Shivakumar Raman

Mechanical evaluation of porous titanium (Ti6Al4V) structures with electron beam melting (EBM).

Patient specific porous implants for the reconstruction of craniofacial defects have gained importance due to their better performance over their generic counterparts. The recent introduction of electron beam melting (EBM) for the processing of titanium has led to a one step fabrication of porous custom titanium implants with controlled porosity to meet the requirements of the anatomy and functions at the region of implantation. This paper discusses an image based micro-structural analysis and the mechanical characterization of porous Ti6Al4V structures fabricated using the EBM rapid manufacturing process. SEM studies have indicated the complete melting of the powder material with no evidence of poor inter-layer bonding. Micro-CT scan analysis of the samples indicate well formed titanium struts and fully interconnected pores with porosities varying from 49.75%-70.32%. Compression tests of the samples showed effective stiffness values ranging from 0.57(+/-0.05)-2.92(+/-0.17)GPa and compressive strength values of 7.28(+/-0.93)-163.02(+/-11.98)MPa. For nearly the same porosity values of 49.75% and 50.75%, with a variation in only the strut thickness in the sample sets, the compressive stiffness and strength decreased significantly from 2.92 GPa to 0.57 GPa (80.5% reduction) and 163.02 MPa to 7.28 MPa (93.54 % reduction) respectively. The grain density of the fabricated Ti6Al4V structures was found to be 4.423 g/cm(3) equivalent to that of dense Ti6Al4V parts fabricated using conventional methods. In conclusion, from a mechanical strength viewpoint, we have found that the porous structures produced by the electron beam melting process present a promising rapid manufacturing process for the direct fabrication of customized titanium implants for enabling personalized medicine.

On the selection of flatness measurement points in coordinate measuring machine inspection

Abstract Inspection of form tolerances using the coordinate measuring machine (CMM) presents two distinct problems: data collection and data fitting. The former problem deals with the selection of the sample size and the sample point location while the latter involves the determination of the tolerance zone enveloping these points. Four types of strategies and five different sample sizes were studied in this work to address the former problem. The accuracy of flatness measurement was investigated using realtime experiments with respect to the above two factors and their respective levels. In addition, the length of the probe path was studied with respect to the two factors using a simulation study. A priority coefficient was developed to combine the influence of accuracy and path while selecting sampling strategies and sample size. Preliminary observations made suggest that any one sampling method may not be the best solution in all cases, while considering the accuracy of flatness and the shortest CMM probe path.

Intelligent Search-Based Selection of Sample Points for Straightness and Flatness Estimation

Form error estimation with a CMM requires prudent sampling and accurate zone fitting. This paper proposes use of optimization search methods for reducing sample size, while maintaining high accuracy. The approach is demonstrated with examples of straightness and flatness. For straightness, region-elimination search is used. For flatness, two pattern search methods: Tabu search and a hybrid search, are employed and their performance is compared. Sampling begins with a necessary number of initial points and a zone is determined. Next points are sampled based on the search methods, with suitable application of intensification/diversification strategies, looking for improvement in the zone fit. Search is conducted in both the +ve and -ve directions from the fit feature and is stopped when a solution for the maximum deviation is realized. The two solution points are added to the initial set and the corresponding tolerance is computed. The tolerance is compared with that obtained for the population of a large sample, to verify the accuracy. It is found that the number of points sampled is potentially less than that typically used to achieve the same accuracy.

Feature-based operation sequence generation in CAPP

Abstract This research explored in depth the integration of a machinists concept of manufacturing precedence among part features with a complete and explicit graph representation for alternative process plans, A precedence graph that captures the implicit predecessor-successor cost for any directed pair of features observed in the part replaces the state space representation adopted by AI based search strategies. The Hamiltonian path (HP) analogy for a process plan was developed and the Latin multiplication method (LMM) for constrained enumeration of all feasible HPs was implemented in a Turbo Pascal program running on a 486/25 PC. The program was tested using several examples of actual industrial parts obtained from the literature. The experimental results showed that severe constraints are imposed on problem size due to memory limitations imposed by the DOS environment, that the computational burden of using constrained enumeration is heavy and that there is need for using randomly generated alternati...

Texture analysis using computer vision

Abstract Surfaces of industrial parts need to be specified based on their utility and application environment. Surface characterization is hence very vital for design, manufacturing and inspection. Current techniques of surface measurement use surface profilometers, coordinate measuring machines and some optical techniques to estimate the nature of the surfaces. With the advent of automation surface characterization needs to be totally computerized so that the task of inspection (of surfaces) is greatly simplified. In the present paper a methodology is presented that uses a computer vision system to characterize the nature of the surface. In doing so, optical principles and lighting are discussed. Simple experiments were conducted using precalibrated gauge blocks to verify the feasibility of the method. A simple discussion is then presented with details on how to use this system for inspection. The advantages of using a vision system over other techniques is also adequately discussed.

Support Vector Regression for Determination of Minimum Zone

Probe-type Coordinate Measuring Machines (CMMs) rely on the measurement of several discrete points to capture the geometry of part features. The sampled points are then fit to verify a specified geometry. The most widely used fitting method, the least squares fit (LSQ), occasionally overestimates the tolerance zone. This could lead to the economical disadvantage of rejecting some good parts and the statistical disadvantage of normal (Gaussian) distribution assumption. Support vector machines (SVMs) represent a relatively new revolutionary approach for determining the approximating function in regression problems. Its upside is that the normal distribution assumption is not required. In this research, support vector regression (SVR), a new data fitting procedure, is introduced as an accurate method for finding the minimum zone straightness and flatness tolerances. Numerical tests are conducted with previously published data and the results are found to be comparable to the published results, illustrating its potential for application in precision data analysis such as used in minimum zone estimation.

Machine vision assisted characterization of machined surfaces

This paper demonstrates the feasibility of using an inexpensive machine vision system to compute non-contact, optical parameters for the characterization of surface roughness of machined surfaces. Two parameters were selected for online analysis, where surface roughness is measured during the rotation of a specimen on a lathe. The sensitivity of the vision-based optical parameters to differences in surface roughness, ambient light and spindle speed of a lathe during measurement was evaluated. Statistical analysis of data collected through experimentation revealed that the vision parameters can discriminate different surface roughness heights and are insensitive to changes in ambient lighting and speed of rotation during measurement. The results of the experimental analysis were used to conclude the feasibility of using machine vision for the evaluation of surfaces.

METEX—An expert system for machining planning

Optimal selection of process parameters is an important problem faced by most process planners and NC programmers. Computerized systems and mathematical models are available in literature for machining parameter selection. These systems are usually constructed with compiled handbook data. The machining process is non-linear and exhibits piecewise behaviour within different cutting ranges. If this piecewise behaviour can adequately be represented in machinability parameter selection systems, more insightful selection of cutting conditions can be achieved. The present paper provides a comprehensive survey of existing literature on machinability parameter selection systems, followed by a literature-based analysis of the anomalies of machining and the process and material effects in machining. A justification of expert systems for qualitative modelling is provided. A prototype system (METEX) for machining parameter selection is then discussed. This system uses production (FF-THEN rules to qualify machining at...

Experimental Analysis of Search-Based Selection of Sample Points for Straightness and Flatness Estimation

In earlier work [Bader et al., ASME J. Manuf Sci. Eng. 125(2), pp. 263-271 (2003); Int. J. Mach. Tools Manuf. 45(1), pp. 63-75 (2005)] the authors have presented an adaptive sampling method utilizing manufacturing error patterns and optimization search techniques for straightness and flatness evaluation. The least squares method was used to compute a tolerance zone. In this paper, experimental analysis is performed to verify the sturdiness of the adaptive sampling procedure. Experiments are carried out to investigate the effects of different factors on the sample size and absolute percent error of the estimated tolerance from that of a large population sample. Twelve 7075-T6 aluminum plates are end-milled and 12 cast iron plates are face-milled. Two sets of four plates from each lot are selected randomly, one each for straightness and flatness estimation. Factor A used in both straightness and flatness analyses is manufacturing process (i.e., surface error profile). Factor B for straightness is step size whereas for flatness it is search strategy (i.e., number of bad moves and restart allowed). Factor C for flatness is search algorithm (i.e tabu and hybrid). Plates are nested within the levels of manufacturing process. The results have been analyzed and compared with other sampling methods. The analyses reveal that the current approach is more efficient and reliable.

From support vector machine learning to the determination of the minimum enclosing zone

The verification of form tolerances requires the determination of the minimum enclosing zone according to the ANSI Y14.5M National Standard on Dimensioning and Tolerancing. However, to date many coordinate measuring machines (CMMs) still employ the least-squares method, which has the economic disadvantage of sometimes rejecting good parts. Support vector machines represent a new approach in the area of machine learning, which has been implemented successfully in pattern recognition and regression estimation problems. This article outlines, how the support vector algorithm, as used in classification problems, can be modified in order to identify the minimum enclosing zone for straightness and flatness tolerances. A gradient ascent approach is proposed to identify the solution of the resulting non-convex optimization problem. Numerical results for evaluating the minimum enclosing zone suggest rather promising properties of the employed gradient ascent method.