Land-Use Change Prediction by CA-Markov Method for Electric Load Density Map

Abstract

This paper presents a method to develop the spatial load density map (SLD) based on the top-down load allocation concept of spatial load forecasting. The yearly power load of electric customers as the top-level information is distributed to the total floor areas as the down-level information. To attain the total floor areas of each land-use category, the CA-Markov technique and Floor Area Ratio (FAR) are applied. The land-use change forecasting is achieved by using the CA-Markov technique to develop the model to describe a change of land-use influenced by the CA theory. This technique consists of 2 important parts including Markov chains and logistic regression methods. FAR index is applied to calculate the total floor area of each category that shows the intensity of different land uses. The amount of load is allocated to the proportion of total floor areas which is mapped to electric customer classes. Then, SLD is derived from the total load divided by the area of the polygon shape. This approach helped confine the results of spatial load forecasting so it is not to be excessive. As a result, the minimum value of Pseudo R2, ROC, and KIA are 0.3471, 0.9024 and 0.7120, respectively. Therefore, CA-Markov is a suitable method to forecast land-use change. During the forecasted year, the peak load in a service area is 222.42 MW, which causes the maximum SLD of 28.9 MW/km2. The spatial load density map developed by the proposed method is effectively used for power system planning and expansion.