Study on On-line Monitoring Method for Grinding Quality

A on-line monitoring new method for grinding quality is presented based on theory analysis and test study, which collects the information of grinding quality from the AE signals produced by friction and grinding process to realize the on-line intelligent detection and prediction of grinding quality by A wavelet neural network & Fuzzy BP algorithm based network. The reliability and feasibility of the method is proved by actual test. The test results indicate that this method can be used to monitor grinding quality on line. Preface The process of intelligence and automation is a general tendency in manufacturing [1]. Grinding, as an important finishing process, is preventing this trend now, because there is not suitable on-line monitor method for grinding quality [2]. Therefore, the development of the intelligent monitoring techniques for permanent on-line system identification is a very important task in order to enable detection of incipient deterioration and remedial action to be taken before the fatal breakdown for grinding process. The majority of standard methods for the grinding surface roughness estimation are not suitable for in process measurements and on-line monitoring. This paper describes an attempt to obtain on-line information about the grinding quality in grinding process. For this purpose, a wavelet neural network & Fuzzy BP algorithm based intelligent method for on-line monitoring grinding surface roughness, grinding burning and grinding chatter is presented based on theory analysis and practical test in this paper. Feasibility Analysis of On-line Detecting Grinding Quality by Acoustic Emission Friction is a complex phenomenon influenced by the combined effects of adhesion, asperity deformation, plowing by hard asperities, and wear of entrapped particles. The changes in the stress distribution around asperity contact at surface traction generate transient stress waves called acoustic emission (AE), which can be detected by suitable AE sensor [3-4]. It is well known that elastic distortion and the plastic deformation become large under same pressure as surface roughness worsening in friction process, grinding burning and grinding chatter states. This will further result in more strength elastic stress wave. Therefore, if the mapping relationship between the some feature value of AE signals and grinding surface roughness, grinding burning and grinding chatter, then AE signals can be used to on-line detect grinding surface roughness, grinding burning and grinding chatter continually. J.S. Kwak and J.B. Song [5] have developed a diagnosis system for grinding process by using acoustic emission signals, and verified this system has high trouble recognition rates by test. E. Susic and I. Grabec presents a newly developed for on-line estimation of the roughness and hardness parameters of surfaces involved in sliding friction process, in which the information about the process is extracted from the acoustic emission signal generated by friction. The system performance was tested in the on-line pin-in-disk experiment for different classes of the surface roughness, and the recognition reliability for placing given cases in correct classes was found to be greater than 94%. Many study results indicate that acoustic emission signals can be used to monitor Key Engineering Materials Vols. 259-260 (2004) pp 287-292 online at http://www.scientific.net © (2004) Trans Tech Publications, Switzerland Online availa le since 2004/Mar/15 All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 130.203.133.33-11/04/08,03:23:21) 288 Advances in Grinding and Abrasive Processes the grinding quality on line. Experimental Investigation Experimental Method. The experimental device is shown in Fig.1. The flat end of pin attach with AE sensor, and spherical end contact with work-piece surface by pressure. The contact pressure between the pin and work-piece is adjusted to about 70 N by knob in measuring process. The AE signal of friction between pin and work-piece is amplified by head amplifier firstly, then loaded TOP data acquisition in computer, and processed by TOPVIEW2000 virtual equipment software. In order to preserve the surface roughness of work-piece and allow only for the wear of the pin surface, the pin hardness was chosen an order of magnitude lower than that of the work-piece material. Here, the copper material is chosen as pin material, the pin diameter is 8 mm, and the surface roughness of spherical end with r=4 mm is 0.16 μm. In order to get the relationship between the feature value of AE signal and grinding quality, a series of tests were conducted on a cylindrical grinder. The test condition is shown in Table 1. Table 1 Test conditions Menu Conditions Grinding wheel [mm] Type: WA60LV, Dimension: 228×24 wheel speed [m/s] vs=26.3 Work-piece Material: 45 Steel, Hardness: HRC45 Work-piece speed [m/s] vw=0.10~0.30 In-feed [mm/min] 0.5, 1.0, 2.0 Grinding fluid Dry grinding Dressing conditions Single particle diamond Dresser, Cutting deep: 0.0125 [mm], Lead: 0.015 [mm/rev] 783 .7 7 83.8 7 83 .9 78 4.0 7 84 .1 78 4.2 -80 0 -60 0 -40 0 -20 0 0 20 0 40 0 A m p lit u d e ( m v ) G r inding tim e (S)