Adaptiveness of a model predictive controller for a thermal-electrical renewable energy system in four different German single-family house energy standards

• Published in: Case studies in thermal engineering 2021-08, Vol.26, p.101118, Article 101118
• Main Authors: Narayanan, Muthalagappan, Mengedoht, Gerhard, Commerell, Walter
• Format: Article
• Language: English
• Subjects: Energy optimization; Home energy management; Intelligent controllers; MPC; Universalization
• ISSN: 2214-157X; 2214-157X
• DOI: 10.1016/j.csite.2021.101118

Model predictive controller (MPC) until now is required to be custom designed for each particular case. A combined system-level controller and universalization of MPC is yet to be achieved. This article focuses on simulatively investigating the adaptability of a previously developed whitebox-model MPC using a generic objective function for increasing renewable energy fraction and self-consumption in a decentralized integrated renewable energy system. The chosen single-family house energy standards are: high renewable fraction house (Sonnenhaus); modern building (KfW55); low primary energy house (passive house); and old building from 1995 (WSchVO95). From the perspective of universal implementation, it could be identified that a change in weightage factors is only required for the old WSchVO95 building. An astonishing reduction in space heating demand: 15% for Sonnenhaus; −30% for KfW55; −30% for passive house; and −2% for WSchVO95 could be accomplished. In terms of energy supply, ca. 10% increase in solar thermal fraction with the new buildings and a 14% increase in PV supply in all buildings could be achieved. Altogether, decentralized system self-consumption fraction increases by 8–12% via MPC. Whitebox MPC is however still a hindrance with huge optimization time making it not suitable for real-time implementation. •Universalization of model predictive controllers.•Universally compatible with changes required in objective function for old buildings.•Whitebox model still requires huge optimization time, especially in winter days.