Search

Load model is one of the most important elements in power system operation and control. However, owing to its complexity, load modeling is still an open and very difficult problem. Summarizing our work on measurement-based load modeling in China for more than twenty years, this paper systematically introduces the mathematical theory and applications regarding the load modeling. The flow chart and algorithms for measurement-based load modeling are presented. A composite load model structure with 13 parameters is also proposed. Analysis results based on the trajectory sensitivity theory indicate the importance of the load model parameters for the identification. Case studies show the accuracy of the presented measurement-based load model. The load model thus built has been validated by field measurements all over China. Future working directions on measurement-based load modeling are also discussed in the paper.

This paper formulates two identification principles of dynamic load models with discrete time form, and emphasizes the efficiency of the model responses feedback identification principle in dynamic load modeling. By combined with artificial neural network, the concept of feedback neural network model is advanced and the training algorithm is presented. The example of load modeling based on field test indicates that the feedback neural network model as presented in this paper is more efficient than the model-based on the system responses feedback identification principle.

This paper elucidates the closed loop structure of 'Load-System' and the characteristics ofrandom signals hi power system, and presents the minimum cross-correlation function method, an on-line identification method for static load modeling. This method is essentially different from the current least squares fitting method in the main idea, and is suitable for the static load modeling under the normal operating conditions with low noise-to-signal ratio. The simulation and field data modeling of this method are carried out and the effects, of this method are also compared with the least squares fitting method.

Power system is approaching its limits due to the deregulation of power industry and unavailability of new transmission corridor. However power system security assessment is carried out and the operating limits are set according to the given contingency list which might happen in very rare frequency. In other words, the frequency of fault occurrence and the risk of blackout is varied when the operating condition (load level, status of power equipment, and weather) changes. How can we do the security assessment based on the operating condition to get the best equilibrium between the security and the profit is the focus of this paper. The idea of risk based on-line power system security assessment is suggested. A precise calculation method of average failure rates of components is presented. Besides, operating condition based failure rates of components are also analyzed systematically in the paper. The Monte-Carlo simulation demonstrates the viewpoint of the paper.

A fault location algorithm using estimated line parameters is provided in this paper. The characteristic of this algorithm is using estimated line parameters. And the influence of the line parameters is eliminated. Support vector machines theory is used to estimate transmission line parameters, which is a nonlinear black box modeling problem. The historical data is used as training sample. EMTP simulation shows that this method is to improve the accuracy of fault location notability.