Manufacturing process variability is a major issue of concern in high value industries. Manufacturing small batches and in some cases batches of one is a very expensive process with specific requirements for manufacturing operations, tooling and fixturing and their level of automation and informatics provision. The automation targets cost reduction and a counterbalancing of the ever lower numbers of skilled shop floor workers. However, these small series typically are products that contain complex and compliant parts, and often also a high number of parts and components. The automation of this type of low-volume high-value production can be a daunting task.\udEach process has its own key parameters that are required to be within a certain tolerance band in order to ensure product quality, such as e.g. the dimensions and location of assembly mating features. Dimensional quality assurance is typically done with in-process measurement, or the measurement of certain key characteristics (KCs) in the current setup, but a special setup may have to be used in a measurement-only step in the manufacturing process. Each manufacturing stage introduces errors stemming from uncertainties in the fixturing, used processes etc. These errors will propagate in downstream stages and can even worsen errors introduced in the latter stages.\udThe paper presents a new generic methodology for the use of stream of variation (SoV) analysis within a Smart Factory environment such as the Evolvable Assembly Systems (EAS) framework. The research is demonstrated using a simplified case study of one of EAS demonstrators for an aircraft wing box assembly. The wing box assembly and its KCs are described using formal representation. The SoV model is applied to model and simulate the assembly process. The simulation results are then analysed to predict, control and minimise the error propagation coming from uncertainties in process and equipment.
展开▼
