A.K. Srivastava

Modelling geometric and thermal errors in a five-axis cnc machine tool

Abstract The total volumetric error within the workspace of a machine tool is induced by the propagation of both scalar and position dependent geometrical errors, as well as time-variant thermal errors. This paper presents a compact volumetric error model which can be used as a basis for a practical compensation scheme. The broad objective is to increase the achievable accuracy of an industrial five-axis CNC machine tool. In place of using Denavit-Hartenberg (D-H) transformations, the method used here directly considers the shape and joint transformations for inaccurate links and joints using small angle approximations and then finds the total volumetric error in the workspace as a function of all the possible errors. The development of the model shows that angular deviations are independent of translational errors. However, the tool point deviations are dependent on both translational and rotational errors. The model has been used for the design and testing of a compensation strategy. The simulation studies indicate that CNC compensation for errors in X , Y and Z axes is possible. However, the capability of the CNC compensation for pitch, roll and yaw errors is dependent on the positioning of the rotary axes on the machine tool. This is shown by an example using the compensation scheme developed.

A Model for Chip Formation During Machining of Hardened Steel

Abstract Saw-toothed chips are formed during machining of hardened steel (H RC ∞ 60–63). This paper presents a new analytical approach for modelling the chip formation mechanism in hard-turning. It has been observed that the chip formation starts with initiation of a crack at the free surface of the workpiece which further propagates towards the cutting edge of the tool. The crack soon ceases to grow at a point where severe plastic deformation of the material exists under higher level of compressive stresses. The chip segment caught up between the tool rake face and the crack is pushed out while the material in the plastic region just below the base of the crack is displaced along the tool rake face thus forming saw-toothed chips. The direction of crack initiation and propagation are predicted using the surface layer energy/strain energy density criterion. The maximum value of surface layer energy, Y emax can be used to evaluate the angle of crack initiation while the strain energy density criterion predicts the corresponding crack propagation angle. Here, the process of chip formation is considered to be a mixed mode crack problem of Mode I and Mode II. The theoretical predictions are verified by the resultant chip contours obtained experimentally. The predictions made are shown to be in good agreement with those measured experimentally.

Adaptive Force Control for Robotic Disk Grinding

Abstract In this paper, the design and implementation of an adaptive predictive force controller for robotic grinding are described. The force regulation is established by modifying, on-line, the robots programmed position control commands. The experiments were conducted using an industrial robot equipped with a pneumatic grinder and a six degrees of freedom force sensor. Two predictive control laws, namely; Generalized Predictive Control (GPC), and Extended Horizon Control (EHC) are evaluated by simulation and experimentally. The control laws are designed based on accurate model(s) for the open-loop dynamic of the robot and grinding process, which are identified from experimental data. The results show that both controllers are able to effectively regulate the normal grinding force and compensate for force errors caused by step force disturbances and robot path tracking errors. A slightly better control performance was achieved, however, using the Generalized Predictive Controller.

Analysis of the robotic disc grinding process

Two major aspects of the robotic rigid disc grinding process, namely, disc wear and grinding forces have been examined in the present study. The disc wear is observed to be nonuniform, being greatest at the outer edge of the disc. A simple wear model has been developed by assuming that the amount of wear can be approximated as having a triangular cross section. The dynamic grinding force model developed includes the effects of disc wear and nonlinear stiffness of the robot system. Experiments have been conducted on an actual robotic grinding system to verify the validity of these system models. Several practical issues which should be considered during robotic grinding are discussed.

Surface finish in robotic disk grinding

Abstract In spite of the various complexities, the ground surface during robotic grinding shows a definite and distinct surface texture. This paper presents a simple model for the surface finish produced during this process. The model is based on two simple assumptions. First is that though many grits remove material, the last grit that cuts largely in the ground surface determines the surface finish value. The second assumption is that the average shape of the active grit is spherical which generates a groove of parabolic cross section. The model is shown to be consistent with surface finish values measured experimentally during robotic rigid disk grinding and is able to predict the effect on surface finish to be expected when a given grinding variable is changed. It can be used for the purpose of process planning to improve the surface finish when desired.

Analysis of rigid-disk wear during robotic grinding

Abstract An investigation of disk-wear during robotic grinding is described. It is shown that the wear of resinoid bonded disks can be described via combined thermally and mechanically controlled mechanisms. The process of accelerated grain wear can be correlated with the rise in temperature at the grain-work interface in a manner predictable from the simple law of adhesive wear in combination with the theory of rate process. Experimental results are shown to confirm the theoretical interpretation.

A simple analysis for evaluating grinding wheel loading

Abstract A simple theoretical model has been presented for evaluating the amount of wheel loading under various grinding conditions based on the adhesion of work material onto the grain surface. Fair agreement is obtained between experimental results and the pattern of wheel loading predicted by theory.

A new technique for evaluating wheel loading

Abstract The X-ray fluorescence technique using high energy Am 241 as X-ray source has been applied for the study of the grinding wheel loading. Experiments with lead and mild-steel workpieces indicate the sensitivity and feasibility of this technique. The quantitative values of loaded material have been evaluated through calibration with standard samples. Results obtained using this and a conventional technique show close agreement.

Workpiece burn and surface finish during controlled force robotic disk grinding

Abstract Workpiece burn and surface finish during robotic disk grinding are studied under controlled force conditions. Workpiece burn occurs due to gradual deterioration in the sharpness of active grains participating in the cutting process. This change in the sharpness of grains can be indirectly monitored by the change in the average coefficient of friction and average depth of cut. Both of these parameters gradually reduce with increasing disk wear due to attrition. In this paper, an analytical expression is derived which can be used to predict the occurrence of such burns during controlled force grinding. The experiments conducted under these conditions using a finely tuned PID controlled showed fair agreement with the predicted results. The experiments also showed that the surface finish gradually improves during successive passes. Hence, an attempt has been made to explain and predict this changing behaviour in the surface finish on the basis of gradual deterioration in the cutting efficiency of the disk. These relations can be used for practical optimization of the robotic grinding process.

Optimal planning of an adaptively controlled robotic disk grinding process

Abstract In this paper, an adaptive control optimization (ACO) robotic grinding system is described. An industrial robot equipped with a pneumatic grinder, a six degree-of-freedom force sensor, and two non-contact inductive proximity sensors inter-linked with a PC are the integrated part of this system. The system is designed to grind the workpiece with maximum possible metal removal rate (MRR) within the constraints on workpiece burn and surface finish which, in turn, depend on the degree of disk dullness. A model-based approach has been used for the optimization. To perform the cutting operation, normal grinding force and robot feed speed have been selected as control parameters. The normal grinding force is effectively tracked on-line using an Adaptive Generalized Predictive Controller (GPC) algorithm, which was experimentally tested. The optimal robot feed speed is tracked due to the robot systems capability to accept on-line end effector path modifications by means of the robots “ALTER” command. This system is able to plan and efficiently execute a multi-pass grinding sequence for removing large amounts of unwanted material. Several illustrative results are presented that confirm the practical feasibility of the optimization concept and demonstrate the performance of the ACO robotic grinding system.