Product Design Optimisation Studies
Product design optimisation studies, from a few weeks to several months, are carried out in close collaboration with our customers in order to ensure that we fully understand their problem and provide a suitable solution. These studies are carried out either on the customer's premises or internally at Eurodecision. They apply our product design optimisation methodology and rely on the Test Management Toolbox (OGE) to co-ordinate numerical simulations (or real tests) of physical phenomena.
A study aims at exploring the interesting regions in the decision space. Ultimately, business experts obtain a set of solutions ranked according to their preferences, representing a panel of various components. Experts can back up their choices with these detailed results.
Optimisation studies may be included in the various milestones of the design process:
Factors to be considered vary according to the stage in the design process. During the preparatory phase, the study explores a wide range of solutions, and modifications may be carried out at any time (changes in geometry, testing of different technologies, etc.), even if some carry-over constraints are important. More options are open during the preliminary design phase. Once the product is in production, modifications are limited to simple variations such as the parts thickness which do not require changes to stamping tools or presses, but which may significantly reduce materials costs.
Studies carried out for Renault
EURODECISION has carried out several optimisation studies for Renault. For example:
- Reducing the weight of future vehicles: a report on the impact of changes to thickness, materials, and the potential presence of parts on the vehicle's superstructure, underbody, seats, cradle, and driver's compartment crossmember. These items are analysed for weight reduction using constraints concerning a dozen services (head-on or lateral crashes, vibrations and acoustics, endurance, hitting the curb, etc.).
- Design of a rear shock absorber: benefit of the free positioning of the part's side-walls
- Seat structure optimisation: evaluation of the contribution of new materials
- Design of a bonnet complying with new pedestrian collision standards: study of the shape of the bonnet's lining
- Study of the air flow for an engine air intake grill: analysis of influential factors
- Benchmarking of a thermal digital model for the projector: determination of the heat flux through the bulb
Challenges
In most cases, the design of a complex system must consider factors within several disciplines (crash, vibration, thermal, aerodynamic, etc.). The generated solutions must meet all the physical constraints involved. This multidisciplinary (or multi-service) aspect requires communications between different parties in the company. The study's definition phase is critical in order to identify all the constraints in the various lines of work and to translate them correctly into optimisation constraints.
For each service, the designers must list the influential factors, their admissible ranges, and their possible nesting (e.g. the choice of the material for a part defines the possible discrete values for its thickness). During this definition phase, designers must keep in mind that increasing the number of design factors also increases the number of simulations required to find solutions. Making variations to too many factors can also impact on services not considered within the study. Yet the more factors you play with, the greater the potential gain is. Our studies cover problems in a great variety of sizes, from just a few design factors to several hundred.
The "study" expert attempts to limit the number of simulation tests (choice of a design of experiments, breakdown of total number of experiments between the initial design and the optimisation iterations, etc.). In a multidisciplinary context, where the duration of simulations may vary considerably (from a few minutes to several hours), this leads to strategies for taking services into account. For instance, if the simulation crash code requires significantly more computational resources than the vibration code, then only designs which meet vibration constraints will be subject to crash simulations runs.
Study prerequisites
Optimisation studies imply a very large number of simulations. Therefore it is necessary to fully automate the steps required to launch those simulations, from entering configuration data in the input files to post-processing and extraction of results in the correct format in the output file, as well as initiating pre-processing and performing the calculations for the simulation itself. EURODECISION also offers assistance for this design phase.
Related topics:
Product design optimisation methodology
The Test Management Toolbox (OGE)
