A New Approximate Linear Programming Algorithm On Cyclic Scheduling Of Boiler Steam System

During the process of heating the boiler water into steam, only littlepart of impurities is carried out the boiler with steam. Then theconcentrations of impurities in the boiler water will increase severaltimes. With the attachment generating on the surface of the boil, thedepth of scale will increase exponentially with time. In addition, thecoefficient of themal conductivity of scale is far less than steel.Therefore, a periodic cleaning and maintenance is very necessary to along-term used boil.This paper introduces the process of building a mathematical modelof boiler steam system optimization scheduling in detail, and analysessome characters which contain the type and amount of the variables,the convexity of the objective function, and the convexity of thefeasible region formed by the constraints. With some comparativeresearches on traditional nonlinear programming optimizationalgorithm, we prepare to use approximate linear programmingalgorithm to solve the boiler steam system optimization scheduling problem.With testing a large number of nonlinear programming problem, wefound some defects in the traditional approximate linear programmingalgorithm. Hence, the paper proposes a new method called the steepestdescent-approximate linear programming (SDALP). After testing theaccuracy and the convergence speed of this new algorithm with severalrepresentative examples, we found that it had overcome the defects oftraditional approximate linear programming algorithm in application.The simulation results also demonstrate that the new method canimprove the convergence speed, and enhance the solution accuracy inmany kinds of nonlinear programming.Finally, this method has been simulated in a real thermal powerplant. Moreover, the usefulness of this method is showed as it achievesremarkable energy saving compared with the original approaches.