Extract Voltage Stability Operating Guidelines using Decision Tree

The operational planners in power grid companies perform a large number of offline voltage stability studies for a variety of operating conditions. How to effectively determine voltage stability limits from massive amount of offline data has become a challenge for operational planners. This paper applies decision tree (DT) to extract operating guidelines from off-line analysis results. Case study results on a practical power system prove that voltage stability operating guidelines, which are extracted by decision tree, can be used by system operators for online monitoring and control of electric power systems. Introduction Voltage stability is the focus of planning and operating electric power system. In order to ensure the power system to maintain acceptable voltage both in normal operation and in the event of severe disruption, the operational planners have done a lot of offline analyses and collected a large amount of offline data [1][2]. The operational planners developed voltage stability operating guidelines based on their experience and subjective judgement. With the development of power system, the structure of the power grid has become very complex and the scale of computational data has increased dramatically. Meanwhile, large scale integration of renewable energy resources increase complexity and uncertainty of power system operation. Once facing with the massive data, the efficiency of manual data processing will be very low and the accuracy cannot meet the needs of system operation [3][4]. Due to the advantages in extracting useful information automatically from massive data, data mining techniques have been applied into power system. Recent applications of data mining techniques, including artificial neural network (ANN), self-organizing map (SOM) and radial basis function (RBF), support vector machine (SVM) [5], Reference [6] applied a multiclass SVM for static voltage stability assessment. Reference [7] proposed a new hybrid classifier for transient stability prediction, which can forecast both TSS of the power system and synchronism state of each generator in response to a disturbance. A new approach to ANN based VCPI prediction method to be used for extracting voltage stability operating guidelines has been proposed in reference [8]. The above mentioned methods all belong to black box model, which cannot reveal the internal rules of the massive data. Decision tree belongs to white box model. One of the advantages of decision trees is that they produce models that are relatively easy to interpret. This paper uses decision tree to extract voltage stability operating guidelines. Decision Tree and C4.5 algorithm A decision tree is a flowchart-like tree structure, where each internal node denotes a test on an attribute, each branch represents an outcome of the test, and each leaf node holds a class label. The topmost node in a tree is the root node [9]. A typical decision tree is shown in Figure 1. 4th International Conference on Electrical & Electronics Engineering and Computer Science (ICEEECS 2016) Copyright © 2016, the Authors. Published by Atlantis Press. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/). Advances in Computer Science Research, volume 50