A methodological approach for optimisation of product development processes by application of design automation

Abstract

In the past decades a clear transition can be seen from fully human-based production techniques towards more automated systems. With the Product Development Process (PDP) being a potential source of competitive advantage the same trend of adopting more automation can be seen in the PDP. The application of automation can lead to large reductions in process lead time. The reduced lead time can be used to enhance product performance by assessing more design options in the same time, or to reduce the cost-of-delay and obtain a larger market share. It is clear that a reduction in lead time is worth an investment for companies and therefore automation can be adopted to improve the PDP performance. Both industry and academia acknowledge the lack of a quantitative method to assess the effect of the application of automation on the performance of a given process. Furthermore no methods are available to assess the effect of incremental automation. Another limitation is that most models investigating automation in the PDP only take into account a binary type of automation: either fully human or fully automated. This research aims at addressing these gaps of knowledge and the goal of this research is to develop a tool that provides more insight in the costs and benefits of automation in the PDP taking into account incremental automation and various levels of automation. To achieve the objectives a novel methodology is developed i) to model any PDP as a combination of a predefined set of specific activities and ii) to define different levels of automation for these activities. Subsequently, metrics are developed to measure the impact of different levels of automation on different activities, both in terms of activity lead time reduction and implementation cost. A simulator using Discrete Event Simulation (DES) is developed which utilises the proposed process model and metrics to analyse, among others, the lead time, automation investment cost and process cost for the overall process (for a given process architecture). Finally, the simulator is connected to an optimiser which tries to find the most convenient level of automation for each of the PDP activities, in order to generate a Pareto front to visualise the trade-off between the lead time reduction and required investment cost. The proposed methodology is extensively verified. Verification is performed for both the simulator and the optimiser. Based on this verification it can be concluded that with appropriate settings the optimiser manages to find Pareto optimal solutions for different levels of automation. Validation however remains limited to positive expert judgement due to scarce validation data. The proposed methodology and the analysis and optimization framework are demonstrated by application to an industrial case study. The case study concerns the conceptual design phase of an aircraft component, performed by a multinational aerospace enterprise. The case study successfully demonstrates the feasibility and applicability of this methodology and accompanying frameworks. Multiple Multi-Objective Optimisations are performed to trade off various objective functions. In this case study, specific activities in the process are identified to be more effective to automate. Results show that compared to the status quo, an investment in the lug sizing tasks of 7,1% of the investment cost of full process automation can lead to a potential lead time reduction of more than 40%. The proposed methodology proves to be successful in objectively quantifying the costs and benefits of automation in the PDP and subsequently selecting the optimal automation level.