Cyclic cleaning scheduling of multiple industrial boilers based on nonlinear programming with trust region filter algorithm

摘要

The cyclic cleaning scheduling of multiple industrial boilers was formulated as a mixed-integer nonlinear programming (MINLP). A sub-section nonlinear programming method with trust region filter algorithm (FNLP), which introduces the trust region filter algorithm into solving stationary point, was proposed for the NLP problems. The FNLP makes sure that the objective function will have efficient decrease while the iterative point is a feasible non-station point, and it also has a great freedom of step computing. The adjustment of iterative direction and step size exhibits that the FNLP has properties of easy implementation and convenient solution. The FNLP has been applied in optimizing a thermal power plant simulation, which is a mathematical model of the cyclic cleaning scheduling of multiple boilers system with decaying performance, and the result shows that it achieves remarkable energy saving compared with the original approaches.