COMPREHENSIVE DESIGN STRATEGY FOR A "RECONFIGURABLE MANUFACTURING SYSTEM

Abstract

Globalization has created a new landscape for Industry, because of fierce competition, short windows of market opportunity and frequent changes in product demand. New and highly customized product types/styles are demanded with ever-shortening life cycles, high quality and low cost. Product mix and quantities might vary dramatically, even on a monthly, weekly or daily basis, depending upon the business portfolio. To cope with such a dynamic market scenario, an industry frequently needs to redesign and re-plan its manufacturing system to reconfigure it within a short time to meet changed production requirements. For a manufacturing system that uses conventional manufacturing resources, the reconfiguration exercise can be exorbitant and highly troublesome, as these resources are designed for a specific product mix and production rate, both assumed to last for a certain long period. The requirement to halt the production during reconfiguration exercise further worsens the monetary and reputation losses of the enterprise. Therefore, Industry requires a manufacturing system that has the technical capabilities to swiftly and smoothly allow its reconfigurations. A new manufacturing system paradigm named as Reconfigurable Manufacturing System (RMS) is envisaged to have capabilities that allow it to be reconfigured within a short time to provide the exact functionality and capacity that is needed, exactly when it is needed. At the heart of RMS philosophy lies its core component Reconfigurable Machine Tool (RMT), which provides RMS its distinct advantages such as changeable structure, customized functionalities, customized capacity etc. The research on development of RMTs has attracted considerable attention, but literature on design of manufacturing systems that uses RMTs (i.e. RMS) is scarce. The present work proposes a generic framework for design of an RMS accommodating all types of changes in the system in the backdrop of prevailing market conditions. The proposed RMS model features sequential manufacturing of a set of different part types in different order quantities within a predefined time period called Reconfiguration Cycle Life Span (RC-Span) using RMTs in a flow line layout. When the whole set of part types is manufactured, it is said that a Reconfiguration Cycle is over and the system can start another reconfiguration cycle to manufacture another set of part types, orders for which were accumulated during execution of the previous reconfiguration cycle. The major postulations and features incorporated into the model are highlighted below. i. Orders for next reconfiguration cycle are accepted to exploit the available capacity during RC-Span as per a first-come-first-serve policy allowing maximum delivery time equal to RC-Span. ii. Each part type has many alternative operation sequences of equal priority to manufacture it. iii. The proposed RMS utilizes a bidirectional material handling system. iv. For each operation there are many possible alternative RMT configurations of equal performance level. An RMT configuration is a combination of one or few structural Basic Modules and some kinematical Auxiliary Modules. v. The industry maintains and regularly updates a library of basic modules and auxiliary modules available with it and/or acquirable from a suitable source and/or commercially available in the market as and when required. The information on various possible permutations into which various basic modules and auxiliary modules can be combined to configure different RMTs and their respective operational capabilities is also the part of the library database. vi. All the P part types to be manufactured during a reconfiguration cycle are grouped into F Part Families so that 1 < F < P. The parts within a family present soft variety in terms of operations required to manufacture them. However, the families themselves present hard variety. vii. Parts belonging to a part family are manufactured one by one using a Primary Configuration that represents a set of RMTs providing all the required operations. After the production of all the part types in the part family is over, the manufacturing system is reconfigured to another primary configuration providing all the required operations of the next part family to be manufactured. This reconfiguration exercise of the manufacturing system required for switching over the manufacturing from one part family to the next is termed as Primary Reconfiguration. To complete a reconfiguration cycle, the system undergoes as iv many primary reconfigurations as the number of families. The sequence, in which various part families are taken up for production one after the other is termed as Primary Reconfiguration Path. viii. The primary configuration (i.e. set of RMTs) of a part family require undergoing Secondary Reconfiguration involving relocations and minor adjustments of RMTs to conform to a favourable RMT-layout (called Secondary Configuration) whenever the next part within the same family is taken up for manufacturing. During manufacturing of a part family, the sequence, in which its member part types are taken up for production one after the other is termed as its Secondary Reconfiguration Path. The work elements for realizing in practice an RMS based on the proposed model are summarized in the form of objectives as enlisted below. i. To develop a methodology to group part types (to be manufactured in a reconfiguration cycle) into Part Families. ii. To develop a methodology to select one operation sequence from the set of alternative operation sequences for each part type and therefore, to find Operation Groups for all part families. iii. To develop a methodology for allocation of RMTs to each operation group and hence to form Primary Configurations for all part families. iv. To develop a methodology to select an optimal Primary Reconfiguration Path for executing the Reconfiguration Cycle so as to maximize the economic objectives and minimize the efforts involved in primary reconfigurations. v. To develop a methodology to select a Secondary Configuration (i.e. an RMT layout) for each part. vi. To develop a methodology to select an optimal Secondary Reconfiguration Path for each part family so as to maximize the economic objectives and minimize the efforts involved in secondary reconfigurations. vii. To develop a methodology to finalize a Comprehensive Manufacturing Plan for v A sample problem of 6 part types requiring total 6 operations has been used for demonstration and validation of the proposed heuristic procedure. It has been assumed that each part can be produced by up to 3 alternative operation sequences and each operation has a set of up to 4 alternative RMT configurations of equal priority. In all there are 13 operation sequences consisting of 2 to 3 operations each and 15 RMTs in the library configurable by using 5 types of basic and 15 types of auxiliary modules. To establish a high face validity of the proposed RMS, based on the potential capabilities which are under development, a number of actions are followed throughout the course of this work. The results of the proposed optimisation procedures are validated module by module. Module-1 algorithms are verified first by running under simplified conditions and matching the results with those obtained by standard hierarchical clustering algorithm. Secondly, the final results are verified by solving the problem through hand calculations. Module-2 results are validated by comparing the Pareto front generated by NSGA-II to the Pareto front generated by solving the problem as a single objective optimisation problem. The results of the remaining modules are verified by brute force search of the solution space corresponding to each solution alternative. Also, it has been shown that the application of the proposed RMS design methodology reduces the size of the search space by ignoring a large chunk of less important solutions. Two larger problems adapted from CMS literature with appropriate assumptions are also demonstrated as case studies. The first problem has 10 parts, 13 operations, 20 operation sequences, 27 RMTs, 10 types of basic modules and 25 types of auxiliary modules. The second problem has 20 parts, 20 operations, 51 operation sequences, 40 RMTs, 16 types of basic modules and 30 types of auxiliary modules.