On-line Monitoring System Of Micro-hole Drilling Based On Rough-Fuzzy Control

With the development of science and technology and high-tech products becoming more and more accurate, integral and miniature, there is an increasing need for the micro-holes and the demand for the quality is higher and higher in the manufacture industry. Nowadays, there are many methods for machining micro-holes, but from comprehensive consideration, machining by using power drills is still most economical and widely used in the world, so it plays a leading role. It can obtain high production efficiency and good processing quality. However, micro-drills have low intensity easily to be broken. Furthermore, because the statistic for the life-span of micro-drills is very dispersive, it is difficult to estimate the working life of a micro-drill. Once a micro-drill break off, it is hard to take it out from the workpiece so that the workpiece is usually discarded. Moreover, because of severe wear-out of the micro-drill, the quality for the holes declines very quickly before the drill is broken. So, state monitoring for micro-drilling has quite important and practical significance. The on-line and real-time micro-drilling monitoring system for triphase currents of the spindle motor based on Rough-Fuzzy control is built up in this paper.Firstly, this paper introduces hardware and software of on-line and real-time micro-drilling monitoring system. The hardware system mainly contains high speed precision drill press, Hall current sensor, A/D data-acquisition card, IPC, transistor manostat and SCM control system. The software system adopt mixed programme of LabVIEW and MATLAB. It has the following function module: serial communications, data acquisition, data read, storing data, data display, numeric filter, fuzzy identification, alarm display and so on. During micro-hole drilling, Hall current sensor transform triphase current signal of the spindle motor into voltage signal which is then acquired into the computer as digital signal. Computer analyses these digital signals timely, and has serial communication with SCM. Then, according to eigenvector, fuzzy control module judges drill state, which determines whether the drill is going on drilling or withdraw. If the judgment resultis equal to or larger than the wear-out threshold value, monitoring system give an alarm signal, and then computer sends out a signal to SCM control system in order to control step motor, which ends drilling. So, it can protect drill and workpiece and implement the aim to monitoring. The main function of SCM control system is that in terms of the signals and the parameters computer gives, it sends out pulse signal to the step motor to control the micro-drill. Because the signal that SCM control system gives is very weak, it has to pass through power amplifier to be magnified, which can improve the drive capability to the step motor.Secondly, this paper analyses rough sets theory and fuzzy control theory as well as the intersectant and synthetical technique. Fuzzy control theory and rough sets theory are to study the uncertainty of knowledge, have each trait and strong complementarity, so it is necessary to unite the two theories partly. Based on the two theories, it focuses on the study about the entire design of fuzzy logic controller, including ascertaining subjection function, instituting fuzzy control rule, fuzzification, fuzzy logic reasoning, defuzzification and so on. At first, a fuzzy logic controller of the drill states is primarily design. The triphase currents of the spindle motor are chosen as input quantity of this controller, the drill wear state as out quantity, and these two quantities are divided as five grade. Primarily establish 53 pieces of control rules. Then, rough sets theory reduces the fuzzy logic controller. Reduction includes attributes reduction and rules reduction. Through attributes reduction, it can wipe off redundant factor attributes; then through rules reduction, it can wipe off the redundant factor attributes in each rule; finally, the least decision table can be attained. During attributes reduction, specification matrix and specification function are used to reduce this fuzzy logic controller. Through this, none of redundant attributes are found. During rules reduction, by comparing importance of the rules, that is assuming that whether there is conflict in decision values when certain attribute value is deleted. If there is a conflict, the deleted attribute value is a core. If there is no conflict, the deleted attribute value is redundant. Through this reduction, 55 pieces of rules are reduced to 33 pieces, and 159 conditional attribute elements are reduced to 89, which accordingly simplifiesthe initial fuzzy logic controller. Fuzzy decision table after reduction is accordant with the initial one.According to experiment data, we can find drill wear-out threshold value and then do the on-line and real-time micro-drilling monitoring experiments. The experimental result indicates that the reliability of on-line and timely monitoring is related with drill wear-out threshold value. When a small threshold value is chosen, we can have a good effect for avoiding drills being broken, but the quantity of holes is less before alarming. So, during manufacture we should properly choose drill wear-out threshold value according to practical situation. Therefore, applying rough-fuzzy control technique into micro-drilling monitoring system can correctly identify the drill wear state and timely give an alarm for quick return which can effectively avoid drill break. It provides a very effectively way for intellectualized on-line monitoring of the drill state in micro-drilling processing.