Multi-objective turbomachinery optimization using a gradient-enhanced multi-layer perceptron
Response surface models (RSMs) have found widespread use to reduce the overall computational cost of turbomachinery blading design optimization. Recent developments have seen the successful use of gradient information alongside sampled response values in building accurate response surfaces. This pap...
Main Author: | |
---|---|
Format: | Journal article |
Language: | English |
Published: |
2009
|
_version_ | 1797077254860177408 |
---|---|
author | Duta, M Duta, M |
author_facet | Duta, M Duta, M |
author_sort | Duta, M |
collection | OXFORD |
description | Response surface models (RSMs) have found widespread use to reduce the overall computational cost of turbomachinery blading design optimization. Recent developments have seen the successful use of gradient information alongside sampled response values in building accurate response surfaces. This paper describes the use of gradients to enhance the performance of the RSM provided by a multi-layer perceptron. Gradient information is included in the perceptron by modifying the error function such that the perceptron is trained to fit the gradients as well as the response values. As a consequence, the back-propagation scheme that assists the training is also changed. The paper formulates the gradient-enhanced multi-layer perceptron using algebraic notation, with an emphasis on the ease of use and efficiency of computer code implementation. To illustrate the benefit of using gradient information, the enhanced neural network model is used in a multi-objective transonic fan blade optimization exercise of engineering relevance. Copyright © 2008 John Wiley and Sons, Ltd. |
first_indexed | 2024-03-07T00:15:18Z |
format | Journal article |
id | oxford-uuid:7aa24d37-23a7-417c-93a2-d7745e13cf49 |
institution | University of Oxford |
language | English |
last_indexed | 2024-03-07T00:15:18Z |
publishDate | 2009 |
record_format | dspace |
spelling | oxford-uuid:7aa24d37-23a7-417c-93a2-d7745e13cf492022-03-26T20:45:14ZMulti-objective turbomachinery optimization using a gradient-enhanced multi-layer perceptronJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:7aa24d37-23a7-417c-93a2-d7745e13cf49EnglishSymplectic Elements at Oxford2009Duta, MDuta, MResponse surface models (RSMs) have found widespread use to reduce the overall computational cost of turbomachinery blading design optimization. Recent developments have seen the successful use of gradient information alongside sampled response values in building accurate response surfaces. This paper describes the use of gradients to enhance the performance of the RSM provided by a multi-layer perceptron. Gradient information is included in the perceptron by modifying the error function such that the perceptron is trained to fit the gradients as well as the response values. As a consequence, the back-propagation scheme that assists the training is also changed. The paper formulates the gradient-enhanced multi-layer perceptron using algebraic notation, with an emphasis on the ease of use and efficiency of computer code implementation. To illustrate the benefit of using gradient information, the enhanced neural network model is used in a multi-objective transonic fan blade optimization exercise of engineering relevance. Copyright © 2008 John Wiley and Sons, Ltd. |
spellingShingle | Duta, M Duta, M Multi-objective turbomachinery optimization using a gradient-enhanced multi-layer perceptron |
title | Multi-objective turbomachinery optimization using a gradient-enhanced multi-layer perceptron |
title_full | Multi-objective turbomachinery optimization using a gradient-enhanced multi-layer perceptron |
title_fullStr | Multi-objective turbomachinery optimization using a gradient-enhanced multi-layer perceptron |
title_full_unstemmed | Multi-objective turbomachinery optimization using a gradient-enhanced multi-layer perceptron |
title_short | Multi-objective turbomachinery optimization using a gradient-enhanced multi-layer perceptron |
title_sort | multi objective turbomachinery optimization using a gradient enhanced multi layer perceptron |
work_keys_str_mv | AT dutam multiobjectiveturbomachineryoptimizationusingagradientenhancedmultilayerperceptron AT dutam multiobjectiveturbomachineryoptimizationusingagradientenhancedmultilayerperceptron |