Extracting anomalies from time sequences derived from nuclear power plant data by using fixed width clustering algorithm

Time series is basically data recorded at successive points in time. In this paper we have analyzed time series data provided to us by Nuclear Power Corporation of India. We aim to find anomalies, correlations and patterns in the time series. In a nuclear reactor, anomalies can be generated due to various reasons, and it is important to identify the anomalies so that the cause of the anomaly can be found and corrective action can be taken. In order to analyze the dataset we have used Fixed Width Clustering Algorithm. While using this algorithm, we have proposed a dynamic method for deciding the cluster width that is used in clustering. We have also identified correlations between parameters in the dataset. We have cross checked all our results.