مدل‌سازي و مطالعه تجربي شبكه مبدل‌هاي حرارتي يك كارخانه قند و استفاده از روش‌هاي يادگيري ماشين و الگوريتم ژنتيك براي انتخاب شرايط بهينه ديناميكي

Simulation an‎d experimental investigation of a sugar factory heat exchanger network an‎d surveying optimal dynamic working condition using machine learning an‎d genetic algorithm methods

مهندسي مكانيك

The current study focuses on modeling an‎d optimizing the heat exchanger network of a targeted sugar factory to achieve dynamic optimal conditions. Dynamic optimal conditions refer to the systemʹs ability to continually reset to the optimum point as various parameters change over time. Analytical models have been employed for simulation of certain elements within the heat exchanger network an‎d their influencing components. For other elements, statistical models that utilize measurement data an‎d are powered by machine learning tools are employed. To achieve this, the relevant elements are measured during the factoryʹs operation, an‎d the collected data are processed through machine learning tools to derive corresponding correlations, which are then integrated into the analytical models. The integration of analytical models with statistical models represents a key innovation of this research. Additionally, the use of machine learning-based image processing to obtain specific measurements, the introduction of engineering assumptions to develop models for the first time, an‎d the creation of a comprehensive model for the heat exchanger network are among the significant innovations of this study. The results generated by the models are compared to existing experimental data to validate their accuracy. Prior to this research, the application of machine learning within the sugar industry was primarily limited to monitoring the sugar crystallization. In light of this, the current work aims to broaden the application of machine learning to other sections of the sugar industry.

4

دكتر جنر كسنجي