Adaptive fuzzy clustering based anomaly data detection in energy system of steel industry

The accuracy of the acquired data is very significant for the decision-making process for the purpose of the safety and reliability of the energy system in steel industry. However, owing to the instability and vulnerability of industrial system of supervisory control and data acquisition (SCADA), the anomaly data usually exist in practice. In this study, considering the data feature of the energy system, we classify the anomalies as the trend anomaly for the pseudo-periodic data and the deviants for the generic data. As for the trend anomaly, a dynamic time warping (DTW) based method combining with adaptive fuzzy C means (AFCM) is proposed by referencing the similar industrial processes; while, as for the deviants detection, a k-nearest neighbor AFCM algorithm (KNN-AFCM) is designed here for the local anomaly detection for the generic data. To verify the effectiveness of the proposed method, the real-world energy data coming from a steel plant are employed to perform the experiments, and the results indicate that the proposed method exhibits a higher precision compared to the other methods for the anomaly detection.