Genetic tool monitor (GTM) for micro-end-milling operations

Abstract

Almost all existing tool condition monitoring methods require either the critical parameters of models which are experimentally found or the self-learning algorithms that are trained with existing data. Genetic Tool Monitor (GTM) is proposed to identify the problems by using an analytical model for micro-end-milling operations and genetic algorithm. The current version of the GTM is capable to monitor the micro-end-milling operations without any previous experience and is able to estimate symmetrical wear and local damages at the cutting edges of a tool. Genetic algorithms (GA) are found as a promising health monitoring tool if an expression exists and the necessary computational time is allowable in that particular application. GTM generates meaningful information about the ongoing operation and allows the establishment of rules based on the operators' experience.

Introduction

The consistency of the machining quality is very important in an automated manufacturing. Many monitoring techniques have been developed to detect tool breakage and to estimate tool wear by evaluating the most important characteristics of the signals coming from sensors such as dynamometer, accelerometer(s), acoustic emission sensors, thermocouples, and microphones. The first approach was to evaluate the characteristics of the signal at different tool conditions and to set the limits or to relate them to tool wear. Later, intelligent computational algorithms such as neural networks were used to automate this task. To use these methods effectively, the operating conditions such as speed, feed rate, depth of cut, and the tool-material couple should be identical or very close to the test conditions. GTM was developed to interpret the cutting force signals by using analytical expressions. Theoretically, once the key parameters are estimated in a single test from the data of any cutting force component, cutting operations can be monitored by measuring any cutting force at any operating condition by using this method. However, GTM should be used very conservatively to eliminate false alarms and to ensure that the wear estimations are performed accurately within an acceptable time frame using a low cost computational hardware. In this study, the performance of GTM will be evaluated on simulated and experimental cutting force data by considering the micro-end-milling of soft electrode materials.

The genetic algorithm [1], [2], [3] estimates any number of parameters of an analytical expression by using a search method. It works with linear and non-linear analytical expressions, and very complex conditional statements can be included in the objective function. The GTM concept was developed to estimate the key parameters such as the cutting force coefficient and run-out of the analytical model [4], [5], [6], [7], [8] during micro-end-milling operations. Other researchers have used GA to determine the optimal cutting conditions [9], [10], [11], [12] in machining and monitoring of turning [13], [14] operations.

Generally, in academic studies, continuous monitoring of the machining operations have been aimed, while the workpiece is cut and the cutting conditions vary. However, GTM was developed for micro-end-milling of the graphite like electrodes of electrical discharge machining (EDM) by considering the following two requirements:

• The magnitude of the cutting forces required to remove the typical electrode material of the EDMs is small with respect to the noise. It is necessary to cut an aluminum type material periodically to collect the data with acceptable signal to noise ratio without wearing the tool.

• A continuous monitoring system inspects the machining operation many times per second during the actual machining operation. Unexpected force profiles are always encountered when the cutting conditions vary. Even 0.01% wrong estimations will give several false alarms every hour. Machining of electrode with the same micro-tool sometimes takes a few days and operation continues during the off-duty hours of the operators. False errors should be completely eliminated.

The proposed GTM estimates the cutting force coefficient from the periodically collected data. In some cases, also the run-out was estimated. Cutting force coefficient indicates the dullness of the cutting edge and continuously increases while the operation continues. Contact length of the cutting edge can be calculated when the run-out is estimated from the same data. Although, the genetic algorithm of the GTM [15] may identify most of the cutting parameters theoretically, the computational time will be unacceptable in the practical applications and the maximum two variables will be estimated in this study.

In the following sections, the theoretical background, the proposed GTM, the simulation, the experimental data collection, results and conclusion will be presented.