Parameter Variability
OPTIMUS allows to easily taking into account the variability of some of your design variables. For each design variable it is possible to choose from 12 distribution functions including Normal (Gaussian), Lognormal, Uniform, Exponential, Triangle, Gamma, Beta, Rayleigh, Erlang, Gumbel, Weibull, User-Defined parameter probability density functions.
Robustness
Based on the assigned parameter variability, the OPTIMUS user can choose between the Monte Carlo Simulation (MCS) or the faster First-Order Second Moment (FOSM) method to predict the resulting distribution or variance of the outputs of the simulation workflow. Both the MCS and FOSM method can be applied on the full simulation process or on a previously calculated response surface model.
Reliability
In addition OPTIMUS contains the First-Order and Second-Order Method (FORM and SORM) to predict the probability of failure and the reliability index of your design. These methods estimate how “far” your given design is located from your constraint boundaries. The closer your design is situated to the failure domain, the higher your probability of failure. In a Design-For-Six-Sigma philosophy a reliable design should located at least six sigmas away from the failure boundary, which results in a low 0.002 defects per million.
Optimization for Robustness & Reliability
In case the initial robustness (output variance) or reliability (failure probability or reliability index) does not meet your design requirements, OPTIMUS enables you to add the robustness and reliability measures to your optimization problem either as constraints or objectives. You can apply all available OPTIMUS optimization methods to identify a new design that meets your robustness and reliability criteria.
|
