基于融合算法的航空发动机涡轮前温度最优控制

Turbine Inlet Temperature Optimization for Aeroengine Based on Fusion Algorithms

关键词:  航空发动机  涡轮前温度  小生境遗传算法  非线性规划算法  混合算法

Key words:  Aeroengine  Turbine inlet temperature  Niche genetic algorithm  Non-linear programming by quadratic lagrangian algorithm  Hybrid algorithm

摘要:

涡轮前温度是航空发动机的关键控制参数之一，在保持发动机推力不变的前提下，降低涡轮前温度可以有效提高发动机使用寿命，涡轮前温度最优控制是降低涡轮前温度的有效技术途径。本文研究了航空发动机涡轮进口温度的在线优化问题，并根据该优化问题的特点，提出了一种基于小生境遗传算法（NGA）与非线性规划（NLPQL）相结合的混合优化算法。数值仿真研究结果表明，虽然NLPQL计算速度快，但对涡轮进口温度的降低效果较差，NGA具有全局收敛能力，优化效果较好，但计算耗时较长。NGA和NGA-NLPQL混合算法在飞机全飞行包线内可分别降低涡轮前温度27.35K和27.19K，但与NGA相比，NGA-NLPQL混合算法节省了74.6%的计算时间。因此，所提出的NGA-NLPQL混合算法是一种效果更好、实时性更优的航空发动机涡轮前温度在线优化方法。

Abstract:

The turbine inlet temperature is a key control parameter of the aeroengine， reducing the turbine inlet temperature with the thrust unchanged can effectively improve the service life of the aeroengine， and the optimal control of the turbine inlet temperature is an effective technical way to reduce the turbine inlet temperature. The online optimization for turbine inlet temperature of the aeroengine was investigated， then a hybrid optimization algorithm based on niche genetic algorithm （NGA） and non-linear programming by quadratic lagrangian （NLPQL） hybrid algorithm was proposed according to the characteristics of this optimization problem. Numerical simulation results show that， although the NLPQL is fast， very little reductions of turbine inlet temperature can be obtained by employing it and the NGA has the ability of global convergence and can achieve good optimization effect but it needs a lot of time. NGA and NGA-NLPQL can reduce the turbine inlet temperature average by 27.35K and 27.19K respectively within the full flight envelope of the aircraft， but the presented hybrid NGA-NLPQL saves 74.6% computational time when compared to NGA. It is shown that the hybrid NGA-NLPQL is a more effective and higher real-time method for on-line turbine inlet temperature optimization of the aeroengine.