Acrel Maximum Demand Below a Power Spike? Check the Averaging Window
A plant trend shows a brief rise to 90 kW, but the meter’s maximum demand remains much lower. That difference can be normal: an instantaneous power value and demand averaged over a time window answer different questions. Before changing a CT ratio, resetting a maximum or blaming communications, identify the quantity and averaging period on each display.
Start with the meter’s definition
Section 7.4 of Acrel’s DTSD1352 installation and operation instruction V2.6 defines demand as average power over the demand cycle, and maximum demand as the highest demand within a specified period. It describes sliding demand and gives defaults of a 15-minute cycle and one-minute slip time. Verify the actual configuration; a manual’s default is not evidence of the settings in a particular installation.
The same manual lists maximum demand and its occurrence time as optional functions. Its display notes exclude maximum demand where the multi-tariff function is absent. Confirm the full model, options and applicable manual before making a demand register or dashboard tile part of the project specification.
The ProSense Acrel smart power meter listing identifies the available product family and electrical configurations. Ask for confirmation of the required demand functions and communication data before ordering. A meter chosen for an inverter interface should not automatically be assumed to provide every energy-management option.
Worked example: a short high load is diluted
Consider a hypothetical 15-minute window with a steady 30 kW load for 14 minutes and a 90 kW load for one minute. Assume the power is constant within each of those intervals. The energy in the window is (30×14+90×1)/60=8.5 kWh. Divide by 0.25 hours and the average is 34 kW.
The highest power in that example is 90 kW, while the window average is 34 kW. Neither value contradicts the other. The first describes the highest instantaneous level in this simplified scenario; the second describes the average over the full interval. These are illustrative figures, not a test result or a forecast for your site.
With a sliding calculation, successive windows overlap. A one-minute slip with a 15-minute cycle means the average is recalculated as the window moves; it does not turn the measurement into a one-minute average. As the high-load minute enters and later leaves the window, its contribution changes.
A fixed block calculation can give a different maximum from a sliding calculation because its boundaries divide the same load history differently. Record the boundary convention as well as the duration whenever comparing two systems. Matching a label such as ‘15-minute demand’ is not sufficient.
Four checks before investigating a mismatch
Identify the exact quantity. Distinguish current active power, current averaged demand and retained maximum demand. Also distinguish kW from kVA and import from export. Use the applicable register description, scaling and units rather than a short dashboard label.
Record the calculation settings. Capture the averaging duration, sliding interval where applicable, time basis and the period covered by the retained maximum. Preserve existing settings and evidence before making changes.
Compare the same time window. Align meter timestamps with the host system, including the timezone and whether timestamps represent the start or end of an interval. Check what the dashboard does with late, missing or duplicated records.
Reconcile using suitable history. For a complete interval, average power equals the interval’s energy divided by elapsed hours. Use data with adequate resolution and confirmed units. Do not assume a few widely spaced power samples reproduce the meter’s internal calculation.
Polling faster does not change the demand cycle
A host reading a demand value every few seconds may receive repeated values between internal updates. That alone does not prove the connection has stalled. Likewise, recording the largest polled power value does not recreate a maximum-demand statistic. Label the two series separately so operators know which one they are using.
If a discrepancy remains, compare the local meter indication with the corresponding host value under the same conditions. Investigate mapping, units and scaling before altering the measurement configuration. Keep this read-only until the meaning of each value is established; a successful communication response is not proof that the intended statistic was selected.
Preserve the maximum and define the next action
Do not clear a retained maximum merely to make it match the current load. Save the value, occurrence time and applicable period first. If a reset or settings change is authorised, record its time and mark the resulting history boundary so later comparisons do not silently combine different definitions.
This guide addresses operational measurement interpretation. It does not establish how a retailer, network or contract calculates a charge; check the applicable metering and tariff documentation separately. Monitoring data also does not replace electrical protection. Installation, access to electrical panels and any control changes require competent personnel and the current requirements applicable to the site.
For a useful support enquiry, send the model and options, manual revision, demand settings, exact register or display name, timestamped readings and the comparison interval. That evidence usually resolves the question more efficiently than a screenshot containing only a number labelled ‘peak’.
Need help with your application?
ProSense Instruments - discuss meters and instruments. ProSense Solutions - discuss monitoring and integration projects. ProSight Software - discuss software requirements; device integration must be confirmed for your application.
Instrument enquiries: sales@prosenseinstruments.com.au | 1800 560 854. Project enquiries: sales@prosense.com.au. Software enquiries: info@prosightsoftware.com.
Comments