Part assurance in a mixed-model assembly line : A case study at Scania Engine Assembly in Södertälje

Abstract: Mass customization is considered the new paradigm for production companies, offering customers an extensive amount of options to choose from. They face increasing complexity that is closely related to the increasing number of part variants. Mixed-model assembly is often used to manage variant com-plexity. However, this can lead to part confusion, which is a source for assembly errors when parts are physically interchangeable. A qualitative case study was conducted during the spring of 2014 at Scania engine assembly (DE), were complex products are designed, driven by a modular structured platform. Engines are assembled on a mixed-model assembly line where assembly errors are common due to physical interchangeable parts that are difficult to distinguish for the assembly personnel at DE. Part assurance can then be adopted in the production to prevent assembly errors from occurring. The study initially confirmed that there was no standardized approach in order to systematically evaluate the need for part assurance during product development process, when products are adapted for the production environment. Part assurance is often implemented as a corrective action when deviations have been recorded. The research question that this paper aims to investigate is how DE can manage quality deviations in-duced by multi variants in a mixed-model assembly system. The department of Global Product Engi-neering requested a method that suggests when part assurance is necessary, and further what kind of measure to implement. Studies were conducted to obtain how risk assessment is performed during product preparation and analyzing the current state of the engine assembly. Data was mainly gathered through interviews and observations. A theoretical framework was developed, where relevant theories were obtained regarding preventive approaches to ensure assembly quality. Concepts such as cognitive automation, poka yoke, automatic identification and Failure Modes and Effects Analysis (FMEA) was considered prominent for this case study. Consequently a method was developed for decision support and can be used for facilitating evaluation of risk regarding assembly of wrong parts, suggesting when part assurance needs to be implemented and what kind of method, according to developed criteria. The method suggests three levels of risk that should be identified during early phases in product development that indicates necessary process and product changes. The study concludes that there are opportunities for improvement regarding risk assessment using P-FMEA. However, DE has the necessary tools and competences, and the suggested method will serve as a necessary complement. It was identified that risk management in early development stages can be complex due to insufficient data or abstract data. A systematically performed P-FMEA is important for efficient decision making, however certain obstacles must be managed. Further a joint and cross-func-tional approach is necessary for managing the identified problems. For instance, cooperation between design engineers and product engineers during risk management has to be enhanced. One suggestion is that Design-FMEA should serve as an input for Process-FMEA to enable efficient risk identification activities. Keywords Part assurance, P-FMEA, management of variants, mixed-model assembly, risk management, assembly error, assembly quality, cognitive automation, poka yoke, case study, production engineering, Scania