Performance prediction of the centrifugal compressor using a long short-term memory-based model with transfer learning and similarity conversion

Abstract

Centrifugal compressors are commonly employed for pressure compensation in long-distance natural gas pipelines. Accurate prediction of compressor performance is essential for safety and economic considerations. However, due to the invariance of process requirements over time, operational data is usually insufficient to establish a comprehensive model that covers varying operating conditions. This paper presents a prediction method based on the long short-term memory (LSTM) network and transfer learning technique, integrating similarity laws to expand the feature domain and address the issue of limited data. In the case study, the constructed TransLSTM model, a fusion of transfer learning and LSTM, is compared with pure LSTM, Gaussian Process (GP), Gradient Boosting Decision Tree (GBDT), Random Forest (RF), and Support Vector Machine (SVM) models. Results indicate that TransLSTM achieves the highest prediction accuracy, with deviations from actual values of only 2.9% for power and 0.971% for compressor ratio. Verification through comparisons between the training and test sets demonstrates its excellent generalization ability, which is particularly crucial when operational data spans a limited range of working conditions. Additionally, its stability is significantly superior to that of benchmark models.

Keywords

Compressor performance / Transfer learning / Long short-term memory / Similarity laws / Feature enriching

References

[1] Lu H, Song K, Shan Y, Cheng Y.F.. Quantifying greenhouse gas emission risks from natural gas pipeline incidents. Nexus, 2025, 2: 4-100097

[2] Wang Y, Wang K, Lin F, Nie C, Abraham E. Performance analysis and improvement of a high flow coefficient centrifugal compressor. Sci China Technol Sci, 2014, 8: 1647-1657

[3] Zamiri A, Lee BJ, Taek Chun J. Numerical evaluation of transient flow characteristics in a transonic centrifugal compressor with vaned diffuser. Aerosp Sci Technol, 2017, 70: 244-256

[4] Niazi S, Stein A, Sankar L (1998) Development and application of a CFD solver to the simulation of centrifugal compressors. https://doi.org/10.2514/6.1998-934

[5] Le Sausse P, Fabrie P, Arnou D, Clunet F. CFD comparison with centrifugal compressor measurements on a wide operating range. EPJ Web Conf, 2013, 45: 01059

[6] Kalinkevych K. Shcherbakov O (2013) Numerical modeling of the flow in a vaneless diffuser of centrifugal compressor stage. ISRN Mech Eng, 2013, 1: 602384

[7] Oh HW, Yoon ES, Chung MK. An optimum set of loss models for performance prediction of centrifugal compressors. P I Mech Eng A-J Pow, 1997, 211: 331-338

[8] Cicciotti M, Xenos DP, Bouaswaig AE, Thornhill NF, Martinez-Botas RF. Physical modelling of industrial multistage centrifugal compressors for monitoring and simulation. Proc Inst Mech Eng C J Mech Eng Sci, 2015, 229(18): 3433-3448

[9] Kang SH, Ryu C, Ko HS. Analysis of performance for centrifugal steam compressor. J Mech Sci Technol, 2016, 30(12): 5521-5527

[10] Gutiérrez Velásquez EI (2017) Determination of a suitable set of loss models for centrifugal compressor performance prediction. Chin J Aeronaut 30(5):1644–1650. https://doi.org/10.1016/j.cja.2017.08.002

[11] Persky R, Sauret E. Loss models for on and off-design performance of radial inflow turbomachinery. Appl Therm Eng, 2019, 150: 1066-1077

[12] Japikse D (1985) Assessment of single- and two-zone modeling of centrifugal compressors, studies in component performance: Part 3. In: Proceedings of the ASME 1985 International Gas Turbine Conference and Exhibit 1: V001T03A023. https://doi.org/10.1115/85-GT-73

[13] Rohne KH, Banzhaf M. Investigation of the flow at the exit of an unshrouded centrifugal impeller and comparison with the “classical” jet-wake theory. J Turbomach, 1991, 113(4): 654-659

[14] Oh HW, Yoon ES, Chung MK. Systematic two-zone modelling for performance prediction of centrifugal compressors. P I Mech Eng A-J Pow, 2002, 216: 75-87

[15] Britton I, Gauthier JED. Performance prediction of centrifugal impellers using a two-zone model. Proceedings of the ASME Turbo Expo, 2008, 6: 1695-1704

[16] Javed A, Pecnik R, Olivero M, Buijtenen JPV. Effects of manufacturing noise on microturbine centrifugal impeller performance. Proceedings of the ASME Turbo Expo, 2012, 5: 847-857

[17] Javed A, Olivero M, Pecnik R, van Buijtenen JPV. Performance analysis of a microturbine centrifugal compressor from a manufacturing perspective. Proceedings of the ASME Turbo Expo, 2011, 7: 2155-2166

[18] Wu S, Ríos-Mercado RZ, Boyd EA, Scott LR. Model relaxations for the fuel cost minimization of steady-state gas pipeline networks. Math Comput Model, 2000, 31(2–3): 197-220

[19] Leufvén O, Erisksson L. Surge and choke capable compressor model. IFAC Proc Vol, 2011, 44(1): 10653-10658

[20] Leufvén O, Erisksson L. A surge and choke capable compressor flow model-validation and extrapolation capability. Control Eng Pract, 2013, 21(12): 1871-1883

[21] Leufvén O, Erisksson L. Measurement, analysis and modeling of centrifugal compressor flow for low pressure ratios. Int J Engine Res, 2014, 17(2): 153-168

[22] Llamas X, Eriksson L. Parameterizing compact and extensible compressor models using orthogonal distance minimization. J Eng Gas Turbines Power, 2017, 139(1): 012601

[23] Llamas X, Eriksson L. Control-oriented compressor model with adiabatic efficiency extrapolation. SAE Int J Engines, 2017, 10: 1903-1916

[24] Casey M, Robinson C. A method to estimate the performance map of a centrifugal compressor stage. J Turbomach, 2013, 135(2): 021034

[25] Japikse D. Centrifugal compressor design and performance, 1996, Wilder, VT, Concepts ETI Inc

[26] Korpela SA. Principles of turbomachinery, 2011, Hoboken, New Jersey, Jonh Wiley & Sons Inc

[27] Balje OE, Japikse D. Turbomachines, a guide to design, selection and theory. J Fluids Eng, 1981, 1034644

[28] Lakshminarayana B. Fluid dynamics and heat transfer of turbomachinery, 1996, New York, John Wiley & Sons Inc

[29] Roberts SK, Sjolander SA. Effect of the specific heat ratio on the aerodynamic performance of turbomachinery. J Eng Gas Turbines Power, 2005, 127: 773-780

[30] Den G, Malyshev A. Modeling of the flow parts of stationary centrifugal compressor machines for compression of real gases. Gasdynamic tests and processing of their results. I. Theoretical calculations. J Eng Phys Thermophys, 2002, 75: 1076-1085

[31] Dufour G, Carbonneau X, Cazalbou JB, Chassaing P. Practical use of similarity and scaling laws for centrifugal com-pressor design. Proceedings of the ASME Turbo Expo, 2006, 6: 1131-1140

[32] Wu X, Li Y (2007) Computationally efficient data-driven surge map modeling for centrifugal air compressors. In: Proceedings of the American Control Conference: 810–815. https://doi.org/10.1109/ACC.2007.4283023

[33] Ghorbanian K, Gholamrezaei M. An artificial neural network approach to compressor performance prediction. Appl Energy, 2009, 86(7–8): 1210-1221

[34] Huang S. Research and application of wavelet neural networks of particle swarm optimization algorithm in the performance prediction of centrifugal compressor. Adv Mater Res, 2011, 187: 271-276

[35] Chu F, Wang F, Wang X, Zhang S. Performance modeling of centrifugal compressor using kernel partial least squares. Appl Therm Eng, 2012, 44: 90-99

[36] Wang X, Zhang K (2013) Modeling of centrifugal compressor using RBF neural network based on cooperative clustering. In: Proceedings of 2013 2nd International Conference on Measurement, Information and Control, pp 1362–1366. https://doi.org/10.1109/MIC.2013.6758212

[37] Fei J, Zhao N, Shi Y, Feng Y, Wang Z. Compressor performance prediction using a novel feed-forward neural network based on Gaussian kernel function. Adv Mech Eng, 2016, 8(1): 1-14

[38] Bao HH, Fu JQ, Sun XL, Sun C, Kuang JB, Wang X, Liu JP. Performance prediction of the centrifugal air compressor for fuel cells considering degradation characteristics based on the hierarchical evolutionary model. Therm Sci Eng Prog, 2023, 46: 102212

[39] Peng Q, Bao R, Li J, Ren J, Tang Jet al.. Centrifugal compressor performance prediction and dynamic simulation of natural gas hydrogen blended. Int J Hydrogen Energy, 2024, 52: 872-893

[40] Chu F, Dai B, Dai W, Jia R, Ma X, Wang F. Rapid modeling method for performance prediction of centrifugal compressor based on model migration and SVM. IEEE Access, 2017, 5: 21488-21496

[41] Chu F, Dai B, Lu N, Ma X, Wang F. Improved fast model migration method for centrifugal compressor based on bayesian algorithm and Gaussian process model. Sci China Technol Sci, 2018, 61: 1950-1958

[42] Luo R, Li Y, Guo H, Wang Q, Wang X. Cross-operating-condition fault diagnosis of a small module reactor based on CNN-LSTM transfer learning with limited data. Energy, 2024, 313: 133901

[43] Yu Y, Chen Y, Huang S, Wang R, Wu Y, Zhou H, Li X, Tan Z. Enhancing the effluent prediction accuracy with insufficient data based on transfer learning and LSTM algorithm in WWTPs. J Water Process Eng, 2024, 62: 105267

[44] Zhang Z, Yang H, Yang X. A transfer learning-based LSTM for traffic flow prediction with missing data. J Transp Eng A-Syst, 2023, 149(10): 04023095

[45] Gers FA, Schmidhuber J, Cummins F (1999) Learning to forget: continual prediction with LSTM. In: 1999 Ninth International Conference on Artificial Neural Networks ICANN 99:850–855. https://doi.org/10.1049/cp:19991218

[46] Pan SJ, Yang Q. A survey on transfer learning. IEEE Trans Knowl Data Eng, 2010, 22: 1345-1359

[47] Hochreiter S, Schmidhuber J. Long short-term memory. Neural Comput, 1997, 9: 1735-1780

[48] Graves A, Mohamed A, Hinton G (2013) Speech recognition with deep recurrent neural networks. In: 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, pp 6645–6649. https://doi.org/10.1109/ICASSP.2013.6638947

[49] How DNT, Loo CK, Sahari KSM. Behavior recognition for humanoid robots using long short-term memory. Int J Adv Robot Syst, 2016, 13(6): 1-14

[50] Buckingham E. On physically similar systems: illustration of the use of dimensional equations. Phys Rev, 1914, 4: 345-376

[51] Kline SJ, Radbill JR. Similitude and approximation theory. J Appl Mech, 1966, 33(1): 238

[52] La L, Guo Q, Cao Q, Wang Y. Transfer learning with reasonable boosting strategy. Neural Comput Appl, 2014, 24: 807-816

[53] Oquab M, Bottou L, Laptev I, Sivic J. Learning and transferring mid-level image representations using convolutional neural networks. IEEE Conference on Computer Vision and Pattern Recognition, 2014, 2014: 1717-1724

[54] Yosinski J, Clune J, Bengio Y, Lipson H (2014) How transferable are features in deep neural networks? In: Proceedings of the 28th Conference on Neural Information Processing Systems 27:3320–3328. https://doi.org/10.48550/arXiv.1411.1792

[55] Ma X, Mei X, Wu WQ, Wu XX, Zeng B. A novel fractional time delayed grey model with Grey Wolf Optimizer and its applications in forecasting the natural gas and coal consumption in Chongqing China. Energy, 2019, 178: 487-507

[56] Wang J, Yang W, Du P, Niu T. A novel hybrid forecasting system of wind speed based on a newly developed mul-ti-objective sine cosine algorithm. Energy Convers Manage, 2018, 163: 134-150

RIGHTS & PERMISSIONS

The Author(s)