论文部分内容阅读
为消除差动变压式传感器零点残余电压及非线性特性,提出基于重新参数化的B样条函数以及粒子群算法β参数B样条神经网络(B-BP-PSO)。由粒子群算法(PSO)取代传统BP算法,并由其搜索最佳β因子,用以得到适合本网络权值搜索的最优重新参数化B样条基函数,从而使得该神经网络可有效克服传统算法易于陷入局部最优的缺点。实验结果表明:经其校正后的差动变压式传感器的最大输出误差为11 mV,最大相对误差为3.7%,零点残余电压为5 mV.该方法可有效消除各种参数对差动变压式传感器输出结果的零点残余电压及非线性影响,且可用于其他各类传感器的非线性校正,具有很大的实际应用价值。
In order to eliminate zero residual voltage and nonlinear characteristic of differential transformer, a B-spline function based on re-parameterization and a B-spline neural network (B-BP-PSO) based on particle swarm optimization algorithm are proposed. The traditional BP algorithm is replaced by Particle Swarm Optimization (PSO), and the best β factor is searched for the optimal re-parameterized B-spline basis function which is suitable for the weight search of the network, so that the neural network can effectively overcome Traditional algorithms tend to fall into the shortcomings of local optima. The experimental results show that the maximum output error of the differential transformer is 11 mV, the maximum relative error is 3.7% and the zero residual voltage is 5 mV.The method can effectively eliminate the influence of various parameters on the differential transformer Sensor output zero residual voltage and nonlinear effects, and can be used for other types of sensors nonlinear correction, with great practical value.