The objective of this research is to develop an effective inspec-tion planning strategy for sculptured surfaces in the OMM(on-machine measurement) process. As a first step, effectivemeasuring-point locations are determined to obtain optimumresults for given sampling numbers. Two measuring-point selec-tion methods are suggested in this study based on newlyproposed CAD/CAM/CAI integration concepts:1. By the prediction of cutting errors.2. By considering cutter contact points to avoid the measure-ment errors caused by cusps.As a next step, the TSP (travelling salesman problem) algorithmis applied to minimise the probe moving distance. Appropriatesimulations and experiments are performed to verify the pro-posed inspection planning strategy in this study, and the resultsare analysed.Keywords: CAD/CAM; CAI (computer-aided inspection);CMM (coordinate measuring machine); Inspection 29183
planning;OMM (on-machine measurement); Sculptured surface1. IntroductionIn sculptured surface machining, the tool path is determinedby the CAD/CAM system, based on given geometric data, andthen a series of operations such as rough cutting and finishcutting is performed on a CNC (computerised numericalcontrolled) machine using appropriately chosen cutters.Subsequently, after terminating the cutting processes, polishingprocesses are required to remove remained cusps and to pro-duce the desired shapes. In order to measure machined sculp-tured surfaces precisely, a computer-based CMM (coordinatemeasuring machine) system is often used. To reduce measuringerrors produced by operating the CMM system manually, toCorrespondence and offprint requests to: Dr M. Cho, Departmentof Mechanical Engineering, Inha University, 253 Yonghyan-dong,Nam-gu, Inchon, Korea. E-mail: chomwnet inha.ac.krimprove measuring accuracy and to obtain optimal measuringconditions from the CAD database, work [1–3] on CAD/CAI(computer-aided inspection integration systems) has been car-ried out and has contributed toward measuring technique devel-opment. It thus becomes possible to inspect machined surfaceaccuracy effectively by analysing given CAD data. In a generalcutting process, the accuracy of the machined part depends onthe effectiveness of the CNC machine. If parts include difficultto machine shapes and imposed difficult tolerance criteria, thecutting process cannot progress favourably because it incurscost and time losses.Generally, this possibility is minimised through simulatingand validating the machining conditions based on variousCAD/CAM software systems. However, in a real cuttingprocess, it is possible to incur machining errors due to uncon-sidered causes, for example, geometric errors of the machineitself or tool deformation. Therefore, these CAM data have tobe taken into account in an accurate measuring process. Inorder to integrate CAD/CAM data into the CAI process, theOMM (on-machine-measurement) system has recently focusedon a new measuring process technique [4].The purpose of this study is to implement an improvedinspection process in an OMM system by integratingCAD/CAM data into the CAI process. Geometric shapes ofthe machined surface are taken into account through amachined surface prediction method. For this purpose, effectivemeasuring-point locations are determined to obtain optimalresults for given sampling numbers. Two measuring-point selec-tion methods are suggested based on the newly proposedCAD/CAM/CAI integration concept:1. By the prediction of cutting errors.2. By considering cutter contact points to avoid the measure-ment errors caused by cusps.As a next step, the TSP (travelling salesman problem) algorithmis applied to minimise the probe moving distance. Appropriatesimulations are performed to verify the proposed inspectionplanning strategy, and the results are analysed. To validatesimulation results, the measuring process is performed with 2. CAD/CAM/CAI Integration ConceptIn general machining processes, product shapes are designedbased on CAD systems, and CAM systems generate the mach-ining data in order to carry out the machining process. In thiscase, it is necessary to inspect the quality of the machinedparts. In the conventional measuring process the machinedsurface accuracy is often measured on a CMM (coordinatemeasuring machine) system, which is controlled by a computer.To improve the measuring accuracy based on the measuringerrors produced by a manual inspection process, work has beencarried out to derive optimal inspection conditions from a CADdatabase [3,5], which is often referred to as a CAI (computer-aided inspection) process.As a result of this work, it is possible to measure thedimensional accuracy of complex shaped products throughappropriate measuring point selection and an automatic analysisof the inspection results. Figure 1(a) illustrates the interrelationbetween CAD, CAM, and CAI in this case. However, aproductivity drop can occur because of the high cost of theCMM system and the increase of measuring process time.Generally, in the case of the CMM process, the locations ofthe measuring points can be obtained by using an “equi-interval” measuring method [6]. A machined part is completedon a CNC machining centre through several steps (roughingand finishing). According to the geometric type of the tooland the cutting path, cusps remain on the final machinedsurface. Therefore, if the geometric form of the machinedsurface is not taken into account, the measuring points maybe located irregularly on the tip or the bottom of the cusps,as shown in Fig. 2. In this case, the equi-interval measuringmethod leads to measuring errors, and when the rough cuttingprocesses are performed using a large diameter tool, the possi-bility of introducing measuring errors increases. This is becausethe measuring process has to be carried out on the basis ofCAD data only. Moreover, if the measuring process intervenesbetween consecutive cutting processes, the machined part must be released for inspection on CMM, and then must be fixedagain on the CNC machine for the subsequent cutting process.In this case, the calibrated workpiece coordinates will be lostbecause of machined part movement; accordingly, the sub-sequent cutting process cannot be achieved accurately.Therefore, we need a new inspection process, which can beperformed directly on the CNC machine and on the basis ofCAM data.In order to solve these problems, the OMM system has beendeveloped. The OMM system makes it possible to carry outthe measuring process directly on the CNC machining centreby exchanging a cutting tool for a measuring probe. In thiscase, both the CAD and the CAM database have to be simul-taneously considered to constitute an inspection databasebecause the machining and measuring processes are beingperformed on the same CNC machining centre.In order to implement an effective measuring process on theOMM system, we propose in this paper an improved inspectionprocess planning strategy based on a concept, called the“CAD/CAM/CAI integration concept” (see Fig. 1(b)). TheCAD/CAM/CAI integration concept means that the CAD/CAMdata are integrated into the CAI process. Compared to theconventional CAI process on a CMM, this makes it possible to establish “in-process-inspection”. Therefore, in this paperwe will show how to realise the CAD/CAM/CAI integrationconcept by using the OMM system.3. Machining Error Prediction ProcessFor the conventional inspection process using the CMM system,some workers have proposed to integrate CAD data into theCAI process [1,3]. However, it is not possible to establish anaccurate CAI process because the geometric shape of themachined part, as the actual measured object, does not conformwith CAD data. Therefore, we try to integrate the CAM datainto the CAI process by taking into account the geometricinformation of the machined surface. For this purpose, anappropriate analysis of the machined surface shape has to beperformed in order to carry out the CAI process effectively.This analysis corresponds to the machining error predictionprocess, which consists of predicting the machined surfaceshape.In order to manufacture sculptured surfaces, it is first neces-sary to calculate CC-point (cutter contact point) locations sothat they correspond to tool size and tolerance criteria. Sub-sequently, CL-points (cutter location point) are determinedfrom these CC-point locations. Generally, linear movements ofthe cutter between consecutive CL-points execute the cuttingprocess. Therefore, it becomes inevitable that some deviationerrors occur between the desired sculptured surface and realmachined surface. There has been some work on improvingthese problems by predicting cutting errors. In this paper, wepropose an improved surface prediction model based on thestudies proposed by Jerard [7,8] and Kim [9]. Figure 3(a)illustrates the proposed approach.In the first step, it is necessary to determine check pointsto estimate the errors on the machined surface. Based on theequivalently sized vectors that are determined at these checkpoints on the X/Y-plane, a virtual workpiece is constructed. Inthe case of conventional 3-axis machining, these vectors areinitially considered as point vectors parallel to the Z-axis.Comparing these point vectors with the tool path generated byCC-points and with the tool shape, it is possible to obtain asimulation of the geometrical form of the machined surface.The machining errors can be predicted by comparing thissimulated machined surface with the designed surface in theCAD system. The details of this concept are described in thefollowing section.3.1 Machined Surface PredictionIn the cutting process, the tool path mainly determines themachined surface shape. The machined surface shape is determ-ined by the amount of the material removed by the cutter, andit can be modelled by considering the contact zone betweenthe tool path and the workpiece. Figure 3(b) illustrates thisconcept. As depicted in the figure, while the cutter moveslinearly between the two consecutive CL-points a and b, theworkpiece is removed by the tool envelope produced by boththe tool shape and the tool path. Thus, the part remainingcorresponds to the machined surface, and this part can berepresented by the following method.As shown in Fig. 3(b), the volume of the tool envelope isdefined by the coordinates of the CL-point a, b and tool radiusR. Here, take into account projected tool movement on theX, Y-plane. First, it is necessary to verify whether the selectedcheck points are inside this projected zone. If they are, thecoordinates of the machined surface can be defined by theintersection points between the tool envelope and the pointvectors (see Fig. 3(a). The mathematical description of thisconcept is presented in the following paragraphs.When an arbitrary check point e is inside the rectangle     , a supplementary point f, which is the intersectionpoint between the two lines   and  , is taken into accountin order to obtain the Z-coordinate of the intersection pointbetween the point vector of the check point e and the toolenvelope. Because the check point e is on the line  , the Z-coordinate zfof the point f is given by:zf= za +(zb − za)(xf− xa)(xb − xa)(1)where xa, xb, za, and zb are the X- and Z-coordinates of theCL-points a and b. As shown in Fig. 3(c), the Z-coordinate zeof the check point e depends on the tool shape used, accord-ingly in the case of a ball-end cutter, ze can be represented by:ze = zf+ R −  (R2− r2e) (2)re =  ((xf− xe)2+(yf− ye)2)When the check point e is inside the circle  or  , ze canbe calculated by substituting xa, ya or xb, yb for xfand yfinEq. (2). In the case of a flat-end cutter, ze becomes the sameas the Z-coordinate of the point f, and in the case of fillet-endcutter, ze can be represented by:ze = zf+ r2 −  (r22 − (re − r1)2)(re   r1) (3)ze = z1 (re   r1)3.2 Machining Error CalculationComparing the predicted milled surface with the designedsurface, it is possible to estimate the machining errors at allthe check points. The machining error of a sculptured surfaceis given by [6]:Error =  ((xm − xs)2+(ym − ys)2+(zm − zs)2) (4)where (xm, ym, zm) is the coordinate of the machined surfaceat the check point and (xs, ys, zs) is the coordinate of anarbitary point on the designed surface, which deviates by aminimum distance from the check point.4. Measuring Point Locations Based onCAM DataWhen measuring the sculptured surface by using a touch typeprobe, the accuracy of the measuring results depends on thenumber of measuring points. In the case of the often-used “equi-interval measuring point method”, if the number of meas-uring points decreases, the geometric information about themachined surface becomes insufficient. Accordingly, the meas-uring accuracy deteriorates. Generally, by increasing the num-ber of measuring points, the measuring accuracy is improved.However, the measuring process becomes very ineffectivebecause the process time and quantity of data increase equally.Menq et al. [2] proposed a methodology that consists ofdetermining the appropriate number of measuring points byconsidering the performance of the measuring instruments usedand the imposed tolerance criteria of the sculptured surface.When the sculptured surface is measured using the samenumber of measuring points, it is possible to measure thesculptured surface more effectively by appropriately selectingthe measuring point locations. Therefore, based on the proposedsurface prediction model and the tool path data, we proposein this study two methodologies to measure the sculpturedsurface effectively for a given number of measuring points.The first method consists of appropriately rearranging so thatthe measuring points are located at positions where significanterrors appear on the basis of the predicted surface errors. Thesecond method consists of selecting the measuring points atthe CC-points to reduce inspection errors due to cusps.4.1 Selection of Measuring Point Locations Basedon Predicted Cutting ErrorsWhen the number of measuring points is fixed, if the measuringpoints are located in the part where significant errors mayappear, it is possible to discriminate more effectively thegeometric errors of the machined parts. For this purpose, ithas been proposed, to locate more measuring points in thehigh curvature region [10,11]. This method can contributetoward the reduction of the measuring error occurring due tounknown geometric data between consecutive measuring points.However, when the machined part is measured, the surfacegeometric errors depend mainly on such machining conditionsas cutting tool type and tool path. In many actual cases, thesemachining error sources have to be taken into account as animportant reference for the selection of the measuring pointlocations, rather than using the surface curvature provided fromCAD data.For this reason, we propose an improved method for theselection of measuring point locations by taking into accountthe predicted machining errors. In this case, simply concentrat-ing the measuring points in the maximal error region, thegenerality of the measuring process is lacking for the whole CAD/CAM/CAI集成过程的检查机器英文文献和中文翻译:http://www.751com.cn/fanyi/lunwen_24319.html