The Study Of Data Processing Technology In Wireless Sensor Network Of Power System

Wireless sensor network (WSN) has become a research hotspot now since it has such advantages as the accuracy of the data collected by wireless sensor, the simple deployment as well as the low cost, and it does not need to on-site maintenance etc. Generally, wireless sensor nodes are directly arranged in exposed geographical environment. Factors like weather conditions, communication ability of sensor nodes, signal failure, and human elements can lead to the frequent breakout of the communication link, thus making the gathered sense data lost or exceptional in the process of transmission. Deleting the data set containing missing data or abnormal data directly can cause the loss of massive information. Therefore, reasonable filling for the missing data and amendment with recognition for bad data can improve the accuracy and reliability of the data analysis results.In order to improve the accuracy of the estimated missing data in Wireless Sensor Network(WSN), a self-decision interpolation algorithm was proposed. The algorithm selected different estimation strategies of missing data according to the spatial correlation of the data sets and the continuity of missing data, than introduced the Auto-Regressive and Moving Average Model(ARMA) into the study of missing data interpolation. In corresponding to the traditional missing value estimation algorithm, the proposed algorithm not only considered the characteristics of wireless sensor networks, but also took the characteristics of the data themselves into account. The experimental results on the real data sets show that the proposed algorithm improves the precision of estimates for missing data.For processing the missing data of WSW instantly, an algorithm which combines the second exponential smoothing method and linear regression analysis is presented. This algorithm can deal with real-time data, and has high estimation accuracy. Many times results show that, the algorithm can estimate the missing data.In order to improve to the accuracy of the estimated bad data in the wireless sensor network, a new bad data detection method was proposed. The algorithm takes full account of the characteristics of wireless sensor networks. The proposed algorithm to model uses the prior data bank and the spatial correlation of WSN, and then it gives the range of the deviation. If the detected data in a certain deviation range that we have given, the data is a normal data, otherwise the data is a bad data. The test results on the data sets show that the proposed algorithm able to detect bad data, and then gives dad data an estimated value.