无磁水表非线性修正策略的研究与实现

The Research and Realization of Nonlinearity Correction in Nonmagnetic Water Meter

针对无磁水表的非线性问题,提出了径向基(Radial Basis Function,RBF)神经网络的曲线拟合方法,并与传统的最小二乘法拟合结果进行对比,发现神经网络方法效果更优.在系统实现方面,通过在单片机内存储所拟合的曲线数据,并采用折半查找的方法快速查得相应瞬时流量,既提高了计量精度,又满足了系统实时性的要求.

In order to resolve the problem of nonlinearity in nonmagnetic waterAneter, the method of RBF neural network, which was testified to be better than least squares here, was proposed in this paper. In the aspect of system realization, saving the data of curve fitting in the memory of singlechip and using the way of binary search to get the Instantaneous flow quickly could not only increase the accuracy, but also meet the need of the Real- time in system.