Operative forecasting of fuel gas consumption in gas transportation companies

In the current context of energy resource conservation and increased effi ciency of gas transportation systems, developing approaches to improve the accuracy of fuel gas consumption forecasting at compressor stations is a pressing task. This paper analyzes approaches and algorithms for forecasting the volumes of gas needed for the technological and internal needs of compressor stations during gas compression within the gas transportation organizations of Gazprom PJSC. A classifi cation of gas consumption for technological needs and losses is presented, emphasizing the importance of managing fuel gas consumption to optimize the cost of natural gas transportation.

The goal of this study is to develop an approach for operational forecasting of fuel gas consumption at compressor stations of gas transportation organizations, aimed at increasing economic effi ciency and reducing operating costs. To achieve this goal, the following tasks were undertaken: analysis of existing forecasting methods, investigation of data processing techniques for detecting errors and anomalies, and comparison of various regression models to ensure high forecast accuracy.

The study employed data cleaning and preprocessing methods, including the Isolation Forest method for anomaly detection, as well as various regression models such as multiple linear regression, RandomForestRegressor, CatBoostRegressor, and XGBoost. Data segmentation was performed using cluster analysis (KMeans), which allowed for improved model accuracy. Forecast accuracy was assessed using t-tests, F-tests, and the mean absolute percentage error (MAPE) metric.

The results of the study confi rmed the high accuracy of the proposed approach, demonstrating its potential for optimizing fuel costs in gas transportation organizations.

1. Посягин Б.С., Герке В.Г. Справочное пособие для работников диспетчерских служб газотранспортных систем. М.: ООО «Газпром экспо»; 2015. 796 с.

2. Khalikova E.R. Fuel gas costs management in gas transmission subsidiaries of PJSC Gazprom. Tekhniko-tekhnologicheskie problemy servisa. 2021;(2((56)):55–62. (In Russ.)

3. Chernyshov Y.Yu. About using of autoencoders for anomaly detection in cyber-physical systems. Bulletin of Perm University. Mathematics. Mechanics. Computer Science. 2022;(4(59)):89–94. (In Russ.). https://doi.org/10.17072/1993-0550-2022-4-89-94

4. Chesnokov M.Y. Time series anomaly detection based on DBSCAN ensembles. Iskusstvenniy Intellekt i Prinyatie Resheniy = Artificial Intelligence and Decision Making. 2018;(1):98–106. (In Russ.)

5. Chandola V., Banerjee A., Kumar V. Anomaly detection: A survey. ACM Computing Surveys. 2009;41(3):1–58. https://doi.org/10.1145/1541880.1541882

6. Dau H.A., Ciesielski V., Song A. Anomaly detection using replicator neural networks trained on examples of one class. In: Dick G. (ed.). Simulated evolution and learning. SEAL 2014. Lecture notes in computer science. Cham: Springer; 2014. Vol. 8886. P. 311–322. https://doi.org/10.1007/978-3-319-13563-2_27

7. Shatalov K.V., Kirillova A.V. Application of Student t-test for evaluation of interlaboratory comparative tests results. Measurement Standards. Reference Materials. 2016;(1):42–49. (In Russ.). https://doi.org/10.20915/2077-1177-2016-0-1-42-49

8. NIST/SEMATECH e-Handbook of Statistical Methods. Exploratory Data Analysis: Measures of Skewness and Kurtosis. https://doi.org/10.18434/M32189

9. Sallehuddin R., Shamsuddin S.M., Hashim S.Z.M. Hybridization model of linear and nonlinear time series data for forecasting. In: Second Asia inter. conf. on modelling and simulation (AMS 2008). Kuala Lumpur, Malaysia, 13–15 May, 2008. P. 597–602. https://doi.org/10.1109/AMS.2008.142

10. Breiman L. Random Forest. Machine Learning. 2001;45(1):5–32. https://doi.org/10.1023/A:1010933404324

11. Dorogush A.V., Ershov V., Gulin A. CatBoost: gradient boosting with categorical features support. 2018. https://doi.org/10.48550/arXiv.1810.11363

12. Friedman J. Greedy function approximation: a gradient boosting machine. The Annals of Statistics. 2000;29(5):1189–1232. https://doi.org/10.1214/aos/1013203451

13. Subramaniyaswamy V., Logesh R. Adaptive KNN based recommender system through mining of user preferences. Wireless Personal Communications. 2017;97(2):2229–2247. https://doi.org/10.1007/s11277-017-4605-5

14. Van der Maaten L., Hinton G. Visualizing data using t-sne. Journal of Machine Learning Research. 2008;9(2605):2579–2605.

15. Ester M., Kriegel H.P., Sander J., Xu X. A densitybased algorithm for discovering clusters in large spatial databases with noise. In: Proc. of the 2nd Intern. Conf. on Knowledge Discovery and Data Mining (KDD-96); 1996. P. 226–231.

16. Arias-Castro E., Chen G., Lerman G. Spectral clustering based on local linear approximations. Electronic Journal of Statistics. 2011;(5):1537–1587. https://doi.org/10.1214/11-ejs651

17. Druzhkov P.N., Zolotykh N.Yu., Polovinkin A.N. Parallel implementation of prediction algorithm in gradient boosting trees method. Vestnik Yuzhno-ural’skogo gosudarstvennogo universiteta. Seriya: Matematicheskoe modelirovanie i programmirovanie = Bulletin of the South Ural State University. Series: Mathematical Modeling and Programming. 2011;(37(254)):82–89. (In Russ.)