HYDROGRID

HYDROGRID — Curated hydropower inflow and operations dataset: quality-corrected plant telemetry (reservoir levels, flows, production) cleaned by explicit data error detection and correction models, transformed into per-basin machine-learned inflow forecasts spanning 1-48 hour nowcasts through to 12-month seasonal horizons, and combined in the HIRO optimiser with environmental and operational release constraints, turbine efficiency curves, energy price forecasts and maintenance windows. The latent asset is the cross-basin training corpus — years of cleaned telemetry paired with realized inflow…

HYDROGRID offers (Alternative, Fundamental, Geospatial, Price) — est. low tens of millions of curated hydrometric records: each plant/basin in a portfolio carries automated ingestion of raw telemetry (level, flow, gate/spill, production) plus hourly nowcast updates across 1-48 hour horizons and separate weekly-to-12-month forecast tracks, so 100-500 monitored plants x 8,760 hourly cycles x several series = millions of rows per year per client, with the cross-basin model-training corpus accumulated over nearly 8 years of model development.

Training and benchmarking of hydrological and ML inflow-forecast models on a cross-basin, quality-corrected corpus; hydrometric sensor QC model development using the labelled error/correction library; imbalance-cost and intraday trading strategy research linking forecast error to realised cost; turbine efficiency and dispatch optimisation model calibration; cascade and reservoir-system planning; flood and drought early-warning model training; hydropower asset valuation and merchant-risk underwriting using historical inflow-outcome series; climate-impact studies on hydro generation across diverse basins and decades.

The data is with 8 years of history.

Coverage spans Other, Europe, Japan; Electric Utilities, Application Software; alternative, fundamental, geospatial, price; commodities, derivatives, equities.