论文部分内容阅读
针对小批量多品种生产模式下的织物热传递性能的预测问题,提出了用支持向量机对织物热传递性能进行预测的方法。以织物的经纬向纱线的粗细、织物的经纬密度及织物厚度为输入变量,织物稳态下的热传递性能指标为输出变量,分别建立了保暖率、克罗值的SVM预测模型。采用网络搜索并结合交叉验证的方法分别对保暖率、克罗值的SVM预测模型的参数进行优化。实验结果表明:从17组样本中随机选取15组为训练样本,2组为预测样本,预测误差都在5%以下,验证了所提出的在小样本时用SVM预测模型预测织物的热传递性能的有效性。
In order to predict the heat transfer performance of fabric under small batch and multi-variety production mode, a method of predicting fabric heat transfer performance by using support vector machine is proposed. Taking the warp and weft of the fabric, the thickness of the yarn, the warp and weft density of the fabric and the thickness of the fabric as the input variables, the heat transfer performance index of the fabric is taken as the output variable, and the SVM prediction model of the warm-keeping rate and the Kerr value are respectively established. The network search and the cross-validation method were used to optimize the SVM prediction model parameters of warm-up rate and croosh value respectively. The experimental results show that 15 samples are randomly selected from 17 samples and 2 samples are used as prediction samples. The prediction error is below 5%. The proposed SVM prediction model is used to predict the heat transfer performance of the fabric Effectiveness.