An Approximate Analytical Method to Evaluate the Performance of Multi-product Assembly Manufacturing Systems

Abstract

The automotive industry is characterized by the continuous and frequent adoption of cutting edge technologies in manufacturing. The introduction of new technology often entails a system level investigation to determine whether it meets expected production requirements. Many system design alternatives are usually considered at this stage, and a fast performance evaluation tool is desired to quickly obtain the optimum system configuration. In this paper, we introduce an analytical tool that was developed for this specific purpose. The investigated manufacturing system is a multi-product, automated, assembly production line composed of unreliable machines decoupled by finite buffers. Each product is an assembly of two components that are produced in sub-assembly lines before being assembled at an assembly machine. The performance of this system needs to be evaluated for different configurations that arise mainly due to the integration of a remote laser welding technology. This new technology promises, among other advantages, a much faster assembly than existing technologies. This makes it feasible to share assembly among many products with different possible production policies. The processing machines including the assembly machine are liable to fail randomly in operation thereby necessitating an evaluation method capable of accounting for stochastic behavior.To solve this problem, we develop a decomposition-based approximate analytical method. We use the continuous material flow approximation, which allows to model machines having deterministic but inhomogeneous processing times as in the investigated real system. We also account for multiple machine failure modes and introduce new techniques to incorporate the assembly of multiple products with different policies of switching between products. Comparison with simulation results shows the good accuracy of the model in estimating system throughput and work-in-progress. The model has been successfully utilized to evaluate the performance of an automotive door assembly manufacturing system and the main results of this case study are reported.

Performance analysis of unreliable manufacturing systems with uncertain reliability parameters estimated from production data

“…Later, these methods were extended to deal with more complex machine models, featuring multiple failure modes (Levantesi et al 2003). Recently, methods for studying complex system layouts (Colledani et al 2005) and different product variants (Colledani et al 2015) have been proposed. These works have been foundation for the development of various decision support system for the design and management of complex manufacturing systems (Colledani et al 2014), (Terkaj et al 2009).…”

Rapid deployment of remote laser welding processes in automotive assembly systems

“…The station models, as well as the system topology to be optimized, are provided as input to the system analyzer sub-module by means of so-called transfer functions. Next, the system analyzer sub-module tests several alternative system configurations before implementation, by exploiting the features of a fast performance evaluation module [6], based on approximate analytical methods. Upon convergence of the selected optimization algorithm, the set of candidate Paretooptimal configurations are visualized to the designer.…”