Fault diagnosis of a cutting machine using convolutional neural network

Mohd Nazim, Muhammad Nazmie (2020) Fault diagnosis of a cutting machine using convolutional neural network. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full Text) Fault diagnosis of a cutting machine using convolutional neural network.pdf - Submitted Version Download (1MB)

Abstract

Based on the signals produced into the filters or decomposer to avoid the features in state control with the result of analysis that corrupted identity. But what is generally avoided is that a seasoned professional can know what's going on by following the signals produced on the oscilloscope with no analysis results. The input of the vision image and the feedback of the experience are two points in this process of human brain data classification. From the experiment can be easily divided into good or bad and used for identify the model constructed. The method perform in this object are plotting the vibration data from excel file in MATLAB. The plotted vibration data set will be classify into two classes, good and bad. The data set are trained and tested by using Convolutional Neural Network to validate the accuracy on classifying the vibration data. In the case of a closed loop control system, it is necessary to have a signal feedback point by point to adjust the required system. Fault analysis and data classifications usually require a pattern that is hidden between the signal points, which is precisely one of the specific fields of image representation to indicate a complex interrelationship. Data detection and fault analysis to explore the possibility of using the imaged signals as feedback for the architecture. From this research, the main result is to achieve highest accuracy for vibration data with low sampling time.

Actions (login required)

View Item