At present, the automation degree of the rolling mill has reached a high level. Its structure is complex, and the actual operation process has the characteristics of high temperature, high pressure, and high speed. Once a failure occurs, a chain reaction will occur, which may cause damage to the entire equipment. Not only cause huge economic losses but also may endanger personal safety. Practical experience and historical lessons have made people more and more aware that in order to ensure the safety of the rolling mill and the stability of product quality, it is necessary to conduct condition monitoring and fault diagnosis of the entire rolling mill. Rolling mill status monitoring is to monitor the main process variables involved in the rolling mill in real-time to determine whether a fault occurs. The fault diagnosis based on the rolling mill mainly studies how to detect, separate, identify and recover the fault of the rolling mill.
01 Rolling mill status monitoring
1) Status monitoring of mechanical parameters of rolling mills: including rolling pressure, rolling torque, looper torque, looper hydraulic cylinder pressure, looper angle, etc. of each stand.
2) Rolling process status monitoring: including finish rolling entrance strip speed, temperature, thickness, roll gap, work roll speed, etc.
3) Condition monitoring of rolling mill equipment: including the temperature of main bearings, oil temperature, vibration of rolling mill components, axial displacement of the motor rotor, and wear of main parts (such as gears and bearings, etc.).
02 Classification of main faults of rolling mills
In order to reasonably classify the condition monitoring and fault diagnosis methods of rolling mills, this paper divides the faults of rolling mills into three levels: equipment faults, system faults, and product quality faults.
1) Equipment failure
It mainly refers to various mechanical failures in the production equipment included in the rolling mill, such as electro-hydraulic servo valves, gears, and bearings.
2) System failure
It mainly refers to various faults in the control system of the rolling mill, which make it difficult for the system to act according to the instructions of the host computer due to process design problems or worker misoperations that deviate from the normal process regulations, thus making the production unable to proceed normally.
3) Product quality failure
It mainly refers to abnormalities in the quality of rolled products due to disturbances and faults in rolling mill equipment or systems, such as failures in the thickness or shape accuracy of strip steel products that do not meet the requirements.
03 Analysis of the characteristics of the variable data in the rolling process and the analysis of the fault characteristics of the rolling mill
There are more than 200 main process variables reflecting the working status of rolling mill equipment, including temperature, pressure, speed, hydraulic parameters, tension, torque, and electrical parameters. These process variables are coupled with each other. For example, the change in the roll gap will affect the exit thickness of the rolled piece, and the change in the exit thickness will affect the forward slip value. At the same time, the factors affecting the forward slip value are the amount of reduction, the rolled piece, and the roll. coefficient of friction etc. In addition to the mutual coupling characteristics, the rolling process variables also have nonlinear, dynamic, large-scale, and multi-scale characteristics. The rolling mill system is a very typical nonlinear system, and there is a strong nonlinear relationship between the process variables.
The dynamic characteristics of data refer to the characteristics of the time series correlation between the observed data at a certain moment and the observation data sampled in the past period. In the steel rolling process, the sampling time interval of each sensor is very short, and there is an energy storage link and a closed-loop control system, so the steel rolling process data has obvious dynamic characteristics.
The large-scale nature of data means that in the steel rolling process, each production process involves a large number of equipment and devices, including multiple subsystems and operating areas.
The multi-scale feature means that the rolling mill system is a multi-level, high-performance complex industrial system. Faults of different natures at different locations generally only occur in a specific frequency band. Therefore, the rolling production process includes multiple frequency features, and the process data has Multi-scale features.
There is an intricate relationship between rolling mill faults and fault symptom signals. For example, roll eccentric faults are caused by the overrun of parameters such as rolling force, roll gap, and product thickness deviation and the overrun of rolling speed parameters can also cause roll surface damage. Wear and tear, motor mechanical failure, abnormal current and voltage, etc., belong to the situation of multiple symptoms and multiple faults. Moreover, when multiple faults occur at the same time, they are often coupled together, and the symptoms corresponding to the same signal of the same fault will be different. Therefore, there is also an unclear corresponding relationship between symptoms and faults.