Quick answer
Commercial solar can reduce demand charges when it lowers grid demand during the intervals that determine your bill. Lower monthly electricity purchases alone do not prove this. Evening peaks, cloudy intervals, historical-demand ratchets and minimum billing demand can limit savings. Before accepting a forecast, require separate calculations for kWh energy savings and kW billing-demand reductions. (nlr.gov)
This guide addresses U.S. commercial electricity bills. The utility examples illustrate particular rules—not terms that apply nationwide. Use the tariff and supply agreement for your account.
1. Separate energy savings from demand savings
Energy, measured in kilowatt-hours (kWh), is electricity used over time. Demand, measured in kilowatts (kW), describes the rate of electricity use. Commercial bills can charge for both, alongside fixed fees and other adjustments. (samrepo.nlr.gov)
A simple hypothetical calculation shows the distinction:
- Equipment drawing 100 kW for one hour uses 100 kWh.
- Equipment drawing 25 kW for four hours also uses 100 kWh.
- The energy totals match, but the demand levels differ.
For a straightforward tariff:
Energy charge = purchased kWh × applicable $/kWh
Demand charge = billing-demand kW × applicable $/kW
Actual schedules may have multiple time periods, tiers or demand components. Billing demand is the tariff-defined quantity used to calculate the charge; it need not equal the highest demand measured in the current month. It may incorporate a time-of-use peak or a historical lookback. (samrepo.nlr.gov)
Ask bidders to show energy and demand savings separately. A proposal based only on annual solar production and your average bill cost per kWh does not demonstrate a reduction in billing demand.
2. Read the rules that determine your billed peak
Start with the exact schedule code on your bill. Then locate its demand definition, measurement interval, time periods and minimums.
Measurement intervals matter. PG&E’s published B-19 schedule defines maximum demand for affected accounts using the highest 15-minute average during the billing month. It also specifies exceptions, including possible five-minute measurement for intermittent or severely fluctuating use. A monthly total cannot reveal those peaks. (pge.com)
Three billing concepts deserve particular attention:
| Rule | What determines the charge | What to investigate |
|---|---|---|
| Non-coincident demand | Your account’s maximum demand, regardless of when the wider grid peaks | Does solar reduce every interval that could become the monthly maximum? |
| Time-period demand | Your maximum within a specified window | Does solar produce during that entire window? |
| Demand ratchet | A floor tied to historical demand | How much measured reduction translates into lower billing demand? |
Research finds that solar’s demand-charge benefit depends strongly on this design, particularly the overlap between production and the relevant peak window. (research-hub.nlr.gov)
A real ratchet example appears in Tucson Electric Power’s published Large General Service tariff. Monthly billing demand is the greatest of:
- The highest measured 15-minute demand during the billing period;
- 75% of the greatest demand used for billing in the preceding 11 months;
- Contract capacity or 200 kW, whichever is greater.
That tariff sheet is effective September 1, 2023 and remained linked from TEP’s LGS page when checked on October 10, 2026. Lower measured demand may therefore leave billing demand unchanged. (tep.com)
Coincident peaks are different again. They concern demand when the broader system peaks, rather than necessarily when your building peaks. PJM’s Manual 19, effective June 30, 2026, describes five summer coincident peaks used in electricity-distribution-company-specific capacity allocation methods. Whether those obligations affect your retail charges depends on your utility and supply contract. (learn.pjm.com)
3. Check whether solar overlaps the peaks that matter
A daytime office is a more promising candidate for solar-only demand reduction than a facility whose maximum occurs after sunset. But the building’s label is not sufficient evidence: its actual load pattern matters. NREL’s commercial-building guidance also identifies diminishing returns when solar reduces midday demand enough for the maximum to move into the evening. (nlr.gov)
Consequently, do not examine only the original peak. After subtracting solar generation, another interval may become the highest remaining grid demand.
Tariff windows can also extend beyond strong production hours. PG&E’s published B-19 schedule defines its peak period as 4 p.m. to 9 p.m. every day, including weekends and holidays. A forecast showing substantial noon generation does not, by itself, establish savings during that window. (pge.com)
Request two plots from the analyst:
- Facility demand and solar output on the same timeline;
- Remaining grid demand, with the highest intervals marked.
Include several high-demand days—not just a clear summer day. The question is not “Does solar generate a lot?” It is “Does solar reduce the demand quantities this account is charged for?”
4. Use interval data to calculate the bill twice
Monthly bills establish charges and tariff codes. Time-stamped interval data establishes when demand occurs. SAM’s official documentation supports hourly and subhourly load inputs and warns that scaling a generic load profile to monthly kWh does not accurately represent commercial peak-demand data. (samrepo.nlr.gov)
Use this review process:
- Collect the baseline. Request at least a full year of bills and interval data, plus enough earlier history to cover any ratchet lookback.
- Check data quality. Confirm units, meter identifiers, missing readings, billing dates, time zone and daylight-saving treatment.
- Match the billing resolution. For a 15-minute demand rule, require analysis capable of representing those intervals. Hourly averages can conceal a shorter peak.
- Align solar and load. Require site-specific AC solar output on matching timestamps. Ask how weather, shading and output variability are represented.
- Apply every demand rule. Recalculate all-hours peaks, time-period peaks, ratchets and floors—not just one demand line.
- Compare complete bills. Show baseline and solar-case charges month by month, with energy, demand, fixed charges and export credits separated.
SAM provides monthly bill-component tables and demand results by time period, making that separation possible. (samrepo.nlr.gov)
Treat this as a validation exercise, not merely a software run. Reproduce historical charges before trusting projected savings. SAM’s rate documentation specifically warns that imported utility-rate data can contain errors, especially for ratcheting demand rates. (samrepo.nlr.gov)
5. Worked example: daytime office versus evening-peaking facility
The following is invented arithmetic for illustration, not measured performance, a project quotation or a forecast for either utility above.
Assume two hypothetical U.S. facilities, each with:
- Baseline monthly grid purchases of 50,000 kWh;
- Baseline maximum demand of 200 kW;
- A 100 kW DC solar array with an 80 kW AC inverter limit;
- Assumed monthly solar production of 12,000 kWh, entirely consumed onsite;
- Flat energy pricing of $0.12/kWh;
- One monthly, all-hours demand charge of $20/kW, measured over 15 minutes;
- No ratchet, demand floor, export credits or tariff change.
The assumed solar output is an input to the example—not an estimate established by location or weather data.
| Illustrative result | Daytime office | Evening-peaking facility |
|---|---|---|
| Demand during original peak | 200 kW | 200 kW |
| Solar output during that interval | 80 kW | 0 kW |
| Grid demand during original peak | 120 kW | 200 kW |
| Highest grid demand anywhere after solar | 160 kW | 200 kW |
| Reduction in billing demand | 40 kW | 0 kW |
For the office, assume another interval becomes the new maximum at 160 kW, and all other post-solar intervals are at or below it. For the evening facility, assume solar reduces daytime purchases but leaves its after-sunset maximum unchanged.
Both facilities’ energy-charge savings:
12,000 kWh × $0.12/kWh = $1,440
Office demand-charge savings:
(200 − 160) kW × $20/kW = $800
Evening facility demand-charge savings:
(200 − 200) kW × $20/kW = $0
Total illustrative monthly bill savings are therefore $2,240 for the office and $1,440 for the evening facility.
Notice that the office’s billing-demand reduction is 40 kW—not the 80 kW of solar output during its original peak. This example deliberately accounts for the replacement peak.
If a separate hypothetical tariff imposed a 180 kW historical-demand floor, the office’s billed reduction would be only 20 kW, worth $400. Neither calculation includes capital cost, maintenance, financing or incentives, so neither establishes project payback.
6. Catch common mistakes before approving the forecast
Applying an energy-saving percentage to demand charges.
Require a monthly demand calculation. A reduction in purchased kWh is not a matching reduction in billed kW. (samrepo.nlr.gov)
Using a generic profile as evidence of your peak timing.
Ask which measured intervals support the conclusion. Matching monthly consumption does not establish an accurate peak profile. (samrepo.nlr.gov)
Treating midday output as a reduction in an evening-window peak.
Overlay production on the tariff’s actual hours. Check weekends and holidays where applicable. (pge.com)
Ignoring the new maximum after solar.
Ask for the highest remaining intervals, including low-solar periods. Increasing array size can eventually move the controlling peak outside useful production hours. (nlr.gov)
Assuming software has captured every tariff provision.
Compare inputs with the published schedule. Flag unsupported billing features, rate changes and differences between the modeled calendar and the actual account. (samrepo.nlr.gov)
Use a conservative demand-savings scenario alongside the central forecast. If the financial case depends heavily on peak reduction, ask which plausible operating or production conditions would erase it.
7. Turn the analysis into a procurement checklist
Before accepting demand-charge savings, require written answers to these questions:
- Which utility territory, tariff version and supply agreement were modeled?
- What interval establishes each demand charge?
- What are the current measured demand and billed demand?
- Which ratchets, contract minimums and floors apply?
- When do the highest original and post-solar intervals occur?
- How much of the forecast is energy savings versus demand savings?
- Were historical bills reproduced, and were discrepancies explained?
- How would new equipment, shift changes or electric-vehicle charging alter the load?
- Who will compare actual bills with the forecast after commissioning?
If solar misses the controlling peak, compare operational load management or solar-plus-storage rather than forcing a solar-only demand benefit into the model. Research supports potential demand-charge benefits from combining PV and storage, but results remain dependent on load and charge design—not a universal savings guarantee. (research-hub.nlr.gov)
The decision rule: credit solar with energy savings under the applicable energy-billing rules. Credit it with demand savings only after a tariff-aware interval analysis demonstrates a lower billable peak.