Contract Capacity vs.
Connected Load: The Two Numbers Owners Confuse

A facility can have more equipment connected than it normally uses at once. It can also have a low electricity-bill peak without having permission or physical capability to add the next process. An energy decision goes wrong when either observation is treated as a complete answer about available power.

Section 01Define the number before using it

Connected load describes the equipment attached to the electrical system, using an explicitly stated rating basis. Measured demand describes what the site actually drew during particular intervals. Documented service capacity describes what the applicable utility records establish about service at a defined point, with whatever operating conditions those records contain. These belong in separate lines of the decision memo.

“Contract capacity” is shorthand in this paper for that documented service capability and its conditions. It is not presented as a universal tariff term, a promise of uninterrupted supply, or a standard field on every California bill. An owner's agreement may use different language. Quote that language and identify the document rather than translating it into a stronger claim.

The distinction is practical. An equipment schedule tells a reviewer what could operate. Meter records show what did operate. Utility correspondence, applicable service documents, and engineering review establish what must be confirmed before the proposed operation proceeds. None can silently substitute for the others. This paper offers a decision framework, not a service determination or an electrical design.

Section 02Keep the units and boundaries attached

Power and energy answer different questions. Kilowatts describe active power; kilowatt-hours describe energy over time. Apparent power, expressed in kilovolt-amperes, also reflects reactive power. An amperage label cannot be converted into a usable active-power limit without the relevant electrical conditions. EIA's electricity glossary distinguishes demand, coincident demand, peak demand, and apparent power.¹

Keep the location beside the number. A main meter, a process feeder, an individual motor, and an upstream utility circuit are different boundaries. A transformer rating is an equipment characteristic; it is not, by itself, a customer-specific service commitment. A building's spare panel positions tell even less about upstream capability. Avoid presenting any of these as “available power” without qualification.

A useful working table has a value, unit, measurement point, interval, operating condition, source, and date for each entry. Where the value is unknown, write unknown. Do not conceal a missing service confirmation behind a precise engineering unit. Precision in notation does not repair an unsupported assumption.

Section 03Connected equipment is an inventory, not a load shape

A connected-load inventory needs enough detail to distinguish normal production from standby equipment, seasonal equipment, maintenance spares, and planned additions. Record whether a rating describes electrical input, mechanical output, or an assembly that already includes its auxiliaries. Counting both an assembly and its internal components can overstate the total.

Then examine coincidence. Independent equipment maximums may occur at different times. A shutdown-and-restart sequence may produce a different combination than steady production. Spare equipment may be interlocked against simultaneous use, or operators may be able to start it manually. The inventory should say which case applies, with the supporting drawing, operating record, or approved procedure.

Do not apply an arbitrary diversity percentage to make the project fit. Historical diversity is evidence about past operation, not automatic evidence about an expanded production plan. If the proposed business depends on new equipment running together, the analysis must represent that operation even when the existing meter history never shows it. Qualified electrical professionals determine the applicable design and safety requirements.

Section 04The billed peak is evidence about a specific past

PG&E's Time-of-Use demand-charge FAQ, checked on August 30, 2026, describes billing based on the highest-use 15-minute interval in the billing month. Its explanation also distinguishes measured demand from connected load in the agricultural context.² This is a sourced example of a billing measurement, not a universal rule for every account. The actual rate schedule and metering arrangement control.

A monthly peak omits much of the operating story. It does not show how long high demand persisted, whether the facility was underused, or whether on-site generation reduced utility purchases. Ask what was running, what was unavailable, and whether the period represents the intended future business. A quiet production year is a weak basis for an expansion.

Interval data are more useful when paired with an operating calendar. Mark shifts, shutdowns, seasonal production, unusual weather, and equipment changes. Separate observed records from forecasts. Also keep billing analysis separate from short-duration electrical behavior: an interval average does not replace evaluation of starting events, protection, or power quality.

Section 05A service confirmation needs a defined scope

PG&E's commercial project guidance asks for project scope, equipment and load information, drawings, and timing. Utility review may determine that no infrastructure upgrade is needed or may produce a design and project-contract documents. The same guidance describes its hosting-capacity maps as illustrative and dynamic.³ A map is a planning input, not a site-specific confirmation.

For an owner-side review, assemble the current service agreement or equivalent records, approved drawings available to the owner, recent utility correspondence, and the proposed equipment schedule. Ask the utility to clarify which documented conditions apply to the proposed operation. Do not assume that a historic drawing, preliminary discussion, or neighboring property's experience remains applicable.

Separate current service from an expansion request. Record what is existing, what is proposed, what depends on unfinished work, and what remains unanswered. An equipment delivery date does not establish an energization date. A preliminary capacity discussion does not establish final service terms. Keep unresolved utility matters visible in the decision instead of burying them in a model footnote.

Section 06Use arithmetic that exposes the missing evidence

Illustrative example only, prepared August 30, 2026: assume an invented facility has 900 kW of connected electrical input ratings, a representative measured import peak of 540 kW, and an assumed documented import limit of 700 kW. A proposed process is expected to add 120 kW during that same peak interval. These are teaching assumptions, not a customer site, utility offer, equipment selection, or design allowance.

The 40 kW result does not establish that the addition can proceed. Other loads may grow, auxiliaries may have been omitted, and the historic peak may not represent the new operating schedule. The assumed service limit may contain conditions. Customer-side equipment also needs review. Calling the difference “spare capacity” would hide all of those dependencies.

Now change one assumption: suppose the earlier 540 kW import peak occurred while 100 kW of on-site generation served the load. Ignoring losses for this illustration, the corresponding facility demand was 640 kW. With that generator unavailable and the new process operating, import would be 760 kW. The same bill now supports a very different question. The model must say whether it represents net imports or total site demand.

Section 07Generation changes the balance, not the paperwork by itself

For a defined interval and electrical boundary, net utility import equals facility consumption plus storage charging, less on-site generation and storage discharge. A negative result represents net export. Treat losses consistently and do not count auxiliaries twice. The accounting is straightforward; establishing credible values for each operating condition is the work.

Generator nameplate output is not interchangeable with dependable net contribution. EIA distinguishes nameplate capacity from tested summer and winter capability measured at the interconnection point.⁴ Maintenance, temperature, fuel conditions, and auxiliary demand should therefore remain explicit screening questions, not hidden adjustments to an optimistic number.

Import and export also need separate answers. PG&E's distribution-generation guidance warns that nearby infrastructure may not have capacity to receive proposed generation.⁵ CPUC's Rule 21 overview explicitly includes non-export facilities within its described applicability.⁶ Neither an import arrangement nor a zero-export intention alone establishes permission to operate generation. Confirm the applicable requirements through the responsible utility and qualified advisors.

Section 08Compare paths against the same boundary

The following table is Bcal's qualitative screening framework, not a technology ranking or a performance forecast. Each path must address the same proposed load, meter boundary, operating conditions, and owner objective. DOE explains why solar timing and storage power and energy limits matter; EPA's technical catalog documents the different operating characteristics of established CHP technologies.⁷⁸ Older technical references support general principles, not current equipment prices or supplier commitments.

A combined path must survive the same review. A battery does not remove a generator's maintenance needs, and a generator does not remove a battery's recharge requirement. Avoid crediting the same flexible load twice. If one assumption supports several benefits, show that dependency so the owner can see what changes when it fails.

Section 09Test the operating states that can reverse the answer

Normal production is only one state. Review the proposed expansion during the site's demanding season, with a generating unit unavailable, after storage has discharged, and during restoration after a shutdown. These are analytical cases to define with the owner and qualified reviewers, not a universal design checklist. Select them because they could change the decision.

For any import-limiting strategy, identify the loads that can actually be reduced, the decision authority to reduce them, and the consequence for the business. A control strategy that fits an electrical limit by stopping essential production may be technically interesting and commercially unacceptable. The decision memo should preserve that trade-off in plain language.

Check whether a proposed operational restriction is documented, enforceable in the relevant equipment, and acceptable under the actual service arrangement. Do not treat an operator's promise to “be careful” as equivalent to a reviewed control scheme. Conversely, do not assume every new load requires generation when credible sequencing or a clarified utility answer could resolve the issue.

Section 10Give the owner a capacity record that can be maintained

The useful output is a short reconciliation with an evidence trail. It should separate established facts, illustrative scenarios, unanswered questions, and actions needed before commitment. The owner should be able to understand why the proposed operation fits, why it does not yet fit, or why available evidence cannot answer the question.

1. Identify the site and decision.Record the meter boundary, site controller, proposed business change, intended operating date, and the person accountable for the decision.
2. Reconcile equipment and operations.Keep connected input ratings, measured import, on-site supply, and future coincidence assumptions separate. Record data dates and missing operating periods.
3. Preserve the utility's exact wording.Identify current documents, stated conditions, unfinished work, and the specific written clarification still required. Do not promote a planning map into approval.
4. Compare credible paths and failure states.Include utility service, load measures, suitable storage and generation, phasing, and no project. Record what would invalidate each path.
5. Assign the next evidence step.Name the unresolved item, its owner, and the event that triggers review. Keep construction, protection, contractual interpretation, and final design with the appropriate qualified parties.

Update this record when equipment, occupancy, production, service documents, or on-site supply changes. A prior conclusion belongs to its stated conditions. If those conditions no longer hold, reopen the question before reusing the answer in an expansion budget or a supplier discussion.

The owner does not need the largest possible number. The owner needs a defensible account of what the facility will require, what the service documents establish, what each proposed path contributes, and which uncertainty still matters. Keeping those statements separate is what makes the eventual decision usable.