Energy management

Renewable energy can fluctuate greatly due to weather conditions, which makes it difficult to achieve a balance in the supply and demand of electric power as the use of this form of energy expands. As a consequence, technologies that use a variety of devices for balancing supply and demand are being actively developed.
Having many power-generation and storage facilities and hydrogen production facilities, the ENEOS Group is developing energy management technology to optimize facility operations.
This technology will make it possible to achieve a stable balance between the supply and demand of power and maximize the use of renewable energy by storing power and producing hydrogen at times of a surplus and discharging power at times of a shortage.
Optimization system for operating storage batteries (hammock® Pro)
The ENEOS Group has installed large storage batteries (55MW in total) at its Negishi Refinery and Muroran Site and began utilizing them in its VPP business* from August 2023. The storage batteries are controlled by an operation optimization system (hammock® Pro) with AI (optimal operation control algorithm) developed in-house and utilized in-house to automate and upgrade operations.
*VPP (Virtual Power Plant): A technology for controlling a variety of devices such as solar power generators and household storage batteries all together to make them appear like a huge power plant.

Optimal control technology for hydrogen production equipment
Hydrogen, which is attracting attention as a next-generation decarbonized energy source, can be produced from surplus electricity from renewable energy sources using water electrolysis equipment, thereby contributing to stabilizing the supply-demand balance of electricity. ENEOS has developed energy management technology to maximize the use of these hydrogen features, simultaneously optimize the supply and demand for power and hydrogen and operate facilities economically.
ENEOS has been using this technology in some of its hydrogen stations since FY2022 and applying it to the production of CO2-free hydrogen with low electricity charges.

Development of forecasting technology
Establishing operational plans based on accurate forecasts of the power generated by renewable energy, the demand for power and hydrogen, and the electricity market price is essential for the optimal operation of facilities and optimal purchasing/selling on the power market.
For each of these forecasting technologies, ENEOS evaluates outside technologies and develops forecasting systems (hammock® forecast) in-house.
For example, to forecast the amount of solar power to be generated, we have developed a highly accurate, hybrid forecasting method that combines a method for forecasting power generation in an engineering manner based on the characteristics of power panels, installation conditions, etc. and a method that forecasts future power generation based on past figures using machine learning. We are expanding the use of this hybrid method to a variety of business applications.

Development of demonstration platform (hammock® EMS)
The Central Technical Research Institute is equipped with a variety of demonstration facilities including solar power generation, storage batteries, hydrogen production equipment, and EV chargers. We are developing a cloud system, hammock® EMS, to monitor and control all these facilities collectively by connecting them to the Internet. This system allows us to immediately apply newly devised control technology to actual equipment and conduct demonstration tests, thereby contributing to commercialization. This system is also used in demonstrations with our external customers.

Technology stack
In developing in-house services, we promote agile development with one team, including engineers from partner companies, and aim to maintain a release cycle of at least weekly, even for mission-critical systems. We also always endeavor to use the latest technologies in selecting an optimal technology stack.
Table: Main technology stacks (as of September 2024)
Backend | Python (FastAPI) |
---|---|
Optimization | PuLP, Gurobi |
Machine Learning | Python, scikit-learn, LightGBM, Optuna, MLflow |
Infrastructure | AWS, Terraform, serverless framework |
Visualization | Streamlit, Reflex |
Database | AWS RDS (PostgreSQL, Aurola serverless), AWS DynamoDB |
Environment | VSCode (devcontainer), docker |
Code Management, CI/CD | GitHub, GitHub Actions |