坏数据的追踪辨识

TRACKING LDENTIFICATION OF BAD DATA

针对目前我国地区电网的特点，提出一种利用时间序列方法建立多个模型进行追踪辨识的坏数据处理方法。由于该方法与网络结构无太大关系，因而不仅适用于环网，也适用于开式网。另外，本方法对量测系统的冗余度要求不高，并且能给出坏数据的替代值，从而保证了系统的可观测性。已用于一实际地区电网，试验结果证明了本方法的有效性。

Many existed methods for bad data identification can’t work well in local system ,because the net-work is always open .In order to solve this problem, this paper presents a new method which is based on the theory of time series analysis. It has the following characteristics : first ,it uses multimodels to trackingly identify the bad data;second ,it can identify multiple interacting bad data. As this method has little relation with the structure of the system ,it can be used in all kinds of networks.It can use the tracking data to re-place the bad on,so ,the observability of the system can be kept.This method has been tested in a real local system ,the results show that it can satisfy the real-time applications.