耦合17级压气机通流模型的燃气涡轮发动机混合维度仿真研究

Hybrid-Dimensional Simulation of Gas Turbine Coupled with 17-Stage Compressor Through-Flow Model

为理清不同数值缩放方法特点,发展适用于多级高压比压气机的燃气涡轮发动机混合维度仿真算法,针对某型分轴式燃气轮机及其17级压气机,分别基于弱耦合、常规迭代耦合、完全耦合、以及本文提出的改进迭代耦合方法建立了耦合整机零维模型与压气机通流模型的整机混合维度仿真模型。对比分析不同数值缩放方法的差异,结果表明,弱耦合、迭代耦合与完全耦合方法基本可实现相同的计算精度,与试验数据的最大相对误差小于10%。其中,弱耦合方法易于收敛,其计算精度与计算效率受特性图质量等因素影响;常规迭代耦合方法计算效率虽高,但其鲁棒性在多级高压比压气机的数值缩放中难以保证;完全耦合方法计算效率较低,且对求解算法与迭代变量初值要求较高;而改进迭代耦合方法有效利用了17级压气机压比范围宽的特点,兼具鲁棒性与计算效率,整体表现最佳。

In order to clarify the characteristics of different zooming strategies,and develop the algorithm of gas turbine hybrid-dimensional simulation for multistage compressor with high pressure ratio.For a split-shaft gas turbine and its 17-stage compressor,the hybrid-dimensional simulation models that coupled zero-dimensional engine model and compressor through-flow model were developed based on the De-coupled approach,conventional iterative coupled approach,fully coupled approach,and the improved iterative coupled approach proposed in this paper respectively.The variations of different zooming strategies were compared and analyzed,the results indicate that the De-coupled approach,iterative coupled approach and fully coupled approach can basically achieve the same simulation accuracy,the maximum relative error between simulation result and experimental data is less than 10%.Moreover,the De-coupled approach tends to converge easily,its simulation accuracy and efficiency are affected by factors such as the quality of characteristic map.Although the conventional iterative coupled approach exhibits high simulation efficiency,its robustness is difficult to guarantee in the zooming of multistage compressor with high pressure ratio.The fully coupled approach has relatively low simulation efficiency and high requirements for both the solving algorithm and the initial value of iterative variables.Meanwhile,the improved iterative coupled approach effectively utilizes the wide pressure ratio range of the 17-stage compressor,exhibiting both high robustness and high simulation efficiency,the effect is optimal.