Build the application

## Build
Model photovoltaic power loss from dust accumulation against site wash water availability and desalination costs. Generate optimal cleaning schedules that balance water transport limits and energy yield.

## User
Utility-scale solar plant maintenance planners and operations supervisors

## Workflow
1. Ingest irradiance telemetry, panel soiling ratio estimates, and available site demineralized water reserves.
2. Calculate daily energy yield losses using PVLib soiling loss and irradiance transposition routines.
3. Evaluate trade-offs between automated dry brush cleaning and high-consumption water wash cycles.
4. Construct a multi-day wash schedule and water truck logistics route plan for site engineering approval.

## Interface
Map-based solar block grid showing soiling heatmaps, water tanker dispatch rosters, and yield recovery curves.

## Data
Leverage PVLib synthetic time-series irradiance datasets and normalized utility-scale solar telemetry mocks. Provided examples are simplified synthetic app records, not complete official schema files.
Starter examples below are synthetic. They are not downloaded records.

## Source
[PVLib solar modeling](https://raw.githubusercontent.com/pvlib/pvlib-python/main/README.md)
Uses PVLib photovoltaic performance and soiling simulation functions to evaluate power generation recovery.

## Tools
optuna, pydantic-ai, pvlib.
optuna executes parameter searches across cleaning intervals, water consumption, and energy price curves.
Local or hosted LLM for generating textual shift plans and constraint explanation reports.

## 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
- Calculates soiling energy loss across individual solar blocks using standard PVLib degradation curves.
- Enforces strict daily site wash water limits based on desalinated supply tanker allocations.
- Generates prioritized cleaning block sequences maximizing recovered megawatt-hours per cubic meter of water.
- Requires field solar asset manager sign-off before dispatching automated wash vehicle routes.

## Starter Examples
```json
[
  {
    "input": {
      "block_id": "PV-NORTH-04",
      "installed_mw": 50,
      "soiling_ratio": 0.82,
      "daily_water_allowance_m3": 40,
      "water_wash_requirement_m3_per_mw": 0.5,
      "dry_brush_available": true
    },
    "expected_output": {
      "action": "dry_brush_clean",
      "water_used_m3": 0,
      "projected_soiling_ratio_after": 0.94,
      "priority_score": 78.5,
      "notes": "Conserve water; dry brush recovery sufficient for current dust profile."
    }
  },
  {
    "input": {
      "block_id": "PV-SOUTH-12",
      "installed_mw": 60,
      "soiling_ratio": 0.71,
      "daily_water_allowance_m3": 15,
      "water_wash_requirement_m3_per_mw": 0.5,
      "dry_brush_available": false
    },
    "expected_output": {
      "action": "partial_wet_wash",
      "water_used_m3": 15,
      "projected_soiling_ratio_after": 0.98,
      "priority_score": 95,
      "notes": "Severe cementation; allocated entire 15m3 quota to recover critical 30MW section."
    }
  }
]
```

## Safety
Cleaning routes are generated as advisory work orders; equipment interlocks and physical vehicle clearances require manual validation.
Use mock external writes. Require human review before any real deployment.