Integrated approach to predictive management of the technical condition of technological equipment at dairy processing enterprises based on machine learning
DOI:
https://doi.org/10.34121/1028-9763-2026-2-134-145Keywords:
information technologies, technical condition management, forecasting, digital twins, machine learning, dairy processing enterpriseAbstract
The article addresses the problem of improving the operational reliability of technological equipment at dairy processing enterprises in the context of the transition to intelligent maintenance strategies. The aim of the study is to develop an integrated approach to predictive management of technical condition of the equipment based on the combination of machine learning methods, multidimensional time series analysis, and the digital twin concept, taking into account the dynamics of degradation processes. The study presents a systematic analysis of maintenance strategies, identifies the limitations of schedule-based approaches with fixed intervals, and substantiates the feasibility of applying a hybrid model that integrates condition-based maintenance with predictive-oriented mechanisms. Based on retrospective operational data, the experimental results demonstrate a statistically significant advantage of the ensemble method XGBoost compared to alternative models in terms of F1-score and ROC-AUC metrics. The proposed approach enables the generation of maintenance recommendations for the next required period, which directly affects operational production planning as equipment in critical condition is excluded from the production cycle until maintenance is performed, thereby minimizing the risks of unplanned downtime, raw material losses, and violations of technological regimes. The developed forecasting module implements real-time assessment of the probability of failure and residual resource, which increases the probability of early detection of potentially emergency conditions. The practical significance of the results lies in the implementation of the proposed approach within a decision support system, enabling the reduction of unplanned downtime, optimization of maintenance schedules, and minimization of technological losses of raw materials. The proposed approach is consistent with the principles of industrial digital transformation and can serve as a foundation for the development of intelligent management systems within the frameworks of Industry 4.0 and Industry 5.0 concepts. Tаbl.: 3. Refs.: 20 titles.
References
1. Dairy Processing Handbook. Tetra Pak / Tetra Pak International S.A. 2025. 468 p. URL: https://dairyprocessinghandbook.tetrapak.com (дата звернення: 13.01.2026).
2. Lee J., Bagheri B., Jin C. Introduction to cyber manufacturing. Manufacturing Letters. 2016. Vol. 8. P. URL: https://www.sciencedirect.com/journal/manufacturing-letters (дата звернення: 13.01.2026).
3. Yavuz O., Yürekli M., Öztaş G. Reliability Centered Maintenance Practices in Food Industry. Procedia Computer Science. 2019. Vol. 158. P. 227–234. DOI: https://doi.org/10.1016/j.procs.2019.09.043 (дата звернення: 13.01.2026).
4. Zhang W., Yang D., Wang H. Data-driven methods for predictive maintenance of industrial equipment: A survey. Expert Systems with Applications. 2020. Vol. 158. DOI: https://doi.org/10.1016/j.eswa.2020.113517 (дата звернення: 13.01.2026).
5. Lee J. Industrial AI: Applications with Sustainable Performance. Singapore: Springer Singapore, 2020. 162 p. DOI: https://doi.org/10.1007/978-981-15-2144-7 (дата звернення: 13.01.2026).
6. Zhang W., Yang D., Wang H. Data-Driven Methods for Predictive Maintenance of Industrial Equipment: A Survey. IEEE Systems Journal. 2019. Vol. 13, N 3. P. 2213–2227. DOI: https://doi.org/10.1109/JSYST.2019.2905565 (дата звернення: 15.01.2026).
7. Що таке data lake? AWS Amazon. URL: https://aws.amazon.com/big-data/datalakes-and-analytics/what-is-a-data-lake/ (дата звернення: 15.01.2026).
8. Seoh R. Solving Bayesian Network Structure Learning Problem with Integer Linear Programming. arXiv. 2020. 104 p. URL: https://arxiv.org/abs/2007.02829 (дата звернення: 15.01.2026).
9. Маслов Ю.В. Використання цифрових двійників у задачах прогнозування зносу електромеханічних систем. Вчені записки ТНУ імені В.І. Вернадського. Серія: Технічні науки. 2025. Т. 36 (75), № 6, Ч. 2. С. 214–220. DOI: https://doi.org/10.32782/2663-5941/2025.6.2/30.
10. Molnar C. Interpretable Machine Learning: A Guide for Making Black Box Models Explainable. 2nd ed. Munich: Christoph Molnar, 2022. 317 p. URL: https://christophm.github.io/interpretable-ml-book/ (дата звернення: 15.01.2026).
11. Scikit-learn: Machine Learning in Python. Common pitfalls and recommended practices. URL: https://scikit-learn.org/stable/common_pitfalls.html (дата звернення: 16.01.2026).
12. Brownlee J. How to Use ROC Curves and Precision-Recall Curves for Classification in Python. MachineLearningMastery.com. 2018. URL: https://machinelearningmastery.com/roc-curves-and-precision-recall-curves-for-classification-in-python/ (дата звернення: 16.01.2026).
13. Friedman J. H. Greedy function approximation: A gradient boosting machine. Annals of Statistics. 2001. Vol. 29, N 5. P. 1189–1232. DOI: https://doi.org/10.1214/aos/1013203451.
14. Clark B., Lee F. Qué es boosting de gradiente? IBM. 2025. URL: https://www.ibm.com/es-es/think/topics/gradient-boosting (дата звернення: 16.01.2026).
15. Breiman L. Random Forests. Machine Learning. 2001. Vol. 45, N 1. P. 5–32. DOI: https://doi.org/10.1023/A:1010933404324.
16. Конотопець М., Туровський О., Аксамітний О., Матійко А. Порівняльний аналіз моделей машинного навчання Random Forest та XGBoost у задачі класифікації інцидентів безпеки. Herald of Khmelnytskyi National University. Technical sciences. 2025. Т. 357, № 29. С. 235–245. DOI: https://doi.org/10.31891/2307-5732-2025-357-29.
17. XGBoost. GeeksforGeeks. 2026. URL: https://www.geeksforgeeks.org/machine-learning/xgboost/ (дата звернення: 17.01.2026).
18. Khallaf M., Mhawi A. Long Short-Term Memory Networks: A Comprehensive Survey. AI. 2025. Vol. 6, N 9. Р. 215. DOI: https://doi.org/10.3390/ai6090215.
19. Khyathi G., Indumathi K. P., Jumana Hasin A., Lisa Flavin Jency M., Sibyl Siluvai, Krishnaprakash G. Support Vector Machines: A Literature Review on Their Application in Analyzing Mass Data for Public Health. Cureus. 2025. Vol. 17, N1. e77169. DOI: https://doi.org/10.7759/cureus.77169.
20. Bovchaliuk S., Haidai Y. Аналіз методу опорних векторів у порівнянні з традиційними методами передбачення ринкових рухів. Системи управління, навігації та зв'язку. 2024. Вип. 3. С. 89–93. DOI: https://doi.org/10.26906/SUNZ.2024.3.089.
157
64
