Modeling And Control Method For Laminar Cooling Processes Of Hot Rolled Strip

In order to enhance the quality and yield of strip, the laminar cooling technique is used to improve the metallurgical properties of the hot rolled strips by water-cooling after rolling in modern iron and steel industry. The strips are cooled from the austenitic finishing temperature (800～900℃) down to the ferrite coiling temperature (550～700℃) by controlling the cooling water quantity in the laminar cooling system. Laminar cooling process is one of the key procedures influencing the microstructure and properties of hot rolled strips. The coiling temperature is one of the important factors that determines the processing properties and mechanics properties of the strip product, and too high or too low of the coiling temperature will downgrade the final strip product quality.The heat transfer in cooling process is extremely complex. What’s more, the heat transfer coefficient and thermal conductivity change nonlinearly along with the operation conditions. It’s difficult to establish the accurate mathematical model for the strip cooling process. The cooling process models reported in the literature so far are only for dynamic cooling process in a single coiling unit. Those models cannot be used to compute the strip coiling temperature directly. The choice of water-cooling heat transfer parameter and thermal conductivity doesn’t consider the influences of varying process condition, and thus results less accurate models.The strip temperature is difficult to be measured online in the cooling zone because of the high temperature vapour generated. The relationship between cooling water quantity and strip coiling temperature is nonlinear. For the above reasons, the existing control methods based on looking up tables and strip temperature models can’t adapt the changing operation condition and lead to low control accuracy.Supported by the national project 973 (No.2002CB312200) of "Research on the real-time, intelligent optimal and control theory and methods for complex production manufacturing process" and aiming at the above problems in the modeling and control of laminar cooling processe, this paper carries out the research on the modeling and control of the laminar cooling process, and the target is to increase the control precision of the strip coiling temperature. The main works of this research are summarized as follows.1. A strip coiling temperature model is proposed consisting of four parts:the status of cooling unit valves calculated model, the strip segment track model, the model switching mechanism and the top surface temperature model of the strip in a cooling unit. The status of cooling unit valves calculated model is used to determine the valve status when the strip segment passes by every cooling unit, according to the setting values given by the control system of the sprayers, such as the number of active valves, the initial active valve location both side of the runnout table, and the spray mode. The strip segment track model is to compute the location of the strip segment, according to the initial velocity and the setting accelerations. The model switching mechanism is to decide the heat transfer style in the cooling unit. The top surface temperature model for the strip is to choose the air-cooling or water-cooling model according to the heat transfer style and to compute the top surface temperature of the strip.2. A method to determine the key coefficient of the water-cooling heat transfer parameter and thermal conductivity is proposed according to the fluctuating operation condition. Especially, the key coefficient of the water-cooling heat transfer parameter varys with the segment strip’s operation conditon. The water-cooling heat transfer parameter and thermal conductivity can be modified and the coiling temperature model precision is improved. Simulation is conducted using the industrial operating data from a large strip plant, and the results show that root mean square (RMS) of the error decreases from 11.85℃to 4.2℃, where the error is between the measured coiling temperature and the computed coiling temperature based on the improved water-cooling heat transfer parameter and thermal conductivity.3. A hybrid intelligent control method for laminar cooling process is developed. The proposed setting model for the number of active valves is consisted of four parts: the number of active valves pre-setting model, coiling temperature prediction model, prediction compensator and batch to batch compensator. The number of active valves pre-setting model calculates the pre-setting values of the number of active valves according to the target coiling temperature, the estimated values at the outlet of the finish train, such as the strip thickness, temperature and the running velocity of the strip head. The coiling temperature prediction model is to predict the coiling temperature according to the setting number of active valves. The prediction compensator computes the active valves compensated according to the predicted coiling temperature deviation, where the case-based reasoning and general PI tuning algorithm are used. The batch to batch compensator is to compute the compensated number of the active valves according the cooled strips’measured coiling temperature deviations. The case-based reasoning technology and PI iteration learning algorithm are adopted to conduct the iterative calculation among different strips, to control the measured coiling temperature deviation in a limited range. The number of active valves setting model can automatically adjust the number of active sparyers according to the varying operating condition. The control system of spary computes the states of the spary head valve according the new setting value and realize the cooling water quantity, in order to control the coiling temperature within the acceptable range of their target values.4. Based on the proposed model of the strip coiling temperature and the hybrid intelligent control method, a virtual object software and a setting control software are designed and developed. A simulation system for laminar cooling system is developed by integrating with an existing distributed experimental platform, which consists of a setting computer, a process control system(such as PLC system and the process monitor system), virtual instruments and actuators computer, and a virtual object computer.5. The proposed hybrid intelligent control method is simulated using the above simulation system. Contrasting with the existing control method, the simulation results show that the hit rate increased from 42.9% to 84.1% where the coiling temperature errors are controlled in the range of±10℃, and the root mean square (RMS) of the error decreases from 21.1℃to 8.21℃.