Digital twin-driven anomaly detection in multimodel combining fusion systems

The application of digital twin technology in the field of intelligent manufacturing is gradually increasing, and this paper proposes a multi-model combining fusion physical system anomaly detection method based on digital twin drive. The method constructs a physical model and integrates it into a unified digital twin model of the multi-domain system, simulates the physical model and iteratively modifies it, applies edge computing technology to model the information, constructs a data-driven neural network intelligent detection model, pre-processes the data, introduces the threshold ranking method to predict the anomalous features, stores the data into the database of the cloud platform for the anomalous feature extraction, constructs the final anomaly detection model, and saves the corresponding anomalous data in to the database as an update and optimization of the anomaly detection model. detection model, and save the corresponding anomaly data to the database as data for updating the optimisation model.