Unit Converters
For energy managers, process engineers, and submeter analysts, convert an ordered series onto one unit basis and audit extrema, center, cumulative energy, and equivalent average power before deciding which records need investigation. The model assumes every retained value represents a comparable measurement boundary and a documented total elapsed time; it does not reconstruct peak demand.
Energy distribution
Ordered energy audit
Convert an ordered set of energy observations, then quantify its minimum, maximum, spread, center, cumulative energy, and time-equivalent average power. Use it for metering, test cycles, or production batches where record order and one common unit basis matter.
Input evidence: retain meter timestamps, interval definitions, multiplier history, missing-period flags, and one energy boundary for the entire sequence.
| Record | Entered energy | Converted energy | Deviation from mean | Cumulative energy |
|---|
Accumulation analysis
The staircase preserves record order and separates a late high-energy cycle from a consistently elevated process.
| Scenario or component | Basis | Calculated value | Decision use |
|---|
How to use
Start with the original export rather than a sorted summary. Confirm that every record covers the same asset, register definition, and interval convention; keep timestamps beside the values so any extreme can be traced after conversion.
Maximum minus minimum measures observed spread but ignores the order and frequency of values.
Mean uses every reading; median resists a small number of extreme cycles.
A running sum reveals when the total accumulated, not only how large it became.
Total joules divided by elapsed seconds connects energy quantity to a time rate.
Preserving acquisition order keeps ramps, interventions, and production lots visible.
Result interpretation
The range reports only the distance between the smallest and largest converted records. The mean describes the arithmetic center, the median shows the center after numerical ordering, and cumulative energy preserves how the total developed through the entered sequence. Equivalent average power divides total energy by the full represented time; it is neither the highest interval demand nor an estimate of instantaneous power.
Interpret differences before ranking records. A mean well above the median suggests a high-side tail, a zero range means every retained value is identical on the displayed precision, and a threshold count of zero means only that no value crossed the entered rule. Negative energy is valid only for a boundary that permits export or regeneration.
Each reading is converted through joules before summary statistics are calculated. A wide range can reflect a genuine load shift, mixed operating states, or a recording defect; it does not by itself identify the cause.
Minimum and maximum should be traced back to timestamps, operating modes, and meter status.
A threshold may represent review, budget, or capacity. Its direction and consequence must be written before flags are counted.
Equivalent average power is valid only when the entered duration covers the same energy records with no gaps or overlaps.
Interval integrity
Daily register differences, cycle-level energy, and submeter intervals can all be expressed in kWh, but they are not comparable unless their start and end boundaries, included loads, and acquisition rules match. A 24-hour reading should not be compared directly with a production cycle merely because both use the same unit.
Split the analysis when a meter reset, tariff-register change, current-transformer multiplier revision, daylight-saving clock change, estimated interval, or missing period occurs. Repairing those events outside the calculator preserves an auditable reason for every retained value and prevents duration errors from contaminating equivalent average power.
Outlier investigation
Link the minimum, maximum, and threshold records back to timestamp, product, batch size, weather, occupancy, operating mode, maintenance event, and meter status. Compare the raw register movement with an independent production or equipment log before changing the dataset.
Exclude a value only when evidence shows a duplicated interval, reset, transcription error, multiplier mistake, or known sensor failure. A genuine extreme may reveal startup, cleaning, rework, abnormal idle consumption, or a new operating regime. Report both the original and justified cleaned result when exclusion changes the decision.
Demand boundary
Two days can consume the same energy while having very different fifteen-minute peaks. Dividing a daily kWh total by 24 hours gives an average rate that can hide motor starts, simultaneous loads, production bursts, and short tariff-demand windows.
Use timestamped interval-demand or waveform data when transformer capacity, demand charges, generator sizing, protection settings, or power-quality decisions depend on short-duration behavior. This page remains useful for energy spread and accumulation, but it cannot reconstruct a maximum that was never measured.
Visual reading guide
Record number is the horizontal axis in both views. The first chart shows converted energy per record with mean and review lines; the second shows the running total. A single tall bar creates one abrupt step, while a sustained operating shift changes the staircase slope across several records.
Changing the source unit rescales labels without changing order; changing the threshold moves only the review reference; changing elapsed time affects equivalent average power but not the bars. Both views become misleading when interval durations differ, records are missing, or acquisition order no longer represents time.
Eᵢ,target = Eᵢ,source × Fsource / Ftarget; R = max(Eᵢ) − min(Eᵢ); P̄ = Σ(Eᵢ × Ftarget) / Δt
E is an energy record; F is joules per declared unit; R is the range; P̄ is average power in watts; Δt is elapsed time in seconds.
| Eᵢ | individual converted energy | target energy unit |
| F | unit-to-joule factor | J/unit |
| R | maximum minus minimum | target energy unit |
| Δt | observation duration | s |
Reconciliation:
Default-value audit
Defaults: 18, 22, 17, 25, 21 and 28 kWh; target kWh; elapsed time 360 minutes; review threshold 24 kWh.
| Symbol | Meaning | Unit | Default role |
|---|---|---|---|
| Eᵢ,s | Source energy record i | kWh | Six entered values |
| Fₛ, Fₜ | Source and target joule factors | J/unit | Both 3,600,000 |
| Eᵢ,t | Converted record i | kWh | Equal to source defaults |
| Emin, Emax | Observed extrema | kWh | 17 and 28 |
| R | Range, Emax − Emin | kWh | 11 |
| EΣ | Sum of converted records | kWh | 131 |
| Δt | Total observation duration | s | 21,600 |
| P̄ | Equivalent average power | W | EΣ in joules ÷ Δt |
| T | Review threshold | kWh | 24 |
Reasonableness check: 131 kWh across 6 hours equals 21.833 kW average; the range remains 11 kWh because source and target units match.
Evidence requirements
Keep raw register values, timestamps, interval duration, meter and channel identifiers, current-transformer multiplier, timezone, reset flags, estimated-reading flags, calibration status, and the equipment boundary. Freeze the unit definition and sign convention before conversion.
Use records from one comparable population and period. Reconcile the accepted row count to the source export, the final cumulative total to an independent utility or historian report, and the represented duration to timestamp differences. Unexplained count, total, or time mismatches invalidate the summary even when the arithmetic is correct.
Limits and exclusions
The model does not estimate meter uncertainty, repair missing intervals, separate active from reactive energy, reconstruct peak demand, allocate shared loads, weather-normalize consumption, or prove why an extreme occurred. Range is a descriptive statistic, not a confidence interval, control limit, or specification tolerance.
Consequently, use the output to organize investigation and reconcile energy—not to approve billing, size electrical equipment, declare a process statistically stable, or attribute savings to an intervention. Those decisions require governed meter data, appropriate time resolution, and a method matched to the causal or compliance question.
Energy-series glossary
Worked cases
Inputs: 18, 22, 17, 25, 21 and 28 kWh across 360 minutes; review line 24 kWh.
Calculation: total 131 kWh, minimum 17, maximum 28, range 11 and equivalent average power 21.833 kW. Two records exceed the review line.
Decision: inspect cycle 4 and cycle 6 timestamps and machine states; do not delete them merely because they are high.
Inputs: 410, 398, 425, 612, 590, 438 and 405 kWh for seven complete 24-hour days with no clock gaps.
Calculation: total 3,278 kWh, minimum 398, maximum 612, range 214, mean 468.286 and median 425 kWh. Across 168 hours, equivalent average power is 19.512 kW.
Decision: the mean-median gap and two high days justify checking occupancy, weather and plant schedules. Use 19.512 kW for weekly energy-rate reconciliation, but obtain shorter demand intervals before reviewing transformer or tariff peaks.
Do not rely on the range alone when safety, billing, compliance, or equipment protection depends on peak demand timing or measurement uncertainty.
Frequently asked questions
Sorting hides ramps, interventions, meter resets, and the point at which cumulative energy changed. Keep acquisition order for the visuals and ledger; calculate the median on a temporary numerical ordering without replacing the source sequence.
Only after tracing it to a timestamp and independent operating evidence. Remove a value for a demonstrated data defect, not because it is inconvenient; report the original and cleaned result when the exclusion changes range or decisions.
No. Equivalent average power spreads total energy across the full entered duration. Peak demand requires shorter interval or instantaneous measurements and can be much higher even when daily or cycle energy is unchanged.
Do not compare them as one series. Normalize to comparable boundaries and durations when evidence supports that transformation, or analyze daily and monthly populations separately so one long period does not dominate interpretation.
The mean uses every magnitude and is pulled toward a small number of large or small observations. The median identifies the middle ordered value and is more resistant to extremes, so their gap helps reveal asymmetry.
A correct linear energy-unit conversion changes displayed magnitude and unit but preserves order, ratios, and relative spread. Early rounding can break that invariance, which is why statistics use unrounded converted values.
Equivalent average power is invalid until gaps are repaired or excluded consistently from both total energy and duration. Preserve gap flags and use a timestamp-aware process rather than assuming the missing period has zero or repeated consumption.
Only when the documented rule defines crossing as failure. A threshold may instead mean investigate, budget review, alarm, or capacity attention; record the direction, owner, and required response before counting flags.
Use a demand or load-profile model when timing, coincidence, tariffs, or equipment capacity matter more than interval-energy spread. Use a statistical process method when the question concerns stability or control limits rather than simple observed range.