基于PSO-ANFIS模型的矩形薄板荷载情况反演
Load inversion of rectangular thin plate based on PSO-ANFIS model
荷载情况反演在汽车工程和医学等领域有重要意义,目前常用的方法普遍存在需要的观测点较多、计算量大和计算效率低的问题。为了通过少量观测点的挠度高效准确地反演矩形薄板荷载情况,建立了基于粒子群算法(particle swarm optimization,PSO)改进自适应神经模糊网络(adaptive network-based fuzzy inference system,ANFIS)的反演模型。利用解析法求解矩形薄板在不同集中荷载作用下的挠度作为训练样本,输入4个观测点的挠度,再利用粒子群算法对自适应神经模糊网络的适应度寻优,反演集中荷载的位置和大小。结果显示,PSO算法可以有效提高ANFIS模型的精度,ANFIS迭代次数的增加能提高精度。PSO-ANFIS模型中荷载位置xF、yF和荷载大小F的最大残差分别为0.027 m、0.025 m和0.126 N,最大相对误差分别为5.00%、4.40%和4.50%。
Load inversion is of great significance in fields like automotive engineering and medical domain,etc.At present,commonly used methods generally have the problems of too many observation points,large amount of calculation and low calculation efficiency.In order to efficiently and accurately invert the load of rectangular thin plate through the deflection of a few observation points,in this paper,a load inversion of adaptive network-based fuzzy inference system(ANFIS)is optimized based on particle swarm optimization(PSO). Analytical method was used to solve the deflections of rectangular thin plates under different concentrated loads as training samples,and the deflections of four observation points were input. PSO was then used to optimize the fitness of the ANFIS,and the position and size of the concentrated load are retrieved. The results show that the PSO can effectively improve the accuracy of ANFIS model,and the increase of iteration times of ANFIS can improve the accuracy. In the PSO-ANFIS model,the maximum residuals of load positions xFyF?and load F were 0.027 m,0.025 m and 0.126 N,and the maximum relative errors were 5.00%,4.40% and 4.50%,respectively.
张宇鹏
工程基础科学公路运输工程
矩形薄板PSO-ANFIS模型荷载情况反演
张宇鹏.基于PSO-ANFIS模型的矩形薄板荷载情况反演[EB/OL].(2024-12-18)[2025-08-23].https://chinaxiv.org/abs/202412.00256.点此复制
评论