Research and analysis of the dynamic weighing system for aggregates in cement stabilized soil mixing plant based on machine learning

The precision of aggregate weighing systems in cement-stabilized soil mixing stations is paramount for the quality and durability of infrastructure projects. This paper presents an advanced design and comprehensive analysis of an aggregate weighing system utilizing MATLAB/Simulink for control design and machine learning techniques for predictive analysis. By optimizing Proportional-Integral-Derivative (PID) controller parameters and implementing a Kalman filter for noise reduction, we achieved a 15% improvement in measurement accuracy and a 20% increase in system stability. Machine learning models, including neural networks, were trained on both real and synthetic datasets to predict system performance under varying conditions, enhancing operational efficiency. The proposed approach was validated through simulations and experimental data, demonstrating significant enhancements over traditional methods.