There was a story in USA Today this week with the headline “American muscle cars outselling fuel-thrifty hybrids despite $4 gas”. It is a pretty entertaining fact and while for the month of May it was true that Ford Mustang, Chevrolet Camaro and Dodge Challenger outsold all Hybrids added together, it is not the whole story. The Car Connection did a good job of breaking down the numbers and putting them in context.
Data without context is rarely useful information. My fellow Altair Blogger Bob Yancey referenced the use of Simulation Data Management (SDM) and the panel discussion at the NAFEMS World Congress I recently sat on. Managing data is one aspect of putting a data point into context. Another that was discussed extensively at the NAFEMS World Congress was stochastics. Although there was a comment made during the congress that maybe the use of the word “stochastic” could be the most significant barrier to understanding this valuable tool (often referred to instead as a Monte Carlo method)!
So how does a Monte Carlo simulation put an analysis result into context? If I perform a single analysis on a model that gives me a deterministic result. That means I’ve applied a single set of assumptions to get my answer. A Monte Carlo simulation will take my model and introduce randomness into those assumptions achieving a probabilistic result. In this case my answer is not a single data point determined by a unique set of values for my initial conditions, but a cloud of data created from probability distributions. The shape of the cloud will tell you far more than that single point ever could.
Study tools such HyperStudy allow engineers to use Monte Carlo methods to assess reliability and robustness of designs and provide qualitative guidance to improve and optimize based on these assessments. The example used to spark discussion during the NAFEMS Stochastics Focus Session was work that Altair performed with EADS Astrium to not only optimize the landing behavior of the ExoMars lander but also to investigate the probability of failure. By the way it uses an airbag, but not quite the kind you’ll find in a car…
Maybe stochastics isn’t quite as much fun as any one of the three best selling American sports cars, but it can ensure that you are drawing conclusions from a set of data and not just reporting on an outlier.
Latest posts by Tony Norton (see all)
- Thought Leader Thursday: The “America Makes” Project for Better Lives - August 13, 2015
- Increase CAE Workflow Efficiency by 74% with Altair Packaged Solutions - April 30, 2015
- Optimize Early and Optimize Often! - December 6, 2013