Optimization of Finite Element Model for Electrical Resistance Tomography

Liqing Xiao,Huaxiang Wang
DOI: https://doi.org/10.11784/tdxbz201210020
2014-01-01
Abstract:Aiming to improve the accuracy of solving the forward problem in electrical resistance tomography (ERT),the new topological finite element model was optimized in this paper. Using the radius of the second outer-most layer of the model as a variable,and the reciprocal of the root mean square value of a homogeneous sensitive field distribution as a fitness function,the improved genetic algorithm for solving low-dimension multi-peak function optimization problem was adopted to optimize the new topological finite element model which simulated the current line density and distribution of the sensitive field. Finally,the root mean square value of the core flow model corre-sponding to each peak point was used as an error function to determine the optimal value of the radius of the second outermost layer of the model. Simulation results demonstrate that,compared to the unoptimized topological finite element model,the conventional finite element model based on uniformly-spaced dissection and its modified version,the optimized model reduce the root mean square value of a homogeneous sensitive field distribution by 32.002 5%,83.395 8%and 44.760 5%,effectively improves both the accuracy of solving the forward problem and the quality of image reconstruction under the same experimental conditions.
What problem does this paper attempt to address?