Abstract
In multi-energy systems the full utilisation of the generated energy is a challenge. Integrating heat and electricity supply at the system level design could provide an opportunity to address this challenge. In this paper we introduce and examine two coupled thermal-electrical dispatch strategies for grid-connected hybrid multi-energy systems supplying electrical and thermal demand loads. The dispatch strategy employs forecasting of energy resources and demand loads to prioritise supplying the thermal load in times of renewable surplus. Four forecasting algorithms, namely, baseline forecast, Facebook Prophet (FBP), Neural Prophet (NP), and Long Short-Term Memory model (LSTM) are implemented and used to generate annual forecast data for solar irradiance, wind speed, and thermal and electrical demand loads. To integrate forecast data within the dispatch strategy, new parameters are proposed to quantify the expected available energy within the forecast time horizon. A building complex for the Department of Education in the UK is used for conducting a system design case study. A genetic algorithm-based multi-objective optimisation with the levelised costs of electricity and heat as two objectives is conducted. The results show that the proposed dispatch algorithm produces systems with reduced levelised costs compared to the base case of using utility gas and electricity. Forecasting is particularly useful in reducing cost of heat, as it can prioritise supplying the thermal load in times of renewable surplus. LSTM proved to be the most accurate forecasting algorithm for this case, where the data has strong seasonality and trends. The main contribution of this work is to propose and demonstrate the effectiveness of tightly coupling thermo-electrical dispatch algorithms of HRES from the design stage, and how to effectively integrate forecast data within such algorithms.
Original language | English |
---|---|
Article number | 117460 |
Number of pages | 25 |
Journal | Energy Conversion and Management |
Volume | 293 |
Early online date | 11 Aug 2023 |
DOIs | |
Publication status | Published - 1 Oct 2023 |
Bibliographical note
AcknowledgementsThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Keywords
- Grid-connected
- Hybrid renewable energy system
- Clean heat
- Supervised machine learning
- Time series forecasting
- Multi-energy systems