称重传感器蠕变补偿的RBF网络模型

Creep Compensation for Load Cell Based on RBF Network

称重传感器的蠕变是影响精度的主要因素之一。针对传感器蠕变的实时性与非线性,建立了称重传感器蠕变补偿的RBF网络模型。设计硬件采集电路并采用低功耗处理器对传感器数据进行软件补偿。仿真结果表明,RBF神经网络具有很强的逼近非线性函数和自学习能力,能够对称重传感器的蠕变误差进行修正。补偿后的蠕变误差减小至0.005%以内,补偿效果明显。

Creep is one of the main factors affecting the accuracy of the load cell. According to the creep＇s real-time and nonlinear property, the RBF （ Radical Basis Function ） network model was established for load cell. Hardware circuit was designed and low power processor was used to compensate the sensor＇s data. The simulation results show that RBF neural network has its self-learning function and its strong ability to approach a nonlinear function. It can correct the creep error of the load cell. The error reduces to O. 005 % after compensation and the effect is obvious.