Temperature Soft Sensing And Oxide Skin Cleaning Of Supercritical Unit Superheater

Though supercritical units can achieve higher thermal conversion efficiency by optimizing the operation condition parameters,this also requires safer and more reliable operation of unit equipment in a longer term.One of the typical problems is that the quick start-stop operation leads to the flaking of the oxide skin on the high-temperature heating surface of superheaters,which can block pipelines to make them burst.It is of great research significance and practical value to monitor the operation of supercritical units on site and study how to identify and forecast the operation state of units,so as to ensure normal and reliable operation.The superheater pipeline of a supercritical unit in Liaoning province burst due to the blockage caused by oxide skin.Against this background,this thesis combined independent component regression with combinatorial optimization,to conduct combinatorial optimization of classical non-quadratic functions in independent component analysis through particle swarm optimization.Compared with traditional methods of independent component regression,the method of this thesis can be used to make more accurate regression predictions of field data,thus achieving better effects of forecasts.On this basis,it was analyzed and discussed how oxide skin grew in the high-temperature and high-pressure environment of the pipeline.Finally,a chemical cleaning solution was presented according to the growth condition of oxide skin in the superheater pipeline.The study of this thesis was conducted mainly in three aspects:(1)Through simulation analysis,the three classical non-quadratic functions G1,G2,G3 in the independent component regression algorithm were compared in terms of their effects on the extraction and prediction of independent components.Moreover,the independent component regression based on traditional non-quadratic functions was transformed into a combinatorial optimization problem,which was solved with the algorithm of particle swarm optimization,to find the combination weight coefficient to make the prediction index optimal.(2)On the basis of the above theoretical analysis,the data of operation acquired on site from the supercritical unit in Liaoning province was analyzed for verification,in which the combinatorial optimization method of this thesis was applied in forecasting the temperature of the supercritical superheater,to verify the effectiveness of this method in combination with data measured on site.(3)Based on the oxide skin growth model,a solution of using citric acid for cleaning was designed under an ideal condition,and an experimental platform was built for cleaning,to verify the analysis on the cleaning of oxide skin in this thesis.