针对传统BP神经网络存在的学习速度慢，容易陷入局部极小的问题，为了提高诊断效率和非线性拟合能力，提出利用粒子群算法、遗传算法和布谷鸟算法优化BP神经网络对燃气轮机积垢性能预测，将各算法所得误差和拟合度进行对比。结果表明：粒子群算法优化BP神经网络的非线性拟合度为0.9344，优于其他算法，相较于其他算法具有更低的绝对误差平均值，有效避免了局部最优情况。
Aiming at the slow learning speed of traditional BP neural network and easy to fall into local minima, in order to improve the diagnosis efficiency and nonlinear fitting ability, it is proposed to use particle swarm algorithm, genetic algorithm and cuckoo algorithm to optimize the BP neural network for gas turbine fouling Performance prediction, comparing the error and fit of each algorithm. The results show that the non-linear fit of the BP neural network optimized by the particle swarm algorithm is0.9344, which is better than other algorithms, has a lower absolute error average value than other algorithms, and effectively avoids the local optimal situation.
2021,43(10): 108-112 收稿日期：2020-07-24
DOI：10.3404/j.issn.1672-7649.2021.10.022
分类号：TK47
基金项目：湖北省自然科学基金资助项目(2017CFB49)
作者简介：聂勇恒(1996-),男,主要从事基于大数据下燃气轮机性能预测与故障诊断研究
参考文献：
[1] 谢春玲, 戴景民. 燃气轮机故障诊断技术研究综述与展望[J]. 汽轮机技术, 2010, 52(1): 1–3
XIE Chunling, DAI Jingmin. Review and prospect of gas turbine fault diagnosis technology[J]. Steam Turbine Technology, 2010, 52(1): 1–3
[2] KURZ R, BRUN K. Degradationin gas turbine systems[J]. Transactions of the ASME, 2001, 123(1): 70–77
[3] YOON J E, LEE J J, KIM T S, et al. Analysis of performance deterioration of a micro gas turbine and the use of neural networkfor predicting deteriorated com-ponent characteristics[J]. Journal of Mechanical Science and Technology, 2008, 22(12): 2516–2525
[4] 谷思宇, 李刚. GE LM2500+SAC型燃气轮机压气机积垢判据的确定[J]. 油气储运, 2016, 35(7): 763–767
[5] 熊蕾. GA-BP神经网络在船舶供电网络评价中的应用[J]. 舰船科学技术, 2020, 42(16): 109–111
[6] 赵丽娟, 金忠峰. 基于GA-BP神经网络薄煤层采煤机装煤性能预测[J]. 机械强度, 2018, 40(3): 620–625
[7] 谢劭峰, 赵云, 李国弘, 等. GA-BP神经网络的GPS可降水量预测[J]. 测绘科学, 2020, 45(3): 33–38
[8] 刘豹, 胡代平. 神经网络在预测中的一些应用研究[J]. 系统工程学报, 1999, 14(4): 338–344
[9] 程向军. 神经网络原理及其应用[M]. 北京: 国防工业出版社, 1995: 1−300
[10] YANG X S, Deb S. Cuckoo search via levy flight [C]// Proceedings of World Congress on Nature&Biologically Inspired Computing. India, Washington: IEEE Publications, 2009: 210−214.
[11] BENVIDI A, ABBASI S, GHARAGHANI S, et al. Spectrophotometric determination of synthetic colorants using PSO–GA-ANN[J]. Food Chemistry, 2017, 220(apr.1): 377–384
[12] CHOUIKHI N, AMMAR B, ROKBANI N, et al. PSO-based analysis of Echo State Network parameters for time series forecasting[J]. Applied Soft Computing, 2017, 55(C): 211–225
[13] 王泓霖, 李伟, 许强, 等. 基于PSO-BP神经网络的HRRP目标识别[J]. 火力与指挥控制, 2019, 44(12): 39–44
[14] 房友龙, 刘东风, 刘永葆, 等. 燃气轮机结垢性能退化特征提取和剩余寿命预测[J]. 海军工程大学学报, 2018, 30(2): 100–104
