Nonlinear Parameter Prediction and Estimation of Fossil Power Plant Based on Kernel Partial Least Squares
In fossil power plant, many variables are online measured. These variables, which are the key indicators of process performance, are normally determined by on-line parameter analyzer. These analyzer are often expensive and require frequent and high cost maintenance. Furthermore, significant delay and measurement error will be incurred, which lead to sampled variables cannot be good enough to act as feedback signals. Such limitations have a severe influence on combustion efficiency and the real-time control. Joseph and Brosillow1 introduced inference control of process to solve the problem. Other methods, such as principal component regression (PCR), partial least squares (PLS), and canonical variate analysis (CVA) address this issue by projecting the original process variables onto a low dimension space with orthogonal latent variables (LVs).
Xi Zhang Yaqing Zhu Weiwu Yan Huihe Shao
Guangdong Electric Power Research Institute China Southern Power Grid No.73,Meihua Road,Guangzhou,51 Department of AutomationShanghai Jiaotong University No.800,Dongchuan Road,Shanghai,200240,China Department of Automation Shanghai Jiaotong University No.800,Dongchuan Road,Shanghai,200240,China
国际会议
2010 IEEE信息与自动化国际会议(ICIA 2010)
哈尔滨
英文
1-4
2010-06-20（万方平台首次上网日期，不代表论文的发表时间）