基于PSO-SVR的植物纤维地膜抗张强度预测研究

发布时间：2018-02-26 19:33

本文关键词： 植物纤维地膜抗张强度预测模型支持向量机回归粒子群算法正交试验设计　出处：《农业机械学报》2017年04期 　论文类型：期刊论文

【摘要】：为快速、准确地对生产过程中植物纤维地膜抗张强度进行预测,降低生产成本,提高原料利用率,以植物纤维地膜中试平台为依托,基于粒子群算法(PSO)优化支持向量机回归(SVR)模型,结合正交试验设计L25(56)方法,以纤维打浆度、施胶剂添加量、湿强剂添加量、地膜定量、混合比作为模型输入参数,以植物纤维地膜抗张强度为输出进行模拟预测,并将模拟结果与SVR、BP、RBF智能算法模型进行对比分析。结果表明:PSO-SVR模型能够较好地表达植物纤维地膜抗张强度与模型参数间的非线性关系,并能根据输入参数快速准确地对植物纤维地膜抗张强度进行预测,测试集样本中预测值与实际值间均方误差、决定系数和均方根误差为0.117 N2、0.915、0.342 N;与其他智能算法(SVR、BP、RBF)相比,PSO-SVR算法模型具有更高的适用性与稳定性。研究结果可为生产过程中不同抄造工艺参数下植物纤维地膜抗张强度的在线监控提供参考依据。
[Abstract]:For fast, accurate of plant fiber in the production process of plastic tensile strength prediction, reduce production cost, improve the utilization rate of raw material, the plant fiber film test platform based on particle swarm optimization (PSO) algorithm based on support vector machine regression (SVR) model, combined with orthogonal design L25 (56) method. The fiber beating degree, sizing agent dosage, wet strength agent addition, film quantitative mixing ratio as the model input parameters, the plant fiber film tensile strength were simulated as output, and the results of the simulation with SVR, BP, RBF intelligent analysis algorithm model. The results show that the PSO-SVR model can the expression of the nonlinear relationship between plant fiber film tensile strength and model parameters of the well, and can quickly and accurately according to the input parameters of plant fiber film tensile strength was predicted, and the actual value between the mean square prediction value in the sample test set Error, coefficient of determination and root mean square error is 0.117 N2,0.915,0.342 N; and other algorithms (SVR, BP, RBF) compared to the PSO-SVR algorithm and the applicability of the model has higher stability. The research results can provide reference basis for on-line monitoring of plant fiber film making process in the production process of different parameters under the tensile strength.

【作者单位】：东北农业大学工程学院;
【基金】：“十二五”国家科技支撑计划项目(2012BAD32B02-5)
【分类号】：TB383.2;TP18

【相似文献】