Abstract: The big data collected during the product operations is of great value for product design, equipment maintenance and health assessment. It aims to optimize the wheel loader gearbox design by exploiting the big data acquired when the wheel loader is operating. A data-driven design optimization model is proposed, including data collection, data analysis, and optimization of the wheel loader gearbox. First, a data acquisition system is designed to collect the operational signals under four typical working conditions, i.e. native soil, iron ore, fine sand and coal cinder. Subsequently, the obtained gear operating signals and hydraulic pump pressure signals are processed and converted to cumulative gear utilization and the hydraulic system diversion power, respectively, which are further deployed as the optimization factors for the wheel loader power and fuel economy models. To solve the optimization models, the MOPSO and NSAG-Ⅱ algorithms are implemented to leverage diverse competing and conflicting goals related to multiple sub-targets of the gearbox design. The objective function is geared towards the optimal power performance while achieving fuel economy. Simulation studies indicate improvements of the resulting optimal design. For the respective MOPSO and NSAG-Ⅱ optimal solutions, the power loss rate is reduced by 8.83% and 4.80%, respectively, whilst the fuel consumption is reduced by 0.19% and 0.34%, respectively. The studies demonstrate the feasibility and potential of design optimization with consideration of feedbacks from big operating data. It is envisioned that product operating data-driven design is promising approach to facilitating continuous design improvement and promoting product innovation.

Key words: big operating data, fuel economy, gear utilization ratio, gearbox, optimal design, product development, wheel loader