Pralay Pal

Feature extraction from large CAD databases using genetic algorithm

Syntactic recognition, Graph based method, expert systems and knowledge-based approach are the common feature recognition techniques available today. This work discusses a relatively newer concept of introduction of Genetic Algorithm for Features Recognition (GAFR) from large CAD databases, which is significant in view of the growing product complexity across all manufacturing domains. Genetic Algorithm is applied in a random search process in the CAD data using population initialisation; offspring feature creation via crossover, evolution and extinction of the offspring sub-solutions and finally selection of the best alternatives. This method is cheaper than traditional hybrid and heuristics based direct search approaches. Case study is presented with simulation results.

A hybrid approach for identification of 3D features from CAD database for manufacturing support

A number of approaches for identification of 3D features are available today including syntactic recognition methods, graph based methods, expert systems and knowledge based approach. Work done on all such approaches shows lack of generalisation, difficulty in implementation or little demonstrated computational viability. In this paper, a hybrid approach for non-interacting 3D-feature identification from CAD database has been proposed, which is partly mathematically iterative and partly based on heuristic rules. Initially a void non-polyhedral pocket feature has been undertaken to establish the procedures and later this is extended for all features in general. A couple of test results are presented with CPU times and inherent heuristics. The process is applicable for process planners and manufacturers in industries.

An easy rapid prototyping technique with point cloud data

In recent days, rapid machining through digital prototyping has been popular for its applicability in a wide range of complex and useful parts. Rapid construction of prototypes from point cloud data based on section plane method is available, which is an approximate method. Discusses some suitable methodology for conversion of point cloud data to a physical prototype where data acquisition is through a mechanical touch trigger probing process using CNC milling machine. The process is quite useful for reverse engineering of complex sculptured parts. A concept called tangent plane method is adopted for the generation of 3D geometry on point cloud data of sculptured parts with due emphasis on probe radius compensation after data capture and tool radius compensation during tool‐path generation. Computer simulated results are presented, based on real‐world point cloud data.

“16T”: toward a dynamic vendor evaluation model in integrated SCM processes

Purpose – The purpose of this paper is to provide a “trait based approach” for vendor selection and evaluation for expensive procurements in large businesses through a simple and easy‐to‐use mathematical model using safety, quality, delivery and cost criteria. Design/methodology/approach – We use 16 traits addressing safety (S), quality (Q), delivery (D) and cost (C) areas for evaluating performance of a vendor on a linear 10 point scale. We start with evaluating each “supplier‐supply item” combination and compute gross averages for each of the SQDC areas and finally arrive at an “Overall Performance Index” for each supplier‐supply item combination. These indices form a “Vendor Performance Dashboard” for decision making. Findings – The case study shows that the proposed method is quick and easy to adopt, and provides a logical framework for vendor selection and management, based on performance in the four critical (SQDC) areas. Research limitations/implications – A lot of monotonous computations are requi...

Object shape reconstruction through NURBS surface interpolation

The skinning of point cloud models has become one of the most popular reverse-engineering exercises for rapid product development in most manufacturing sectors. Common methods such as ‘polygon generation and segmentation’ and ‘direct surface fitting on segmented data’ exhibit shape loss, distortion, continuity problems and a large number of orphan surface patches. The current work discusses data processing through ‘data rarefaction’ to optimize a data set for true object shape and NURBS-based surface interpolation on rarefied data for a reconstruction with good accuracy. For originally rarefied data, on the other hand, conventional subdivision techniques are applied to construct logical points before surface fitting. NURBS interpolation technology ensures accurate surfaces within a limited computation of knot vectors, weight vectors, bi-directional control net and IGES interface file writing. A few reconstruction examples with their results are shown with computation issues and analysis.