From centralized towards distributed and renewable energy systems

Transition towards distributed and renewable energy systems can succeed only if passive consumers are motivated and enabled to become active prosumers. The main driver for the distributed energy system is renewable energy production, which is geographically spread and requires intelligent demand-side management to match the available renewable energy production capacity continuously.

AI and DLT (Distributed Ledger Technologies)/Blockchain pilot, currently ongoing at Smart Otaniemi innovation ecosystem, concentrates on enabling the mentioned transition. Unlike traditional databases, distributed ledgers have no central data store or administration functionality. Blockchain is one type of a distributed ledger. Distributed ledgers use independent computers (nodes), to record, share and synchronize data in their respective electronic ledgers instead of keeping data centralized.

‒ AI and DLT technologies have great potential in energy management of future distributed and renewable energy system. For forerunner companies in energy management, now is the time to start exploiting the benefits of these technologies in product and service offerings, says Principal Scientist Daniel Pakkala from VTT Technical Research Centre of Finland Ltd.

Pakkala leads the activities in the pilot enabling the mentioned transition and focusing on the related research on applying AI and DLT technologies. The research results include concept and first prototype of Energy Management Agent (EMA) for implementing automated and user governed demand flexibility at individual sites. Further results include also modeling and simulation of EV charging for building and district grids, as well as experimenting blockchain based CO2 transparency in embedded generation and local consumption.

Change from passive consumers into active prosumers

New technology and artificial intelligence support consumers in turning into prosumers.

‒ In our research we have for example contributed to new state-of-the-art neural network architecture for energy loads forecasting in buildings, as well as experimented with the EMA concept on automating energy flexibility at sites for the prosumers. These kind of forecasting capabilities are needed in continuous energy optimization of buildings, and for identification of the dynamically varying and situation-dependent energy flexibility potential in buildings, Pakkala points out.

Besides that, the pilot members have worked on forecasting models, simulations and automated demand-flexibility control via EMA for facilitating energy management problems related to use of variable renewable energy production. Potential of AI based prediction models has now been confirmed in cases where large amounts of relevant data already exist as a starting point for training the prediction models.

‒ There are many good applications out there already of AI based prediction of energy production and consumption. I would expect the development to continue in the future, providing new applications and their further use, says   pilot member Lauri Paavola, Researcher from Aalto University. His work in the pilot has focused on business modeling and future scenarios planning related to energy markets.

Further research is still needed in improving exploitability of AI based prediction models, especially in data-efficiency, adaptivity and robustness of the models in different use cases.

AI in energy and flexibility management

According to Pakkala the data-driven models work well in forecasting and achieve good results, which can be continuously visualized for supporting human users in decision making and control. When AI is used in context of automation of the energy and flexibility management, further research is still needed on situational-awareness, planning, automated decision making and control capabilities of intelligent energy management agents. The energy demand forecasting model, created for Otakaari 4 in the pilot, is ready to be applied in other buildings via initial training on new buildings’ historical energy consumption data.

‒ We are currently searching for concrete use cases and piloting opportunities to continue the research on AI and DLT technology enablers for energy management. We are also looking for companies interested in exploitation of the research results as part of their product and service offerings for energy management, Pakkala concludes.


More information:
Enabling Technologies: AI & DLT/Blockchain pilot’s page
Daniel Pakkala, Principal Scientist, Senior Project Manager, VTT, +358 40 587 6261,


Text: Sirpa Mustonen, Motiva