Real-time simulation modeling method for variable geometric split axle type combustion gas turbine

Abstract

The invention relates to a real-time simulation modeling method for a variable geometric split axle type combustion gas turbine. The combustion gas turbine suitable for the method has the structure characteristics of adjustable compressor guide vane and split form and is composed of a fuel gas generator and a power turbine. The method comprises the following steps: selecting a sample point; acquiring a training sample point; utilizing the acquired sample point to train a RBF neural network; utilizing a part modeling method to establish a power turbine calculating module; connecting the module with the RBF neural network, wherein the RBF neural network is responsible for outputting the related parameters of the fuel gas generator and the power turbine calculating module is responsible for outputting the related parameters, such as, power, thereby acquiring a real-time simulation model of the combustion gas turbine. According to the modeling method for a real-time simulation model provided by the invention, the condition of angle variation of an entry guide vane of the air compressor is considered, the model is reasonably simplified, the advantages of the RBF neural network and the module method are combined with each other, the high instantaneity and high precision of the model are guaranteed while the sample size is greatly reduced.

Description

technical field [0001] The invention relates to the field of gas turbines, in particular to a modeling method for real-time simulation of variable geometry split-shaft gas turbines. Background technique [0002] In the process of gas turbine semi-physical simulation experiments and gas turbine control law research, the real-time performance of the gas turbine simulation model is high. Especially for split-shaft gas turbines with variable geometry, such gas turbines include gas generators with adjustable vane compressors and power turbines. Due to the large number of variables, the nonlinear model established based on the working mechanism of gas turbines has poor real-time performance. Although the linearized modeling method has better real-time performance, it is difficult to guarantee the simulation accuracy for the situation far away from the steady-state working condition. Although the neural network-based system identification method can achieve high accuracy and real-...

AI Technical Summary

Problems solved by technology

Especially for split-shaft gas turbines with variable geometry, such gas turbines include gas generators with adjustable guide vane compressors and power turbines. Due to the large number of variables, the nonlinear model established based on the working mechanism of the gas turbine has poor real-time performance.

Although the linearized modeling method has good real-time performance, it is difficult to guarantee the simulation accuracy for the situation far away from the steady state

Although the system identification method based on neural network can achieve high accuracy and real-time performance, it is difficult to directly establish a simulation model for split-shaft gas turbines with variable geometry due to the variety of dynamic conditions and the large sample size required.

Images