Portland General Electric just launched their new Schedule 83 TOU rate for large commercial customers. I'm analyzing 24 months of 15-minute interval data for a data center client to see if they should switch from Schedule 85 demand-based pricing. The TOU periods are 6am-10pm weekdays for on-peak, with summer super-peak 2pm-7pm. Anyone have experience modeling PGE's TOU rates? The rate differential is significant - $0.087/kWh on-peak vs $0.041/kWh off-peak.
TOU rate optimization using load profile analysis
Wendy, I just finished a similar analysis for We Energies here in Milwaukee. Key thing is to look at load factor by time period, not just total consumption. Data centers usually have flat load profiles which might not benefit as much from TOU rates. What's your client's load factor during those super-peak hours? If it's above 80%, you might want to stick with demand billing.
Mia's right about load factor. CPS Energy down here has similar TOU structure and I've seen data centers actually pay more on TOU because of their consistent baseload. Have you factored in the demand charges on Schedule 83? PGE still has coincident peak demand even on TOU rates. Pull their system peak data and see how often your client's peak aligns with grid peak.
Jorge makes a good point about coincident peak. Here in Memphis with MLGW, we see summer system peaks around 3-4pm. If your data center is running chillers during those hours, the coincident peak charge could wipe out any TOU savings. I'd recommend running a 12-month projection using actual interval data mapped to both rate schedules.
Good insights everyone. Client's load factor is actually 95% year-round, very flat profile. Looking at PGE's system peak data, our facility peaks align with grid about 60% of summer days. The coincident peak charge is $14.50/kW vs $12.80/kW on current schedule. Might be close, but I'm leaning toward staying on Schedule 85. The demand predictability gives us better budget certainty.
Wendy, have you looked at partial-peak pricing too? Some utilities offer mid-peak rates that might work better for data centers. Black Hills Energy here in South Dakota has a three-tier TOU that splits the difference. Also consider if your client has any plans for battery storage - that could completely change the TOU economics.
Joanne brings up a great point about battery storage. If your client is considering demand response programs, TOU rates become much more attractive. You can shift charging to off-peak hours and discharge during super-peak. I've got a client saving $40K annually doing exactly that with We Energies Schedule Tg-4.
Battery economics are getting better but still need 7-10 year payback analysis. CPS Energy offers rebates for demand response but you need consistent load shifting capability. Data centers are tricky because you can't just shut down servers during peak hours. Thermal storage might be better option if they're running significant cooling loads.
Client is actually looking at thermal storage for their cooling system upgrade. Current chiller plant is 15 years old and due for replacement anyway. If we can size the thermal storage to shift 200kW of cooling load off-peak, the TOU savings would be around $25K annually. That changes the economics significantly.
200kW shift would definitely make TOU worthwhile. Make sure to account for thermal losses and pump energy though. We had a client whose thermal storage system consumed 15% more total energy but still saved money on TOU rates. The key is optimizing the charge/discharge cycle to avoid partial-peak periods.
Also check if PGE offers any incentives for thermal storage installation. Many utilities are pushing demand flexibility programs now. MLGW offers up to $200/kW for qualified thermal storage systems. Between utility rebates and TOU savings, payback could be under 5 years.
Wendy, if you move forward with TOU, make sure to negotiate a pilot period with option to switch back. NorthWestern Energy here in Montana allows 12-month trials on new rate schedules. That way if load patterns change or thermal storage doesn't perform as expected, you're not locked in. Document baseline costs thoroughly before making the switch.