CO-EVOLUTION AND ALGORITHMIC GENERATION OF MACHINING PROCESS PLANS AND KINEMATIC CONFIGURATIONS

Abstract

The field of manufacturing is one of the oldest and the most established industry of the world. It has been evolving and developing for many ages. The concept of using machines to develop the products started in the early twentieth century. At the end of the twentieth century a new manufacturing system known as RMS (reconfigurable manufacturing systems) was introduced. It completely revolutionized the manufacturing setup with reconfigurable machines which can adapt to the changes in the products while accommodating the machinability issues as well. As the ability to adapt increased, the demands for changes in the products increased as well. Due to the rapid changes in the products these days, the process plans and the structures required to develop the products following the process plans are changing at a rapid pace as well. Traditionally, the approaches developed for the process plans and kinematic configurations of RMS either focus on the process plans and then develop the corresponding kinematic configurations or develop the kinematic configuration and then the process plan. Furthermore the issues that these approach address make them unilateral: i.e. the approaches either minimize the initial cost of production or the overall production rate while completely ignoring the overall quality of the product. This thesis is focused on improving the previous approaches and to address certain issues regarding co-evolution. The approaches that have been improved are algorithmic in general, however, a Petri-net model for a particular system has been proposed as well. Furthermore, an algorithm for changing parts within the same part family is presented. It explores the complete solution space for a particular part group. The application of the proposed algorithm is illustrated by its implementation on automotive parts, having a set of machining features to be realized.