本文選題：果蠅算法 + RBF神經(jīng)網(wǎng)絡(luò)��； 參考：《傳感技術(shù)學(xué)報》2017年11期

【摘要】：針對應(yīng)變片式銷軸傳感器井下工作過程中溫度發(fā)生變化產(chǎn)生溫度漂移,導(dǎo)致測量精度降低的問題,提出一種果蠅算法優(yōu)化RBF神經(jīng)網(wǎng)絡(luò)的溫度補償模型,采用果蠅算法對神經(jīng)網(wǎng)絡(luò)的擴展參數(shù)進行全局優(yōu)化,利用應(yīng)力測試平臺實測參數(shù)及神經(jīng)網(wǎng)絡(luò)非線性映射能力訓(xùn)練溫度補償模型。為驗證溫度補償模型補償效果及訓(xùn)練效率,對35℃下傳感器進行實驗測試。結(jié)果表明:35℃下,溫度補償模型補償平均誤差遠小于單一算法補償效果,驗證了此方法具有較高的訓(xùn)練效率及補償效果,能夠提高傳感器在不同溫度、載荷作用下測量精度,同時將本文模型應(yīng)用采煤機截割煤壁工作中,得到導(dǎo)向滑靴在采煤機行走截割煤壁過程中受力,為導(dǎo)向滑靴結(jié)構(gòu)優(yōu)化及提高采煤機可靠性和使用壽命提供依據(jù)。
[Abstract]:In order to solve the problem of temperature drift caused by temperature change in the working process of strain gauge pin shaft sensor, a temperature compensation model of Drosophila algorithm for optimizing RBF neural network is proposed. The expanded parameters of neural network are optimized globally by Drosophila algorithm, and the temperature compensation model is trained by using the measured parameters of stress test platform and the nonlinear mapping ability of neural network. In order to verify the compensation effect and training efficiency of the temperature compensation model, the sensor was tested at 35 鈩，

本文編號：1984373