论文部分内容阅读
本文针对压电陶瓷执行器的迟滞非线性问题,以实验数据为基础,采用双线性插值法建立离散化Preisach模型。为进一步提高建模精度,提出一种神经网络Preisach模型,并利用果蜗算法(Fruit fly optimization algorithm,FOAM*化BP神经网络的权值和阈值,避免其陷人局部最优解的风险,显著提高了BP网络寻优的准确性。对比采用双线性插值法建立的离散Preisach模型发现,经优化后的BP神经网络寻优效率和预测精度显著提高。实验结果表明,通过果蝇算法优化BP神经网络来对迟滞非线性位移进行预测建模是正确有效的,为压电陶瓷驱动器静态迟滞建模提供了一种新思路。
In this paper, aiming at the hysteresis nonlinearity of piezoelectric actuator, a discrete Preisach model was established by using bilinear interpolation based on experimental data. In order to further improve the modeling accuracy, a Preisach model of neural network is proposed. By using the weight and threshold of FOAM * BP neural network, the risk of avoiding local optimal solution is avoided Improve the accuracy of BP network optimization.Compared with the discrete Preisach model established by bilinear interpolation method, it is found that the optimized BP neural network optimization efficiency and prediction accuracy are significantly improved.Experimental results show that the optimization of BP Neural network to predict the nonlinear hysteresis displacement modeling is correct and effective, which provides a new idea for the static hysteresis modeling of piezoceramic actuators.