Autonomous observer of hydrogen storage to enhance a model predictive control structure for building microgrids

• Published in: Journal of energy storage 2022-09, Vol.53, p.105072, Article 105072
• Main Authors: Yassuda Yamashita, Daniela, Vechiu, Ionel, Gaubert, Jean Paul, Jupin, Samuel
• Format: Article
• Language: English
• Subjects: Autonomous observer; Building microgrids; Engineering Sciences; Hydrogen storage; Self-consumption
• ISSN: 2352-152X; 2352-1538
• DOI: 10.1016/j.est.2022.105072

Hydrogen energy storage has emerged as a promising technology to improve the integration of renewable energy sources in building microgrids. However, inaccuracies in the modelling of fuel cells and electrolysers reduce the performance of building microgrids' energy management system. To improve the flexibility of building microgrids, this paper proposes to associate a two-level hierarchical model predictive controller empowered with an Autonomous Observer of Hydrogen Storage (AOHS). This novel observer evaluates the hydrogen production and consumption rates, storing little past data and needing no tuning of the parameters. Relying only on instantaneous data measurement, the algorithm can estimate the tank's level of hydrogen with an average relative error inferior to 2 %, even under measurement noise. A case-study based on a building microgrid currently under construction serves as the basis for all simulations. The performance of the AOHS is evaluated by comparing the self-consumption rates of the case-study when governed by two-level energy management system: one level using a fixed parameters model and the other one equipped with the proposed AOHS algorithm. Results show that the microgrid associated to the AOHS has better self-consumption compared to the microgrid with fixed parameters, as well as a better robustness regarding the measurement noise and modelling error. Furthermore, this algorithm demonstrates a planning function as it facilitates the energy planning from the aggregator's point of view and the external grid management. •Automatic parameter identification of the hydrogen energy storage models to improve the model predictive controller•Design of a hierarchical model predictive controller to maximise the photovoltaic self-consumption of a building microgrid•Evaluation of the robustness of the proposed hydrogen autonomous observer against measurement noise