Smart-engineering Tools Dedicated to Sports Product Design

Abstract

Make product design easier, faster and smarter, that’s the ultimate goal of many companies. Moreover, companies involved in sports design and manufacturing require fast innovations to remain competitive in their business. In pursuit of this goal, Decathlon has developed a simple interface called “Technical Configurator”, accessible for the simulation engineers, non-expert in design optimization and predictive modeling, allowing its development team to be autonomous in optimizing their product easier and to accelerate overall time to market. Once developed and validated, each “Technical Configurator” is released and hosted in a “Design Application Store” to be used by the whole authorized engineering team. It is easily accessible worldwide through a simple web browser and absolutely user-friendly. While developing the “Technical Configurator”, Decathlon has explored the idea of server-based solution for predictions and optimization powered by approximation models. A true cloud collaborative platform developed by DATADVANCE was used as enabler. Inputs are provided by domain experts through a simple web page, and back-end server sends the request to the expert module of Design of Experiment, Optimization, Surrogate modelling and Simulation. As it has full interoperability with other engineering tools, the platform allows to combine in-house routines related to performance (mechanical and feeling), cost estimation, and environmental impact, to ultimately validate design alternatives with respect to multiple complex parameters, exposed as tunable variables.. All of it is possible thanks to use of external APIs (REST, GraphQL). The immediate benefit is to obtain an almost real time prediction of product performance taking into account a growing number of design parameters. By democratizing access and usage of its configurator, Decathlon enables capitalization of knowledge from all configurations analysed by all users. Opportunity is given to transparently collect results of all design alternatives and enrich a common database to improve the quality of predictive models and finally the knowhow on product design.