Physically enhanced prediction method of extrapolating low speed characteristics of compressor based on PSO-SVR
-
摘要:
针对航空发动机压气机低转速部件特性外推精度较低且等熵效率不连续的问题，提出一种基于物理增强的PSO-SVR压气机低转速特性预测方法。该方法结合了几种物理外推方法并采用粒子群算法优化支持向量回归模型参数对压气机低转速特性进行回归预测。将传统压气机单级特性表达参数转换成换算扭矩，同时对压气机零转速特性线进行外推计算，将已知高转速特性数据作为训练集建立支持向量回归模型，对模型参数的变化域进行寻优建立精度较高的回归模型，并采用此模型对压气机低转速特性数据进行预测。研究表明：该方法可以解决在压气机进行特性数据外推时效率不连续的问题，单级压气机特性转换成换算扭矩后具有明显的规律性，建立的回归预测模型具有较好的预测效果。基于物理增强的PSO-SVR方法的预测结果中，平方相关系数为最小为0.976，最大均方误差为0.06%，最大平均相对误差为8.96%。利用换算扭矩工作参数可以预测出压气机在低转速运行下的压气机态、搅拌机态和涡轮态3种工作状态。所建立的预测方法能够充分挖掘高转速特性数据中所包含的信息，提高压气机低转速特性的预测精度，为发动机的起动模拟提供数据支撑。
Abstract:A physically enhanced prediction method based on particle swarm optimization (PSO) optimization support vector regression (SVR) was proposed to address the issue of low accuracy and discontinuous isentropic efficiency of compressor performance at low speed. This method combined several physical extrapolation methods and utilized particle swarm optimization to optimize the support vector regression model parameters for the regression prediction of compressor’s low speed characteristics. The traditional single-stage characteristic parameters of the compressor were converted into corrected torque, and the compressor zero speed characteristic line was extrapolated. The high-speed characteristics were employed as the training set to establish the support vector regression model, PSO was used to obtain the high-precision regression SVR model, and the model was applied to predict the low speed characteristics of compressor. Results showed that the minimum determination coefficient was 0.976, the maximum mean square error was 0.06%, and the maximum average relative error was 8.96% in the prediction results based on PSO-SVR model. The three working states of compressor were predicted at low speed based on PSO-SVR model, including compressor, stirrer and turbine states. Therefore, the prediction method can fully mine the information contained in the data, improve the prediction accuracy of compressor at low-speed characteristics and provide data support for engine starting simulation.
-
图 1 压气机3种工作状态的焓-熵示意图
Figure 1. Enthalpy-entropy schematic diagram of compressor for three operating states
图 2 压气机3种工作状态的等熵效率变化示意图
Figure 2. Sketch map of compressor isentropic efficiency in three operations
图 3 压气机叶栅受力和速度三角形示意图
Figure 3. Schematic diagram of the forces and velocity triangles of compressor blade
图 4 基于PSO-SVR的压气机低转速部件特性预测流程图
Figure 4. Flow chart of prediction method of compressors low speed characteristics based PSO-SVR model
图 5 压气机已知转速特性数据转换前后对比
Figure 5. Comparison of compressor high speed characteristics before and after conversion
图 6 压气机不同转速特性线二次系数变化拟合结果
Figure 6. Fitting curve of quadratic coefficient of the characteristic line for different speeds of compressor
图 7 第1级压气机无量纲特性参数扩展结果
Figure 7. Expansion results of the first stage compressor dimensionless characteristics
图 8 第2级压气机无量纲特性参数扩展结果
Figure 8. Expansion results of the second stage compressor dimensionless characteristics
图 9 无量纲特性参数等熵效率扩展结果
Figure 9. Expansion results of dimensionless characteristic parameters isentropic efficiency
表 1 压气机级特性预测准确性结果
Table 1. Prediction accuracy of compressor stage characteristic data
级数 评估指标 R2 Ems/% Emr/% 第1级 Tor1 0.9986 0.06 6.91 π1 0.9763 0.03 8.96 第2级 Tor2 0.9994 0.01 3.05 π2 0.9948 0.02 6.44 
下载: 导出CSV
-
[1] JONES G,PILIDIS P,CURNOCK B. Extrapolation of compressor characteristics to the low-speed region for sub-idle performance modelling[C]//Proceedings of ASME Turbo Expo. Amsterdam,The Netherlands: ASME,2002: 30649. [2] 施洋,屠秋野,严红明,等. 一种航空发动机全状态性能模型[J]. 航空动力学报,2017,32(2): 373-381. SHI Yang,TU Qiuye,YAN Hongming,et al. A full states performance model for aero engine[J]. Journal of Aerospace Power,2017,32(2): 373-381. (in ChineseSHI Yang, TU Qiuye, YAN Hongming, et al. A full states performance model for aero engine[J]. Journal of Aerospace Power, 2017, 32(2): 373-381. (in Chinese) [3] WANG Zhitao,LI Yiguang,MENG Hui,et al. Method of extrapolating low speed compressor curves based on improved similarity laws[C]//Proceedings of ASME Turbo Expo 2015: Turbine Technical Conference and Exposition,Montreal,Canada: ASME,2015: 44108. [4] 郑绪生. 某型涡轴发动机起动建模技术研究[D]. 南京: 南京航空航天大学,2005. ZHENG Xusheng. Research on starting modeling technology of a turboshaft engine[D]. Nanjing: Nanjing University of Aeronautics and Astronautics,2005. (in ChineseZHENG Xusheng. Research on starting modeling technology of a turboshaft engine[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2005. (in Chinese) [5] AGRAWAL R K,YUNIS M. A generalized mathematical model to estimate gas turbine starting characteristics[J]. Journal of Engineering for Power,1982,104(1): 194-201. [6] SEXTON W R. A method to control turbofan engine staring by varying compressor surge valve bleed[D]. Blacksburg,US: Virginia Polytechnic Institute and State University,2001. [7] 郝旺,王占学,张晓博,等. 变循环发动机地面起动建模及控制规律设计方法[J]. 航空动力学报,2022,37(1): 152-164. HAO Wang,WANG Zhanxue,ZHANG Xiaobo,et al. Ground starting modeling and control law design method of variable cycle engine[J]. Journal of Aerospace Power,2022,37(1): 152-164. (in ChineseHAO Wang, WANG Zhanxue, ZHANG Xiaobo, et al. Ground starting modeling and control law design method of variable cycle engine[J]. Journal of Aerospace Power, 2022, 37(1): 152-164. (in Chinese) [8] GUAN Cong,THEOTOKATOS G,ZHOU Peilin,et al. Computational investigation of a large containership propulsion engine operation at slow steaming conditions[J]. Applied Energy,2014,130: 370-383. [9] JOHNSEN I,BULLOCK R. Aerodynamic design of axial-flow compressors: NASA SP-36[R]. Ohio,US: NASA,1979. [10] STEINKE R J. A computer code for predicting multistage axial flow compressor performance by a meanline stage stacking method: NASA-TP-2020 [R]. Ohio,US: NASA,1982. [11] KURZKE J. SmoothC 8.3 user’s manual[M]. Aachen,Germany: GasTurb GmbH,2015. [12] KURZKE J,MARIE D W. Generating compressor maps to simulate starting and windmilling[C]//Proceedings of International Symposium of Air Breathing Engine,Canberra: ISABE,2019: 24299. [13] KURZKE J. How to get component maps for aircraft gas turbine performance calculations[C]//Proceedings of International Gas Turbine and Aeroengine Congress and Exhibition. Birmingham,US: ASME,1996: 164. [14] ZACHOS P K,PACHIDIS V,CHARNLEY B,et al. Flowfield investigation of a compressor cascade at high incidence: part 1: pneumatic probe measurements[C]// Proceedings of ASME Turbo Expo: Power for Land,Sea,and Air. Orlando,US: ASME,2009: 59906. [15] ZACHOS P K,PENGUE F,PACHIDIS V,et al. Flowfield investigation of a compressor cascade at high incidence: Part 2 numerical analysis[C]// Proceedings of ASME Turbo Expo: Power for Land,Sea,and Air,Orlando,US: ASME,2009: 59908. [16] ZACHOS PAVLOS K,IOANNA A,VASSILIOS P,et al. A sub-idle compressor characteristic generation method with enhanced physical background[J]. Journal of Engineering for Gas Turbines and Power,2011,133(8): 081702. [17] 饶高,苏三买,翟向博. 指数外推法和支持向量机相结合的压气机特性扩展方法[J]. 航空动力学报,2017,32(3): 749-755. RAO Gao,SU Sanmai,ZHAI Xiangbo. Method of compressor characteristic extension combining exponent extrapolation method with support vector machine[J]. Journal of Aerospace Power,2017,32(3): 749-755. (in ChineseRAO Gao, SU Sanmai, ZHAI Xiangbo. Method of compressor characteristic extension combining exponent extrapolation method with support vector machine[J]. Journal of Aerospace Power, 2017, 32(3): 749-755. (in Chinese) [18] 杨欣毅,沈伟,刘海峰,等. 一种应用滑动最小二乘求取压气机特性的方法[J]. 航空动力学报,2009,24(8): 1741-1746. YANG Xinyi,SHEN Wei,LIU Haifeng,et al. Compressor characteristics generation method using moving least square[J]. Journal of Aerospace Power,2009,24(8): 1741-1746. (in ChineseYANG Xinyi, SHEN Wei, LIU Haifeng, et al. Compressor characteristics generation method using moving least square[J]. Journal of Aerospace Power, 2009, 24(8): 1741-1746. (in Chinese) [19] WALSH P P,FLETCHER P. Gas turbine performance[M]. 2nd ed. Malden,US: Blackwell Science,2004. [20] FERRER-VIDAL L E,PACHIDIS V,TUNSTALL R J. Generating axial compressor maps to zero speed[J]. Proceedings of the Institution of Mechanical Engineers: Part A Journal of Power and Energy,2021,235(5): 956-973. [21] SONG Fu,ZHOU Li,WANG Zhanxue,et al. Integration of high-fidelity model of forward variable area bypass injector into zero-dimensional variable cycle engine model[J]. Chinese Journal of Aeronautics,2021,34(8): 1-15. [22] KENNEDY J,EBERHART R. Particle swarm optimization[C]//Proceedings of IEEE International Conference on Neural Networks. Piscataway,US: IEEE,1995: 1942-1948. [23] GAUDET S R,DONALD GAUTHIER J E. A simple sub-idle component map extrapolation method[C]// Proceedings of ASME Turbo Expo: Power for Land,Sea,and Air,Montreal,Canada: ASME,2007: 27193. [24] 欧阳慈. 某型燃气轮机建模与自适应控制研究[D]. 南京: 南京航空航天大学,2018. OUYANG Ci. Research on modeling and adaptive control of a gas turbine[D]. Nanjing: Nanjing University of Aeronautics and Astronautics,2018. (in ChineseOUYANG Ci. Research on modeling and adaptive control of a gas turbine[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2018. (in Chinese) [25] ZACHOS P K. Gas turbine sub-idle performance modelling: altitude relight and windmilling[D]. Cranfield,UK: Cranfield University,2010 -
点击查看大图
计量
- 文章访问数: 384
- HTML浏览量: 191
- PDF量: 40
- 被引次数: 0