Products are successful because they meet customer needs. However, many customer needs are not expressed in measurable terms. In addition, when such needs are achieved by a complex system made of hardware parts and software, decomposing customer needs to part-level specification is not a trivial task. This paper presents a model-based approach to address such problems. In the case study, the customer need was the noise and vibration level of an unconventional gasoline engine system when running at idle. The hardware component whose performance tolerance needed to be specified was a new type of fuel injectors. These new fuel injectors had higher piece-to-piece performance variations than the conventional fuel injectors. It was unclear whether such variation was acceptable for customer perceived powertrain quality. A virtual powertrain system simulation model was used to analytically evaluate the impact of the fuel injector performance variability. Monte Carlo simulation was carried out to assess the impact of injector variability. The results from the simulation were further refined using engine hardware testing. This study made recommendations for the acceptable level of hardware tolerance, which was different from what the supplier of the injectors had suggested.

This content is only available via PDF.
You do not currently have access to this content.