW output versus scheduled MW output across each dispatch interval, expressed as deviation or accuracy percentage.
What Happens if Missed: Settlement charges accumulate quietly, and chronic deviation can disqualify units from future ancillary services participation.
Indicator Type: Current. Compliance is judged per dispatch interval and aggregated daily for settlement and market reporting.
Unit of Measure: Percent (and MWh deviation)
Ideal Visualization(s): KPI trend with real-time alerts, SPC trend (control chart), Bar chart, Pareto chart when ranking units by deviation magnitude
Frequency: Per dispatch interval (typically every 4 to 5 minutes)
Data Required: Scheduled MW setpoint, actual MW output, dispatch interval timestamps
Pro Tip: Tag deviations with root cause when they happen. Reconstructing why a unit underdelivered at 14:22 three weeks ago is nearly impossible.
Red Flag: Compliance trending down only during high-head or low-head extremes. The unit’s operating curve has shifted and the bid model hasn’t caught up.
Pump-Turbine Bearing Vibration
Why it Matters: Bearing vibration is the earliest warning of mechanical trouble. Catching a rising trend prevents a multi-month forced outage to replace a thrust bearing.
What it Measures: Vibration amplitude at upper guide, lower guide, and turbine guide bearings, captured continuously during operation.
What Happens if Missed: A failing bearing seizes, damages the shaft, and pulls the unit out of service during peak market hours.
Formula: Vibration Amplitude = RMS displacement or velocity per ISO 7919 / ISO 10816 standards
Indicator Type: Leading. Vibration trends climb well before a bearing fails, giving the operations team weeks of warning when the data is watched.
Unit of Measure: Micrometers (displacement) or mm/s (velocity)
Ideal Visualization(s): KPI trend with real-time alerts, SPC trend (control chart), Bullet chart, Pareto chart when ranking bearings by amplitude
Frequency: Real-time during operation
Data Required: Bearing vibration sensor readings, shaft speed, MW load, mode, head
Pro Tip: Normalize vibration against load and head. Raw amplitude looks alarming at full output and benign at part-load when the underlying condition is identical.
Red Flag: Vibration rising at the same load and head conditions over consecutive starts. The mechanical condition is changing, not the duty.
Generator and Motor Stator Temperature
Why it Matters: Stator insulation life halves for roughly every 10 degrees Celsius of sustained overheat. Hot operation today is a rewind invoice tomorrow.
What it Measures: Temperature at stator winding RTDs and core iron, monitored across all phases, slots, and cooling circuits during operation.
What Happens if Missed: Insulation degrades, a phase fault eventually develops, and the unit goes offline for a stator rewind costing months of downtime.
Formula: N/A
Indicator Type: Current. Stator temperature reflects real-time thermal balance between losses and cooling, and trends surface slow cooling-system degradation.
Unit of Measure: Degrees Celsius
Ideal Visualization(s): KPI trend with real-time alerts, Bullet chart, SPC trend (control chart), Pareto chart when ranking slots by temperature
Frequency: Real-time
Data Required: Stator RTD readings per phase and slot, cooling water inlet and outlet temperature, MW load, ambient temperature
Pro Tip: Track the spread between hottest and coldest slot, not just the peak. A widening spread reveals cooling problems before any RTD alarms.
Red Flag: One slot consistently 5 degrees hotter than its neighbors. A blocked cooling duct or a developing local fault is the usual cause.
Penstock Pressure Transients
Why it Matters: Water hammer events damage penstocks, valves, and turbine runners. A single severe transient can crack a pipe or warp a wicket gate.
What it Measures: Peak and trough pressure deviations in the penstock during load changes, mode transitions, and emergency wicket gate closures.
What Happens if Missed: Cumulative transient damage goes undetected until a leak appears or a valve seat fails, both of which take the unit offline.
Indicator Type: Leading. Transient severity correlates directly with fatigue accumulation in penstock welds and turbine components.
Unit of Measure: Bar or kPa (and percent of design pressure)
Ideal Visualization(s): KPI trend with real-time alerts, Histogram, SPC trend (control chart), Pareto chart when ranking events by severity
Frequency: Per transient event, with real-time capture
Data Required: Penstock pressure sensors at multiple stations, wicket gate position, valve position, mode flag, MW output
Pro Tip: Tie every transient back to the operator action or trip that caused it. Operator habits and control tuning are easier to fix than steel.
Red Flag: Transient peaks rising over time at the same operating sequence. Guide vane closure rate, surge tank condition, or relief valve performance has shifted.
Starts and Stops per Unit
Why it Matters: Each start consumes a measurable fraction of bearing, shaft, and stator life. Cycling assets are bought and sold on this number.
What it Measures: Cumulative count of unit starts and stops by mode and start type, tracked against design-life budgets per component.
What Happens if Missed: Starts accumulate faster than maintenance plans assume, and the next major overhaul arrives before the budget cycle does.
Formula: Cumulative Starts = Σ (mode transition events per unit per period)
Indicator Type: Lagging. The count itself is historical, but tracked against design budgets it becomes a forward-looking maintenance and capital signal.
Unit of Measure: Count (and percent of design budget consumed)
Ideal Visualization(s): Bar chart, Bullet chart, KPI blocks, Pareto chart when ranking units by cumulative starts, Group rollup bars
Frequency: Real-time event, daily rollup
Data Required: Unit status transitions, mode flag, start type (cold, warm, hot), event timestamps
Pro Tip: Weight starts by severity. A cold start does more damage than a hot restart, and treating them as equal undercounts real wear.
Red Flag: One unit absorbing a disproportionate share of starts because dispatch logic always picks it first. Rotate duty before the asset pays the bill.
Auxiliary Load Ratio
Why it Matters: Aux loads quietly eat into RTE. Small motors and pumps can consume one to three percent of gross output and rarely get attention.
What it Measures: Auxiliary power consumption as a percentage of unit gross output during generation or gross input during pumping.
What Happens if Missed: Plant-wide efficiency drifts down, and the cause hides inside dozens of small motors and pumps nobody trends individually.
Formula: Aux Load Ratio = (Auxiliary Power Consumption / Gross Unit Power) × 100
Indicator Type: Current. The ratio updates continuously during operation and aggregates cleanly per cycle for efficiency analysis.
Unit of Measure: Percent
Ideal Visualization(s): KPI trend with real-time alerts, Bullet chart, Pareto chart when ranking auxiliary loads by consumption, SPC trend (control chart)
Frequency: Real-time
Data Required: Auxiliary power meter readings, gross unit power, operating mode, ambient conditions
Pro Tip: Break aux load into cooling, lubrication, drainage, and ventilation buckets. A creeping number on one bucket is a clear path to the root cause.
Red Flag: Aux load ratio elevated only during pump mode startup. Dewatering or air injection systems are running longer than they should.
Why Real-Time Visibility Matters
Pumped hydro storage is most lucrative when it acts fast and most expensive when it acts late. A unit that takes ninety seconds longer to synchronize, a bearing that nobody noticed climbing, a transient that crossed a threshold last Tuesday, these are not small problems. They are the difference between a profitable cycle and a missed market window, between a routine overhaul and a stator rewind.
Operating these assets without continuous, alerted KPIs means trusting that nothing important happened between morning reports. In a plant where every dispatch interval is a financial event and every start consumes measurable equipment life, that trust is not supported by the math. Real-time visibility, with alerts that fire on the conditions that actually predict trouble, turns a multi-hundred-megawatt machine from a black box into something you can run on intent rather than reaction.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
Underground at the working face: a Sandvik LH625iE electric LHD loads ore beneath ventilation duct and service cabling, the choke point where airflow, fleet availability, and ore production all meet in a single heading.
In an underground mine, every meter of advance is a trade against air quality, ground stability, and water inflow you cannot see from the surface. The faces that pay the bills are also the ones most likely to surprise you.
When data takes hours to reach the operations leader, decisions get made on yesterday’s reality. A blocked airway, a stalled pump, or a creeping convergence trace becomes a shift-ender instead of a fix. These ten KPIs are the ones that should be in front of operations leaders in real time, every shift, every panel.
Ventilation Air Volume at Working Face
Why it Matters: Adequate airflow dilutes diesel particulate, gas, and dust. Without it, the face shuts down or workers get hurt.
What it Measures: Volumetric flow of fresh air reaching each active heading, compared to regulatory and design minimums.
What Happens if Missed: Workings exceed exposure limits, production halts under regulatory order, or workers face acute health events that trigger investigations.
Formula: Measured airflow at heading / Required airflow per regulation
Indicator Type: Leading. Airflow drops precede gas accumulation, so a falling trace gives operators time to act before exposure becomes a problem.
Unit of Measure: m³/s or cfm
Ideal Visualization(s): KPI blocks for active headings, KPI trend with real-time alerts, Pareto chart when ranking headings by airflow deficit
Frequency: Real-time
Data Required: Heading airflow, fan duty, regulator position, design ventilation requirement
Pro Tip: Tie airflow alerts to mobile fleet position so operations sees airflow per occupied heading, not per nominal heading.
Red Flag: A heading whose airflow drifts down across consecutive shifts without a known fan or door change. Something has shifted in the network.
Hazardous Gas Concentration at Face
Why it Matters: Methane, CO, and NO₂ spikes can ignite, intoxicate, or trigger evacuation. Real-time visibility is the difference between a controlled response and a casualty.
What it Measures: Concentration of regulated gases at active headings, return airways, and bleeders relative to action and statutory limits.
What Happens if Missed: Ignition sources meet flammable atmospheres, crews are exposed beyond TLV, or false alarms erode trust in the monitoring system.
Formula: Measured concentration / Statutory limit
Indicator Type: Current. Gas readings reflect what is happening now, but the rate of change also functions as a leading signal.
Unit of Measure: ppm or %
Ideal Visualization(s): KPI blocks per zone, KPI trend with real-time alerts, KPI Map showing concentrations by location, Pareto chart when ranking zones by exceedance frequency
Frequency: Real-time
Data Required: Sensor concentrations, sensor calibration status, fan operating state, blast schedule
Pro Tip: Alert on rate of rise, not just absolute level. A slow methane crawl matters before it crosses the action threshold.
Red Flag: Repeated CO spikes correlated with diesel equipment cycles. The fleet is undermaintained or the airway is undersized.
Development Advance Rate
Why it Matters: Development feeds future production. If meters per shift slip, the production schedule slips with them, often by months.
What it Measures: Linear advance achieved per active development heading per shift or day, against the planned advance rate.
What Happens if Missed: Stope sequences run out of prepared ground, mill feed gaps appear, and contractors fall behind on milestone payments.
Formula: Meters advanced / Planned meters per period
Indicator Type: Lagging. Advance is reported after the round is complete, but cycle-time analysis turns it into a leading signal.
Unit of Measure: m/shift or m/day
Ideal Visualization(s): KPI trend with real-time alerts, Group rollup bars by crew or contractor, Pareto chart when ranking headings by advance shortfall
Frequency: Per round or per shift
Data Required: Round length, rounds completed, scheduled meters, drilling and blasting cycle times, support installation status
Pro Tip: Track cycle component times (drill, charge, blast, muck, support) separately. Aggregate advance hides which step is bleeding hours.
Red Flag: A heading that meets advance targets but with deteriorating support installation rate. You are buying meters with future ground problems.
Ore Production Rate per Stope
Why it Matters: Stope production is the revenue line. Real-time visibility lets the operations leader rebalance the fleet before a slow stope becomes a missed shift.
What it Measures: Tonnes mucked from each active stope or production heading per shift, compared to the production plan.
What Happens if Missed: Mill feed runs short, blending strategy collapses, and idle fleet hours compound across the shift.
Formula: Tonnes mucked per stope / Planned tonnes per stope
Indicator Type: Current. Updates as bucket counts or weightometer data arrives, giving the shift control room time to redirect units.
Unit of Measure: t/shift
Ideal Visualization(s): KPI blocks per stope, KPI trend with real-time alerts, Pareto chart when ranking stopes by production deviation
Frequency: Real-time
Data Required: LHD bucket counts or payload, conveyor weightometer, ore pass levels, planned tonnes per stope
Pro Tip: Calibrate bucket factors per LHD operator. The same machine moves different tonnes in different hands, and the variance is not noise.
Red Flag: A stope hitting tonnes but trending down on grade. The crew is mucking dilution to make rate.
Ore Dilution
Why it Matters: Dilution turns paying tonnes into hauling cost. Every percent of waste in the ore stream is mill capacity, energy, and reagent burned for nothing.
What it Measures: Proportion of waste rock in the ore stream, calculated against the geological block model and reconciled with mill grade.
What Happens if Missed: Mill recovery drops, concentrate quality fails contract specs, and reconciliation variances destroy confidence in the resource model.
Indicator Type: Lagging. Dilution is confirmed after reconciliation, but draw point grade trends provide a leading signal during the shift.
Unit of Measure: % dilution
Ideal Visualization(s): KPI trend with real-time alerts, Group rollup bars by stope, Box plot for dilution distribution, Pareto chart when ranking stopes by dilution percentage
Frequency: Per shift, with reconciliation per cycle
Data Required: Mucked tonnes, planned ore tonnes, draw point grades, mill head grade, block model estimates
Pro Tip: Track unplanned dilution separately from planned dilution. The operations leader can only act on the unplanned portion.
Red Flag: Steady dilution creeps on a single stope while drill and blast parameters have not changed. Wall slough or oversize draw is in progress.
Mobile Equipment Availability
Why it Matters: Underground fleet is expensive, hard to swap, and bottlenecks the entire production cycle. Availability gaps translate directly into missed tonnes.
What it Measures: Ratio of time mobile equipment is available to operate against total scheduled time, by unit and by class.
What Happens if Missed: A single down LHD or truck stalls a stope sequence. Hot-seat plans collapse and contractors charge standby.
Formula: (Scheduled hours – Down hours) / Scheduled hours
Indicator Type: Current. Status updates from telemetry let operations and maintenance coordinate before the next shift change.
Unit of Measure: % availability
Ideal Visualization(s): KPI blocks per unit, Status history trends, Group rollup bars by equipment class, Pareto chart when ranking units by downtime hours
Frequency: Real-time
Data Required: Equipment operating state, maintenance work order status, fault codes, scheduled hours
Pro Tip: Distinguish between mechanical downtime, waiting on service, and waiting on heading. They have different owners and different fixes.
Red Flag: Repeated short-duration stoppages on a single unit across shifts. A failure is forming and the long stop is coming.
Shaft Hoisting Throughput
Why it Matters: The shaft is the choke point. Every tonne of ore and every person in or out of the mine moves through it.
What it Measures: Tonnes hoisted per hour and skip cycles per hour against rated capacity for production and service hoists.
What Happens if Missed: Ore stockpiles build at the loading pocket, shifts run short on workers underground, and emergency egress capacity becomes uncertain.
Formula: Skips hoisted per hour × skip payload
Indicator Type: Current. The hoist control system reports each cycle, allowing immediate response to load times or queue conditions.
Unit of Measure: t/h
Ideal Visualization(s): KPI blocks per hoist, KPI trend with real-time alerts, Histogram of cycle times, Pareto chart when ranking causes of hoist delay
Frequency: Real-time
Data Required: Skip cycle counts, payload per skip, loading pocket level, hoist availability state, cage occupancy
Pro Tip: Watch loading pocket fill rate as a leading indicator of upstream production. An empty pocket means a stope is in trouble.
Red Flag: Stable cycle count but rising mean cycle time. The hoist is masking a developing brake, rope, or skip-loading issue.
Mine Water Inflow vs Pumping Capacity
Why it Matters: Pumping water is one of the largest fixed costs underground. Losing the dewatering balance floods working areas and idles the lowest, often highest-grade levels.
What it Measures: Ratio of measured water inflow to active pumping capacity across the entire dewatering chain, level by level.
What Happens if Missed: Sumps overtop, electrical infrastructure floods, and lower levels are evacuated until pumping capacity is restored or augmented.
Formula: Total inflow / Active pumping capacity
Indicator Type: Leading. Inflow trends rise before sump levels become critical, giving time to start standby pumps or call out crews.
Unit of Measure: % of pumping capacity used
Ideal Visualization(s): KPI blocks per pump station, KPI trend with real-time alerts, Status history trends for pump state, Pareto chart when ranking stations by capacity used
Frequency: Real-time
Data Required: Pump flow rates, pump operating state, sump levels, motor current, inflow estimates by level
Pro Tip: Treat pump VFD output and motor current as health signals. A pump moving rated water at higher current is wearing out, not working harder.
Red Flag: Inflow climbing on a level that has not seen new development. You have intercepted a structure.
Seismic Activity and Ground Movement
Why it Matters: Rockbursts and falls of ground turn a shift into a fatality investigation. Real-time seismic data is the only early warning available.
What it Measures: Frequency, magnitude, and location of seismic events plus convergence and extensometer readings against ground control thresholds.
What Happens if Missed: Crews are exposed to unstable ground, exclusion zones are imposed reactively, and production areas are sterilized for re-entry protocols.
Indicator Type: Leading. Event clustering and convergence acceleration precede major ground failures by hours to days.
Unit of Measure: Event count, local magnitude, mm/day
Ideal Visualization(s): KPI Map showing event location and magnitude, KPI trend with real-time alerts, Histogram of magnitudes, Pareto chart when ranking zones by event count
Frequency: Real-time
Data Required: Seismic event waveforms, event locations and magnitudes, extensometer and convergence readings, blast schedule
Pro Tip: Filter blast-induced events out of the trend. Otherwise alarm fatigue makes operators distrust the genuine signals when they arrive.
Red Flag: A migrating cluster of small events trending toward an active stope. Stress is redistributing and the next event may not be small.
Energy Consumption per Tonne Hoisted
Why it Matters: Underground mines are among the most energy-intensive operations on a per-tonne basis. Energy waste signals equipment problems and erodes margin.
What it Measures: Total electrical and diesel energy consumed per tonne of ore hoisted, broken down by major consumers (ventilation, pumping, hoisting, fleet).
What Happens if Missed: Demand charges spike, carbon reporting drifts, and degrading equipment runs unflagged because nobody is watching the consumption signal.
Formula: Total energy consumed / Tonnes hoisted
Indicator Type: Lagging at the period total, but real-time at the consumer level, where deviations indicate developing equipment issues.
Unit of Measure: kWh/t or MJ/t
Ideal Visualization(s): KPI trend with real-time alerts, Group rollup bars by consumer class, SPC trend (control chart) for stable consumers, Pareto chart when ranking consumers by share of total load
Frequency: Real-time at consumer level, daily and per-shift roll-ups
Data Required: Substation power readings, diesel fuel consumption, hoisting tonnes, consumer-level submetering
Pro Tip: Benchmark per consumer, not per mine. Total kWh/t is too noisy to drive decisions, while a fan or pump trending up gives a clear maintenance signal.
Red Flag: Hoisting energy per tonne climbing while tonnes per hour stays flat. Friction or rope condition is changing.
Why Real-Time Visibility Matters
Underground operations cannot be walked. The operations leader does not see the airflow dropping at a remote heading, the convergence opening on the back of a stope, or the pump current creeping up on a sump 800 meters below surface. Without real-time KPIs, those signals reach the surface as incidents instead of decisions.
The mines that run well are the ones where ventilation, ground, water, fleet, and production data converge in front of one shift control screen. When a leading indicator drifts, a person makes a call before the lagging indicator becomes a lost shift, a flooded level, or a rescue operation.
How Transpara Can Help
If real-time operational visibility is a challenge you’re facing, you’re not alone. At Transpara, we help teams like yours gain clarity from complex systems without the need to centralize or overhaul your data stack. Learn more about Transpara Browse our documentation Contact us
