A prediction method for urban heat supply based on grey system theory

Abstract

Urbanization in China is bringing distinct growth in heating energy consumption. Hence, the key to ensuring the responsibility of energy structure reform is to reasonably predict urban heat supply. This paper presents a top-down prediction method based on the grey system theory. The prediction method is embedded in a genetic algorithm optimization with the objective function of a posterior difference ratio in order to achieve the highest accuracy. In the methodological framework, the first step is to screen the explanatory variables from the preliminary factors by employing grey relational analysis. Then, the grey models of the dependent and explanatory variables are combined to establish the grey system state equation. Finally, the posterior difference of the results is examined to determine the optimization process. The case study focuses on 16 preliminary factors and 3 cities in China. According to the prediction results, the improved prediction accuracy shows the advantages of optimization process as well as updated training data. The comprehensive annual growth rate of the heat supply is finally proposed to integrate all of the prediction possibilities. The top-down prediction method reveals the future tendency of annual heat supply so as to assist urban planning and energy structure reform.