Prediction models and demand characterization

SunHorizon will demonstrate up to TRL 7 innovative, reliable, cost-effective coupling of solar and HP technologies. It addresses three main research pillars that interact each other towards project objectives achievement, demonstration and replication: i) optimized design, engineering and manufacturing of SunHorizon technologies, ii) smart functional monitoring for H&C, iii) KPI driven management and demonstration. D5.1 reports those advances during the first year of the project concerning modelling and simulation of the SunHorizon solutions aiming at building energy demand and systems' performance prediction. Such predictive capabilities will contribute to the development of a smart-integrated control system that will be fully deployed and demonstrated in real conditions within two representative SunHorizon demo sites and partially validated in simulation in some other. Key objectives covered by D5.1 are to (1) select those two demo cases where the SunHorizon Controller will be fully deployed in real conditions (2) develop simulation models of the building for demand characterization and energy systems models for renewable energy sources contribution, (3) develop prediction algorithms for demand and production forecasting, (4) define a weather forecast service, and (5) define the necessary interfaces for model and service interaction. In this sense, the following main activities have been conducted: ? Collection of details from SunHorizon demo sites, which are relevant for the selection of demonstration cases and the generation of suitable energy models ? Definition of a common simulation methodology to provide suitable prediction capabilities to the SunHorizon advanced controller thanks to the integration of the building digital twin (in IESVE software) and energy systems' models (in TRNSYS software) into the overall control workflow. ? Definition of requirements for a weather forecast service and identification of relevant input/output variables to enable the interaction of the prediction models and algorithms with the self-learning and end-user feedback controller features. The output of the task is a first step for the implementation of prediction algorithms to estimate the energy demand and the renewable energy contribution into the SunHorizon controller. The algorithms will feed tasks 5.3 and 5.4, for self- learning and optimization strategies capabilities. This study provides a complete and adapted methodology for demand and RES contribution prediction of the project for two different cases, residential and tertiary building, which sets a robust basis for replication and scalability in future developments that will contribute to a more efficient energy supply and the decarbonisation of heating and cooling applications.