CIESC Journal ›› 2018, Vol. 69 ›› Issue (S2): 291-299.DOI: 10.11949/j.issn.0438-1157.20181211

Previous Articles     Next Articles

Research on temperature prediction model for microwave heating based on wavelet denoising and improved PSO-SVM

ZHOU Xinzhi1, SHAO Lun1, LI Rongkun1,2, ZHAO Chengping1, DONG Chenlong1   

  1. 1 Electronic Information School, Sichuan University, Chengdu 610065, Sichuan, China;
    2 Ziyang Army Reserve Corps of Engineers, Ziyang 641300, Sichuan, China
  • Received:2018-10-16 Revised:2018-10-26 Online:2018-12-31 Published:2018-12-31
  • Supported by:

    supported by the National Basic Research Program of China (2013CB328903).

基于小波去噪和改进型PSO-SVM的微波加热温度预测模型研究

周新志1, 邵伦1, 李荣昆1,2, 赵成萍1, 董晨龙1   

  1. 1 四川大学电子信息学院, 四川 成都 610065;
    2 资阳陆军预备役工兵团, 四川 资阳 641300
  • 基金资助:

    国家重点基础研究发展计划项目(2013CB328903)。

Abstract:

Aiming at the nonlinear correlation and time-delay of microwave heating process, a microwave-heated lignite temperature prediction method based on wavelet denoising optimized by thresholds through correlation principles and support vector machine optimized by improved particle swarm optimization is proposed. The temperature for microwave heating of lignite is acquired via wavelets with an optimized threshold based on relevant principles. The particle swarm optimization algorithm is improved by introducing the similarity function and the control parameter α. It is used to optimize the penalty coefficient C, the insensitive loss function parameter ε and the kernel function parameter g of the microwave heating lignite temperature support vector machine prediction model parameters. As the similarity function increases, the random variability of the particles increases; the number of iterations increases, and the control parameter α decreases. The simulation results show that the optimized support vector machine regression prediction model improves the accuracy of microwave heating lignite temperature prediction, and provides a mathematical model for the study of microwave drying lignite control method.

摘要:

针对微波加热过程的非线性相关性和时滞性,提出基于相关原则优化阈值的小波去噪和改进型粒子群算法优化支持向量机相结合的微波加热褐煤温度预测方法。采用相关原则优化阈值的小波对微波加热褐煤温度进行提取,通过引入相似度函数和控制参数α改进粒子群算法,用于优化微波加热褐煤温度支持向量机预测模型的惩罚系数C、不敏感损失函数参数ε、核函数参数g等3个直接影响温度预测精度的参数。相似度函数增大,粒子随机变异率增大;迭代次数增加,控制参数α减小。仿真实验结果表明,优化后的支持向量机回归预测模型提高了微波加热褐煤温度预测的准确率,为微波干燥褐煤控制方法的研究提供一种可供参考的数学模型。

CLC Number: