Temperature Control Method for Electric Heating Furnaces Based on Auto-Encoder and Fuzzy PI Control.

Area of Science:

• Industrial Engineering
• Control Systems Engineering
• Applied Mathematics

Background:

• Precise temperature control in electric heating furnaces is critical for industrial production and research.
• Challenges include system nonlinearity, time-varying dynamics, and significant time delays, hindering accurate control.
• Existing methods struggle to maintain optimal performance under complex operational conditions.

Purpose of the Study:

• To develop a robust and adaptive composite control method for electric heating furnaces.
• To improve temperature control accuracy and system performance in the presence of disturbances.
• To validate the proposed method's effectiveness in complex industrial environments.

Main Methods:

• A discrete-time temperature model was constructed for the electric heating furnace.
• An auto-encoder based prediction model was integrated to learn system dynamics and predict future temperatures.
• A fuzzy proportional-integral (PI) controller was employed to adaptively tune PI parameters in real-time.

Main Results:

• The proposed composite control method achieved high modelling accuracy and system adaptability.
• Simulations demonstrated temperature overshoot within 2% under various disturbances.
• The system showed strong robustness against maximum delays of 243 s, measurement noise, voltage fluctuations, and gain variations.

Conclusions:

• The integrated auto-encoder and fuzzy PI control method offers a robust solution for precise temperature management in electric heating furnaces.
• The approach significantly enhances control system adaptability and performance in challenging industrial settings.
• This method is suitable for advanced control design in complex industrial applications requiring high precision and reliability.