Case Studies
What Ampgrove looks like inside a real grid operation.
Three early-access utilities ran 90-day pilots with Ampgrove on a subset of their feeders. Here's what the integration looked like, what they measured, and what changed in their operations.
Ridgecrest Public Power
Municipal Electric Utility- Challenge
- Managing summer peak demand across 18 residential feeders with legacy SCADA that only gave 24-hour ahead forecasts at the substation level. Operators carried a 25% peak reserve because they couldn't trust feeder-level estimates during afternoon convective storm periods.
- Approach
- Ampgrove connected to their GridSense SCADA system over a 12-day integration. 18 feeders went live on the Monitor tier. We configured per-feeder thresholds and tuned the weather model for their semi-arid climate and afternoon storm pattern.
- Result
- Forecast accuracy on peak summer feeders averaged 92% at the 4-hour horizon. Peak reserve dropped from 25% to 17% across the pilot feeders. Zero substation trips during the pilot period compared to 2 in the prior equivalent period.
Clearwater Electric Cooperative
Electric Distribution Cooperative- Challenge
- Rural distribution cooperative with 42 feeders covering a mix of agricultural and residential load. Irrigation season created sharp load ramps that their existing demand response program couldn't predict accurately enough to pre-position for. They were activating DER resources reactively rather than 3-4 hours ahead.
- Approach
- Pilot covered 12 feeders with the highest irrigation season variability. Ampgrove integrated with their NovaStar EMS and added agricultural load pattern features specific to their service territory. Irrigation schedule data from their AMI system fed into the weather-adjusted model.
- Result
- Dispatch recommendations arrived 3.5 hours ahead on average during peak irrigation periods. DER activation shifted from reactive to pre-positioned in 78% of alert events during the pilot. Capacity reservation costs on pilot feeders estimated down 28% annualized.
Mesa Valley Utilities
Small Investor-Owned Utility- Challenge
- Small IOU with a mixed feeder topology: commercial downtown feeders with predictable weekday load profiles alongside fast-growing suburban residential feeders with EV charging load that was making existing 24-hour forecasts increasingly inaccurate.
- Approach
- 8 feeders on the Operate tier. Ampgrove incorporated EV charging telemetry from their ClearPath DMS alongside standard SCADA interval data. Commercial and residential feeder models were trained separately on 30 months of history. Automated dispatch recommendations tied to their existing DER portfolio.
- Result
- EV-affected feeder forecast accuracy improved from an estimated 79% (prior 24-hour model) to 96% at the 4-hour horizon during pilot period. Operators reported reducing manual load checks from 8 times per shift to 2. Alert fatigue score (self-reported) improved from 3.2 to 1.7 on a 5-point scale.
Aggregate Results
Combined data from 3 early-access utility partners.
Figures represent combined results across all three 90-day pilot deployments. Results vary by grid topology, feeder type, climate, and operational context. These are pilot-period estimates, not production guarantees.
94%
Average feeder forecast accuracy at the 4-hour horizon across all pilot feeders
Combined pilot data, 38 feeders, 90 days each
32%
Average reduction in peak-period over-provisioning reserve across pilot utilities
Operational reports from pilot partners
2 weeks
Average integration time from SCADA connector setup to first forecast output
Implementation logs, all 3 pilot utilities
Start Your Pilot
We configure Ampgrove for your feeder topology in 2 weeks.
Start on a subset of feeders. No upfront commitment. We scope the pilot based on your SCADA platform and feeder count.