轴流涡轮变工况性能预测分析及校准方法研究

Study on Performance Prediction Analysis and Calibration Method of Axial Flow Turbine Under Variable Conditions

叶片流道内复杂的三维和非定常粘性流动大大增加了轴流涡轮高精度变工况性能预测的难度。尤其针对燃气轮机涡轮一维性能分析,在中径处基元叶栅中应用经验模型代替流道内真实流动特性,显然具有一定的局限性,并存在较大的变工况性能计算误差。为此,采用智能优化算法,针对性能分析应用的7类经验损失模型中的33个经验系数,以实验特性或三维仿真结果获得的效率为优化目标,实现经验模型系数的自动校准,从而构建涡轮变工况性能的准确预测方法。经实例验算后发现原经验损失模型校准后,不同工况下的涡轮特性与实验结果相对误差均保持在1%以内,可较好的满足工程应用要求。

The complicated three-dimensional unsteady viscous flows in axial flow turbines greatly increases the difficulty of high-precision performance prediction under off-design conditions.In particular for the one-dimensional performance analysis of gas turbine turbines,the application of empirical models instead of realistic flow characteristics in the flow channel is clearly limited and has large errors in the calculation of variable performance conditions.Therefore,an intelligent optimization algorithm was used to automatically calibrate 33 coefficients of 7 categories of empirical loss models,and an accurate prediction method for turbine performance under off-design conditions was established as for a goal of optimal efficiency.It is found that after calibration of the original empirical loss models,the relative errors between turbine characteristics and experimental results are kept within 1%,which can better meet the requirements of engineering application.