基于特征相似性迁移学习的燃气轮机动态仿真方法

doi:10.1007/s11708-020-0709-9

Frontiers in Energy

2020, Vol. 14

Issue (4): 817-835 https://doi.org/10.1007/s11708-020-0709-9

研究论文

本期目录

基于特征相似性迁移学习的燃气轮机动态仿真方法

周登极, 郝佳瑞, 黄大文, 贾星云, 张会生(

)

上海交通大学教育部动力机械与工程重点实验室，中国上海 200240

Dynamic simulation of gas turbines via feature similarity-based transfer learning

Dengji ZHOU, Jiarui HAO, Dawen HUANG, Xingyun JIA, Huisheng ZHANG(

)

Key Laboratory of Power Machinery and Engineering (Ministry of Education), Shanghai Jiao Tong University, Shanghai 200240, China

全文: PDF(3133 KB) HTML

摘要:

燃气轮机是重要的发电设备之一，其安全性和可靠性要求越来越高。随着大量可再生能源接入电网，燃气轮机在维护整个电网稳定起重要作用，参与负荷深度调峰以满足动态变化的用户需求，故其工况高度不稳定。设备启停和负荷波动是影响燃气轮机运行状态的主要因素。因此，仿真和分析燃机在不稳定工况下的动态特性，对于改进设备设计、运行和维护具有重要意义。然而，传统基于机理微分方程的动态仿真方法难以处理实际运行工况中的不确定性和噪声。尽管数据驱动建模方法在一定程度上可以缓解这一问题，但在数据不足的情况下也难以准确仿真。为解决这一问题，本文提出了一种新的迁移学习框架，将机理微分方程域的知识迁移到实际运行域，以弥补数据的不足。在机理微分方程的基础上，建立包含各类信号的强动态运行数据集，构建并训练结合编码器和解码器结构基于特征相似度的迁移学习模型，实现特征自适应知识迁移。与基准模型相比，所提出的方法仿真精度显著提高了24.6%，预测误差降低了63.6%。与其他经典转移学习模式相比,提出的方法能最好地模拟现场的燃机动态变工况过程。通过进行迁移学习模型的超参数分析,证明所提出的方法能够自适应地平衡从机理方程到实际运行的知识迁移权重。

Abstract：

Since gas turbine plays a key role in electricity power generating, the requirements on the safety and reliability of this classical thermal system are becoming gradually strict. With a large amount of renewable energy being integrated into the power grid, the request of deep peak load regulation for satisfying the varying demand of users and maintaining the stability of the whole power grid leads to more unstable working conditions of gas turbines. The startup, shutdown, and load fluctuation are dominating the operating condition of gas turbines. Hence simulating and analyzing the dynamic behavior of the engines under such instable working conditions are important in improving their design, operation, and maintenance. However, conventional dynamic simulation methods based on the physic differential equations is unable to tackle the uncertainty and noise when faced with variant real-world operations. Although data-driven simulating methods, to some extent, can mitigate the problem, it is impossible to perform simulations with insufficient data. To tackle the issue, a novel transfer learning framework is proposed to transfer the knowledge from the physics equation domain to the real-world application domain to compensate for the lack of data. A strong dynamic operating data set with steep slope signals is created based on physics equations and then a feature similarity-based learning model with an encoder and a decoder is built and trained to achieve feature adaptive knowledge transferring. The simulation accuracy is significantly increased by 24.6% and the predicting error reduced by 63.6% compared with the baseline model. Moreover, compared with the other classical transfer learning modes, the method proposed has the best simulating performance on field testing data set. Furthermore, the effect study on the hyper parameters indicates that the method proposed is able to adaptively balance the weight of learning knowledge from the physical theory domain or from the real-world operation domain.

Key words： gas turbine dynamic simulation data-driven transfer learning feature similarity

收稿日期: 2020-04-08 出版日期: 2020-12-21

通讯作者: 张会生 E-mail: zhslm@sjtu.edu.cn

Corresponding Author(s): Huisheng ZHANG

引用本文:

周登极, 郝佳瑞, 黄大文, 贾星云, 张会生. 基于特征相似性迁移学习的燃气轮机动态仿真方法[J]. Frontiers in Energy, 2020, 14(4): 817-835.
Dengji ZHOU, Jiarui HAO, Dawen HUANG, Xingyun JIA, Huisheng ZHANG. Dynamic simulation of gas turbines via feature similarity-based transfer learning. Front. Energy, 2020, 14(4): 817-835.

链接本文:

https://academic.hep.com.cn/fie/CN/10.1007/s11708-020-0709-9
https://academic.hep.com.cn/fie/CN/Y2020/V14/I4/817

Components	System equations
Gas compressor	$T 2 = T 1 (η c + π c k c / (k c − 1) − 1) η c$
	$P 2 = P 1 π c$
	$η c = M A P c, η (T 1, P 1, π c, n c)$
	$Q c = M A P c, Q (T 1, P 1, π c, n c)$
High pressure turbine	$T 34 = T 3 1 − (1 − π h t k h t / (k h t − 1)) η h t$
	$P 34 = P 3 / π h t$
	$η h t = M A P h t, η (T 3, P 3, π h t, η h t)$
	$Q h t = M A P h t, Q (T 3, P 3, π h t, η h t)$
Low pressure turbine	$T 4 = T 34 1 − (1 − π p t k p t / (k p t − 1)) η p t$
	$P 4 = P 34 / π p t$
	$η p t = M A P p t, η (T 34, P 34, π p t, η p t)$
	$Q p t = M A P p t, Q (T 34, P 34, π p t, η p t)$

Tab.1

Fig.1

Fig.2

Tab.2

Fig.3

Tab.3

Fig.4

Fig.5

Fig.6

Fig.7

Fig.8

Fig.9

Tab.4

Fig.10

Fig.11

Fig.12

Fig.13

Fig.14

Fig.15

Fig.16

No.	Similarity metrics	Formulations
0	–	–
1	Pearson correlation	$d c o r r (V 1, V 2) = ∑ i = 1 m (V 1 i − V ¯ 1) ? (V 2 i − V ¯ 2) ∑ i = 1 m (V 1 i − V ¯ 1) 2 ? ∑ i = 1 m (V 2 i − V ¯ 2) 2$
2	Manhattan distance	$d m (V 1, V 2) = ∑ i = 1 m (V 1 i − V 2 i)$
3	Euclidean distance	$d e (V 1, V 2) = ∑ i = 1 m (V 1 i − V 2 i) 2$
4	Cosine distance (ours)	$d cos ? (V 1, V 2) = ∑ i = 1 m V 1 i ? V 2 i ∑ i = 1 m (V 1 i) 2 ? ∑ i = 1 m (V 2 i) 2$

Tab.5

Fig.17

R_g	Gas constant
T	Gas temperature/K
V	Volume/m³
HV	Fuel heating value/(kJ·kg^–1)
h	Enthalpy/(kJ·mol^–1)
c_p,g	Heat capacity/(J·K^–1)
$ρ$	Combustor gas density/(kg·m^–3)
NET_sim	Neural networks trained by simulation data set
NET_real	Neural networks trained by real-world data set
d	Distance metrics
MSE	Mean square error
Error	Relative error
I	Inertia moment/(kg·m^–2)
P	Pressure/Pa
N	Rotation speed/(r·min^–1)
D	Data set
X	Input signal tensor of simulation model
Y	Output signal tensor of simulation model
V	Encoded vector
L	Latent vector
W	Parameters of neural networks
β	Weighting factor
R²	R² score for regression
t	Turbine
c	Compressor
g	Gas
in	Inlet
out	Outlet
GG	Gas generator
PT	Power turbine
1	Inlet of the gas generator
2	Inlet of the combustor
3	Inlet of the high pressure turbine
34	Outlet of the gas generator
4	Outlet of the power turbine
field	Field data
sim1	Field signal-simulation data
sim2	Transient process simulation data

1	D J Zhou, T T Wei, S X Ma, et al. Study on meta-modeling method for performance analysis of digital power plant. Journal of Energy Resources Technology, 2020, 142(4): 042005 https://doi.org/10.1115/1.4044765
2	International Energy Agency. Electricity statistics. 2018, available at website of International Energy Agency
3	H L Wang, J K He. China’s pre-2020 CO2 emission reduction potential and its influence. Frontiers in Energy, 2019, 13(3): 571–578 https://doi.org/10.1007/s11708-019-0640-0
4	Z R Chong, S H B Yang, P Babu, et al. Review of natural gas hydrates as an energy resource: prospects and challenges. Applied Energy, 2016, 162(1): 1633–1652 https://doi.org/10.1016/j.apenergy.2014.12.061
5	D J Zhou, T T Wei, D W Huang, et al. A gas path fault diagnostic model of gas turbines based on changes of blade profiles. Engineering Failure Analysis, 2020, 109: 104377 https://doi.org/10.1016/j.engfailanal.2020.104377
6	Z Ling, X Yang, Z L Li. Optimal dispatch of multi energy system using power-to-gas technology considering flexible load on user side. Frontiers in Energy, 2018, 12(4): 569–581 https://doi.org/10.1007/s11708-018-0595-6
7	J Li, G D Liu, S Zhang. Smoothing ramp events in wind farm based on dynamic programming in energy internet. Frontiers in Energy, 2018, 12(4): 550–559 https://doi.org/10.1007/s11708-018-0593-8
8	D J Zhou, Z Q Yu, H S Zhang, et al. A novel grey prognostic model based on Markov process and grey incidence analysis for equipment degradation. Energy, 2016, 109: 420–429 https://doi.org/10.1016/j.energy.2016.05.008
9	E Tsoutsanis, N Meskin, M Benammar, et al. A dynamic prognosis scheme for flexible operation of gas turbines. Applied Energy, 2016, 164(2): 686–701 https://doi.org/10.1016/j.apenergy.2015.11.104
10	D W Gao, Q Wang, F Zhang, et al. Application of AI techniques in monitoring and operation of power systems. Frontiers in Energy, 2019, 13(1): 71–85 https://doi.org/10.1007/s11708-018-0589-4
11	D T Zeng, D J Zhou, C Q Tan, et al. Research on model-based fault diagnosis for a gas turbine based on transient performance. Applied Sciences (Basel, Switzerland), 2018, 8(1): 148 https://doi.org/10.3390/app8010148
12	C Wang, Y G Li, B Y Yang. Transient performance simulation of aircraft engine integrated with fuel and control systems. Applied Thermal Engineering, 2017, 114: 1029–1037 https://doi.org/10.1016/j.applthermaleng.2016.12.036
13	A Chaibakhsh, S Amirkhani. A simulation model for transient behaviour of heavy-duty gas turbines. Applied Thermal Engineering, 2018, 132(3): 115–127 https://doi.org/10.1016/j.applthermaleng.2017.12.077
14	Z W Xie, M Su, S L Weng. Extensible object model for gas turbine engine simulation. Applied Thermal Engineering, 2001, 21(1): 111–118 https://doi.org/10.1016/S1359-4311(00)00007-7
15	E Tsoutsanis, N Meskin, M Benammar, et al. Dynamic performance simulation of an aeroderivative gas turbine using the Matlab Simulink environment. In: ASME 2013 International Mechanical Engineering Congress and Exposition, San Diego, California, USA, 2013: 56246
16	H Wang, X S Li, X Ren, et al. A thermodynamic-cycle performance analysis method and application on a three-shaft gas turbine. Applied Thermal Engineering, 2017, 127(12): 465–472 https://doi.org/10.1016/j.applthermaleng.2017.08.061
17	M Badami, M G Ferrero, A Portoraro. Dynamic parsimonious model and experimental validation of a gas microturbine at part-load conditions. Applied Thermal Engineering, 2015, 75(1): 14–23 https://doi.org/10.1016/j.applthermaleng.2014.10.047
18	A Mehrpanahi, G Payganeh, M Arbabtafti. Dynamic modeling of an industrial gas turbine in loading and unloading conditions using a gray box method. Energy, 2017, 120(2): 1012–1024 https://doi.org/10.1016/j.energy.2016.12.012
19	H Asgari, X Q Chen, M Morini, et al. NARX models for simulation of the start-up operation of a single-shaft gas turbine. Applied Thermal Engineering, 2016, 93(1): 368–376 https://doi.org/10.1016/j.applthermaleng.2015.09.074
20	H Nikpey, M Assadi, P Breuhaus. Development of an optimized artificial neural network model for combined heat and power micro gas turbines. Applied Energy, 2013, 108(8): 137–148 https://doi.org/10.1016/j.apenergy.2013.03.016
21	E Tsoutsanis, N Meskin. Derivative-driven window-based regression method for gas turbine performance prognostics. Energy, 2017, 128(6): 302–311 https://doi.org/10.1016/j.energy.2017.04.006
22	T Baklacioglu, O Turan, H Aydin. Dynamic modeling of exergy efficiency of turboprop engine components using hybrid genetic algorithm-artificial neural networks. Energy, 2015, 86(6): 709–721 https://doi.org/10.1016/j.energy.2015.04.025
23	S L Weng, C H Gu, Y W Weng. Energy internet technology: modeling, optimization and dispatch of integrated energy systems. Frontiers in Energy, 2018, 12(4): 481–483 https://doi.org/10.1007/s11708-018-0604-9
24	S S Zhong, S Fu, L Lin. A novel gas turbine fault diagnosis method based on transfer learning with CNN. Measurement, 2019, 137: 435–453 https://doi.org/10.1016/j.measurement.2019.01.022
25	D J Zhou, Q B Yao, H Wu, et al. Fault diagnosis of gas turbine based on partly interpretable convolutional neural networks. Energy, 2020, 200: 117467 https://doi.org/10.1016/j.energy.2020.117467
26	S X Tang, H L Tang, M Chen. Transfer-learning based gas path analysis method for gas turbines. Applied Thermal Engineering, 2019, 155: 1–13 https://doi.org/10.1016/j.applthermaleng.2019.03.156
27	M Klenk, K Forbus. Analogical model formulation for transfer learning in AP Physics. Artificial Intelligence, 2009, 173(18): 1615–1638 https://doi.org/10.1016/j.artint.2009.09.003
28	Z H Jiang, Y M Lee. Deep transfer learning for thermal dynamics modeling in smart buildings. In: 2019 IEEE International Conference on Big Data (Big Data), Los Angeles, CA, USA, 2019: 2033–2037
29	Y Yao, G Doretto. Boosting for transfer learning with multiple sources. In: 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Francisco, CA, USA, 2010: 1855–1862
30	D J Zhou, Z Q Yu, H S Zhang, et al. A novel grey prognostic model based on Markov process and grey incidence analysis for equipment degradation. Energy, 2016, 109: 420–429 https://doi.org/10.1016/j.energy.2016.05.008
31	S X Ma, S N Sun, H Wu, et al. Decoupling optimization of integrated energy system based on energy quality character. Frontiers in Energy, 2018, 12(4): 540–549 https://doi.org/10.1007/s11708-018-0597-4
32	W A Little. The existence of persistent states in the brain. Mathematical Biosciences, 1974, 19(1–2): 101–120 https://doi.org/10.1016/0025-5564(74)90031-5
33	F A Gers, J Schmidhuber, F Cummins. Learning to forget: continual prediction with LSTM. In: 9th International Conference on Artificial Neural Networks, Technical report, 1999
34	S J Pan, Q Yang. A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering, 2010, 22(10): 1345–1359 https://doi.org/10.1109/TKDE.2009.191
35	F Z Zhuang, Z Y Qi, K Y Duan, et al. A comprehensive survey on transfer learning. arXiv preprint, 2019: 1911.02685
36	D E Rumelhart, G E Hinton, R J Williams. Learning internal representations by error-propagation. Parallel Distributed Processing: Explorations in the Microstructure of Cognition, 1986, 1: 318–362
37	D P Kingma, J Ba. Adam: a method for stochastic optimization. arXiv preprint, 2014: 1412.6980

Viewed

Full text

Abstract

Cited

Shared

Discussed