称重传感器蠕变误差的神经网络补偿方法

针对电阻应变式称重传感器存在严重的蠕变误差直接影响称重结果准确度的问题,提出了一种基于神经网络的称重传感器蠕变误差自动补偿模型,并给出了模型的训练算法。对量程为50 kg,C3等级的电阻应变式称重传感器进行了实验,实验结果表明,使用神经网络蠕变误差补偿方法补偿后,称重传感器加载标准砝码30 min内的蠕变误差最大变化量为0.0108 kg,小于国家标准GB/T 7551-200《称重传感器》规定的允许值。

Compensation Method for Creep Error of Load Cell Based on Neural Networks

The resistance strain load cell has heavy creep error, which will lead to the low accuracy of weighing results. To reducing this error, a model compensation on the load cell’s creep error based on neural networks(NNs) is proposed, and the training algorithms of the model is given. The experimental results of the load cell which capacity is 50 kg and its accuracy class is C3 show that the creep errors of the load cell with this proposed model are less than those without compensation, and the maximal change of its creep errors is 0.0108 kg, it is less than that of the maximal permissible error defined by Chinese National Standard, GB/T 7551-2008 “Load cell”.