OPTIMIZATION OF MULTIPLE PANEL FllTlNG IN AUTOMOBILE ASSEMBLY

A systematic approach is presented to obtain improved panel tit quality through the use of an optimum panel fitting strategy. ‘The objective of the optimal1 panel fitting strategy is to determine the location of the panels on the automobile body such that the gap and flush variation of the panel fit are mmimized. This approach uses measurement data from both the panels and the body-inwhite (BIW) to determine the optimum position of multiple panels in an automobile body opening. First, some indices are defined to quantify the quality of a panel fit. Second, the sources of variation in the gap and flush are presented. Then the multiple panel fitting problem is formulated into a constrained optimization model. Tbe effects of the optimization process for multiple panels are then demonstrated and validated through the use of computer simulation. The computer simulation demonstrates the optimization model and algorithm by reducing the within-car gap and tlush variation on average by 24.3% and by as much as 43.4% in the case study presented. INTRODUCTION In today’s highly com,petitive automobile market the quality of the product is becoming increasingly important. Among the quality concerns of the automobile manufacturers is the quality of the fit and finish. Specifically, the quality of the panel fits of the automobile ranks high in concern due to the high warranty costs. An inadequate panel fit will contribute not only to functional problems such as water leakage and wind noise but aesthetic problems such as uneven gap and flush. The indices that are used to determine the quality of the panel fit are the gap and flush variation, gap range and flush range. These terms will all be defined more rigorously later in the paper. The dimensional variations of the gap and flush between the BIW and the panel, or between panels, arise from four sources which are (figure 1): (1) dimensional variation of the panels; (2) dimensional variation of the BIW; (3) variation of the panel fitting process; and (4) effects of painting and general assembly. Each branch of the tishbone diagram shown bellow has various sources of variation which effect the gap and flush variation. RodR@a~ ‘+i&,,, Asse,,,b,y / /” FIGURE 1: FISH BONE CHART OF VARIATION COMPONENTS IN PANEL FITTING lhe first two variations are subassembly variations and need to be unproved at the subassembly level. lhese issues have been addressed by Wu and Hu (1990). ‘The fourth source of variation is a topic of research outside the range of this paper. This paper focuses on the variation induced by the panel fitting process, In this paper, a special example of panel fitting, automobile doors, is used to demonstrate the optimization process ‘Ihe control of dimensional deviation in the automobile industry has been mainly done by feeding Statistical Process Control (SPC) data back to manufacturing processes such as stamping or part subassembly, Wu and Hu (1990). Fitting of rigid bodies in general was suggested by Bonna and Menga (1984) which presents a general constrained optimization process of a single body. An approach to integrate the data directly to the fitting Transactions of NAMRUSME 241 Volume XXIII, 1995 Copyright © 1995 Society of Manufacturing Engineers. All rights reserved. process was proposed by Wu, Hu, and Wu (1994) where the best fitting parameters are transformed to a door fixture location. However, Wu et al. (1994) described the best fitting process in which only one door was fit at a time. Recently, Schuler, Zussman, and Seliger (1994) and Sakai and Yasumatsu (1993) presented an approach that is based on 100% measurements to determine the position of a single panel based on different location criterion. However, little literature exists in the area of multiple body fitting optimization. In recent years, the implementation of the in-line Optical Coordinate Measurement Machine (GCMM) in the automotive industry and the increasing accuracy of industrial robots provide new opportunities for the development of a flexible assembly system for the fitting of automobile panels. This flexible assembly system attempts to find the optimum position of the doors in the BIW such that the within-car gap and flush variation is minimized. In this approach, multiple doors are to be tit simultaneously so that the dimensional deviation in the various parts can be distributed evenly over the entire door tit of the automobile by the optimization process. However, effective implementation of this multiple panel fit optimization remains a challenge. Some of these challenges can be summarized as follows: 1. Transformation of the discrete OCMM measurement data to a geometric model of the panel fit. 2. Formulation of an optimization objective function to minimize the within-car gap andflush variation. 3. Formulation of constraints to restn’ct the optimization problem. In this paper, a systematic approach to obtain optimum multiple panel fitting is presented based on a constrained optimization problem. The approach is demonstrated by locating the front and rear doors of an automobile simultaneously. Constraints that are taken into account are Euler parameter constraints, minimum gap constraints and minimum parallelism constraints. The organization of the paper is as follows: Section 2 describes the geometric model of multiple panel fitting. Section 3 presents the constrained multiple panel optimization scheme. Section 4 verifies the proposed method through computer simulations. In section 5, sensitivity analysis was performed on the optimization constraints. The conclusions of the paper are discussed in the section. GEOMETRIC MODEL The objective of the optimization model is to minimize the variation in the gap and flush. Essentially, the gap and flush are projections of the vector, r, from the Door to the BIW at the measurement points. The projection of this vector, J, onto the vectors GPi and FPi at the sphfic point will result in the measure of the gap and flush respectively. The vectors GPi and FPi are unit vectors which detine the direction of the gap and flush at a specified point i (figure 2). The direction of these vectors is governed by the design of the gap and flush. Essentially, the gap vector, GPi, is the unit vector tangent to the design direction of the vector r, Similarly, the flush vector, FPi, is the unit vector perpendicular to the design direction of the vector v ,. An illustration of the gap and flush measures is shown in figure 2.