Optimum design of multi-body systems (MBS) is a common problem in industry for large companies and SMEs and is present in a wide range of industrial applications: automotive, railways, aerospace, heavy machinery, etc. MBS optimum design requires important computational resources and heavily relies on designers’ experience. This problem is solved using numerical optimization and sensitivities are computed numerically. This approach can largely benefit from the use of cloud-based HPC tools.
This experiment proposes the use of software tools for multi-body systems (MBS) simulation and optimization in a cloud computing environment for the optimum design of the rear suspension system of a racing motorbike. This is a real industrial problem, which consists in obtaining the optimum values for the design parameters of the suspension: geometry (part dimensions), inertial characteristics (masses and inertias), elastic elements (springs/bushings) and hydraulic components (shock absorbers). The cloudified version of the software tools will be tested by the end user in real working conditions assessing their usability and applicability in industrial environments.
The proposed experiment pursues two main goals: on one hand assessing the time (and cost) reduction of the overall design process of a rear suspension system for a racing motorbike; on the other hand evaluating the computational speedup factor of the numerical optimization problem derived from the use of the HPC/Cloud implementation.
The design of a whole suspension system for a vehicle is a highly complex task that is carried out by experienced engineers following an iterative procedure. In the first steps, numerical simulations are carried out on the computer. With these results first physical prototypes are built and tested using testing machines. The feedback of those tests is used to improve the design, perform new simulations, build new prototypes and carry out new tests both in test bed and in a vehicle (usually a configurable vehicle called ‘muleto’). The whole process strongly depends on the of the design engineer who, following this trial-and-error approach, obtains an optimum design of the suspension system. The design of the shock absorber is a part of the whole design process. Considering the rear suspension of a racing motorbike the whole process lasts an average time of 2 to3 months. It requires building an average of 3 to 6 physical prototypes of the shock absorber and 1 to 2 prototypes of the whole suspension system for use in testing machines and in-field trials. The whole design process (including tests and prototypes) has a total cost which varies between 200,000€ and 400,000€.