CAJ | 학술논문

为了提高焊接工时定额的计算效率和精确度，消除传统定额方法过度依赖线性公式进行计算而忽略工时定额与其加工工艺参数之间的非线性关系的不足，将BP神经网络技术应用到焊接工时定额中，建立多输入单输出的神经网络模型，并运用实际数据对建立的神经网络进行训练，通过测试样本验证了该网络模型输出的仿真值与实际值之间的误差在5％以内。验证结果表明，用BP神经网络方法建立的焊接工时定额模型能够对焊接工时进行准确的预测。
위료제고한접공시정액적계산효솔화정학도，소제전통정액방법과도의뢰선성공식진행계산이홀략공시정액여기가공공예삼수지간적비선성관계적불족，장BP신경망락기술응용도한접공시정액중，건립다수입단수출적신경망락모형，병운용실제수거대건립적신경망락진행훈련，통과측시양본험증료해망락모형수출적방진치여실제치지간적오차재5％이내。험증결과표명，용BP신경망락방법건립적한접공시정액모형능구대한접공시진행준학적예측。
In order to improve the calculation efficiency and accuracy of welding man-hour quota ,and to eliminate the traditional quota method relied on linear formula excessively but ignored the lack of a non -linear relationship between man-hour quota and their processing technical parameters .The BP neural network technology is applied to the welding man-hour quota ,the establish-ment of a neural network model of multi-input and single-output,and the training on the establishment of the neural network ap-plying actual data ,the error between the simulation output values of the network model and actual values is validated within 5%by testing samples .The verification results show that establishing welding man-hour quota model based on the BP neural network method can predict welding man-hour accurately .