Monitoring & Targeting

Monitoring and Targeting (M&T) gives your operators one dashboard with five dials: kWh against the monthly target, power factor, specific energy in kWh per ton, predicted demand against your notified maximum demand, and plant efficiency against the baseline.

Behind every dial sits the data, the calculation and the evidence that produced it. The dashboard is the front end of one measurement system: measure, compare with expected performance, act and verify.

Start here

Further down this page: baseline scatter · running empty · service-output baselines · demand control · savings and M&V · energy balance · action and integration · technical notes

Dial 1 · Monthly energy target

The operator’s question: how much of the monthly allowance have we used, and will we stay within it?

The monthly energy target is the approved kWh allowance for the production and operating conditions planned for the month. The dial shows how much has been used and how much remains. Its supporting forecast shows whether month-end energy is likely to exceed the allowance.

Read the dial with its forecast

• Used: month-to-date metered energy as a percentage of the approved monthly target.
• Remaining: the target less the energy already consumed. A negative balance means the allowance has been exceeded.
• Forecast: actual energy to date plus the energy expected for the remaining production plan.

Used (%) = actual month-to-date kWh ÷ monthly target kWh × 100
Forecast month-end kWh = actual to date + expected remaining kWh

Worked example: 72% used does not tell the whole story

Set the target from the production plan

Apply the approved baseline model to planned output and operating periods, then include any separately agreed improvement target. Compare actual use with the expected trajectory for actual production, rather than elapsed calendar days alone.

If the production plan, product mix or operating boundary changes, record and approve the adjustment. Do not increase the target simply to absorb inefficient operation. Baseline expectation and a tighter improvement target must be clearly distinguished.

Operator response

If the forecast is above target, open the interval exceptions and baseline scatter. Identify running empty, unstable loading or persistent excess use, and assign actions before the remaining allowance is consumed.

Read the energy balance together with the production forecast. “72% used” alone is not an efficiency judgement.

Dial 2 · Power factor

The operator’s question: how effectively are we using the available electrical capacity?

Power factor (PF) is the ratio of real power, which performs useful electrical work, to apparent power, which determines the loading on the electrical supply. For a normal importing load, a value closer to 1.00 means less apparent power is required for the same real power.¹

Power factor = real power (kW) ÷ apparent power (kVA)

Use the meter’s true power-factor value. In a system with waveform distortion, power factor is not simply the cosine of the phase angle. PF is an electrical-capacity indicator; it is not the same as production efficiency or kWh per ton.

Read the dashboard example

Confirm before commissioning: these colour bands are design settings, not universal tariff thresholds. Confirm the site’s requirements, importing/exporting sign convention and leading/lagging indication before commissioning.

Worked example: the same 1,000 kW load

The reduction in kVA can release capacity and affect demand charges, subject to the tariff. It is not an equivalent reduction in kWh: real energy and electrical losses must be assessed separately.

Operator response

Check whether poor PF coincides with lightly loaded equipment, a failed correction stage or a change in plant operation. Refer correction settings, harmonics and the risk of leading PF or overcorrection to the responsible engineer.

PF explains electrical loading. The next dial explains energy used for useful production.

Dial 3 · Specific energy

The operator’s question: how much energy does each ton require, and what should it require?

Specific energy, expressed as kWh/t, is the electrical energy used to produce one ton of output. The numerator and denominator must cover the same plant boundary and the same time interval. This page uses metric tonnes (t).

Specific energy (kWh/t) = interval energy (kWh) ÷ interval output (t)

Worked example: compare actual with expected

How the baseline provides the target

A simple production model may take the form below. The fixed term represents energy that does not vary directly with output within the modelled operating state; the variable term describes the additional energy associated with production.

Expected interval energy = fixed term + variable rate × tonnes

For example, an illustrative model of 2,200 + 5.0 × tonnes predicts 7,200 kWh at 1,000 t. The expected kWh/t is therefore 7.2 at that output. At a different output the expected ratio changes, because the fixed term is spread over a different number of tonnes.

Data rules for a trustworthy dial

• Match timestamps and intervals. Instantaneous MW and cumulative tonnes cannot directly produce interval kWh/t.
• For a shift or month, divide total energy by total tonnes; do not take an unweighted average of interval ratios.
• At zero or very low output, show an idle or running-empty flag. Do not divide by zero or show a misleading zero kWh/t.
• Display the interval, data status and baseline version. Confirm the scale and target bands from site data.

kWh/t is meaningful only when output, operating conditions and the expected baseline are understood together.

Dial 3 continued · Building the baseline

Each point on a baseline scatter pairs energy and production from one interval. A validated model describes expected energy across the operating range. The scatter shows where performance departs from that expectation and where further investigation is needed.

Create the baseline in four steps

1. Collect representative, time-aligned energy, production and operating-state records. Cover the normal production range and relevant seasonal or product-mix effects.
2. Check missing values, meter resets, shutdowns and abnormal states. Keep an audit trail of exclusions, and separate different operating modes where required.
3. Fit an appropriate model and examine residuals, bias and validity range. A visually neat line or a high correlation alone is not sufficient validation.
4. Approve the baseline, its uncertainty and tolerance, and its adjustment rules before using it for the live dial or savings verification.⁴

Read the points

Above the baseline: more energy than expected for the same output. Near or below it: potentially normal or improved performance, subject to data checks. At little or no output: investigate running empty separately, even if the points lie below the productive-operation baseline.

For example, 7,900 kWh ÷ 500 t = 15.8 kWh/t, compared with 8,200 kWh ÷ 1,000 t = 8.2 kWh/t. Specific energy is about 48% lower at the higher output. This demonstrates a loading effect, not verified savings from an intervention.

Dial 3 continued · Running empty

A scatter of power against production makes running empty visible: a cluster at zero production shows substantial electrical input while no useful output is recorded. This is a priority for scheduling and operating-state investigation.

Energy associated with a power difference = kW difference × duration (hours)

A vertical gap on a power-versus-production graph is in kW. Convert it to kWh using the relevant duration, then judge how much is avoidable. How often a plant runs empty must be confirmed from timestamps and operating-state records, not from the appearance of the scatter alone.

Use running-empty periods to review shutdown discipline, interlocks and sequencing, while preserving essential services and safe process conditions.

Dial 3 continued · Service-output baselines

Not every plant produces tonnes. Compressors, pumps and other utilities use the same baseline principle with an appropriate service measure, such as delivered air volume or water pumped. Compare equivalent service and operating conditions.

What a lower curve means

At comparable airflow, a lower kW curve indicates less electrical input for the same service, provided pressure, air quality and other material conditions are equivalent. Integrate the difference over operating time to estimate the associated kWh reduction.

In one historical compressor example, an average reduction of 138 kW (37%) was reported. It is shown here as a reported result, not a guaranteed saving for any site.

Use the comparison to direct work

• Check unloaded running, leaks, setpoints, sequencing and operation away from the efficient service range.
• Compare baseline and post-action performance at equivalent airflow and pressure, with documented adjustments where conditions changed.
• Choose a consistent energy-per-service unit for the detailed view. Do not substitute instantaneous kW per flow unit for a measured energy-per-volume result without matching the time basis.

Whether the useful output is tonnes, air or water, the question is the same: how much energy should this service require?

Dial 4 · Predicted demand

The operator’s question: where will this demand interval finish against the 4.5 MVA limit?

Predicted demand estimates the average demand that will be recorded when the current demand period closes. It gives the operator time to respond before the final interval value is established. The dashboard compares the prediction with the notified maximum demand (NMD) limit, 4.5 MVA in this example.

Keep three quantities distinct

Why the interval matters

A high completed demand interval can set a monthly maximum. The amount billed depends on the site’s tariff: eligible periods, kW or kVA basis, minimum demand, ratchets and NMD provisions must all be confirmed. A brief instantaneous peak is not automatically equal to a high 30-minute average.

How is demand predicted?

A simple fixed-block forecast adds the demand accumulated so far to the demand expected over the remaining time, assuming the recent load continues, and divides that total by the full block duration.²

Predicted kVA = (accumulated kVAh + recent kVA × hours remaining) ÷ block hours

Example: 12:45 remains in a 30-minute block. That is 0.2125 hours remaining and 17:15 elapsed. The accumulated value is 1,100 kVAh; recent load is 4,000 kVA.

This is an illustrative forecast, not a guarantee. Recalculate as load changes and keep an operating margin. Use the meter’s approved demand method and clock alignment; some tariffs derive kVA from demand kW and kvar rather than a direct kVAh average.

Respond before the interval closes

Warn early, then shift, shed or sequence approved non-critical loads. Protect critical processes, and use agreed deadbands, minimum run times and a controlled restoration sequence.

Meter note: the SATEC PRO series distinguishes accumulated block demand from predicted sliding-window demand. Confirm the actual register, interval and controller algorithm against the billing method before commissioning.³

Read the forecast and the time remaining together. A 3.9 MVA prediction is an estimate of the closing interval, not a fixed allowance to add load.

See also: Maximum Demand Control.

Dial 5 · Plant efficiency

The operator’s question: are we using more or less energy than the baseline expects?

The Plant Efficiency dial summarises whether the plant used more or less energy than expected for the output achieved. Here, “efficiency” is an energy-performance index against the approved baseline, not a thermodynamic conversion efficiency.

Baseline performance index (%) = actual kWh ÷ expected baseline kWh × 100

For the same interval, the same positive output and the same boundary, this is also actual kWh/t divided by expected kWh/t, multiplied by 100. The kWh/t dial gives the specific value; Plant Efficiency summarises its position against the baseline.

Worked example

Translate the index into a clear status

The status is always calculated from the approved index and tolerance, never set independently. A tighter improvement target is kept distinct from the baseline, and missing data never defaults to a green status.

The link back to action

Use the same production, operating-state and baseline records for both dials. An adverse index leads back to its cause: running empty, part loading, process changes or equipment condition. Confirm the cause before assigning corrective work.

From dashboard to results · Measurement & verification

The dashboard points to a problem; the baseline quantifies the energy difference. A positive variance is an investigation opportunity, not proof that every excess kWh is avoidable. Savings are verified after action against an appropriately adjusted baseline.⁴

Illustrative interval calculation

Verify the result after the action

• Record the intervention, date, boundary and expected effect. Agree the comparison period and any baseline adjustments.
• Compare measured post-action energy with the baseline energy predicted for the same output and relevant conditions. Check missing data and changed operating states.
• Report validated kWh savings, the applicable energy cost and any separately calculated demand-cost effect. Keep the model version and evidence.

Use the same data stream to identify the opportunity and verify the result. Do not count a kVA reduction as an equivalent kWh saving.

Supporting measurement · Energy balance & data

Meters at the incoming supply, major plants and utilities create a reconciled energy balance. It identifies the largest users and directs plant-level investigation. Reconcile the same period and explain unmetered use, losses and differences in meter boundaries.

Build one trusted data foundation

Keep the engineering detail behind the simple operator view. Every dial supports drill-down to the source trend, scatter, interval record and data-quality flags that created its indication.

Operating response · Action & integration

Monitoring & Targeting connects the symptom to its operating cause. Assign an owner and a due date, record what changed, then return to the same baseline to check the result.

Commission the live dashboard carefully

• Match timestamps, meter boundaries, output units and intervals. Publish a visible timestamp and data-status flag, and mark stale values clearly.
• Approve the model, validity range and adjustments. Set kWh/t bands and efficiency tolerance from the same baseline and operating period.
• Confirm PF bands, the NMD limit, demand-window settings, billing alignment and the monthly production-based allowance.
• Test calculations and alarm transitions with recorded scenarios: normal operation, no output, missing data, peak demand and changed production.
• Start with passive visibility. Add approved automatic actions only after controls, safety priorities, fallbacks and restoration logic have been tested.

Measure continuously, integrate safely and verify every improvement against the agreed baseline.

Supporting views · Technical notes

The opening dashboard supports immediate decisions. The following views keep the engineering detail and the audit trail needed to explain each indication.

Technical references

1. SATEC — Power Meters: Power Factor. Real and apparent power definitions.
2. SATEC — Power Meters: Predicted Demand. Forecast principle using accumulated energy and present load.
3. SATEC — EM235/PM335 PRO Installation and Operation Manual, Rev. A8, pp. 62–63. Accumulated block and predicted sliding-window demand.
4. US Department of Energy FEMP — Measurement and Verification Activities. Baseline conditions, M&V planning and documented adjustments.

All figures on this page are illustrative examples, not results for a specific site. Live values, targets, tolerances and tariff logic are configured from each site’s approved records.

Why National Power

M&T only works if the metering, the software and the engineering judgement come from people who understand all three. We supply the meters as the Southern African exclusive distributor of Satec products, we run the analysis in ExpertPower, and we have spent 30 years doing the engineering behind the numbers for clients including Sun International, Valterra Platinum and De Beers Group.

That means one accountable party for the whole chain, with no gaps between the hardware vendor, the software vendor and whoever is meant to interpret the result.