Cooling load prediction in a district heating and cooling system through simplified robust filter and multilayered neural network

Cooling load is a heat value of cold water used for air conditioning in a d istrict heating and cooling system. Cooling load prediction in a district h eating and cooling system is one of the key techniques for smooth and econo mical operation. In this article, cooling load prediction in such a distric t heating and cooling system is considered. Unfortunately, since actual coo ling load data usually involve measurement noises, outliers, and missing da ta for several reasons, a prediction method considering the effect of the o utliers and missing data is desirable. In this article, a new prediction me thod using a simplified robust filter to improve a numerical stability prob lem of a robust filter and a three-layered neural network, is proposed. App lications of the proposed method and some other methods to actual cooling l oad data in a district heating and cooling system involving outliers and mi ssing data show the usefulness of the proposed method.