Self-organizes fuzzy neural network and its application to build modeling of the ratio of fuel to water control system in the ultra supercritical unit

For ultra supercritical unit concurrent boiler with characteristics of parameters distribution, nonlinear and coupling tightly multivariable, this paper proposed a method based on self-organizes fuzzy neural network to build model for the ratio of fuel to water control system. The self-organizes fuzzy neural network with better non-linear approximation ability, good user-friendly, better forecast precision and generalization ability and other advantages, is able to solve the nonlinear and dynamic lag characteristics of control object. Use this method to build model for fuel-water ratio control system of the ultra supercritical concurrent boiler, and exert multi-step prediction for the intermediate point temperature; the results show that the model has good prediction ability, can reflect the dynamic characteristics of intermediate point temperature well, proving the feasibility of this method, and has very good practical significance and application value for controlling the ratio of fuel to water of ultra supercritical concurrent boiler.