I'm analyzing Arizona Public Service interval data for a large Schedule E-32 manufacturing customer to challenge their coincident demand billing. APS charges based on the customer's demand during the system peak hour, but their interval data shows our client's peak doesn't align with the APS system peak. Customer peaked at 2,850 kW at 2:30 PM but APS system peak was at 4:45 PM when our load was only 1,920 kW. The coincident demand charge is $847,000 annually based on the wrong methodology. Has anyone successfully challenged APS coincident demand calculations using detailed interval analysis?
APS demand coincident factor analysis using interval data
Sarah, that's a huge discrepancy! TEP here in Tucson has similar coincident demand provisions and we've found success challenging them when the utility can't properly document their system peak methodology. The key is getting APS to provide their actual system load data for the billing period and proving when their true peak occurred. Sometimes utilities use estimated or averaged system peaks instead of actual real-time data, which can shift the coincident demand calculation significantly.
We had a similar case with NorthWestern Energy where their coincident peak methodology was wrong. The tariff specified demand during the "system peak hour" but they were using a monthly average of daily peaks instead of the single highest system demand hour. Took six months of back-and-forth but we eventually got $340,000 in refunds for the client. Document everything about how APS defines and calculates their system peak versus what they're actually billing.
Sarah, make sure you're looking at the correct coincident peak methodology in the APS tariff. Dominion Energy here in Virginia has different definitions for transmission-level coincident peaks versus distribution-level peaks. Some customers get billed based on regional transmission organization (RTO) peaks while others use local distribution system peaks. The timing can be completely different and the financial impact is exactly what you're seeing - hundreds of thousands per year.
This reminds me of a case with BGE where they were using PJM transmission peak data for coincident billing instead of their local distribution peak. The PJM regional peak occurred at a different time than the BGE local peak, creating exactly the situation you're describing. Customer demand was low during the PJM peak but high during local peak periods. We got BGE to recalculate using the correct peak definition and saved the client about $180,000 annually in coincident demand charges.
Great insights everyone. I've requested the APS system load data and their specific methodology for identifying the monthly system peak hour. The Schedule E-32 tariff language is ambiguous about whether they use transmission-level CAISO peaks or local APS distribution peaks. If they're using CAISO regional data instead of local APS system data, that could explain the timing discrepancy and give us grounds for a billing correction.
Sarah, also check if APS is properly adjusting for transmission losses when calculating coincident demand. MLGW had an issue where they were applying loss factors incorrectly for transmission-level coincident peaks, which shifted the effective peak timing and inflated customer demand charges. The coincident peak methodology should account for line losses between the generation source and customer location, especially for large industrial loads.
Out here with Rocky Mountain Power, we learned that coincident demand calculations can vary significantly based on how the utility defines their "system." Some use generation peaks, others use transmission import peaks, and some use local distribution peaks. Each methodology produces different peak timing and different customer coincident demand charges. The tariff language usually specifies which system they're supposed to use, but billing departments don't always implement it correctly.
Connie makes a good point about system definition. Georgia Power down here in Savannah has had similar issues where their billing system was programmed to use the wrong peak identification methodology. Schedule PL customers were getting charged based on Southern Company regional peaks instead of local Georgia Power system peaks. The timing difference was significant - sometimes 2-3 hours apart - creating massive billing errors for large industrial customers.
Sarah, definitely push for the raw system load data from APS. Duke Energy here in Charlotte initially claimed they couldn't provide system peak data due to "confidentiality" but eventually released it after we escalated through regulatory channels. The data showed their billing system was using estimated peaks instead of actual real-time system data. Once we proved the discrepancy, Duke corrected the methodology and issued significant refunds to affected customers.
Update: APS provided their system load data and confirmed they're using CAISO regional transmission peaks instead of local APS distribution system peaks for coincident demand billing. This appears to contradict their Schedule E-32 tariff language which specifies "APS system peak demand." I've drafted a formal billing dispute citing the tariff language and requesting recalculation using actual APS local system peaks. The potential refund could be substantial given the timing differences we've identified.
Excellent work Sarah! This is exactly the kind of detailed tariff analysis that saves clients serious money. Black Hills Energy up here had a similar tariff interpretation issue where "system peak" was being applied incorrectly. The regulatory commission ultimately ruled in favor of the customer interpretation and required the utility to use local system data instead of regional transmission data. Document everything and be prepared to escalate to the Arizona Corporation Commission if APS won't voluntarily correct their methodology.
This case highlights why interval data analysis is so critical for large demand customers. The difference between regional transmission peaks and local distribution peaks can be enormous - both in timing and financial impact. Dominion Energy had to revise their coincident demand methodology for several large industrial customers after we proved their billing system was using the wrong peak identification criteria. Keep us posted on how APS responds to your dispute.
Sarah, this is an outstanding example of how detailed interval data analysis can uncover major billing errors. The $847,000 annual impact makes this worth fighting all the way to the commission if necessary. Here in Memphis, we've seen similar cases where utilities misinterpret their own tariff language around coincident demand calculations. The key is proving that their billing methodology doesn't match the tariff requirements. APS should voluntarily correct this once they realize the tariff compliance issue, but be prepared for a long fight if they dig in their heels.