摘要
通过分析传统的钢芯铝绞线综合弹性系数计算方法的不足,并考察主要影响因素,建立了钢芯铝绞线综合弹性系数估算的BP神经网络模型。依据现有的综合弹性系数理论以及现行各规格产品的综合弹性系数实际值,对主要因素进行了定量化,并结合所建立的BP神经网络模型,运用C语言程序来快速估算钢芯铝绞线综合弹性系数。此BP网络模型采用了现行标准中51种规格的参数,其中21种规格作为训练样本,余下的30种规格作为检测样本。经输入检验样本检验,证明该方法为一种快速、准确的钢芯铝绞线综合弹性系数估算方法。
Through analyzing deficiencies of traditional methods calculating ACSR combined elastic coefficient and investigating major factors affecting the coefficient, this paper establishes BP neural network models applicable to ACSR combined elastic coefficient estimation. Based on existing theories and actual values from current standards, the effects of major factors affecting ACSR are quantified, and based on BP network model, the coefficient is estimated using C code. The BP network model adopts parameters of 51 specs of ACSRs in existing standards, among them, 21 are used as training samples, while the remaining 30 are used as testing samples. Through testing sample inputs, this approach is proved to be fast and accurate in ACSR combined elastic coefficient estimation.
出处
《电力建设》
2008年第3期6-7,13,共3页
Electric Power Construction