1.2 Force Prediction as Basis for Flat Rolling Draft Scheduling
Significance of Level 2 model parameter prediction for the product quality and productivity was discussed in depth in a paper . Among the predicted parameters, roll separating force is the most important one. Force prediction is the basis for flat rolling draft scheduling, such as those in plate and strip mills.
Level 2 model creates pass schedule based on predicted parameters, mainly the roll separating force and temperature, for the purpose of meeting various targets. If one or more targets are missed, it indicates possible product defects. If the parameter prediction is wrong, the targets cannot be met unless two or more wrong factors make a right one. Due to the difficulty in measuring temperature, the temperature is usually calculated from the measured force. Commonly there are three sets of major targets, all of which are tightly linked with the roll force, as discussed below.
- Equal Deformation Targets. The Level 2 model compensates mill and roll deformations, especially roll deflection, with the roll crown, to form an environment that the draft across the stock width is equal. Unequal deformation across width would cause either center buckle (if the draft in the width center is too high) or edge wave (with too high draft in the edges). The roll and mill deformations (e.g. deflections) are mainly caused by roll separating force. Error in the force prediction would easily let this target be missed. In addition, smaller roll gaps are needed for the head end and tail end due to the higher roll forces.
- Metallurgical Temperature Targets. While the force is checked for mill capacity limit and product shape, the temperature is closely watched for product property. When the temperature error is high, the system would miss the metallurgical targets such as those for controlled rolling. Since temperature is usually calculated from the force, force prediction error causes temperature prediction error.
- Mill Capacity and Productivity Targets. As long as the Metallurgical Temperature Targets and the Equal Deformation Targets are met, the Level 2 model would schedule as few passes as possible within the mill capacity, in order to achieve high productivity. For shaping stage except the last two to three finish passes, the highest force possible within the force limit is pursued; for roughing pass, highest torque within torque limit is usually applied. Torque is the product of rolling force, contact length and a factor called lever arm ratio. This general rule applies unless the steel has low formability.
Theoretically, for a given material in a given temperature and a given speed, corresponding to a specific roll crown, there is only one best-fit reduction (腍/H0) and thus force, in which the roll deflection caused by the force can be precisely compensated by the roll crown. However, in the real rolling practice, due to a long list of operational factors involved, usually there is a range of reductions that can be used for the rolling. The Level 2 model抯 job is to find this optimal range and value under certain mill equipment and mill technologies. For example, past experiences may allow a wider range of reduction be applied when roll bending system is in use. Many draft scheduling logics are behind those considerations.
1.3 Mill Production and Product Quality Issues
In Level 2, draft distribution is usually tied to the force distribution, and a certain force distribution over passes leads to good flatness of the plate/coil and high utilization of the equipment. A rolling schedule based on inaccurate force prediction may have more passes than needed; this leads to lower productivity, poor property and higher costs (energy, equipment, labor, etc.). When predicted force is wrong, the predicted mill stretch should be incorrect. This causes error in rolled thickness. Though the thickness eventually will be corrected by AGC, it causes other problems as described later.
Steel Property and Alloy Consumption. Due to missing temperature set points, etc., an inaccurate Level 2 model may lead to poor mechanical properties of the rolled product. To assure the satisfaction of the property requirements, the steel mill with inaccurate Level 2 model usually has to schedule higher grade than required, e.g. by adding more alloys. This causes unnecessary waste. If the level 2 model were more accurate, less alloys should have been consumed to achieve the required properties.
Center Buckle and Edge Wave. Roll deflection, stand deflection and roll flattening are all affected by the roll separating force, so force prediction error causes the errors for all those deformations. If the force is estimated too much lower than the real value, the predicted roll deflection is too much lower than the anticipated one. An equal draft has been planned based on the predicted roll deflection; but the actual deflection is higher than the planned one, so the draft in the width center region is actually smaller than that in the edges and consequently, an edge wave could occur. On the other hand, if the force is predicted too much higher than the true value, the draft in the width middle will be higher compared with that in the strip/plate edges. This would lead to Center Buckle. Mathematical models can be easily developed to predict center buckle and edge wave.
Head End and Tail End Cambers. Force error may lead to head end and tail end cambers, or similar defects. With a large force prediction error, initially estimated roll gap based on roll separating force and mill stretch, etc., also has significant error. In adjusting roll gap, AGC would have a large movement and would therefore be instable. In this process, any minor factor could cause shape defect in the head end, or unfavorable movement of the steel that leads to head camber and body camber.
High-strength Grades. In the production of high-strength grades such as API X80, X100 and X120, there are very strict requirements in the production procedure, such as temperature control. It thus requires precise force and temperature prediction. Its draft schedule does not allow wide rang of parameters (force, temperature). One cannot imagine if the force and temperature predictions are with high errors.
Thermomechanical Rolling. Success of the thermomechanical controls depends on not only the design for the strategy (rolling temperature, draft schedule, cooling speed and temperature profile, etc.), but also production execution system (primarily Level 2 model) – whether it can actually achieve it. If the Level 2 model doesn抰 have sufficient prediction accuracy, it would simply miss the targets of e.g. temperature set points.
Hard and Thin Material Rolling. Two challenges exist when rolling hard and thin materials. Because it is thin, roll gap must be very flat, so accuracy for the rolling force and roll deflection prediction should be both very high. Besides, thin material is more vulnerable to temperature loss, so the temperature prediction should be very accurate; this again, put high requirement for the force prediction because temperature is calculated from force.