Elektrik Güç Dağıtımında Akıllı Sayaç Verileri için Anomali Tespiti ve Tahminleme

Year 2022, , 72 - 85, 28.02.2023

Serhat Yarat , Zeynep Orman

https://doi.org/10.54047/bibted.1224628

Abstract

Nüfus yoğunluğu ve ekonomik büyümenin etkisiyle enerji talebi hızla artmaktadır. Bu talep karşısında enerji ve elektrik şebekeleri daha fazla zorlukla karşı karşıya gelmektedir. Enerji tüketiminin sıkı bir şekilde izlenmesi ve kontrol altında tutulması önem arz etmektedir. Enerji dağılımını düşündüğümüzde akıllı sayaçlar bu enerjinin kontrolünde kilometre taşı rolü oynamaktadır. Enerji tüketim ölçümlerinin yapıldığı sayaçlarda meydana gelebilecek herhangi bir elektrik kesintisi, bir hata veya yanlış ölçüm, dağıtım şirketlerinden son kullanıcılara kadar birçok tarafı etkilemektedir. Enerji sektöründeki bu tür anomalilerin tespiti için gerçekleştirilen veri analitiği çalışmaları ve büyük veri teknolojileri, sensörlerden ve sayaçlardan toplanan zaman serisi verilerini gerçek zamanlı veya toplu olarak analiz ederek verimliliği ve tasarrufu arttırmayı amaçlayan net ve eyleme geçirilebilir çıktılar üretmede önemli rol oynamaktadır. Bu çalışmada, akıllı elektrik sayaçları ile ölçülen aylık tüketim değerlerine dayalı olarak enerji tüketimindeki olası anomalilerin tespit edilmesi ve farklı makine öğrenmesi yöntemleri kullanılarak gelecek tüketiminin tahmin edilmesi amaçlanmıştır. Sonuç olarak; enerji sektöründe genel aydınlatma sayaçları üzerinde yapılan uygulamalarda İzolasyon Ormanı (Isolation Forest-IF), Yerel Aykırı Değer Faktörü (Local Outlier Factor-LOF) ve FbProphet algoritmalarının anomali tespitinde olası uç anomali noktalarını başarılı bir şekilde tespit edebildiği ve FbProphet algoritmasının XGBoost algoritmasına göre sayaç verileri üzerinde zaman serileri ile yapılan tahminlemelerde ortalama olarak daha iyi sonuç verdiği tespit edilmiştir.

Keywords

Anomali Tespiti, Tahminleme, Zaman Serileri, Elektrik Güç Dağıtımı, Akıllı Sayaç, Anomaly Detection, Electrical Power Distribution

Anomaly Detection and Prediction for Smart Meter Data in Electrical Power Distribution

Abstract

With rapidly increasing energy demand as a result of increasing population density and economic growth, energy and electricity grids are now facing more challenges. In the face of this demand, energy consumption should be kept under strict monitoring and control. When we consider energy distribution, smart meters play a milestone role in the control of this energy. Any error or an incorrect measurement in the meters in which energy consumption measurements are made affect many stakeholders, from the distribution company to the end users. Identifying such anomalies in the energy sector, data analytics studies, and big data technologies play an important role to produce solutions. In this study, it is aimed to detect possible anomalies in energy consumption and forecasting consumption using different machine learning methods, based on monthly consumption values measured by smart meters. As a result; the experiments and observations on general lighting meters showed that Isolation Forest (IF), Local Outlier Factor (LOF), and FbProphet were successful in detecting the potential extreme anomaly points. Additionally, it has been found that when using the FbProphet and XGBoost algorithms to forecast data from time series, FbProphet outperforms XGBoost.

Keywords

Anomaly Detection, Forecasting, Time Series, Electrical Power Distribution, Smart Meters

References

• Alfares, H.K. ve Nazeeruddin, M. (2010). Electric Load Forecasting: Literature Survey and Classification of Methods, International Journal of Systems Science, 33, 23–34. Doi: 10.1080/00207720110067421
• Al-Ghaili, M.A., Ibrahim, Z.A., Hairi, S.A.S.S., Rahim, F.A., Baskaran, H., Ariffin, N.A.M., ve Kasim, H. (2021). A Review of Anomaly Detection Techniques in Advanced Metering Infrastructure, Bulletin of Electrical Engineering and Informatics, 10(1), 266-273. Doi: 10.11591/eei.v10i1.2026
• Breunig, M.M., Kriegel, H., Ng, R.T., ve Sander, J. (2000). Lof: Identifying Density-Based Local Outliers, Acm Sıgmod Record, 29(2), 93–104. https://doi.org/10.1145/335191.335388
• Bansal, A., Rompikuntla, S.K., Gopinadhan, J., Kaur, A., ve Kazi, Z.A. (2015). Energy Consumption Forecasting For Smart Meters, ArXiv abs/1512.05979, Cornell university. Elsevier.
• Braei, M. (2019). Anomaly Detection of Time series: A Comparison of Statistical vs Classical Machine Learning vs Deep Learning Approaches, Master Thesis in The Department of Computer Science, Kauschke, Technical University Of Darmstadt. Doi: 10.13140/rg.2.2.17687.80801
• Braei M., ve Wagner S. (2020). Anomaly Detection In Univariate Time-Series: A Survey on The State-Of-The-Art, Cornell University, https://doi.org/10.48550/arxiv.2004.00433
• Borges, H., Akbarinia, R., ve Masseglia, F. (2021). Anomaly Detection in Time Series, Transactions On Large-Scale Data-and Knowledge-Centered Systems L, Lecture Notes in Computer Science, 12930, 46-62.
• Cook, A., Mısırlı, G., ve Fan, Z. (2020). Anomaly Detection For Iot Time-Series Data: A Survey, IEEE Internet Of Things Journal, 7 (7), 6481 – 6494. Doi: 10.1109/jıot.2019.2958185
• Chandola, V., Banerjee, A., ve Kumar, V. (2009). Anomaly Detection: A Survey, Acm Computing Surveys, 41(3), 1–58. https://doi.org/10.1145/1541880.1541882
• Chen, T., ve Guestrin C. (2016). Xgboost: A Scalable Tree Boosting System, arXiv:1603.02754.
• Cheng, Z., Zou, C., ve Dong, J. (2019). Outlier detection using isolation forest and local outlier factor. In Proceedings of the Conference on Research in Adaptive and Convergent Systems, 161-168. https://doi.org/10.1145/3338840.3355641
• Domingues, R., Filippone, M., Michiardi, P., ve Zouaoui, J. (2017). A Comparative Evaluation Of Outlier Detection Algorithms: Experiments And Analyses, Pattern Recognition, 74, 406-421. https://doi.org/10.1016/j.patcog.2017.09.037
• Fang, Z., Yang, S., Lv, C., An, S., ve Wu, W. (2022). Application of a Data-Driven Xgboost Model For The Prediction of Covıd-19 in The Usa: A Time-Series Study, BMJ Open, 12:e056685. Doi: 10.1136/Bmjopen-2021-056685
• Falcão, F., Santos, A., Zoppi, T., Fonseca, B., Bondavalli, A., Silva ,C.B.V., ve Ceccarelli,A. (2019). Quantitative Comparison of Unsupervised Anomaly Detection Algorithms For Intrusion Detection, Sac '19: Proceedings Of The 34th Acm/Sıgapp Symposium On Applied Computing, 318–327.
• Goldstein, M. ve Uchida, S. (2016). A Comparative Evaluation of Unsupervised Anomaly Detection Algorithms for Multivariate Data, Plos One, 11(4), E0152173.
• İşyapar, M.T. (2013). Classification of Electricity Customers Based On Real Consumption Values Using Data Mining And Machine Learning Techniques And Its Corresponding Applications, M.Sc. thesis, Middle East Technical University, M.Sc., Department Of Computer Engineering. http://etd.lib.metu.edu.tr/upload/12616261/ındex.pdf.
• Jha, B.K., ve Pande, S. (2021). Time Series Forecasting Model For Supermarket Sales Using Fb-Prophet, 5th International Conference On Computing Methodologies And Communication, Isbn:978-1-6654-0360-3.
• Liu, F.T., Ting, K.M., ve Zhou, Z. (2008). Isolation Forest, IEEE International Conference On Data Mining, Electronic Issn: 2374-8486,15-19. doi: 10.1109/ıcdm.2008.17
• Liu, F.T., Ting, K.M. ve Zhou, Z. (2012). Isolation-Based Anomaly Detection, Acm Transactions On Knowledge Discovery From Data, 6(1), 1–39. https://doi.org/10.1145/2133360.2133363
• Maatug, F. (2021). Anomaly Detection Of Smart Meter Data, Master’s Thesis, University of Stavanger, Faculty of Science And Technology Department of Electrical Engineering and Computer Science.
• Nakayama, K. ve Sharma, R. (2017). Energy Management Systems with Intelligent Anomaly Detection and Prediction, Published in: 2017 Resilience Week (Rws), 24-29.
• Resulaj R. (2019). Smart Meter Based Load Forecasting For Residential Customers Using Machine Learning Algorithms, Master’s Thesis , University of Stavanger, Faculty of Science and Technology.
• Schmid, S., Wenig, P., and Papenbrock, T. (2022). Anomaly Detection in Time Series: A Comprehensive Evaluation, Proceedings of the Vldb Endowment, 15(9), 1779–1797. https://doi.org/10.14778/3538598.3538602
• Shaukat, K., Alam, T.M., Luo, S., Shabbir, S., Hameed, I.A., Li, J., Abbas, S.K., and Javed, U. (2021). A Review of Time-Series Anomaly Detection Techniques: A Step To Future Perspectives. In: Arai, K. (Eds) Advances In Information And Communication. 1363, Springer, Cham. 13 pages. https://doi.org/10.1007/978-3-030-73100-7_60
• Tajraq, F.G. (2020). Electricity Consumption Forecasting Of Turkey Usingrecurrent Neural Networks, Master Thesis, Istanbul Technical University, Informatics Institute, Department of Informatics Applications.
• Vitale, F. (2021). Run-Time Anomaly Detection With Process Mining: Methodology and Railway System Compliance Case-Study, Master Thesis, Linnaeus University, Faculty Of Technology, Department Of Computer Science And Media Technology.
• Vafeiadis, T., Alexiadis, A., Dimaridou, V., Krinidis, S., Kitsikoudis, K., Makris, L., Davidović, D., Ioannidis, D., ve Tzovaras, D. (2019). Anomaly Detectıon In Smart Meters, 14 Th Conference Of Slovenian Electrical Power Engineers, 6-08.
• Zhang, J.E., Wu, D., ve Boulet, B. (2021). Time Series Anomaly Detection For Smart Grids: A Survey, Ieee Canadian Electrical Power And Energy Conference (Epec2021).
• Zhu, Z. (2022). Anomaly Detection over Time Series Data, Preprints, 2022070407, Doi: 10.20944/preprints202207.0407.v1.
• Yu,E., Wei, H., Han, Y., Hu ,P., ve Xu, G. (2021). Application Of Time Series Prediction Techniques For Coastal Bridge Engineering, Advances in Bridge Engineering. 6. https://doi.org/10.1186/s43251-020-00025-4
• Qin ,Y., ve Lou, Y. (2019). Hydrological Time Series Anomaly Pattern Detection Based on Isolation Forest, IEEE 3rd Information Technology, Networking, Electronic And Automation Control Conference, 15-17.
• URL-1: https://www.toroslaredas.com.tr/yasal-bildirim/aydinlatma-tuketim-bilgileri [Erişim Tarihi: 15.01.2022]
• URL-2: https://www.sedas.com/tr-tr/DagitimHizmetleri/Pages/Genel-Aydinlatma.aspx [Erişim Tarihi: 15.01.2022]
• URL-3: https://www.firatedas.com.tr/BilgiDanisma/GenelAydinlatmaTutarlari?id=1010 [Erişim Tarihi: 15.01.2022]
• URL-4: https://www.baskentedas.com.tr/yasal-bildirim/aydinlatma-tuketim-bilgileri [Erişim Tarihi: 15.01.2022]
• URL-5: https://www.kcetas.com.tr/genel-aydinlatma/ [Erişim Tarihi: 15.01.2022]
• URL-6: https://www.cedas.com.tr/tr/sayfalar/206/aydinlatma-tutarlari-ve-komisyon-kararlari [Erişim Tarihi: 15.01.2022]
• URL-7: https://www.ayedas.com.tr/yasal-bildirim/aydinlatma-tuketim-bilgileri# [Erişim Tarihi: 15.01.2022]
• URL-8: https://www.tedas.gov.tr/#!dagitim_srkt [Erişim Tarihi: 15.01.2022]
• URL-9: http://www.tredas.com.tr/icerik/genel-aydinlatma-bilgileri-165 [Erişim Tarihi: 15.01.2022]
• URL-10: https://www.ayedas.com.tr/yasal-bildirim/ticari-kalite [Erişim Tarihi: 15.01.2022]
• URL-11: https://www.meramedas.com.tr/tr/2022-13.html#detay [Erişim Tarihi: 15.01.2022]
• URL-12: https://www.tedas.gov.tr/#!tedas_tarifeler_1 [Erişim Tarihi: 15.01.2022]
• URL-13: https://scikit-learn.org/stable/modules/outlier_detection.html [Erişim Tarihi: 12.20.2022]
• URL-14: https://xgboost.readthedocs.ıo/en/stable/ındex.html# [Erişim Tarihi: 12.20.2022]
• URL-15: https://facebook.github.io/prophet/docs/quick_start.html#python-api , OpenSource Facebook Fbprophet, 2021 [Erişim Tarihi: 12.20.2022]