LAUSR

Thermal conductivity is employed as a measure of the strength of heat transfer from a substance. There is a lack of accurate methods for solving thermal conductivity in heat transfer inverse problems due to the ill-posedness and non-linearity of the problem. In this paper, an intelligent solution algorithm based on the artificial fish swarm algorithm is proposed and a proportional control algorithm is introduced to invert the heat transfer coefficient of a transient heat conduction system. In the forward heat conduction process, a finite difference algorithm is utilized to solve for the required measurement point temperature of the heat conduction system. In the inverse problem solution, the performance of the algorithm is compared before and after its improvement, and the inversion is compared in the rectangular and circular heat transfer domains under the influence of different heat transfer materials, interference sources and errors. The results show that in the absence of disturbances, the inversion errors of the improved artificial fish swarm algorithm in the rectangular and circular domains are 0.4% and 0.1%, respectively. In the presence of large measurement errors ( $$\sigma = 0.5$$ σ = 0.5 ), the inversion errors in the rectangular and circular domains are 8.7% and 2.4%, respectively, which validate the accuracy of the algorithm.