Multi-agent failure prediction based on data assimilation theory

In the context of preventive maintenance being valued, focusing on the defect of closure and low problem solving ability presented by single failure prediction system, combined with meteorological data assimilation theory and multi-agent technology, failure prediction of steel continuous casting equipment was researched. A distributed failure prediction system based on ensemble Kalman filter (EnKF) and multi-agent technology was developed, which overcomes the inelasticity of conventional prediction method used in a nonlinear environment. A prediction model with higher precision and higher efficiency was built, whose feasibility and effectiveness were verified by an actual case.