Typical daily load curve is of great significance to load scheduling plan and operation control. To solve the problem that the conventional daily load curve selection method does not consider environmental factors

a new method for selecting the typical daily cooling and heating load curve is proposed

based on k-means clustering algorithm and average value algorithm. In this method

the temperature parameter is used as the influence factor to participate in the clustering process

and the average value algorithm is used to extract the average load as load curve for each type. Then

the Pearson correlation coefficients of various load curves and their average load curves are calculated

and the load samples with Pearson correlation coefficients in [–0.2

0.2] is removed according to the nature of coefficients. After that

the remaining data is clustered again by the k-means clustering algorithm until the value of Pearson correlation coefficient does not land in [–0.2

0.2]. The verification based on the load data of an energy station in a district of Shanghai shows that

the method proposed in this paper can accurately obtain the typical daily cooling and heating load curve

which is feasible and effective.