Svensen Jan Lorenz, Bergsteinsson Hjörleifur G, Madsen Henrik
Department of Applied Mathematics and Computer Science, Technical University of Denmark, Richard Petersens Plads 324, 2800 Kongens Lyngby, Denmark.
Heliyon. 2024 May 14;10(10):e31027. doi: 10.1016/j.heliyon.2024.e31027. eCollection 2024 May 30.
This study presents data-driven modeling and nonlinear model predictive control of solar thermal plants in district heating, for the purpose of operation optimization. The study considers the efficient operation of a solar thermal plant in Hillerød, Denmark. A dynamic model is estimated as a system of stochastic differential equations using grey-box modeling and real-world data. The presented nonlinear model predictive controller design is based on repeated trajectory linearization and the dynamic model. Several objective designs are considered, e.g., maximizing energy or temperature. The study provides simulations for analyzing the model fitness and controller performances. The model is shown to fit the daytime operation of the plant. The designed controllers are shown to improve efficiency, increasing the transported energy up to 28%.
本研究提出了用于区域供热中太阳能热电厂运行优化的数据驱动建模和非线性模型预测控制方法。该研究考虑了丹麦希勒勒市一座太阳能热电厂的高效运行。利用灰箱建模和实际数据，将动态模型估计为随机微分方程组。所提出的非线性模型预测控制器设计基于重复轨迹线性化和动态模型。考虑了几种目标设计，例如能量最大化或温度最大化。该研究提供了用于分析模型拟合度和控制器性能的仿真。结果表明，该模型能够拟合该厂的日间运行情况。所设计的控制器被证明能够提高效率，使输送能量增加高达28%。
