Application Of On-line Classification Of Substation Daily Load Curve For Load Modeling

With the realization of China's major power grid interconnection and grid scale expanding, real-time simulation on the grid is the new requirements of the practical application. It's a new challenges for the power system composite load modeling that composite load model of power system be able to accurately describe the time-varying load characteristics and real-time simulation speed to be consistent with the power actual dynamics of system in the condition of load composition and operation state changing randomly. Load characteristics online modeling is an ideal way to resolve the conflict between load random changes and the requirements of real-time simulation to the integrated load model.The requirements of load characteristics online modeling to the data is that accurately reflect the essential characteristics of time-varying load. Substation daily load curve provided by the SCADA system is based on the actual measured data and contains a wealth of information which reflects the characteristics of load constitute. It can be applied to composite load online-modeling. Series theoretical and applied research has been established against the actual need of using the substation SCADA system daily load data to the on-line load modeling considering the load characteristics classification method based on the daily load curve, pre-processing method of daily load curve data and application platform of Load online classification of load characteristics in the substation.Unsupervised classification technology is commonly used in the load characteristics classification and methods based on the fuzzy clustering analysis are widely used. This paper analyzes the existing methods of load characteristics classification and a new classification based on the improved FCM algorithm combining subtractive clustering method has been proposed which can overcome the shortcomings of traditional FCM clustering algorithm like be more sensitive to initial values and convergence results depend on the cluster prototype parameters. Based on the classification results of daily load curve of a substation in Hunan Province and 14 districts substation, it can be seen that the improved FCM algorithm is better than the traditional FCM clustering algorithm in the classification accuracy, category center description ability and algorithm efficiency. At the same time, the improved FCM algorithm classification result is proved to be reasonable from the similarity and dissimilarity analysis on the daily load curve classification results of 14 districts substation.Substation daily load curve being decrypted of the measured data from the SCADA system interfere normal load variation and effect accuracy of load model should process before use. This paper analyzes the features of substation daily load curve and proposes a new system for outlier automatic processing of daily load curve. Completing the missing data use the Neville algorithm firstly, identifying the outlier by horizontal similarity of the load using various types of characteristic curves generated by improved FCM algorithm for cluster of Substation daily load curve.Finally, adjusting outlier by characteristic curves. Case analysis has achieved good results.In order to impel application of on-line classification of load characteristics, this paper exploits an application platform of on-line load characteristics on substation. This application platform data Sources of Oracle Database SCADA system in Hunan province based on modular design consists of automatic Interface platform and load characteristics classification module of substation. Implementation process uses VC++as a development language and stores processed data in SQL Server 2000 back-database by remote access and automatic process daily load curve data.Substation load characteristics online category application platform developed in this paper is an important part of "Real time online load modeling technical support system" and has been in actual running in Hunan Provincial electric power dispatching and communication center.