Pinpointing hidden overnight electricity waste with AI
At an education facility monitored through the Renergia platform, the AI analysis flagged a persistent consumption increase outside core operating hours — isolated the HVAC-related component as the main suspect, and quantified what automated scheduling could save over the warm season.
Context
The case study covers an anonymized education facility monitored through Renergia. The analysis used consumption, import, export, PV production, outdoor temperature, and HVAC disaggregation data from a real monitored site.
Core operating hours are 07:00–15:00, with the main out-of-hours interval running 19:00–05:00. Renergia compared the monitored load profile against the technical reference consumption required under similar warm-season conditions, using 01–10 May as the reference window.
01–10 May
Reference period
11 May 2026
First clear deviation detected
1.6 Lei/kWh
Import price used in the analysis
What Renergia found
Renergia automatically detected an abnormal increase by comparing the monitored load profile with the technical reference consumption required under similar warm-season conditions. The increase was persistent outside core operating hours — especially overnight.
The AI-based disaggregation algorithm isolated the HVAC-related component as the main suspected driver of the increase. A daily correlation of r = 0.79 with outdoor temperature signals seasonal cooling behaviour. Final root-cause confirmation still requires a technical inspection on site.
What the data showed
From 11 May onward, the monitored profile diverged clearly from the reference — and the increase sat strictly on imported energy.
+134.9% versus the reference average.
+117.2% — a 2.17× increase in the main out-of-hours interval.
The increase sits strictly on imported electricity.
Daily correlation with outdoor temperature — a seasonal behaviour signal.
How the issue was identified
The platform combined AI-based consumption profiling, hourly behaviour analysis, and a technical reference consumption adjusted for warm-season conditions — separating suspected unnecessary HVAC operation from consumption caused by warmer weather.
What was proposed
Renergia proposed HVAC automation through the Renergia platform to reduce unnecessary operation outside occupancy-related periods, targeting the avoidable imported energy in the 19:00–05:00 interval.
The savings estimate deliberately excludes two operational windows:
05:00–07:00
Pre-cooling before the facility opens
15:00–19:00
The after-office transition period
The recommendation has not yet been accepted by the client. All values represent estimated potential impact, not an implemented result.
Estimated potential
The calculation is conservative: it is limited to energy considered avoidable and to imported electricity only. It is indicative, not a guaranteed saving.
Avoidable imported energy identified in the June data.
Indicative seasonal potential if the automation is implemented.
Of total June consumption, calculated on avoidable imported energy.
Financial values use an import price of 1.6 Lei/kWh and an exchange rate of 1 EUR = 5 Lei.
Why it matters
Renergia turned a subtle seasonal drift into a specific, quantified opportunity. Instead of only showing that consumption was rising, the platform pinpointed when the waste occurred, isolated HVAC as the main suspect, and put a concrete number on what automated scheduling could save.
Because the analysis runs continuously on monitored data, the potential can be verified the moment the recommendation is implemented — without affecting normal facility operations.