CAJ | 학술논문

为了快速合理地估算航空发动机叶片数控加工工时，在分析影响零件工时的主要因素（特征参数）的基础上，提出了基于SLFM神经网络模型的计算工时定额的思路。以影响零件工时的特征参数作为网络输入、以匹配的相似零件集作为训练样本，采用C++语言构建了定额计算网络模型。与BP网络对比试验表明，这种方法具有计算速度快，估算准确的优点。
위료쾌속합리지고산항공발동궤협편수공가공공시，재분석영향령건공시적주요인소（특정삼수）적기출상，제출료기우SLFM신경망락모형적계산공시정액적사로。이영향령건공시적특정삼수작위망락수입、이필배적상사령건집작위훈련양본，채용C++어언구건료정액계산망락모형。여BP망락대비시험표명，저충방법구유계산속도쾌，고산준학적우점。
In order to estimate the numerical human-hour of aero-engine blade rapidly and reasonably, an approach for estimating man-hour using a SLFM ANN model is proposed on the basis of analyzing the main in-lfuencing factors of man-hour. By using C++program,the man-hour calculation model is constructed with character-istic parameters of determinant of man-hour as the input and matched similar parts as the train set. It is indicated that the method calculates faster and estimates more accu-rately by comparison with BP neural network.