Entwicklung und Bewertung eines flexiblen und dezentral gesteuerten Fertigungssystems für variantenreiche Produkte

Abstract

Decreasing product lifecycles, increasing product variety, and volatile market demands accelerate activities that aim at increasing flexibility and adaptability in manufacturing. Smart factories, capable of handling high product varieties and volatile demands, have gained increasing attention and interest within the context of Industry 4.0. Essential features of smart factories are the replacement of static material flow paths by introducing autonomous guided vehicles (AGVs) for transportation of products and goods and flexible processing orders among products and product variants. The goal of this thesis is the development and evaluation of a planning method for flexible and decentrally controlled manufacturing systems for highly variant products. Therefore, expert interviews at a German OEM were conducted to identify challenges in today’s manufacturing systems and revealed visions for future manufacturing systems. Afterwards, a concept for a manufacturing system addressing the findings of the expert interviews was developed and a design method for the technical configuration of the newly developed manufacturing was elaborated. Special focus of the design method is put on layout optimization, production scheduling, and path finding as well as collision avoidance. In order to evaluate the developed manufacturing system as well as the design method, an integrated optimization and simulation software was created. Key characteristics of the conceptualized manufacturing system are spatial distributed multifunctional manufacturing stations, flexible and non-predefined material as well as product flow paths, and autonomous material flow units. Derived from the characteristics, the developed manufacturing system is defined as flexible and decentrally controlled manufacturing system. The design method for an integrated optimization and configuration of the manufacturing system solves a variant of the quadratic assignment problem (QAP) by using a genetic algorithm (GA) in order to determine an optimized layout. The optimized layout also considers the solution for the production scheduling derived with a tabu search (TS) applied on an extended flexible job shop scheduling problem (FJSP). With the combined solutions, a simulation-based evaluation of a production cycle is performed that takes also path finding of AGVs based on local clearance triangulation (LCT) in combination with a modified A*-algorithm and collision avoidance of dynamic objects into account. The results of the simulation-based evaluation are used for the derivation of a multi-criterial fitness function with application specific weights that is fed back to the GA for the integrated design optimization of the manufacturing system. A validation of the design method as well as the resulting manufacturing system was carried out by considering the use case of a cockpit pre-assembly at a German OEM. The practical applicability was verified, manufacturing specific key performance indicators, e.g. station utilization and lead times, were determined, and due to reviewing different weight settings of the fitness function the design range triggered by the increased system flexibility was shown.