Back to field overview

Energy Optimization Bot

AI 93%Live in demo

Aeration is typically 50–60% of a water reclamation facility's total electricity spend and is the largest controllable opex line in the plant. The Energy Optimization Bot builds a rolling model of blower power versus dissolved-oxygen demand, influent load and utility tariff windows, then continuously recommends the lowest-energy operating envelope that still holds the permit margin.

Coastal Bend Regional Water Authority · Gulf Coast Metro Service Area · TX

Current specific energy
1,412 kWh/MG
Target 1,240 kWh/MG
Avoidable energy identified
11.4%
Rolling 30-day
Annualized opportunity
$214,600
At current $/kWh blend
Actions verified YTD
37
Of 44 issued
Live model output

Specific energy vs target · rolling 24 hours

Every point is normalized to treated volume (kWh per million gallons), so a load swing doesn't read as an efficiency change.

Bot actions

Recommended energy actions

Each action carries an expected kWh delta, a dollar value, and the reasoning behind it.

Trim Basin 3 DO setpoint 2.4 → 1.9 mg/L (00:00–05:00)

AI 94%

Overnight ammonia load is 38% below daytime peak; nitrification margin holds at 1.9 mg/L.

Energy delta 412 kWh/day$18,400/yr

Rotate blower lead from #2 to #3

AI 91%

Blower #2 specific power has degraded 9% since bearing wear onset; #3 is 6.2% more efficient at this duty.

Energy delta 268 kWh/day$11,900/yr

Shift Lift Station 7 & 14 pump-down to off-peak window

AI 87%

Wet-well capacity allows a 90-minute deferral without SSO exposure; avoids 4–7 pm demand tariff.

Energy delta $7,300/yr
Where the power goes

Energy spend breakdown

Aeration blowers58% · $742,000
Raw & finished water pumping21% · $268,000
Lift stations12% · $154,000
UV disinfection6% · $77,000
Buildings & other3% · $38,000

Aeration dominance is why this bot targets blowers and DO control first.

Guardrails

  • Never trims DO below the biological floor required for nitrification.
  • Advisory only — no closed-loop writes to the DCS unless the utility explicitly enables it.
  • Savings claims use a weather- and load-normalized baseline, not a raw year-over-year delta.

Data in · results out

Consumes
  • Blower kW / VFD speed feedback (per unit)
  • Basin dissolved oxygen and ammonia probes
  • Influent flow, MLSS and SRT
  • Lift station pump run hours and wet-well level
  • Utility interval data + time-of-use tariff schedule
Produces
  • DO setpoint trim schedule by basin and hour
  • Blower sequencing / lead-lag rotation plan
  • Off-peak pump-down windows for the collection network
  • kWh/MG scorecard per facility with verified savings
  • Energy anomaly alerts routed to the Alerts feed

DeploymentRequires blower-level power metering and at least one DO probe per aeration basin. Typical stand-up 2–3 weeks on existing SCADA tags.

Switch console

Other active bots

Demo environment · values shown are realistic estimates, not audited financials