Research on Anomaly Detection Algorithm of Chemical Storage Tanks Based on FCM-LSTM

How to accurately detect the abnormal state of chemical storage tanks is the core problem of industrial control system, and real-time detection of abnormal changes of chemical storage tanks is very important for safety production management. However, traditional anomaly detection models are often based on content characteristics, and the accuracy of unsupervised learning is insufficient. Supervised learning requires a large number of learning samples. Therefore, in order to improve the accuracy of anomaly detection, this paper proposes an anomaly detection method for chemical storage tanks based on FCM-LSTM (fuzzy c-means optimization of long-term and short-term memory neural networks). This method establishes a framework for feature extraction, anomaly detection, pattern recognition and prediction analysis of chemical storage tank level, and can be used in anomaly detection of related businesses. Finally, the simulation experiment of abnormal detection of chemical storage tanks is verified and analyzed. After simulating the collected data of chemical storage tanks, the results show that the detection accuracy of this algorithm can reach 95.87%, which is 10.08% higher than that of PSO algorithm. Moreover, the detection error of this algorithm is small, and it can be completely used in the abnormal detection of chemical storage tanks. It provides a new idea and reference for abnormal detection of chemical storage tanks.