Partial Least Squares for Power Plant Performance Monitoring

Abstract The problems faced in monitoring power plant performance are outlined and consideration is given to more effective utilization of the data generated on-line by the distributed control system (DCS). Partial least squares (PLS) is identified as a method ideally suited to coping with the large number of correlated and collinear signals available on a power plant, for process monitoring and performance related analysis. Non-linear PLS models are trained using archived data from a local utility to predict quality measures of thermal efficiency, and NO x and SO x emissions for generating units dual-firing on both oil and gas. The performance and diagnostic capabilities of the resulting models are examined, illustrating the simplicity of operation of the proposed monitoring tool.

[1]  Seongkyu Yoon,et al.  Relationships Between Statistical and Causal Model Based Approaches to Fault Detection and Isolation , 2000 .

[2]  John F. MacGregor,et al.  Process monitoring and diagnosis by multiblock PLS methods , 1994 .

[3]  D. Broomhead,et al.  Radial Basis Functions, Multi-Variable Functional Interpolation and Adaptive Networks , 1988 .

[4]  Thomas F. Edgar,et al.  Identification of faulty sensors using principal component analysis , 1996 .

[5]  E. Swidenbank,et al.  On-line optimization and monitoring of power plant performance through machine learning techniques , 1998 .

[6]  S. Wold.  Cross-Validatory Estimation of the Number of Components in Factor and Principal Components Models , 1978 .

[7]  S. Qin.  Recursive PLS algorithms for adaptive data modeling , 1998 .

[8]  S. Joe Qin,et al.  Subspace approach to multidimensional fault identification and reconstruction , 1998 .

[9]  Soon Heung Chang,et al.  Development of an on-line fuzzy expert system for integrated alarm processing in nuclear power plants , 1995 .

[10]  B. Kowalski,et al.  Partial least-squares regression: a tutorial , 1986 .

[11]  A. J. Lichnowski.  Developments in electricity generation plant control management , 1999 .

[12]  A. J. Morris,et al.  Non-linear projection to latent structures revisited (the neural network PLS algorithm) , 1999 .

[13]  Juan Pedro Caraça-Valente,et al.  Combining expert knowledge and data mining in a medical diagnosis domain , 2002, Expert Syst. Appl..