Multivariate Simulation of Manufacturing Processes Based on Fleet Simulation: Case Study in Metal Sawing

Multivariate simulations have proven valuable in automotive and light electric vehicle (LEV) fleet analysis for reproducing usage patterns and forecasting future developments regarding functional and safety related conditions and failures. Based on these experiences, this paper evaluates how the approach can be adapted to the reliability analysis of manufacturing processes, using metal sawing as a case study. The main challenge is that saw blade wear results from the interaction of multiple variables rather than a single variable. In practice, maintenance decisions are often based on experience or subjective impressions such as increased noise or visible defects. As a consequence, saw blades are sometimes maintained or replaced too early, leaving remaining tool life unused and increasing costs, or too late, leading to quality losses, unplanned downtime, and more expensive refurbishment up to full re-tipping of carbide teeth. A multivariate fleet simulator is applied to generate consistent synthetic datasets from usage and operating data. These data capture the behavior of the sawing process and measurable process parameters, including cutting speed, feed rate and material properties. The method is first illustrated with automotive fleet data and subsequently adapted to the metal sawing process to model operating behavior and load profiles in a data-driven manner. Cutting energy emerges as a practical indicator of saw blade wear. On this basis, reliability analyses can be performed both retrospectively and predictively. The study shows how a proven method from automotive fleet simulation can be adapted to manufacturing processes. This provides a decision basis for data-driven maintenance strategies, timely maintenance, cost reduction and improved efficiency in metal sawing.

Keywords: Reliability Modeling, Multivariate Simulation, Fleet Simulation, Metal Sawing, Multivariate Analysis.