摘要
为保证不同工况下重型燃气轮机发电机组的高效、稳定运行,需要对其轴流式压气机性能进行高精度建模分析。本文以某F级重型燃气轮机为研究对象,有效结合实测数据和基元叶栅法理论计算值,采用随机森林方法,建立其压气机高精度特性计算模型,从而分析不同工况下的压气机性能。结果表明:模型计算的总压比和多变效率的平均相对误差分别为-0.13%和0.04%;最大相对误差分别为2.11%和1.90%。现有文献的模型最大相对误差分别为3.1%和4.4%,本文方法得到的最大误差仅分别为其68.06%和43.18%。
In order to keep the heavy-duty gas turbine power unit to work efficiently and stably under different conditions, we need to build the high accurate performance model of the axial compressor. This paper aims at the axial compressor of a type of F level gas turbines. Based on measured data and calculated data through cascade element method, the random forest method is applied to build the characteristics calculation model with high accuracy. Then the performance of compressor can be analyzed under the off-design condition. The results show that the average error of the total pressure ratio and polytropic efficiency obtained through the calculation model is-0.13% and 0.04%, respectively. While the maximum relative error of the total pressure ratio and polytropic efficiency is 2.11% and 1.90%, respectively. The maximum relative error of the existing literature are 3.1% and 4.4%, respectively. The maximum error obtained by this method are only 68.06% and 43.18% of the former respectively.
引文
[1] AGRAWAL R K,YUNIS M.A generalized mathematical model to estimate gas turbine starting characteristics[J].Journalof Engineering of Power,1982,104(1):194-201.
[2] CHANGDUK K, SEONHEE K, JAYOUNG K. Component map generation of a gas turbine using genetic algorithms[J].Journal of Engineering for Gas Turbines and Power,2006,128(1):92-96.
[3] KONG C,KI J,KANG M.A new scaling method forcom ponent maps of gas turbine using system identification[J].Journal of Engineering for Gas Turbines and Power, 2003, 125(1):979-985.
[4] 张学镭. 焦炉煤气合成甲醇和发电系统关键技术研究[D]. 河北保定: 华北电力大学, 2007.
[5] MA W T, LIU Y W, SU M. Multi-stage axial flow compressors characteristics estimation based on system identification[J]. Energy Conversion and Management, 2008,49(2):143-150.
[6] 马文通,刘永文,苏明.多级轴流式压气机平均模型级特性获取办法[J].上海交通大学学报,2007,41(12):1954-1958.
[7] BREIMAN L. Random forests[J]. Machine Learning, 2001, 45(1):5-32.
[8] 于晓虹, 楼文高. 基于随机森林的P2P网贷信用风险评价、预警与实证研究[J]. 金融理论与实践, 2016(2):53-58.
[9] MICHAEL K. Peer to peer lending: auctioning microcredits over the internet. proceedings of the international conference on information systems [J].Technology and Management, 2008(2):1-8.
[10] 李欣海. 随机森林模型在分类与回归分析中的应用[J]. 应用昆虫学报, 2013,50(4):1190-1197.
[11] GOBRAN M H. Off-design performance of solar centaur 40 gas turbine using simulink[J].Ain Shams Engineering Journal,2013,4(2):285-298.
[12] 邱超, 宋华芬. 变几何燃气轮机性能的计算分析[J]. 热能动力工程, 2010,25(4):377-380.
[13] WU X, LI Y. Self-learning based centrifugal compressor surge mapping with computationally efficient adaptive asymmetric support vector machine[J]. Journal of Dynamic Systems Measurement & Control, 2012,134(5):51008.
[14] GE B, ZANG S S, GUO P Q. Experimental study on turbulent structure of humid air flame in a bluff-body burner[J]. 热科学学报, 2009,18(2):185-192.
[15] ZHANG X W, WANG Y R, XU K N. Flutter prediction in turbomachinery with energy method[J]. Proceedings of the Institution of Mechanical Engineers Part G Journal of Aerospace Engineering, 2011,225(9):995-1002.
[16] 陈靖华, 吴哲明, 屠宝锋. 组合压气机性能和气动稳定性计算分析[J]. 航空计算技术, 2014(6):41-43.