An Artificial Neural Network Model for Prediction of the Operational Parameters of Centrifugal Compressors: An Alternative Comparison Method for Regression

Document Type : Original Research Articles

Authors
seyed hossain ebrahimi 1
Ahmad afshari 2
1 industry engineering ,shomal university,amol,iran
2 industrial engineering,MehrAlborz university,tehran,iran
10.22059/jsciences.2020.297045.1007495
Abstract
Nowadays, centrifugal compressors are commonly used in the oil and gas industry, particularly in the energy transmission facilities just like a gas pipeline stations. Therefore, these machines with different operational circumstances and thermodynamic characteristics are to be exploited according to the operational necessities. Generally, the most important operational parameters of a gas pipeline booster station includes the compressor's input and output pressures, input and output temperatures and also the flow rate passing from the compressors. Different values of those parameters related to every point of operational conditions will exactly affect on the compressor poly-tropic efficiency and their driver fuel consumption. Although, calculating of the poly tropic efficiency and fuel consumption using the existing thermodynamic relations, would need to apply rather awkward equations for each operating point. In this research, a feed forward perceptron artificial neural network is presented to predict the output operational conditions. The network would be trained at least in two scenarios applying by practical data in the neuro solution software version.5 using the Levenberg-Marquadt algorithm and the optimum model is experimentally selected according to R2, MSE and NMSE.
Keywords
Centrifugal compressor
Artificial neural network
Ridge regression
Performance prediction
Pipeline gas booster station
  1. Lauden K.C. and Lauden J.P., Management information system, 13th edition, published by Pearson Education, 258 (2014).
  2. Hagan M., Neural Network DeSign,2nd edition, ebook: 1-3.
  3. Moraal P. and Kolmanovsky I., Turbocharger modeling for automotive control application. SAE Trans,108:1324–38 (1999).

4.Ghorbanian K. and Gholamrezaei M., An artificial neural network approach to compressor  performance prediction. Applied Energy. JAE, 86:1210–1221(2009).

  1. Bao C., Ouyang M. and Yi B., Modeling and optimization of the air system in polymer exchange membrane fuel cell system. Journal of Power Sources. JPS, 156 (2):232–43(2006).

6.Yu Y., Chen L., Sun F. and Wu C., Neural network based analysis and prediction of a compressor’s characteristic performance map. Applied Energy. JAE, 84(1):48–55(2007).

7.Torabian. and Karimian., Prediction of compressor’s map using the artificial neural network(Into Persian). First conference of developing of civil engineering ,architecture and mechanic, iran. DCEAEM01, 1: 111-117(2014).

8.Sanaye S., Dehghandokht M., Mohammadbeigi H. and Bahrami S., Modeling of rotary vane compressor applying artificial neural network. International Journal of Refrigration. Int. J. Refrig, 34: 764-772(2011).

9.Soo-Yong C., Kook-Young A., Young-Duk L. and Young-Cheol K., Optimal Design of a Centrifugal Compressor Impeller Using Evolutionary Algorithms.Mathematical Problems in Engineering. Math Probl Eng, 2012:1-22 (2012).

10.Shaojun L. and Feng L., Prediction of Cracking Gas Compressor Performance and Its Application in Process Optimization. Process systems engineering. Chinese Journal of Chemical Engineering. CJCE, 20:1089-1093(2012).

  1. Palmé T., Waniczek P., Hönen H., Assadi M. and Jeschke P., Compressor Map Prediction by Neural Networks. Journal of Energy and Power Engineering. JEPE, 6:1651-1662 (2012).

12.Chen P., Chang H. and Armada H., A Study of Using Artificial Neural Network in a Non-linear Centrifugal Compressor System. International Journal on Computer Science and Engineering. IJCSE, 4:1890-1896 (2012).

13.Yang L., Zhao L., Zhang, C. and Gua B. ,Loss-efficiency model of single and variable-speed compressors using neural networks.International Journal of Refrigretaion. Int.J.Refrig, 32:1423-1432 (2009).

  1. Vilalta P.C., Wan H. and Patnaik S.S., Centrifugal compressor performance prediction using gaussian process regression and artificial neural networks. Proceedings of the ASME 2019. International Mechanical Engineering Congress & Exposition. IMECE, 2019. 11936:1-10 (2019).
  2. Fei J., Zhao N., Shi Y., Feng Y. and Wang Z., Compressor performance prediction using a novel feed-forward neural network based on Gaussian kernel function. Advances in Mechanical Engineering. AIME, 8: 1-14 (2016).
  3. Li X., Yang C., Wang Y., Wang H., Zu X., Sun Y. and Hu, S., Compressor map regression modeling based on partial least squares. Royal Society Open Science. RSOS, 5:1-11 (2018).

17.Walker E., Ridge Regression as an Alternative to Ordinary Least Squares: Improving Prediction Accuracy and the Interpretation of Beta Weights., Association for Institutional Research Enhancing knowledge. Expanding networks. Professional Development, Informational Resources & Networking. AIR, 92:1-12 (2004).

  1. Sana M. and Eyup C., Efficient Choice of Biasing Constant for Ridge Regression. International Journal of Contemporary Mathematical Sciences. IJCMS, 3:527–536 (2008).
  2. Green W., Econometric Analysis, 7th edition,Mac-Millan, New York University:1039 (2012).

20.Walker E., Detection of Collinearity–Influential observations. Journal of Communications in Statistics, Theory and Methods. JCS-TM , 18:1675–1690 (1989).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Nomenclature

A,B,C,D,E,F

Regression coefficients

Pout

Gas station outlet pressur

Tsuc

Compressor input temperature

FC

Fuel consumption

Psuc

Compressor suction pressure

Z

Compressibility coefficient

HP

Compressor poly tropic head

Pt.rpm

Lp turbine speed

α

Power corrective coefficient

HV

Heating value

q, Q

Flow rate

ηm

Lp tubine mechanical efficiency

k

Poly tropic index

Qac

Real flow rate

ηp

Compressor poly tropic efficiency

MSE

Mean of standard error

Qst

Standard flow rate

ηt

Hp turbine thermal efficiency

MAE

Maximum absolute error

R2

Regression R square

λmax

Maximum eigenvalue

NMSE

Normal mean of standard error

S

Compressor speed

λmin

Minimum eigenvalue

no

Number of compressor

Tamb

Ambient temperature

 

 

Pdis

Compressor output pressure

Tdis Tout

Compressor output temperature

 

 

Pin

Gas station inlet pressure

Tin

Gas station inlet temperature

 

 

 

21.Vinod H. and Ullah A., Recent Advances in Regression Models. International Journal of Forecasting. IJF, 2:246-361 (1981).

22.Hoerl A. and Kennard R., Ridge Regression: Biased Estimation for non orthogonal Problems. Journal of Statistics for the Physical, Chemical, and Engineering Sciences. Technometrics, 42:80-86 (2000).

  1. Hanlon, P.C., Compressors Handbooks manuals, McGraw-Hill publication. 4:22 (2001).
Journal of Sciences, Islamic Republic of Iran
Volume 31, Issue 3
Summer 2020
Pages 259-275