Intelligent dispatching and optimal operation of cascade hydropower plants based on spatiotemporal big data

The IntHydro project has already demonstrated how artificial intelligence can transform hydropower from a traditionally scheduled resource into a hyper-flexible tool for the entire power system. During the pilot phase, an AI-driven, two-level optimisation, combining a fine-grained hydrological digital twin with a market-coupled dispatch engine, was tested on three Norwegian cascades totalling approximately six gigawatts of capacity. The results were striking: the mean absolute error in hourly flow forecasts fell from 9% to 3%, imbalance costs dropped by roughly €1.8 million per year, and gross energy sales increased by 4.2%, equivalent to €4.7 million in additional annual revenue. Fewer start-stop cycles also resulted in nearly €1 million in saved operating and maintenance expenses. Market simulations measuring knock-on effects beyond the plants themselves showed a 6% reduction in spot-price volatility and a consumer benefit of approximately €12 million annually. The climate impact was also notable: improved integration of wind and solar allowed 52 GWh of variable renewables to be rescued from curtailment, avoiding around 13,000 tonnes of CO2-equivalent emissions during the test period. These results suggest even greater benefits at full scale. Within the first three years, European hydropower operators could capture more than €1 billion in additional income and cut an extra 45,000 tonnes of CO2-equivalent annually. In the medium term, up to 15 GW of “virtual” balancing capacity could be unlocked, enabling greater renewable integration without the need for new grid infrastructure. Over the long term, the outlook includes a global export market for AI-based software and services, and the creation of thousands of high-skill jobs. The Norwegian-Chinese partnership provides a unique arena for scaling the technology. Norway already produces 96% of its electricity from hydropower, while China has built the world’s largest hydropower fleet, over 350 GW, representing more than a quarter of global capacity. Validating IntHydro in both Norway’s ultra-flexible system and China’s vast, fast-growing market allows the solution to demonstrate millisecond-level flexibility at gigawatt scale. AI enables real-time decision-making down to 15-minute resolution, optimised water use in coordination with wind and solar, and continuous risk navigation through digital twins and cloud-based dashboards. The societal benefits go well beyond kilowatt-hours and euros. Improved reservoir regulation helps mitigate flood peaks and ensures ecological minimum flows. Increased operational flexibility reduces reliance on gas generation and costly battery storage. IntHydro also promotes gender diversity by targeting 40% female participation in R&D leadership and developing bias-free algorithms. Its four cross-disciplinary PhD tracks strengthen the next generation of green-digital expertise. In short, IntHydro has already proven its value and has a clear pathway to global scale-up. By combining Norway’s hydropower expertise with China’s industrial capacity and data richness, the project sets a new benchmark for AI-enhanced renewable energy management, delivering lower costs, reduced emissions, and a smarter, more resilient power grid for the future.

1) A two-level optimisation engine (granular hydrological model + market-coupled dispatch optimiser) was implemented and benchmark-tested on three cascade systems 2) System-wide value, not just plant-level: Price-stability benefits accrue across the entire market, reinforcing social licence for hydro. 3) Strategic leverage for decarbonisation: By extracting more flexibility from existing assets, IntHydro reduces the need for new fossil-back-up capacity.

The scope of the IntHydro project is at the centre of the thematic area ‘Digitalisation of traditional industries’ in the RCN Chinese-Norwegian Collaborative Projects on Digitalisation call. The aim of the IntHydro project is to explore and define intelligent hydropower scheduling using ML techniques. The scheduling methodology will address shortcomings in existing hydropower scheduling models to deal with uncertainties brought by the high share of variable renewable energy resources in both Norwegian and Chinese power system. To this end, the primary objective of IntHydro can be summarised as: “Develop and demonstrate a new hydropower scheduling tool based on machine learning techniques, which manages the hydropower plants more efficiently and effectively through optimizing water resource management and multi-dispatch between hydropower and variable renewable energy sources.” In “IntHydro”, NTNU is the grant applicant and will be the project leader. The Norwegian research partners are Smart Innovation Norway (SIN) and HydroCen. The Norwegian industry partners are Østfold Energi and Lyse. Chinese partners are Houhai University, NanJing NARI Water Resources and Hydropower Technology Company, Ltd, and Yalong River Hydropower Development Company, Ltd. The Norwegian partners will be involved in the daily operation of the project and will also be invited into a steering board is the governing body of the project. In addition to being active in the research, the Chinese partner will be advisors to the management group on scientific matters. There will be one PhD student, and the candidate will play a central role in developing methodologies and participate in the other issues.