Title: Machine Learning Based Electric Energy Consumption Prediction of a Large-Scaled Production Plant with Small-Scaled Data

Abstract: This report covers the statistical approach to predict consumed energy for a tire production plant. The reasons behind this study are also to optimize the energy consumption budget and to follow the production area wised KPIs which is also vital for ISO 50001 Energy management system standard. In order to make it happen, writers clarify the main problem, then start to apply the steps of the cross industry standard process for data mining (CRISP-DM) [1] methodology. The most important point of this study was that although the historical data is small scaled, the parameters have a higher dimension according to input examples. Hence, the data to be used as input could be explained with simple variables to be used in the budget period. The study introduces data preparation steps based on the production area, grid search for best regression algorithm, comparison of models, and seven-month validation results.

Keywords: Machine learning; regression; small scaled data analytics; energy consumption prediction

Başlık: Küçük Ölçekli Verilerle Büyük Ölçekli Bir Üretim Tesisinin Makine Öğrenimine Dayalı Elektrik Enerjisi Tüketim Tahmini

Özet: Bu çalışma bir lastik üretim fabrikası için tüketilen enerjiyi tahmin ve takip etmeye yönelik istatistiksel bir yaklaşımı kapsamaktadır. Çalışmanın arkasındaki nedenler, enerji tüketim bütçesini optimize etmek ve ISO 50001 Enerji Yönetim Sistemi Standardı için de kritik önem taşıyan üretim proses bilgisine dayalı performans indikatörlerini belirlemek ve takip etmektir. Bunun gerçekleşmesi için endüstriler arası veri madenciliği standart prosesinin (CRISP-DM) [1] adımları takip edilmiştir. Bu çalışmanın en önemli noktası, tarihsel verilerin küçük ölçekli olmasına rağmen, parametrelerin girdi örneklerine göre daha yüksek bir boyuta sahip olmasıdır. Böylelikle girdi olarak kullanılacak veriler bütçe döneminde kullanılacak basit değişkenlerle açıklanabilir. Çalışma, üretim alanına dayalı veri hazırlama adımlarını, çapraz geçerleme ile en iyi bağlanım algoritması ve parametrelerinin seçimini, makine öğrenimi modellerinin karşılaştırmasını ve sonuçlarını takip eden yedi aylık doğrulama adımını tanıtmaktadır.

Anahtar kelimeler: Makine öğrenimi; bağlanım; küçük ölçekli veri analizi; enerji tüketim tahmini

• Wirth R, Hipp J. CRISP-DM: Towards a standard process model for data mining. In Proceedings of the 4th International onference on the Practical Applications of Knowledge Discovery and Data Mining, 2000, pp. 29-39.
• Gorenstein S. Planning tire production. Management Science 1970; 17(2): B72–B82.
• International Organization for Standardization. Energy management systems-Requirements with guidance for use (ISO/DIS Standard No. 50001), 2018, Retrieved from iso.org/publication/PUB100400.
• Chiu TY, Lo SL, Tsai YY. Establishing an integration-energy-practice model for improving energy performance indicators in ISO 50001 energy management systems. Energies 2012; 5: 5324-5339.
• Mckane A, Desai D, Matteini M, Meffert W, Williams R, Risser R. Thinking Globally: How ISO 50001-Energy management can make industrial energy efficiency standard practice. Technical Report from U.S. Department of Energy Office of Scientific and Technical Information 2009; Retrieved from https://doi.org/10.2172/983191
• Rocha BC, de Sousa Jr RT. Identifying bank frauds using CRISP-DM and decision trees. International Journal of Computer Science and Information Technology 2010; 2: 162-169.
• Rivo E, de la Fuente J, Rivo A, Garcia-Fontan E, Canizares MA, P. Gil P. Cross-Industry Standard Process for data mining is applicable to the lung cancer surgery domain, improving decision making as well as knowledge and quality management. Clinical and Translational Oncology 2012; 14(1): 73–79.
• Huber S, Wiemer H, Schneider D, Ihlenfeldt S, DMME: Data mining methodology for engineering applications – a holistic extension to the CRISP-DM model. Procedia CIRP 2019; 79: 403–408.
• Musil CM, Warner CB, Yobas PK, Jones SL. A comparison of imputation techniques for handling missing data. Western Journal of Nursing Research 2002; 24(7): 815–829.
• Lakshminarayan K, Harp SA, Samad T. Imputation of missing data in industrial databases. Applied Intelligence 1999; 11(3): 259–275.
• Sinharay S, Stern HS, Russell D. The use of multiple imputation for the analysis of missing data. Psychological Methods 2001; 6(4): 317–329.
• Hawthorne G, Elliott P. Imputing cross-sectional missing data: Comparison of common techniques. Australian & New Zealand Journal of Psychiatry 2005; 39(7): 583–590.
• Donald BR. Inference and missing data. Biometrika 1976; 63(3): 581–592.
• Belsley DA. Conditioning diagnostics: Collinearity and weak data in regression. Wiley, New York, 1991.
• Farrar DE, Glauber RR. Multi-collinearity in regression analysis: The problem revisited. The Review of Economics and Statistics 1967; 49: 92–107.
• Alin A. Multicollinearity. Wiley Interdisciplinary Reviews: Computational Statistics 2010; 2(3): 370–374.
• Mansfield ER, Helms BP. Detecting multicollinearity. The American Statistician 1982; 36(3a): 158–160.
• Abdi H, Williams LJ. Principal components analysis. Wiley Interdisciplinary Reviews: Computational Statistics 2010; 2: 433-450.
• Alpaydin E. Introduction to Machine Learning. The MIT Press, 2014, pp. 105-130.
• Bergstra J, Bengio Y. Random search for hyper-parameter optimization. Journal of Machine Learning Research 2012; 13: 281–305.