# GridLoss AI

> AI platform for an Indian electricity distribution company (DISCOM). It shows where every unit of energy and every rupee is lost, explains why, forecasts what happens next and prioritises the field actions that recover revenue. Submitted to the Rodic InfraAI Innovation Challenge 2026, Theme 1: AI for India Energy Stack. All data is synthetic with injected, labelled anomalies.

The product follows one story: See, Trace, Predict, Act, Prove.

## Read this first
- [Full description](http://poc.nidhitek.com/llms-full.txt): product, numbers, method, every API endpoint, limitations
- [Interactive API reference](http://poc.nidhitek.com/docs): OpenAPI documentation
- [OpenAPI schema](http://poc.nidhitek.com/openapi.json): machine-readable endpoint list
- [Submission pack data](http://poc.nidhitek.com/submission/all): dataset description, testing report, documentation as JSON

## Dashboard pages (JavaScript application; use the JSON endpoints for content)
- [Overview](http://poc.nidhitek.com/overview): See. How much energy and revenue is lost, split by cause, with the feeder watchlist.
- [Energy flow](http://poc.nidhitek.com/energy-flow): See. Sankey diagram from generation sources through the 220 kV substations and feeders to consumers, with conservation checks.
- [Feeder analytics](http://poc.nidhitek.com/feeder-analytics): Trace. Ranked feeders and a substation to feeder to transformer to meter drill-down where parts add to the whole.
- [Transformers](http://poc.nidhitek.com/transformers): Trace. All 88 distribution transformers with loss breakup, loading and meters.
- [Consumers](http://poc.nidhitek.com/consumers): Trace. Consumer investigation with evidence, peer comparison and case workflow.
- [Load forecast](http://poc.nidhitek.com/forecast): Predict. Demand forecast at feeder, substation, transformer and category level, with congestion and growth outlook.
- [Renewables and DSM](http://poc.nidhitek.com/renewables): Predict. Solar, wind, hydro and biomass forecasts, deviation-settlement exposure in rupees.
- [Asset health](http://poc.nidhitek.com/asset-health): Predict. Transformer failure risk with lead time and drivers.
- [Action queue](http://poc.nidhitek.com/queue): Act. Prioritised cases with reason flagged, rupee value, action, owner and outcome tracking.
- [Revenue leakage](http://poc.nidhitek.com/revenue-leakage): Act. Meter to bill to cash reconciliation and flagged accounts.
- [Load shedding](http://poc.nidhitek.com/load-shedding): Act. Optimised load-shedding plan against blind rotation, protecting critical loads.
- [Field programs](http://poc.nidhitek.com/field-programs): Act. Loss-reduction levers and expected recovery.
- [Collections and subsidy](http://poc.nidhitek.com/collections): Act. Collection shortfall and agricultural subsidy tracking.
- [Copilot](http://poc.nidhitek.com/copilot): Act. Question-answering over the ledger and cases, with guardrails.
- [Model evidence](http://poc.nidhitek.com/model-evidence): Prove. Accuracy against injected ground truth, with limitations.
- [Raw data](http://poc.nidhitek.com/raw-data): Prove. Every dataset the proof of concept runs on, paged, searchable and downloadable.
- [Challenge submission](http://poc.nidhitek.com/submission): Prove. Dataset description, testing report and documentation for the Rodic challenge.
- [Deployment](http://poc.nidhitek.com/deployment): Prove. Dataset contracts, ingest validator, footprint and scale path.

## Key data endpoints
- [Loss ledger](http://poc.nidhitek.com/ledger/network): network loss in kWh and rupees by cause, feeders and substations that sum to it
- [Energy flow](http://poc.nidhitek.com/energy/flow): Sankey nodes, links and conservation checks
- [Evaluation](http://poc.nidhitek.com/evaluation): accuracy against injected ground truth
- [Dataset catalogue](http://poc.nidhitek.com/data/tables): every table with rows, columns and size

## Optional
- [Sitemap](http://poc.nidhitek.com/sitemap.xml)