Well Production Balance Worksheet

Well ProductionBalance Worksheet

Reconcile the water produced, stored and delivered—before turning an unexplained difference into a performance claim.

1. Engineering answer: account for the water before blaming the well

A well production balance worksheet reconciles water discharged from a well with the water transferred, stored, used and otherwise removed over the same reporting period. It combines totaliser readings, flow observations, pump operating records and opening-to-closing storage changes. Its most important output is not a reassuring percentage: it is an explanation of what is accounted for, what remains unresolved and which evidence is needed next.

A pump running for eight hours does not prove that eight hours of useful water reached the process. Equally, process delivery exceeding today's wellhead production is not automatically a meter error: yesterday's stored water or another supply may have contributed. The worksheet keeps production, delivery and storage movement separate so these ordinary operating conditions do not become false performance claims.

Start with one physical boundary and one clock interval; establish each volume from identified evidence; reconcile inventory movement; retain a signed residual; then assign a proportionate investigation. A reconciled production account is not a pumping-test interpretation, a groundwater recharge assessment or permission to increase abstraction.

2. The decision this worksheet owns

This worksheet answers: how much water crossed the selected boundary, how much remained in storage, and why does the documented account differ from the observed result? It supports daily or shift reporting, production reconciliation, meter investigations, transfer reviews and the evidence needed for a subsequent supply-plan decision.

Production and Availability Forecasting addresses future delivery under an approved duty and schedule. Forecast Variance and Production Control explains differences between forecast and actual results. Pumping Test Data Screening checks whether test observations can be interpreted. It does not reproduce the hydrogeological analysis or determine a new operating duty.

Two reconciliations should remain visible. The first compares a runtime-based production estimate with measured wellhead volume. The second follows measured water through transfer, storage and delivery boundaries. Combining the two too early can conceal whether the discrepancy originates in the operating record, the production meter or the downstream account.

3. Draw the physical boundary before entering numbers

Describe every connection crossing the boundary: individual well discharges, purchased water, tanker unloading, treatment outlets, process supply, drains, overflows and external transfers. Identify storage within the boundary. A sketch is useful, but a numbered connection list is sufficient when the installation is simple and the list can be checked on site.

Choose whether the account covers the well-to-tank transfer, the tank alone or the entire well-and-delivery system. An intermediate meter can support diagnosis without becoming an additional inflow. Water measured at a wellhead and again at the tank inlet is the same stream moving through two boundaries, not twice the production. Meter identities and connection direction should make this impossible to overlook.

For recirculation, distinguish a movement wholly inside the boundary from a genuine external return. Internal recirculation is neither new source water nor final delivery in the combined account. A return from a process outside the boundary is an inflow and needs its own evidence. Keep water types and quality status separate where mixing affects the meaning of usable supply.

4. Use one reporting clock and one interval convention

Record start and end timestamps, timezone and whether the report is a shift, operating day or calendar period. Use a half-open interval, from the start inclusive to the end exclusive, when integrating event data. This avoids counting a midnight observation or pump event in two adjacent reports. Keep the original timestamps when creating a corrected analysis view.

Opening and closing totaliser readings must refer to the same interval as storage observations and outlet readings. A 06:00 wellhead reading and a 07:00 tank reading are not directly comparable daily boundaries. Where interval logs permit, calculate the intervening volume explicitly. Otherwise, flag the mismatch and qualify the balance rather than silently assuming the missing hour has no effect.

A timing correction is not a physical loss. Record which meter was shifted, the evidence for the offset, the affected interval and the transformation. Separate a known clock offset from uncertainty about when a photograph was taken. Recurring timing uncertainty is a collection-process problem; multiplying more daily readings together will not resolve it.

5. Establish the minimum evidence set

A usable account needs stable well and meter identities, units, reading times, meter configuration, operating state and the meaning of each volume. Preserve source photographs, logger exports or field sheets. Record whether each result is a direct volume reading, integrated flow, runtime estimate, tank calculation or other estimate. Several methods can coexist, but their evidence status should remain visible.

The well identity must survive pump replacement. The meter identity must change when the instrument changes, even when the display occupies the same panel position. Record the effective time of each configuration change. A new scaling factor, pulse setting or register unit can create a numerical discontinuity without any change in the water supply.

The worksheet's evidence register should describe the actual installation, not require irrelevant fields merely to fill a table. A manual field sheet may not have a logger channel, but it still needs an identifiable observer, time, location and quantity. Missing evidence remains a named gap rather than an invented default.

6. Calculate totaliser volume without hiding resets

For a continuous, correctly scaled cumulative register, period volume is the closing reading minus the opening reading, multiplied by the documented conversion factor. The result inherits the meter's boundary and reading interval. Preserve the displayed register values separately from converted cubic metres so an incorrect unit or multiplier can be investigated later.

V = k × (R_close − R_open)

Here k converts one displayed register unit into cubic metres. If readings are already cubic metres, k is one. A negative difference should trigger review of flow direction, reset, rollover, replacement or data error. Do not apply an absolute-value function simply to make the result positive.

For one confirmed rollover at modulus M, use V = k × (M − R_open + R_close). M is the value at which the register returns to zero, not its highest displayed value. If the count of rollovers is uncertain, the full-period volume is unresolved. For a replacement or reset, calculate each supported segment separately and add the segment volumes. A gap between the final old-meter reading and first new-meter reading remains a gap unless separately supported.

7. Integrate flow over actual water-producing time

For documented interval-average flow, calculate V = sum(Q_i × Δt_i), with flow in m³/h and duration in hours. For instantaneous observations, a trapezoidal estimate uses the average of adjacent flows multiplied by the interval duration, but only where that interpolation represents the operating state. Split intervals at pump stops, rate changes and valve changes instead of smoothing across them.

USGS guidance distinguishes direct volume measurement from rate-time estimation and notes that discharge can vary during operation; multiple flow observations may be needed to represent the period. This supports using state-matched rates rather than multiplying one convenient instantaneous reading by a month of runtime. [R1]

Separate motor-energised time from flow-confirmed time. Starting, priming, a closed discharge path, loss of prime or an instrumentation fault can break a presumed runtime-to-volume relationship. An hour meter is operating evidence, not independent proof of delivered volume. Do not use dry running or unsafe valve operation as a test method; use existing records and approved measurement arrangements.

Where reliable totaliser volume exists, retain the runtime calculation as a cross-check. Do not average the two merely because they differ. They may share a data source or carry different errors, and the discrepancy itself can help identify what needs investigation.

8. Keep planned, available and actual hours separate

The supplied forecasting method uses approved rate, scheduled hours and an explicitly defined availability factor. Actual-period accounting should instead use observed volume or actual supported flow-producing intervals. Multiplying actual runtime by another availability factor usually deducts interruption a second time. Preserve the forecast as a separate comparison, not as an input that reshapes the measured result.

Classify operating time into mutually understood states: producing water, energised without confirmed delivery, tank-full standby, planned maintenance, unavailable power, other confirmed interruption and unknown. Concurrent events need a stated precedence or union-of-intervals rule. A two-hour power failure wholly inside a three-hour maintenance stop does not create five hours of lost operating opportunity.

For several wells, individual runtime totals may exceed the elapsed reporting duration because assets operate simultaneously. That is not an error. The rule preventing time from exceeding the reporting window applies to one asset's non-overlapping state timeline, not the sum of independent well-hours. Conversely, measured combined production does not prove that future simultaneous operation is hydraulically appropriate.

9. Translate tank observations into storage change

Use opening and closing stored volumes on the same physical reference. Define ΔS = S_close − S_open: a positive value means storage increased, and a negative value means stored water supported outflow. The simplest account uses actual water volume within the vessel. Record minimum reserve and inaccessible volume separately from the conservation calculation.

For a rectangular tank with constant internal horizontal area A, ΔS = A × (h_close − h_open). For a vertical cylindrical tank with constant internal diameter D, A = πD²/4. These are geometric relationships, not substitutes for verified internal dimensions or a calibrated tank curve. Nonuniform tanks, horizontal cylinders and irregular sumps require their applicable level-to-volume relationship.

Do not treat a percentage-level signal as a percentage of volume unless its scaling establishes that relationship. Record whether level is measured upward from the floor or downward from a reference. With depth measured downward from the same upper reference, a smaller depth means more water. Review sensor offsets, turbulence and the timing of the level reading before calculating a change.

If usable volume is reported above a fixed reserve, its change equals physical storage change only while the reference and reserve remain fixed and the calculation is not clipped at zero. A revised reserve is a management change, not water entering or leaving the tank.

10. Use a signed water balance with an explicit residual

Choose mutually exclusive external outflow categories. Let W be well inflow, I other imports or external returns, D metered useful delivery, K other documented external outflows and ΔS the change in included storage. K may contain process use not already in D, reject, backwash, drain, overflow or separately measured leakage. Every included movement appears once.

R = W + I − D − K − ΔS

R is the unresolved balance residual, positive for an apparent excess of accounted inflow over accounted outflow and accumulation. Equivalently, S_open + W + I = S_close + D + K + R. This form makes the sign of storage movement easy to check.

A positive residual can be consistent with unrecorded outflow, over-recorded inflow, understated storage rise or overstated storage fall. A negative residual can be consistent with omitted inflow, overstated delivery or the opposite storage errors. These are investigation possibilities, not diagnoses. Never replace a negative residual with zero, and never automatically rename a positive residual as measured leakage.

Define a residual percentage only with a stated denominator. A useful worksheet convention is 100 × R / (W + I) when external inflow is positive. Preserve the signed cubic-metre result even when a percentage is shown. When external inflow is zero, the percentage is not applicable; a storage-only balance can still be meaningful.

11. Distinguish known use, physical loss and accounting difference

Backwash, flushing, sampling, treatment reject and tank cleaning may be legitimate withdrawals, not useful process delivery and not unexplained leakage. Record each where it crosses the selected boundary. An estimated backwash volume remains estimated even when the operation itself is confirmed. A treatment recovery ratio should use compatible feed and product volumes and account for relevant intermediate storage.

EPA water-loss guidance distinguishes physical losses from apparent losses associated with metering and data handling. This is useful context for investigating a discrepancy rather than treating every unaccounted cubic metre as a located leak.

A loss found upstream of the production meter may not appear in the downstream balance. A recirculation flow can inflate a gross-flow reading without increasing net source production. These examples show why a low residual does not prove that the installation has no losses. The account closes only for the boundary, measurements and categories actually used.

12. Reconcile multiple wells, imports and internal transfers

Add measured well volumes over the common interval, keeping each well's evidence and configuration separate. When individual well meters and a common header meter both exist, choose one level as the combined account input and use the other level for reconciliation. Adding both sets inflates production. A common header reading alone does not identify individual contributions.

Where individual allocation is needed but only a common meter exists, distinguish measured total volume from modelled shares. Runtime multiplied by representative rates may support an allocation estimate, but those shares are not independently metered well production. Label the method and uncertainty. Do not conceal uncertainty by allocating the final residual to the largest well.

For a two-tank system, a transfer is an outflow from tank A and an inflow to tank B. It cancels when the system boundary includes both tanks. External tanker receipts, municipal top-up or a process return remain external inputs.

13. Preserve timing and inventory across connected boundaries

A pumping main, treatment vessel or intermediate sump can accumulate or release water between readings. Include material inventory changes inside the selected boundary. A pipeline that remains full with negligible volume change may reasonably have an immaterial inventory term, but document that basis rather than assuming every installation behaves the same way.

Ordinary travel time through a continuously full pipe is not by itself a volume loss. Differences between upstream and downstream totals over a short window can arise from actual accumulation, filling or draining, branches, compressibility, meter timing or measurement error. Do not shift a downstream series by an arbitrary travel time just to force equal totals. Any time alignment must preserve the control-volume interpretation.

For batch treatment or tank filling, close sub-balances at the appropriate event boundaries as well as the reporting boundary. A treatment batch crossing midnight may create a large daily residual when work-in-progress storage is omitted. The monthly total may appear acceptable while daily operating decisions remain misleading. Keep both views when operations depend on the shorter interval.

14. Report useful metrics, not a single percentage

Report metered well production, estimated production separately, external imports, delivered volume, documented other outflows, storage change and the signed residual. Add matched-period data coverage and the evidence status of material terms. A headline ratio without those quantities is difficult to diagnose and easy to overinterpret.

Average producing-period discharge is supported well volume divided by matched flow-producing hours. Dividing the same volume by the full day instead gives a calendar-average supply rate; both can be useful but must have different labels. A delivery-to-well ratio may exceed 100% when imports or storage drawdown support delivery. It should not be labelled pump efficiency or recovery without the relevant boundary definition.

Keep a cumulative residual alongside interval residuals. Positive and negative daily discrepancies can cancel in the monthly total. Show the sum of absolute interval residuals as an additional diagnostic when useful, explicitly noting that it is not a cumulative water-loss estimate. Likewise, record minimum observed storage or a modelled interval minimum: opening and closing inventory alone can hide an intraday reserve breach.

15. Build a production-variance bridge on the same meter boundary

First compare forecast and actual at the same wellhead or delivery boundary. For forecast Q_fH_f and actual average Q_aH_a, the supplied convention allocates rate effect as (Q_a − Q_f)H_f and hours effect as Q_a(H_a − H_f). Their sum equals actual minus forecast volume. This accounting decomposition does not establish the physical cause of either effect.

After the wellhead comparison, separately bridge to downstream delivery using imports, documented outflows, storage movement and the residual. This prevents a tank drawdown from being reported as recovered well performance. Preserve original forecast, revised forecast and actual readings where they answer different management questions.

Do not change the issued forecast retroactively to improve apparent performance. A revised supply plan may be justified, but it receives its own version and effective date.

16. Quantify uncertainty where the evidence supports it

A balance subtracts measured quantities, so its residual can be small relative to uncertainty in much larger input and output volumes. Record uncertainty on period volumes and storage change, not merely the number of decimals displayed. Distinguish standard uncertainty, expanded uncertainty and a conservative engineering bound. A manufacturer's accuracy limit is not automatically a standard uncertainty.

For independent standard-uncertainty contributions, u_R = square root of the sum of the squared uncertainties of W, I, D, K and ΔS. Where inputs are correlated, include their signed covariance contributions. In matrix form, u_R² = cᵀΣc, with coefficients c = [1, 1, −1, −1, −1]. If separate opening and closing storage terms are used, their coefficients are +1 and −1; common level offsets may therefore partly cancel.

Do not apply a universal acceptable residual percentage. Define the investigation criterion from measurement capability, operating consequence, account completeness and the decision margin. A result compatible with the stated uncertainty is not proof of zero leakage. A result exceeding a justified criterion is a reason to investigate, not proof of a particular cause.

17. Keep incomplete periods incomplete

A full-period totaliser difference can establish total volume despite a gap in instantaneous flow logging, provided the register remained valid and both boundary readings are supported. In that situation, total volume and intraday profile have different completeness states. Do not discard valid volume merely because the trend has missing points; equally, do not claim that the missing trend was observed.

If a reset occurs during a data gap, the full-period total may be unresolved. Report supported segment volumes and identify the missing interval. A bounded estimate may be added as a separate scenario using documented possible flow and duration, but it must not silently become measured production. Zero flow with corroborated pump-off state differs from zero transmitted during a communications failure.

For balance reporting, define a matched interval where all essential terms are supported, or qualify the whole-period result with explicit estimates. Keep record completeness, time coverage and quantity completeness separate. Missing a brief high-flow transfer may matter more than missing many repeated low-flow records.

18. Investigate the discrepancy where it first appears

Begin with identity, interval, units and totaliser events. Then compare runtime and operating-state evidence. Next reconcile storage, external sources and known withdrawals. Only after those checks should the account be used to prioritise physical inspection or a deeper well-performance investigation. The supplied variance guidance follows this evidence-first approach rather than assuming groundwater decline.

Assign each material exception to an owner with a specific discriminating check. A meter discrepancy needs measurement evidence; a transfer concern needs a boundary check; a suspected performance change needs comparable water level, head, flow and equipment conditions. Reconcile the following period after corrective work, without deleting the original account.

19. Worked example A: a daily balance with storage accumulation

All examples are synthetic calculation cases, not customer measurements or approved operating duties. A well totaliser rises from 12,450 to 12,546 m³ across one 24-hour interval, so measured well production is 96 m³. A separately recorded tanker adds 12 m³. Metered process delivery is 88 m³, and documented flushing uses 5 m³. Tank volume rises from 30 to 43 m³.

The storage change is +13 m³. Total external inflow is 108 m³. Applying the worksheet equation gives R = 96 + 12 − 88 − 5 − 13 = +2 m³. The residual is 1.85% of external inflow. The statement supported by the arithmetic is an unresolved positive balance difference of 2 m³, not a measured 2 m³ leak.

The closing-volume cross-check is 30 + 96 + 12 − 88 − 5 − 2 = 43 m³. If an independent operating record shows eight flow-producing hours at a representative 12 m³/h, its 96 m³ estimate agrees numerically. Agreement helps the review, but it does not validate both methods if the representative rate was itself derived from the same totaliser.

20. Worked example B: electricity-on time is not delivered-water time

A ten-hour operating window contains six hours of confirmed water production at 12 m³/h, two hours at 9 m³/h, half an hour with the motor energised but delivery confirmed as zero, and 1.5 hours stopped. The interval-based production estimate is (12 × 6) + (9 × 2) = 90 m³. Water-producing time is eight hours; motor-energised time is 8.5 hours.

Multiplying 8.5 hours by a 12 m³/h reference gives 102 m³, overstating the supported estimate by 12 m³. Multiplying the full ten-hour window by that rate gives 120 m³. Neither shortcut represents the stated operating record. Suppose a valid wellhead totaliser reports 89 m³: the difference against the state-based estimate is −1 m³, or −1.11% of the 90 m³ estimate.

The zero credited to the no-delivery interval is an explicit, confirmed condition in this example; in a real record with unknown delivery it would remain unknown. The measured 89 m³ is not increased to 90 to make the table close. Investigate the difference using rate representativeness, meter uncertainty and operating evidence, and retain both results with their status.

21. Worked example C: two wells and a storage drawdown

During a common interval, well A supplies 60 m³, well B supplies 45 m³ and an external source adds 15 m³. Tank A opens at 40 m³, receives the two wells' combined 105 m³, transfers 110 m³ to tank B and has a documented 5 m³ drain. It closes at 30 m³, satisfying 40 + 105 − 110 − 5 = 30.

Tank B opens at 20 m³, receives the 110 m³ transfer and the external 15 m³, supplies 140 m³ to the process and closes at 5 m³. Its account also closes: 20 + 110 + 15 − 140 = 5. In the combined boundary, storage falls from 60 to 35 m³, so ΔS = −25 m³.

Combined R = 105 + 15 − 140 − 5 − (−25) = 0 m³

The internal 110 m³ transfer does not appear in the combined external-flow account. External inflow is 120 m³; delivery is 140 m³ because storage supplies part of the difference. Delivery divided by well production is 133.33%, but this is not a production efficiency. It reflects other source water and inventory use.

A zero residual establishes arithmetic closure for the supplied figures. It does not prove meter accuracy, suitable water quality or sustainable future well duty. If tank B has an independently specified reserve above 5 m³, that operating concern remains even though the volume balance closes. Inventory sufficiency and accounting closure are separate outputs.

22. Worked example D: rollover, missing production and a residual range

A well meter counts in cubic metres and has a confirmed modulus of 100,000. Its opening reading is 99,980 and closing reading is 35. One documented rollover gives 100,000 − 99,980 + 35 = 55 m³. This result would be unsupported if an unknown reset or multiple rollover could have occurred during the same interval.

Suppose those 55 m³ cover only the supported part of a longer reporting period. A separate one-hour interval has no volume reading, but documented operating evidence bounds its possible discharge between 8 and 12 m³. Whole-period well production is therefore a scenario range of 63–67 m³, comprising 55 measured and 8–12 estimated. It is not a measured 65 m³ result.

For the full interval, delivery is 50 m³, documented other outflows are 5 m³ and storage rises 10 m³, with no other imports. The residual range is 63−50−5−10 = −2 m³ to 67−50−5−10 = +2 m³. Its span includes zero, but the missing observation has not been resolved.

Separately, consider a complete-data balance with R = 2 m³ and independent standard uncertainties of 1.0 m³ on inflow, 0.8 on delivery, 0.3 on other outflows and 0.6 on storage change. The combined standard uncertainty is √(1² + 0.8² + 0.3² + 0.6²) = 1.446 m³. An explicitly selected coverage factor of two gives 2.891 m³; the approximate displayed interval is −0.891 to +4.891 m³. No confidence percentage is asserted without the distributional basis. This uncertainty calculation is separate from the missing-data scenario and should not be added to it without a defined model.

23. Editable worksheet: definition, movements and reconciliation

Use the following three tables as manually editable working sheets. Duplicate movement rows as necessary. They contain no automatic Word calculations. Attach source evidence through stable references rather than pasting unrelated photographs into every row. Enter a reason for exclusions and keep original and corrected values distinguishable.

Worksheet A fixes the meaning of the report before numbers are entered. It should identify the actual connection boundary, all included storage and the responsible reviewer. A change in boundary or time convention requires a revised record, not an unexplained change to the total.

Worksheet B records each independent boundary movement once. A paired upstream/downstream meter belongs to a diagnostic comparison, not automatically to two inflow rows. For an estimated term, record the estimation method and affected interval. A blank is not zero.

Worksheet C preserves the result and what it can support. The reviewer should be able to reproduce the equation from Worksheet B and identify any dependency that makes two apparently independent quantities circular.

24. Issue an account that the next shift can use

The report should show the major volumes, storage movement, residual and data status on its first result panel. Put the evidence ledger and calculation trace behind that summary without hiding material limitations. A concise conclusion might state that well production is metered, tank delivery is matched-period, storage increased, and an identified residual requires a timing check.

Separate arithmetic validation from field acceptance. Reproducing a sum confirms the calculation, not meter accuracy or the completeness of the physical connection list. A reviewer should check one source-to-result chain, inspect the meaning of material estimates and verify that any unresolved condition appears in the issued conclusion. Do not mark a measured loss as recovered merely because the account was edited.

Carry outstanding exceptions into the next reporting period. After meter replacement, tank-reference correction or a changed operating arrangement, record the new configuration and preserve the old account. A trend across the change should identify it. The final operational question is whether the evidence supports a specific action: improve measurement, inspect a transfer, reconcile a known withdrawal, revise supply planning or investigate comparable well performance.

For a review request, provide current operating instructions, meter readings and units, pump-state logs, tank geometry or volume curves, alternative-source records and the unresolved difference. Do not change pumping controls or extend abstraction solely to eliminate an accounting shortfall.

25. Questions operators and project teams ask

Is pump runtime enough to calculate production?

It can support an estimate when the discharge during that runtime is adequately represented. An energised motor does not always mean useful delivery. Prefer supported volume readings where available, and retain runtime as a separate check rather than quietly substituting a rated flow.

Why can tank delivery exceed wellhead production?

Storage may fall, another supply may contribute or water may return from outside the boundary. Identify those terms before questioning the meters. A delivery-to-well ratio above 100% is not automatically an error and is not pump efficiency.

Does a positive residual prove a water leak?

No. It identifies a difference in the defined account. Missing outflows, meter bias, incompatible intervals and storage error are among possible explanations. Physical leakage needs supporting investigation; retain the result as unresolved until evidence distinguishes the cause.

What does a negative residual mean?

Accounted outflow plus accumulation exceeds accounted inflow under the chosen sign convention. Investigate omitted imports, excessive recorded delivery, inventory errors and timing. Keep the sign visible. Taking an absolute value loses information that can help locate the problem.

Should a full tank be counted as today's production?

No. Opening inventory belongs to stored water already present at the start. The account uses the change in storage, not its entire capacity as new inflow. Tank reserve and usable operating range remain separate service-planning information.

Can a rollover be corrected automatically?

Only when the register modulus, direction and rollover count are supported and a reset or replacement is excluded. Otherwise request review. Preserving a negative raw difference is preferable to inventing an apparently plausible positive volume.

Can a complete totaliser account coexist with missing flow data?

Yes. Valid opening and closing cumulative readings can support the period total while intraday flow remains incomplete. Report both states. The cumulative readings cannot reveal the timing of an unobserved high-flow event or interruption.

Can several wells be added together?

Their measured volumes can be summed across compatible periods and boundaries. Forecasting future simultaneous performance is different and requires an assessed operating arrangement. A common header meter must not be added again to the individual well totals.

What residual percentage is acceptable?

There is no universal threshold in this worksheet. The criterion should reflect uncertainty, completeness, consequence and the decision being made. A small percentage can still matter commercially or operationally, and a larger one can result from poor time alignment.

Does a balanced account prove sustainable yield?

No. It reconciles the recorded movements and inventory for the stated period. Sustainable operating duty, water quality, pump suitability and applicable permissions require their own evidence and responsible decisions. Arithmetic closure cannot supply missing hydrogeological information.

26. Calculation logic, validation and interface behaviour

Implement separate input objects for assets, meters, reporting windows, movements, storage observations and exceptions. Every movement needs a stable record ID, boundary role, direction, unit, interval, source and quality state. Store numeric values independently of presentation formatting. Preserve original entries, conversion rules and correction revisions so an export can reproduce the reported account.

Use cubic metres and a consistent timestamp basis internally. Support direct period volume, segmented totaliser differences and flow integration as separate modes. A register correction requires its event evidence. A missing essential term blocks an unqualified whole-period balance but may permit a labelled partial account or range.

Compute matched intervals, then sum external inflows and mutually exclusive outflows. Calculate storage change with the documented curve or directly entered supported volumes. Exclude paired internal movements from the combined boundary. Return the signed residual, its declared percentage, evidence composition and all unresolved flags. Numerical zero must never imply approved operating duty or no physical leakage.

For input checks, reject non-finite numbers and invalid timestamps; distinguish missing from zero; review negative cumulative increments; reject unresolved unit changes; detect overlapping records for the same source and double-counted parent/child meters. Permit signed net-flow measurements only in an explicit directional mode. Require nonnegative physical tank volumes within documented capacity, while allowing negative storage change. Do not silently clamp a prediction at a tank limit; show overflow, depletion or an invalid assumption.

The interface should offer boundary definition, reading import, storage, reconciliation and exception views. On mobile, use readable result cards and local table scrolling; keep labels and units attached to values. Keyboard users need labelled fields, visible focus and error messages linked to the affected input. Export the ledger, formula version, quality dictionary and report revision. Do not expose customer well coordinates, photographs or operating records publicly through an example dataset.