摘要
针对建立热电厂磨煤机一次风量软测量模型训练样本多、样本特征维数大等特点,考虑到现场测量所需的实时性和准确性,提出了基于样本优化、核主成分分析(KPCA)和最小二乘支持向量机(LSSVM)相结合的方法进行建模。运用某电厂历史运行数据对模型进行仿真验证,结果表明:基于样本优化的KPCA-LSSVM软测量模型在精确性、跟踪能力和运行速度上均要优于LSSVM、BP和KPCA-BP模型,这为现场磨煤机一次风量的准确、实时测量提供了一定的理论依据。
In accordance with the features of establishing soft sensing model for primary air （PA） flow of pulverizers in cogeneration power plant, e.g. , more training samples and sample characteristics with large dimension, and considering the requirements of real time performance and precision in field measurement, the method of establishing model based on the combination of sample optimization, kernel principal component analysis （ KPCA ） , and least square support vector machine （ LSSVM } is proposed. The simulation verification of the model is conducted using historical operating data of certain power plant ; the results indicate that the method based on sample optimization KPCA-LSSVM soft sensing model is better than LSSVM, BP or KPCA-BP model in accuracy, tracking capability and operating speed, this provides certain theoretical basis for real time and accurate measurement of PA air flow of the pulverizers.
出处
《自动化仪表》
CAS
2015年第3期62-67,共6页
Process Automation Instrumentation
关键词
软测量
样本优化
核主成分分析
最小二乘支持向量机
BP神经网络
Soft sensing Sample optimization Kernel principal component analysis （KPCA） Least square support vector machine{ LSSVM ）BP neural network
