基于偏最小二乘法的锅炉飞灰含碳质量分数预测模型研究

Study on PLS-based Prediction Model for Mass fraction of Carbon Content in Boiler Fly Ash

针对目前锅炉飞灰含碳质量分数测量方法的滞后性及误差大等问题,根据某电厂大量运行(样本)数据,提出基于偏最小二乘法(PLS)的飞灰含碳质量分数的数学预测模型,并对模型的回归效果进行了分析和检验。研究结果表明:该方法能够克服数据间的多重相关性,具有预测速度快、计算精度高的特点。

Aiming at the problems such as lagging, large error, and so on of the existing measurement method for mass fraction of carbon content in boiler fly ash, a Partial Least Squares （PLS）-based prediction model was put forward, based on which analysis and testing on the regression effect of the model have been made upon a large amount of operating （sample） data of a power plant. Results show that the method can overcome the multiple correlations among data and is characterized by high prediction speed and calculation accuracy.