Computer Optimized Control of Heating Furnaces in Seamless Steel Pipe Hot Rolling Production Lines

Literature Overview

This paper, authored by Jiang Zeyi and colleagues from University of Science and Technology Beijing and Baoshan Iron and Steel Group, published in 2003 in the journal "Industrial Heating" (Volume 32, Issue 2, pages 24-25), addresses the computer-optimized control of heating furnaces in seamless steel pipe hot rolling production lines. The work focuses on two critical thermal processing stations: the ring-type heating furnace used for billet preheating and the blank tube reheating furnace used for rethermalization of the intermediate blank tube before piercing and rolling operations. The authors developed mathematical models for both furnaces and implemented a computer-optimized control system that achieved effective temperature tracking throughout the seamless steel pipe hot rolling process.

Core Technical Content and Mathematical Modeling

The seamless steel pipe hot rolling production line involves a series of interconnected thermal and mechanical processes where temperature control is paramount. The ring-type heating furnace heats solid steel billets to the required piercing temperature, typically in the range of 1200 to 1250 degrees Celsius for carbon steel grades, while the blank tube reheating furnace rethermalizes the hot-rolled blank tube to the appropriate rolling temperature before the rolling mill stand. Both furnaces present unique control challenges due to their geometry, fuel distribution patterns, and the dynamic nature of the heating loads.

The mathematical models developed in this study incorporate heat transfer mechanisms including conduction within the steel workpiece, convection between the furnace atmosphere and the steel surface, and radiation exchange between furnace walls, burners, and the workpiece. The ring-type furnace model accounts for the continuous rotation of the hearth plates carrying the billet through the furnace zones, creating a moving-boundary problem that must be discretized for real-time control. The blank tube reheating furnace model addresses the cylindrical geometry of the blank tube and the varying wall thickness that affects the internal temperature gradient.

Key Modeling Parameters

The mathematical model for the ring furnace typically employs a finite difference approach where the billet is discretized into radial and longitudinal elements, with the heat balance equation solved iteratively at each time step. The control algorithm uses the model predictions to adjust fuel flow rates in each furnace zone to maintain the target temperature profile along the furnace length. This is essentially a model-predictive control (MPC) strategy where the furnace model serves as the predictor and the optimization objective is to minimize the deviation from the target temperature while minimizing fuel consumption.

Control System Architecture and Implementation

The computer-optimized control system described in the paper integrates process instrumentation, data acquisition, mathematical modeling, and control actuation into a unified architecture. The system architecture follows a hierarchical structure with a process level, a control level, and a monitoring level. At the process level, thermocouples measure furnace gas temperatures at multiple axial stations, and infrared pyrometers monitor the surface temperature of the billet or blank tube at furnace entry and exit. The control level executes the mathematical model and computes the required control actions. The monitoring level provides operators with real-time visualization of temperature profiles, model predictions, and control adjustments.

The implementation at Baoshan Iron and Steel Group demonstrated that the computer-optimized control system achieved significant improvements in temperature control accuracy compared to conventional PID-based control. The model-based approach allowed the system to anticipate the thermal lag inherent in furnace heating and to make proactive adjustments rather than reactive corrections. This proactive control strategy is particularly important in the blank tube reheating furnace where the heating time is relatively short and the temperature must be achieved rapidly and uniformly.

Control Performance Comparison

Integration with Engineering Practice

In my experience with seamless steel pipe production, the quality of the finished pipe is fundamentally dependent on the thermal processing history of the steel. The austenitization temperature, the cooling rate after piercing and rolling, and the temperature uniformity across the pipe cross-section all directly influence the final grain structure, mechanical properties, and service performance of the pipe. The computer-optimized control system described in this paper represents a significant step toward achieving the precise thermal control required for high-grade seamless pipes.

From a metallurgical perspective, the control of the ring heating furnace outlet temperature is critical because it determines the austenite grain size at the time of piercing. Overheating leads to coarse austenite grains that persist through the subsequent deformation and cooling, resulting in poor transverse toughness and susceptibility to intergranular cracking. Underheating results in insufficient austenite transformation and incomplete recrystallization during rolling, leading to poor dimensional accuracy and surface quality. The model-based control system helps maintain the outlet temperature within a narrow window that balances these competing requirements.

For the blank tube reheating furnace, the control challenge is different. The blank tube is a hollow cylinder with a relatively thin wall, and the goal is to achieve uniform temperature throughout the wall thickness before rolling. The computer-optimized control system must account for the thermal inertia of the steel, the geometry of the blank tube, and the furnace heat transfer characteristics to ensure that the wall temperature is uniform at the time of rolling. Temperature gradients across the wall thickness during rolling can lead to differential deformation, residual stresses, and dimensional defects in the finished pipe.

Practical Considerations for Control System Implementation

1. Model calibration: The mathematical models must be calibrated against actual furnace performance data. This requires extensive data collection during normal production operations, including fuel flow rates, gas temperatures, workpiece temperatures, and production parameters. The calibration process should be performed for each furnace zone and for each product grade.
2. Sensor reliability: The accuracy of the control system depends critically on the reliability of temperature measurement instruments. Infrared pyrometers are susceptible to scale, oxide, and furnace atmosphere effects, and must be regularly calibrated and maintained. Thermocouples in the furnace atmosphere provide more stable readings but do not directly measure the workpiece temperature.
3. Disturbance handling: The control system must be robust to disturbances such as variations in billet size, changes in furnace door opening frequency, and fluctuations in fuel gas pressure and quality. A well-designed control system should incorporate feedforward compensation for known disturbances and adaptive control for unknown ones.
4. Integration with downstream processes: The furnace control system should be integrated with the downstream piercing and rolling processes to ensure that the thermal conditions at the furnace exit are compatible with the deformation requirements of the downstream operations. This requires a coordinated control strategy across the entire production line.

Key Questions and Reflections

The paper raises several important questions that remain relevant in modern seamless steel pipe production. First, how accurate must the mathematical model be for the control system to achieve satisfactory performance? The answer depends on the specific application and the tolerance for temperature deviation. For general-purpose carbon steel pipes, a model accuracy of ±20 degrees Celsius may be sufficient, while for high-grade alloy pipes with strict microstructure requirements, the model accuracy must be better than ±10 degrees Celsius.

Second, the paper does not extensively discuss the handling of model uncertainties and parameter variations. In practice, furnace models degrade over time due to refractory wear, burner fouling, and changes in furnace geometry. The control system must incorporate mechanisms for model updating and parameter adaptation to maintain long-term performance.

Third, the paper focuses on temperature control but does not address the control of the furnace atmosphere composition. For certain steel grades, particularly those susceptible to decarburization or oxidation, the control of oxygen potential and hydrogen content in the furnace atmosphere is equally important as temperature control. A comprehensive control system should integrate both temperature and atmosphere control.

Study Insights and Implications

The work by Jiang Zeyi and colleagues represents an early but significant contribution to the computer-optimized control of seamless steel pipe production furnaces. The fundamental approach of developing mathematical models and implementing model-based control remains valid and relevant in modern production environments. The key insight is that furnace control is not merely a process control problem but a metallurgical control problem where the objective is to achieve a specific thermal history that produces the desired microstructure and properties in the finished pipe.

For engineers involved in seamless steel pipe production, the implications are clear: investment in furnace modeling and computer-optimized control systems is justified by the improvements in product quality, fuel efficiency, and production flexibility. The control system enables tighter temperature windows, which in turn enables the production of higher-grade pipes with more consistent properties and fewer quality defects. The approach also provides a framework for integrating furnace control with the broader process control system, enabling end-to-end quality optimization across the production line.