Data-driven modeling based on kernel extreme learning machine for sugarcane juice clarification.

Area of Science:

• Agricultural Engineering
• Chemical Engineering
• Data Science

Background:

• Sugarcane juice clarification is crucial for sugar production.
• Key quality indicators include gravity purity and color value.
• Current offline measurement methods hinder timely process adjustments.

Purpose of the Study:

• To develop a data-driven model for predicting sugarcane juice gravity purity and color value.
• To optimize model parameters using particle swarm optimization.
• To validate the model's effectiveness against existing methods.

Main Methods:

• Kernel extreme learning machine (KELM) for predictive modeling.
• Particle swarm optimization (PSO) for parameter tuning.
• Experimental validation and comparison with BPNN, RBFNN, and SVM.

Main Results:

• The proposed KELM-PSO model accurately predicts gravity purity and color value of sugarcane juice.
• The model demonstrates superior performance compared to BP neural network, radial basis neural network, and support vector machine.
• Experimental results confirm the reliability and effectiveness of the data-driven approach.

Conclusions:

• The KELM-PSO model offers a reliable and efficient solution for real-time monitoring of sugarcane juice quality.
• This approach facilitates timely adjustments in the sulphitation clarification process, improving sugar production efficiency.
• The study highlights the potential of advanced machine learning techniques in the sugar industry.