JUSTIFICATION AND IMPROVEMENT OF THE PARAMETERS OF A TECHNICAL FACILITY FOR ARTIFICIAL INTELLIGENCE-BASED CONTROL OF THE SELF-HEATING STATE OF COTTON HEAPS
Keywords:
cotton stacks, self-heating, artificial intelligence, temperature sensor, LSTM, early warning, cotton ginning plant, monitoring.Abstract
The article analyzes the problem of self-heating of raw material stacks in the cotton ginning industry of Uzbekistan. The parameters of the control tool based on a sensor system that measures temperature, humidity, and gas content in real time, as well as artificial intelligence (AI) algorithms, are substantiated. Experimental calculation methods determined the sensor placement step, measurement frequency, warning limits, and forecasting horizon. The proposed LSTM + Random Forest ensemble model allows for the early detection of hotspots compared to traditional threshold control: the average detection time in the model assessment is reduced to 2.9 times, and the proportion of false signals does not exceed 6.5%. The results can be used to reduce losses in cotton storage enterprises.
References
1. O'zDSt 604. Raw cotton. Technical specifications. - Tashkent: Uzstandard Agency.
2. Hochreiter S., Schmidhuber J. Long Short-Term Memory // Neural Computation. - Тошкент. Vol. 9, No. 8. - P. 1735-1780.
3. Breiman L. Random Forests // Machine Learning. - Ташкент. Vol. 45, No. 1. - P. 5-32.
4. Kingma D.P., Ba J. Adam: A Method for Stochastic Optimization // arXiv:1412.6980. - Тошкент.
5. Pedregosa F. et al. Scikit-learn: Machine Learning in Python // Journal of Machine Learning Research. - 2020. Vol. 12. - P. 2825-2830.
6. Educational and methodological manuals on the technology of storage and primary processing of raw cotton (higher educational institutions of the textile and cotton industry of Uzbekistan).
7. Isayev B.M. "Research and development of a system for automatic detection and monitoring of cyberattacks in the information and communication systems of industrial enterprises," ADPI Scientific Bulletin, 2026, No. 5 https://eduservis.uz/, pp. 151-154.
Scientific Arena is an international peer-reviewed online scientific journal that publishes original research articles across a broad range of academic disciplines.