Machine-learning Optimizer to Scale Down Product Design Cost

Posted On November 19, 2020     

Computer simulations are a significant part of the product design optimization process, allowing engineers to test various configurations and select the best design methods among the numerous alternatives. But even at an institute such U.S. Department of Energy's (DOE) Argonne National Laboratory, simulations can be very costly and take a long duration to run.

Intending to augment this design process, a research team in Argonne's Energy Systems (ES) division recently developed an innovative design optimization tool called ActivO. The new tool can radically decrease the time required to find the best design.

It employs a new machine learning technique that assists users in focusing on how to most competently target computational resources. Machine learning refers to applying artificial intelligence that permits systems to learn and get better from experience automatically.
ActivO runs the simulations in a brilliant means and rapidly identifies the design space parts. Usually, any process used to take two to three months to present the optimum design can now be completed in a week.

The ActivO approach was productively verified for use in optimizing combustion engines.

ActivO is a fusion algorithm that grabs the strengths of two diverse machine learning substitute models to get more outstanding performance. The machine learning models are intended to work courteously. Rather than run simulations that are sampled arbitrarily, one of the models helps us to explore the design space adaptively, which fundamentally guides about the areas that are most possible to hold the global optimum.

This type of approach supports machine learning surrogates to "exploit" and "explore" the design space in a more balanced and efficient manner than standard evolutionary techniques used in the industry, like genetic algorithms.

It runs in small groups of simulations, making it mainly valuable for industrial users since they frequently don't have the computational power to run a large group of simulations.

ActivO is suitable for discovering the design space typical of combustion engines in the aerospace and automotive industries. But it could also be used for creating optimization for an all-purpose product. It can be voluntarily adopted by industry to better their design workflows, which would cut product design costs. There are significant commercialization opportunities in this regard.
It will not just cut the design costs for the industry but has even more benefits. The major goal is to facilitate energy efficiency and lesser the environmental impact of these engines.

Related Report:

Global Machine Learning As A Service Market Report

Why Choose Us

Client First Policy
Excellent Quality
After Sales Support
24/7 Email Support

Buy Chapters or Sections

Avail customized purchase options to meet your exact research needs:

  • Buy sections of this report
  • Buy country level reports
  • Request for historical data
  • Request discounts available for Start-Ups & Universities