Multi-objective optimization of a motorcycle composite swing-arm

Motorbike swing-arm. Source: iChrome

Composite materials are rapidly supplanting metals in racing and sport vehicles, providing comparable strength and stiffness at much lower weight. This project was a feasibility study to replace the single-sided swing-arm in MV Agusta’s high-performance F4 1000R and Brutale 990R/1090RR motorcycles, originally made of aluminum alloy, with a new design consisting of resin transfer molding (RTM) carbon composite. Injection-based technologies for long-fiber reinforcement such as RTM have proven particularly effective for motorsport cars and motorbikes.

Methodology—After an initial reassessment of shape and design of the part to make it better suited to RTM fabrication, iChrome’s Nexus optimization and process integration software was used to guide an Abaqus finite element model toward identification of optimal solutions.

Engineering the swing-arm required defining constraints on structural strength (maximum stress/strain levels) and structural stiffness (vertical, lateral and, most important, torsional). This was done by launching an Abaqus finite element analysis of the swing-arm and reading relevant results from within Nexus. Additional manufacturing constraints were defined to control change in thickness between different areas of the laminate while changing local lamination attributes such as number of plies and orientations.

Multi-objective optimization was the method applied. Solutions that would guarantee minimum weight for a given torsional stiffness while meeting all other requirements were searched for by minimizing weight while maximizing torsional stiffness. Due to the relatively large number of discrete variables such as number of plies in the composite laminates, Genetic Algorithms were judged to be the preferred choice for this application. Accordingly, the NSGA-II algorithm was selected from Nexus’ library of solution algorithms.

Flowchart of optimization process. Source: iChrome

The optimization process involved 60 design variables used to describe composite laminate evolution in the swing-arm, together with 18 constraint functions including ramp-rate manufacturing constraint on the overall structure.

Principal benefits of using Nexus to guide the optimization process were easy integration of external solvers; parallel and concurrent evaluations to best exploit available hardware and software resources, with a scalable architecture tunable by application; access to all results via organized tables and SQL external databases; advanced DOE and statistical tools to explore and analyze results; and state-of-the-art libraries for single- and multi-objective optimizations, DOEs and response surface modeling.

Pareto frontier. Source: iChrome

Results—The resulting Pareto frontier showed that torsional stiffness increased together with weight, as expected. However, the Pareto frontier quantified this relationship to reveal that a composite swing-arm would give a reduction in twisting angle of up to 19%. Alternatively, if the original torsional was preserved, a weight reduction of about 18% could be achieved.

Stress analysis of the optimized configurations showed that stress levels were below the imposed constraint values—100MPa in absolute value. Accordingly, strength and damage tolerance requirements were satisfied for the selected configuration.