Aiming at the measuring precision will be decreased due to the temperature drift of the strain-gage pin sensor resulted from the temperature variation during the underground working process,a temperature compensation model of the RBF neural network optimized by the drosophila algorithm is proposed. The extended parameters of the neural network are globally optimized by employing the drosophila algorithm,the parameters are measured by using the strain test platform,and the temperature compensation model is trained by employing the nonlinear mapping ca-pability of the neural network. To validate the compensation effect and the training efficiency of the temperature compensation model,the test is performed by using the sensor under 35 ℃. The result shows that the average error of the temperature compensation model is far less than that of the single algorithm compensation,the method is of the high training efficiency and the good compensation effect,the measuring precision of the sensor can be increased under the different temperatures and the different loads. The model of the paper is employed in the shearer working process,and the forces of the guiding sliding boots during the cutting process of the moving shearer are obtained. The research result of the paper provides the basis for the structure optimization of the guiding sliding boots,the shearer reliability improvement and the lifetime of the shearer.%针对应变片式销轴传感器井下工作过程中温度发生变化产生温度漂移,导致测量精度降低的问题,提出一种果蝇算法优化RBF神经网络的温度补偿模型,采用果蝇算法对神经网络的扩展参数进行全局优化,利用应力测试平台实测参数及神经网络非线性映射能力训练温度补偿模型.为验证温度补偿模型补 偿效果及训练效率,对35 ℃下传感器进行实验测试.结果表明:35 ℃下,温度补偿模型补偿平均误差远小于单一算法补偿效果,验证了此方法具有较高的训练效率及补偿效果,能够提高传感器在不同温度、载荷作用下测量精度,同时将本文模型应用采煤机截割煤壁工作中,得到导向滑靴在采煤机行走截割煤壁过程中受力,为导向滑靴结构优化及提高采煤机可靠性和使用寿命提供依据.
展开▼
