Multi-Sensory Data Fusion System for Tool Condition Monitoring Using Optimized Artificial Fuzzy Inference System

Currently, most of the available tool wear condition monitoring systems are based on the signal features quantities that are correlated with tool wear. The evaluation of tool wear based on one sensor is not reliable because the measured features depends not only on tool wear but also on other process parameters and random disturbances. For solving this problem, multi sensor data fusion is used to combine disparate sensory data. The obtained information are more accurate and reliable. In this research, combination of vision, current, strain and vibration sensors for predicting flank wear land is proposed. An optimized adaptive neuro-fuzzy inference system (ANFIS) model is developed to fuse the surface image, motor current, strain and vibration signal features. The structure of proposed ANFIS model has four inputs and one output. The inputs of the model are entropy of surface image (which is filtered by undecimated wavelet transform), time-frequency marginal integral of the motor current, Shannon entropy of strain and Shannon entropy of vibration signals, while output of the model is the flank wear. The results showed that the optimized ANFIS model can be used to fuse the signal features and predict tool flank wear with high accuracy.