Multi-objective Intelligent Optimization Model on Dynamic Error Measurement and Fault Diagnosis for Roll Grinder NC

The error measurement and diagnosis process of roll grinder NC has dynamic complexity, non-linearity, and comprehensive characteristics. However, presently roll error measurement examination mostly uses the manual examination or single parameter optimization, and the efficiency of fault diagnosis is also inefficient. In this study, the multi-objective intelligence optimization model (MIOM) is applied to the roller error measurement and diagnosis. The algorithms are hybrid with modern intelligent ones, such as Artificial Neural Network, Fuzzy Logic Inference and Genetic Algorithm, etc. Fuzzy control rules are created base on expert knowledge. Multi-objective parameters can be simultaneously optimized in the same process. Meantime, by analyzing the optimized results of each error parameter, the state space observation equation model can be established, and the stability of the system can be calculated by NN. Therefore, the fault spot can be inferred out. Finally, according to the error diagnosis results, the diagram of curves is drawn by the 840D HMI. Through the experimental simulation tests, the application of MIOM can simplify roll error measuring and diagnosing processes, and the operations for roll grinder NC are more intellectualized.