ESTO conducted the 32nd APEC Energy Modeling Seminar in Tokyo

Tokyo, Japan, 9 March 2026 – The Energy Statistics and Training Office (ESTO) successfully conducted the 32nd APEC Energy Modeling Seminar from March 2 to 6, 2026, at the APERC office in Tokyo, Japan. This year, 10 participants from seven economies (Indonesia, Malaysia, Papua New Guinea, the Philippines (4), Russia, Thailand, and Viet Nam) attended the training course.

The seminar is a hands-on program designed for researchers, officers, and policymakers who are working on energy system modeling within APEC member economies. It aims to enhance participants’ understanding of the role of energy models as analytical tools for developing energy policies to achieve sectoral goals, identify potential risks, and estimate infrastructure requirements. The seminar used OSeMOSYS (Open-Source Energy Modeling System) as the main modeling framework, as it can represent the data required for long-term sectoral energy system modeling.

The seminar began with an overview of energy modeling as a general introduction, followed by the presentation of OSeMOSYS as a tool for long-term modeling. The OSeMOSYS sessions followed a spiral learning approach, starting on Day 1 with the Utopia model for initial exercises. Day 2 focused on timeslice representation of demand, solar, and wind, and Day 3 centered on the power sector using the Atlantis Model. By Day 4, participants applied their learning to develop models for China, Papua New Guinea, and the Philippines, and on Day 5, they presented their work and received feedback.

During the feedback session, participants shared that the seminar deepened their understanding of energy modeling, particularly power sector modeling. While many found OSeMOSYS intimidating at first, they gained confidence and built their own models. They appreciated Google Colab for making the process accessible and for its ready-to-use notebook. Participants also expressed interest in more practical sessions, clearer visualization of outputs, and further training. ESTO responded positively, noting that these suggestions would be incorporated in the next iteration.