基于支持向量机的超超临界锅炉受热面污染监测模型研究

Research on Fouling Monitoring Model for Heating Surface of Ultra-supercritical Boiler Based on the Support Vector Machine

通过某发电集团科技(创新基金)项目"超超临界锅炉吹灰优化试验研究",利用电厂已有的DCS采集系统,得到实时数据样本,采用ε-支持向量机回归机(ε-SVR)来进行模型学习预测,结果表明基于支持向量机的预测模型的均方误差很小,能够准确快速得跟踪实际过程,取得了很好的预测效果。

Through the study of ＂Experiments on Soot-blowing Effect of uhra-supercritical Boiler＂ from a science （innovative funding）project in a electricity generation group, a novel method was presented in the following part. To begin with, the real-time data sample was gathered by the DCS data acquisition system from the power plant. Furthermore, the ε-support vector regression （ ε -SVR）was used to build the learning and prediction model. Eventually, the consequence confirms that the predictive model based on the support vector machine could achieve a minimal mean square error, which means this approach will perform well in actual process tracking and forecast results.