← Cases
JapanEnergy and utilities2024 – ongoing

Hokkaido Electric Power

AI engine for thermal and hydro generation planning, built with Grid

Hokkaido Electric Power and Grid estimated about 600 million yen a month in thermal fuel cost savings from an AI planning engine

In progressCompany's own claim

Problem

Experienced staff planned thermal generation by hand the day before, which took 4 to 6 hours.

Approach

Hokkaido Electric Power and Grid developed an engine that combines machine learning and optimization AI for thermal and hydro supply-demand planning. The thermal part went into operation in March 2024. Adding hydropower was planned, with full operation targeted for around spring 2025.

Results

約6億円estimated monthly thermal cost reduction, from comparing January 2024 manual results with AI-calculated costs in a trial calculation, as announced by the companiesCompany's own claim[1]
約3%share of cost the 6 hundred million yen represents, as printed in the article; the base was not specifiedCompany's own claim[1]

Takeaway

The savings figure came from a one-month comparison with manual results, and hydro and the shorter calculation time were still expectations at the time.

Categories

Business function
Production and quality
Innovation type
Process innovation
Evidence
Company's own claim

Sources

Source summaries

[1]Robot Start via livedoor News6 Jun 2024

Robot Start, syndicated on livedoor News, reported the 3 June 2024 announcement by Hokkaido Electric Power and Grid. The thermal engine ran from March 2024, and a comparison with January 2024 manual results gave about 600 million yen a month. The article said hydro would add an expected further saving and a calculation time of about 10 minutes was expected.

Full details are in the original sources.

Added 2 Nov 2026 · Updated 2 Nov 2026
Open case page ↗