Performance Monitoring of CNC Machine Using Modelsim

Abstract

Tool condition monitoring is one of the emerging areas in the manufacturing industry. This paper proposes HDL simulator-based simulation model using Modelsim to detect and classify the tool condition using the hybrid network. The system uses multiple sensors for data collection from the CNC machine. Multiple sensor data such as vibration and temperature as well as actual machine parameters are taken into consideration for the system design. The data collected is pre-processed and fed to the self-organizing map (SOM) and a Hebbian network which is a hybrid model. The data is classified according to its range, and which are mapped to get the SOM neurons. The Hebbian network designed is the single-layer feedforward neural network. The recognition process is robust to the number of changes in the input samples. The system operates in training and testing modes. The tool condition is indicated in Matlab with a message window. The system correctly detects the tool condition, with a simulation accuracy of 97.16% which is promising and insists on hardware model development.