A TRACKING STATE ESTIMATOR FOR INDUSTRIAL-POWER STATION BOILER USING LINEAR-PROGRAMMING METHOD

This paper presents a new tracking estimator using linear programming (LP) method with an exponential forecaster prior to the estimation, an d r(n) statistical test following the estimation. While the forecaster is used to verify the new incoming measurements based on previous est imates, r, test is employed to refine the reliability of the estimator to eliminate bad data (ED) influence on resultant estimates. This pro posed configuration plus the LP approach as the estimation algorithm w hich can automatically delete ED from the measurements provides tile e stimator with a good performance in tracking system variations in both normal and abnormal situations. Detail considerations on tracking est imator and its applications are included in this paper. Extensive expe riences show this concept is correct. The developed estimator has been applied to the boiler system of a conventional power station.