Multi-Disciplinary Design Optimization
Multi-disciplinary design optimization (MDO) is the process of using optimization methods and technologies to solve design problems which incorporate a number of design considerations and disciplines simultaneously. The MDO process centers on the idea that considering a number of design constraints at the same time as a basis to perform structural optimization will result in a superior solution than if they were viewed individually. As such, MDO can be described as the process of ‘finding the best compromise’.
Although MDO will potentially offer better design solutions, considering multiple disciplines simultaneously can increase the complexity of the design challenge considerably hence why it is not widely used in all industries.
Altair’s Jeff Wollschlager Briefly Speaks about the use of MDO in the Aerospace Industry
Whereas traditional design optimization processes will attempt to generate the best performing, minimum mass design based on a specific set of loads, use conditions or material properties, robustness optimization acknowledges that these ‘ideal’ set of factors (i.e. nominal) are unlikely to occur in reality. The optimization process must therefore factor in a level of uncertainty and variation to create a design solution which is more robust and able to withstand a wider variety of forces.
To use an analogy of a mountain, there is no practical point trying to get to the peak to get the best view when a slight gust of wind can blow you off, what is practical is to find the highest plateau where the view is unaffected. The same is true for engineering design, there is no point in coming up with a design which is optimized for a set of ideal conditions when in reality there exists uncertainty in the materials, manufacturing and operating conditions.
Achieving robust design is inherent in the quality philosophy of many companies. It will become an increasing requirement to demonstrate that digital designs achieve the required quality levels. This will initially be achieved on a component level and gradually migrate to complex systems. The initial requirement will be to understand the probabilistic variation of various parameters. This will require an increasing amount of measurement and an increased understanding of the physical drives of the component / system. Robustness can only be achieved by understanding the variation of the various factors.
Adding noise factors during optimisation is the best way in obtaining a robust solution the use of Design for Six Sigma (DFSS) principle helps identify failure modes and eliminate them earlier in the design process.
Richard Slade from EADS Astrium discusses the application of robustness optimization the ExoMars lander
Technical Paper - Jaguar Land Rover – Robust Design Optimization of a Knee Bolster