Theory-Lesson 9(Reliability) Flashcards
What is the main idea of reliability?
When designing we have to think if the results of the optimization can be realistically implemented.
The robustness of a system, which is usually the main goal for manufacturers is affected by the stochastic nature and usually provides better results, since optimal designs based on deterministic approaches are often more prone to be most sensitive to parameter uncertainty.
Also, in the real-world the optimization is not based on deterministic approach rathen than a stochastic approach. That means that the systems are subject to variations that should be taken into account when trying to formulate and solve the problem.
Some DVs might be deterministic (number of gear) but most of the DVs we are dealing with are stochastic (value of a load, meaning that a deviation must be taken into account. Also, distrurbances should be accounted.
A gaussian distribution can be used to describe the stochastic nature of some of the DVs.
Pareto optimal solution in a stochastic framework
In this case, while computing the Pareto-points, we should take into account the variance and the expected value.
DRAWBACK: Increase of the number of OFs and thus we get a huge number of Pareto solutions.
What’s the most common approach for stochastic optimality?
We add a risk coefficient 0<a<1 with the goal of improving the robustness of the system. We don’t really care about maximizing the performance.
The idea is to transform the stochastic nature of the problem into a deterministic problem, by intoducing the expected value and the variance of the variables involved.
