Model-based a priori analysis of line-less mobile assembly systems
Hüttemann, Guido; Schmitt, Robert H. (Thesis advisor); Lanza, Gisela (Thesis advisor)
Aachen / Apprimus Verlag (2021) [Book, Dissertation / PhD Thesis]
Page(s): 1 Online-Ressource (xix, 293 Seiten) : Illustrationen, Diagramme
The continuing trend towards producing individualised products leads to process time fluctuations, different process sequences and continuous integration planning in industrial assembly. Furthermore, in the assembly of large and complex components, there is a desire to increase automation for quality and cost reasons. This is often not economically feasible within the framework of conventional, line-based assembly systems with fixed installations. It requires an increase in the changeability of assembly systems. Within this thesis, a new form of organisation for industrial assembly and a method for a priori performance evaluation is developed. Based on a comparison of requirements for changeable production systems with existing changeability paradigms in manufacturing and assembly, Line-less Mobile Assembly Systems (LMAS) are defined. Three core principles characterise LMAS: (1) clean floor approach, (2) mobilisation of all assembly relevant assets, and (3) the highly flexible assignment of resources to jobs and jobs to locations considering process requirements. LMAS are characterised by job routes describing a tuple of resource allocation, order of assembly processes, time sequence, and spatial allocation on the shop floor for all steps necessary for assembly within sequence restrictions. The mastery of complex systems requires a comprehensive model-based understanding. A reference architecture model is developed for the design, evaluation, and control of LMAS. The model provides additional usage levels and a hierarchy for the model's assets through abstract modelling. Based on this reference architecture model, a mathematical formalisation of an LMAS is provided. During rough planning of a new assembly system, the possibility of a low-effort assessment based on typical production key figures such as throughput, lead time, and resource utilisation is required. A method is presented, which transfers an LMAS into a closed queueing network and determines production-relevant key figures using an analytical solution procedure. Topology conflicts are resolved by assuming stochastic process distribution instead of process sequences for long periods. Thereby it is not necessary to consider the assembly sequence. The occurring project conflicts are solved within scheduling by Monte Carlo simulations with randomly generated stochastic schedules. Evaluation criteria are used to determine the quality of the result. The developed evaluation method is validated utilising two case studies. This is facilitated by model investigations regarding internal validity, convergence, and the plausibility of the system behaviour compared to a real system's behaviour using a sensitivity analysis. Additionally, a comparison with a discrete event simulation model is carried out. The validation shows that the developed model is suitable for estimating the suitability of an LMAS a priori in the context of rough planning.