Work examples

Watch a task move, step by step.

Five everyday tasks from factories, utilities and solar and storage sites. Each shows what the software prepares, where people decide and act, and how the result is checked. All records are simulated teaching data: they explain the method, not customer results.

Factories · energy records

Energy records: same period first, then explain.

Bills, the main meter, sub-meters and production records come from different places. The bill says 144 MWh, the calendar-month meter 150 MWh. Is the 6 MWh gap waste?

Simulated records for illustration.

Inputs

Record	Period	Energy and status
B-01 bill	26 Aug → 26 Sep	144.0 MWh, settled
M-01 main meter, calendar month	1 Sep → 1 Oct	150.0 MWh, complete
M-02 main meter, billing period	26 Aug → 26 Sep	144.6 MWh, re-summed from hourly data
S-A workshop A	Billing period	80.0 MWh, complete
S-B workshop B	Billing period	50.0 MWh, 8 h missing on 12 Sep, not estimated
S-C shared load	Billing period	10.0 MWh, complete, no double metering

The application puts every record on the same clock. The calendar-month reading covers another period, so 150.0 − 144.0 = 6.0 MWh is not a like-for-like difference.
Same period: 144.6 − 144.0 = 0.6 MWh, kept as a metering-scope or record difference to check. Main meter against sub-meters: 144.6 − (80.0 + 50.0 + 10.0) = 4.6 MWh, not a confirmed loss while S-B is missing data.
Site staff supply S-B’s missing hours with their source; a specialist settles the metering-scope question. The attachment joins the existing energy report, and the next cycle reuses the confirmed basis.

What you get: An energy report with sources: comparable records, explained changes, open differences and the next steps.
What stays open: Fill S-B’s missing data, check the metering scope of the bill and the meters, and keep who changed what, when and why. A faster report is not an energy saving.
Existing systems: Existing EMS, reports and expert tools stay. Inputs come from authorised exports or interfaces.
How to judge it: Compare the time to prepare, review and correct a full report, and whether unexplained differences shrink.

Equipment · specialist handover

An anomaly handed to a specialist, with the records ready.

Unit UNIT-07 (series M-A, configuration F-02) raises three alarms. Instead of everyone explaining it again, the product gathers the records, the applicable manual and the checks already done into the existing service ticket SVC-DEMO-17.

Simulated records for illustration.

Inputs

Field	Teaching record
Event	Alarms A-01, A-02 and A-03 at 07:55–08:00 (UTC+08:00), linked to SVC-DEMO-17 to be reviewed together; not a claim of one root cause
Equipment and manual	UNIT-07 / M-A / F-02; manual MAN-A-R3 confirmed applicable by a specialist
Checks done	CHK-DEMO-01: all three records point to UNIT-07. CHK-DEMO-02: time zones agree. Record checks only, not a diagnosis

The product matches equipment and manual by confirmed rules, lists the alarms, checks and sources, and attaches them to the existing ticket.
Condition data for 07:55–08:00 are missing. The product marks the gap as open and hands the attachment to specialist support.
At 08:10 the specialist acknowledges (ACK-DEMO-01) and asks for the missing condition data. The ticket stays open: the product never closes it.

What you get: A handover record: what is known, its sources, what is missing, the open questions, and who has taken it.
What stays open: The cause is not known yet: condition data are still missing. Had the manual’s applicability been unconfirmed, the records could still be organised but marked “manual to be confirmed”, with no operating advice from it. If nobody takes the handover, it shows as waiting.
Existing systems: Work continues in the existing service ticket. Equipment protection and site responsibilities do not change. Whether a notice was sent and whether someone took it are recorded separately.
How to judge it: Check that records are matched correctly, that fewer things are missed, and how much preparation the site team and specialist save together. If the specialist has to redo everything, nothing is proven.

Solar · performance and maintenance

Low output: prepare the check, then confirm the result.

At about 800 W/m², inverter INV-07 delivers 76 kW against a 100 kW teaching reference. The owner needs to know whether intervention is worthwhile; the O&M team needs a clear task and a documented result.

Simulated records for illustration.

Inputs

Field	Teaching record
Object	Site A / INV-07; existing work order MT-024
Observed	About 800 W/m², comparable temperature; 76 kW against a 100 kW reference
First judgement	76 ÷ 100 − 1 = −24 %: a deviation, not a root cause
Known context	No curtailment recorded in this example

−24 % under comparable conditions, with the readings and the reference attached to work order MT-024.
An authorised O&M lead sends the site check. The team finds soiling and cleans it per the maintenance procedure, recording what they found and did.
96 kW afterwards: 96 ÷ 100 − 1 = −4 %, kept as a residual difference to watch for recurrence.

What you get: The event, the task and the re-check in one record, as evidence for maintenance decisions and service reporting.
What stays open: No return amount: repair cost, integrated energy and settlement terms are missing. One fewer site visit is not automatically O&M profit, and recovered power is not settled revenue.
Existing systems: Existing monitoring, performance models and work orders stay. Storage needs its own checks of dispatch commands, state of charge, reserve and equipment limits; the solar method does not carry over.
How to judge it: Total handling time, rework, missed repairs and review effort, and whether that lowers the buyer’s cost or raises revenue. Where current practice is enough, no product is needed.

Utilities · supply and demand

Two processes, one steam header: find a schedule that fits.

Two processes share an 8 t/h steam header. Each is fine on its own; the question is whether the schedule as a whole can be supplied. This example compares fixed loads and times only: no pressure transients, thermal inertia or start-up.

Simulated records for illustration.

Inputs

	Process A	Process B
Earliest start	08:00	08:10
Duration	30 min, uninterrupted	20 min, uninterrupted
Steam demand	4 t/h	6 t/h
Finish by	08:40	09:00
Header capacity	8 t/h, shared	8 t/h, shared

Original plan: A 08:00–08:30, B 08:10–08:30. While they overlap, 4 + 6 = 10 t/h, more than the 8 t/h header: not feasible.
Candidate 1, B at 08:30–08:50: peak 6 t/h, meets start, deadline and supply. Candidate 2, A at 08:30–09:00: supply is fine, but A finishes after 08:40: excluded.
Production and utility leads check site constraints not listed here, then adopt it. Actual times, demand and deviations feed the next calculation. Steam use is unchanged (2 t + 2 t): a lower peak, not an energy saving.

What you get: An option comparison grounded in inputs and constraints: the schedule that fits, who confirms it, and how to check it afterwards.
What stays open: Whether unlisted site constraints exist is for the responsible people to check.
Existing systems: First see whether existing control, scheduling and expert tools can do it by configuration; only then add an application or model. No second control platform by default.
How to judge it: For the plant: supply, production and cost. For an equipment maker: whether the capability helps sell, deliver or service. Judge the two separately.

Storage · command and limits

Requested 60 kW, delivered 40 kW: fault or limit?

Battery BESS-DEMO-01 receives a 60 kW discharge request at 14:00 (UTC+08:00) and delivers 40 kW. Before anyone reports a fault, the request, the limit in force and the measurement are checked against each other.

Simulated records for illustration.

Inputs

Source	Teaching record
Request REQ-DEMO-01	Discharge 60 kW; valid 13:55–14:05
Limit LIMIT-DEMO-01	Controller-reported discharge limit 40 kW, configuration L-DEMO-1, same window
Measurement MEAS-DEMO-01	At 14:00:00: output 40 kW, state of charge 32 %
Reserve RES-DEMO-01	Reserve floor 30 %

Request, limit and measurement for the same object at the same moment. The 40 kW comes from the limit record; it is not inferred from the state of charge.
Requested 60, limit 40, delivered 40: consistent in this case, so not reported as a fault. It does not prove the limit strategy is optimal.
If a source is missing, stale or for another object, the explanation is withdrawn and the controls or equipment support party checks. Had output been 20 kW, a new record BESS-CHECK-02 keeps the gap to the 40 kW limit open, waiting for the responsible party.

What you get: A timed, sourced check record: which differences are explained, what still needs checking, and who carries on.
What stays open: Device rules, command authority, models and how results are written back need a site plan. Nothing changes dispatch or equipment commands automatically.
Existing systems: Existing EMS, PCS/BMS, dispatch and vendor services stay; add only the data and checking application you need.
How to judge it: Fewer repeated questions and misjudgements, and open issues reaching the right party. Technical O&M effects and market revenue are judged separately. This is a rule check, not a battery optimiser, a trading strategy or a live control interface.

Bring one task your team repeats.

A short description of the work, the records it uses and the systems involved is enough to start. No full production data is needed for a first conversation.

Start a conversation