Build a District Cooling Plant Sub-Meter Anomaly Detector — implementation

## Build
Analyze district cooling plant tenant sub-meter data to identify bypass valve leakage and faulty energy meters. Generate maintenance tickets to recover lost cooling tonnage billing revenue.

## User
Energy Services Retrofit Engineer

## Workflow
1. Ingest tenant cooling energy sub-meter telemetry and central plant chilled water flow rates.
2. Calculate expected energy consumption versus billed usage to identify usage discrepancies.
3. Classify anomalies as either meter calibration drift or physical bypass valve leakage.
4. Output a verified revenue recovery report and maintenance inspection ticket.

## Interface
Financial and engineering dashboard displaying thermal storage sizing options, peak kW savings, and payback curves.

## Data
Use synthetic normalized app records representing tenant billing sub-meters and central plant flow logs.
Starter examples below are synthetic. They are not downloaded records.

## Source
[EnergyPlus](https://raw.githubusercontent.com/NREL/EnergyPlus/develop/README.md)
Extracts expected flow and thermal load properties using EnergyPlus central plant loop configurations.

## Tools
dspy, pandas, scikit-learn.
DSPy optimizes the prompt pipelines to match cooling load profiles with optimal ice storage dispatch strategies.
Local or hosted LLM with strong mathematical reasoning and JSON formatting capabilities.

## Deliverables
Submit runnable source, pinned dependencies, setup commands, a working interface, example outputs, and a recorded end-to-end demo. Include tests and an action log.

## Acceptance
- Identifies discrepancies between totalized plant cooling output and aggregated tenant sub-meter billing.
- Classifies faults between meter drift and physical bypass valve leakage based on load profiles.
- Produces an itemized revenue loss report for facility management.
- Outputs simplified synthetic JSON maintenance tickets without triggering live billing freezes.

## Starter Examples
```json
[
  {
    "input": {
      "tenant_id": "BLDG-12",
      "billed_ton_hours": 1200,
      "measured_flow_lps": 45,
      "delta_t_c": 1.5,
      "_schema_type": "simplified synthetic app record"
    },
    "expected_output": {
      "anomaly_detected": true,
      "classification": "BYPASS_VALVE_LEAKAGE",
      "estimated_lost_revenue_usd": 450,
      "action": "DISPATCH_INSPECTION"
    }
  },
  {
    "input": {
      "tenant_id": "BLDG-14",
      "billed_ton_hours": 3500,
      "measured_flow_lps": 40,
      "delta_t_c": 8.5,
      "_schema_type": "simplified synthetic app record"
    },
    "expected_output": {
      "anomaly_detected": false,
      "classification": "NORMAL",
      "estimated_lost_revenue_usd": 0,
      "action": "NONE"
    }
  }
]
```

## Safety
Provides advisory financial modeling; requires stamped professional mechanical engineering validation before final procurement.
Use mock external writes. Require human review before any real deployment.