Application of a Transition Graph-Based Predictive Algorithm to a Solar Air Conditioning Plant

Abstract

This brief presents the development of a transition graph-based predictive algorithm and its application to a solar air conditioning plant. The process considered is a hybrid system of variable configuration subject to disturbances in its main energy source, namely solar radiation. The predictive algorithm switches between the operating modes of the plant on the basis of minimizing the performance cost function associated to each operating mode and the switching cost function associated to the transition between operating modes. The cost functions depend on the predicted states of the plant. A transition graph is used to reduce the computational cost of the optimization problem; the graph is plotted using the concept of reachability. The main objective of the predictive algorithm is to find the optimal solution to the configuration problem and the optimal control signal of the plant throughout the day. Experimental results show that solar energy was used during most of the test and cooling demand was satisfied.