Combined peak reduction and self-consumption using proximal policy optimisation

作     者：Thijs Peirelinck Chris Hermans Fred Spiessens Geert Deconinck 

作者机构：ESATKU LeuvenKasteelpark Arenberg 10 bus 24453001 LeuvenBelgium AMOFlemish Institute for Technological Research(VITO)Boeretang 2002400 Mol Belgium AMOEnergyVilleThor Park 83103600 GenkBelgium 

出 版 物：《Energy and AI》 (能源与人工智能（英文）)

年 卷 期：2024年第16卷第2期

页      面：24-31页

核心收录：

学科分类：081203[工学-计算机应用技术] 08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Vlaamse Instelling voor Technologisch Onderzoek  VITO 

主　　题：Demand response Reinforcement learning Electric water heater Peak shaving Transfer learning 

摘      要：Residential demand response programs aim to activate demand flexibility at the household *** recent years,reinforcement learning(RL)has gained significant attention for these type of applications.A major challenge of RL algorithms is data *** RL algorithms,such as proximal policy optimisation(PPO),have tried to increase data *** tionally,combining RL with transfer learning has been proposed in an effort to mitigate this *** this work,we further improve upon state-of-the-art transfer learning performance by incorporating demand response domain knowledge into the learning *** evaluate our approach on a demand response use case where peak shaving and self-consumption is incentivised by means of a capacity *** show our adapted version of PPO,combined with transfer learming,reduces cost by 14.51%compared to a regular hysteresis controller and by 6.68%compared to traditional PPO.