A demand-response provider may report the sum of commanded reductions. The transformer sees the devices that actually respond, mixed with changing background demand and generation. Verification needs an observed local result and a defensible estimate of what would have happened without the event. [1]
Define the claim and its measurement boundary
State whether the claim is active-power peak, phase current or modelled thermal exposure. Evaluate uncertainty of measured and baseline inputs in the actual comparison [2]. A kW reduction is not itself a hot-spot temperature measurement.
Use time-aligned transformer measurements and resource response records. Retain voltage, sampling interval, quality flags and the electrical boundary. A portfolio meter cannot establish the result at an individual transformer whose participating resources and background load differ from the portfolio.
| Evidence | What it establishes | What it does not establish alone |
|---|---|---|
| Dispatch instruction | Requested action and time | Actual local load reduction |
| Device acknowledgment | Receipt or applied state, as defined | Transformer peak relief of equal magnitude |
| Transformer measurements | Observed local electrical trajectory | The unobserved no-event trajectory |
| Baseline with uncertainty | A conditional counterfactual comparison | Exact avoided demand or a guaranteed life benefit |
Expose the baseline uncertainty
Describe how the no-event estimate was constructed and what histories were used. Weather, customer behavior, generation and participation can differ on the event day. Avoid selecting a reference period merely because it maximizes the apparent reduction.
Report a range or sensitivity where the counterfactual is uncertain. If the estimated benefit is smaller than that uncertainty, the evidence cannot resolve the claimed magnitude confidently. This does not erase the measured event; it limits what can be attributed to the control action.
Keep recovery in the same verification window
Extend the evaluation through the defined recovery period. Check whether postponed energy creates a new peak, prolongs loading or leaves services incomplete. A favorable event-only metric can conceal a larger later constraint.
Retain phase and neutral currents where relevant. A thermal claim needs the complete electrical/ambient/cooling history and a stated model; IEC 60076-7 addresses mineral-oil loading/temperature/ageing [3]. Label predicted temperature and ageing separately from measurements.
Assume baseline/controlled powers 420/380 kW for a 30 min event and 410/450 kW for 30 min recovery. Initial energy reduction is 20 kWh; recovery adds 20 kWh, so net saving is zero. The hour’s peak increases from 420 to 450 kW. Against an assumed 440 kW active-power envelope, baseline passes and the intervention fails in recovery. Reject that release pattern and reshape recovery before claiming relief. Reactive/phase current and thermal checks remain additional constraints, not consequences of this kW-only screen.
Check whether a shortfall comes from command execution, the counterfactual estimate or the local electrical boundary. Retain the recovery period in that comparison. An aggregate reduction percentage can hide all three problems when its measurement location and baseline range are unspecified.
References
- A. Jain and M. Karimi-Ghartemani, Mitigating Adverse Impacts of Increased Electric Vehicle Charging on Distribution Transformers, Energies 15, 9023, 2022. DOI: 10.3390/en15239023.
- B. N. Taylor and C. E. Kuyatt, NIST Technical Note 1297, Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results, 1994, sections 5 and 6.
- IEC, IEC 60076-7:2018, Power transformers – Part 7: Loading guide for mineral-oil-immersed power transformers.