IMPROVEMENTS TO PORT TALBOT HOT STRIP MILL S CAPABILITY TO PRODUCE PROFILE CRITICAL LOW CROWN STRIP STEEL. N. D. Beynon*, M. Veryard* * Port Talbot Hot Mill, Corus Strip Products, Port Talbot Works, SA13 2NG. Correspondence Author Abstract This paper presents the findings of a series of plant trials to reduce the strip profile obtained during the hot rolling of a thin gauge (2.5mm) Boron treated full finished formable (0.02wt% Carbon) strip steel product. The work was undertaken at the Corus Strip Products UK, Port Talbot seven stand four-high hot strip finishing mill. Modifications to Work Roll Cambers (WRC) were made to achieve a flatter strip profile to meet the requirements of the customer s tighter edge gauge tolerance. How close the strip profile, or the distribution of the thickness of the hot rolled product across the width, is to the desired profile is extremely important for the final application of the product. Strip profile targets were achieved through working with the internal downstream customer to define their requirements. The final WRC modifications were achieved in five steps, with each step trial led and closely monitored. The final WRC were convex in nature whilst maintaining the crown ratio throughout the seven stands. The significant outcome achieved was that the customer requirement of a 20µm crown was, and continues to be, obtained on a consistent basis. The final WRC are more robust in that they allow the strip to be rolled with a consistent profile performance regardless of small changes in temperature, rolling rate, work roll bending forces and engineering issues. In addition, an Intranet based performance page was created to help operators monitor their performance relative to customer requirements. This work allowed us to obtain a greater understanding of the influence of the WRC on the final strip profile, the development of shape, and the interactions at the roll gap. Key Words: hot rolling; strip profile; crown; work roll cambers. Introduction The current steel market demands for flat rolled products with better dimensional accuracy have resulted in a remarkable reduction in strip crowns. Key driving forces behind this requirement include increasingly stringent tolerances on dimensions (BSEN 10051:1992 [1] and customers imposing tighter edge gauge tolerances in order to reduce edge discards and increase yield. Canstock customers require tight dimensional tolerances across the whole strip width, not just the centreline; this requires profile control through configuration of the roll gap (loaded gap) across the width of the strip [2]. The strip profile will change as material passes through each of the loaded gaps in the finishing mill. The profile of each loaded gap directly affects the interstand and exit mill strip profiles. The loaded gap profile is contributed to by a number of important factors such as the complex nature of the strip deformation, the initial ground work roll crowns (cambers), accumulated work roll wear and thermal build up, parabolic deflection of the work roll and back up rolls due to roll force, and roll flattening between the work rolls and the strip [3]. This paper presents the findings of a series of hot strip mill plant trials, the aim of which was to reduce the strip profile obtained through modifications to work roll cambers (WRC). The work was undertaken at Corus Port Talbot seven stand 4-high hot strip finishing mill.
Background A significant effort had been put in to understand and improve the hot rolling process for WideDWI (Table 1), with respect to the final strip profile. A significant amount (>70%) of the edge gauge tolerance on this product is made up by the strip crown and wedge. Table 1. Product Overview and Dimensions. Typical Slab Parameters (in mm unless specified) Width Length Thickness Boron treated full finished formable (0.02 wt% C). 1280 9800 234 Final Strip Parameters Width Gauge Target Crown ~ 1/3 Gauge Tolerance 1236 2.3 2.5 20µm Strip crown can be defined as: the difference between the strip thickness at the centre and the average thickness (X) mm from both edges (Eq.1). Wedge can be defined as: the difference between operator side and drive side strip thickness, (X) mm from the strip edge (Eq.2). For this report crown and wedge have been defined at 40mm [1] from the strip edge (Fig.1). ho + hd C = hc (1) 2 W = h d h o (2) Distance from strip edge (40 mm) Rolling Direction Thickness Drive Side = h D Thickness Operator Side = h O Thickness Centre = h C Fig.1 Strip profile variation in thickness across the width DWI canstock customers require tight profile control due to the nature of the two-piece canning operation. The forming process involves feeding tinplate strip (~0.02mm gauge) into a cupping press, which blanks and draws multiple shallow cups for each stroke. The cups are fed to parallel bodymaking machines that convert the cups into tall cans. This is the drawing and ironing process where the cups are redrawn to the final can diameter and then rammed through a series of rings which thin (iron) the can walls whilst at the same time increasing the can height [5]. Finally the uneven top edge of the can is trimmed to leave a clean edge and a can of the correct overall height. Cans that are heavy-to-gauge (overweight) can cause increased die wear (due to higher loading forces) or even failure to form. In addition, excessive trimming can lead to yield loss, whereas, underweight cans may fail to draw to the desired height. Canstock customers weigh the stamped disks on calibrated scales and apply the following formula to calculate the thickness Eq.3: Weight( g) Thickness ( mm) = h = Where d = diameter of disk (3) 2 πd 4ρ
Canstock customers compare disks stamped at the edge of the strip to the centreline disk, and will typically expect the supplied tinplate to have a maximum difference of 3µm, referred to as a rundown of 3. This approach to the measurement of profile effects on the final strip gauge was alien to the current hot mill assessment of crown and wedge. Hence, it was determined that to monitor profile performance for canstock customers an approach was needed that would better correlate hot mill profile performance with the final tinplate profiles. Since canstock customers measure rundown (difference between centreline thickness and edge thickness), then the same approach should be applied to the hot rolled coil (HRC). With the HRC this rundown measurement is made up of the crown and ½ wedge (half wedge since only concerned with centre to edge) Eq.4: R = Max[ C + 0.5W, C 0. 5W ] (4) The value quoted is the maximum rundown value determined for either the operator or drive side. Since HRC is subjected to cold reduction from 2.5mm to 0.20mm for the final tinplate application. Assuming no change in the percentage crown during reduction, then to supply tinplate with a rundown of 3 or less, requires the HRC to have a rundown of less than 40µm. Investigation It was clear that the original low crown WRC struggled to obtain the crown targets specified by the customer and were unable to cope with variable mill conditions, with work roll bending (WRB) forces (Fig.2) always on limits. WRB is where positive bending forces at the ends of the rolls cause the rolls to bend away from the strip as though there were more crown on the rolls. The finishing mill control system adjusts the WRB force in response to changes in the rolling force, in order to try and maintain a constant roll gap [6]. This should enable the strip to be rolled with a constant profile regardless of small changes in temperature, tension, metallurgical anomalies and other variations. Back Up Roll Roll Force (From screws and/or capsules) Work Roll Work Roll Work Roll Bending Back Up Roll Ground Crown + Thermal Crown + Wear Fig.2 Schematic diagram of the roll configuration in 4-high mill stand. The crown40 performance per round rolled of WideDWI was consistently over the desired 20µm target. On the rollings where the system was free to set WRB to lower crown (no operator interventions) it was seen that the average WRB tons were very high across the first 6 stands and in most cases WRB was at the maximum in the first 3 or 4 stands. This indicated that the profile control system had very little control and that the first step must be to lower the overall crown on the strip. Thus, it was decided to make to modifications to WRC in order to achieve a flatter strip and to lower work roll bending forces, thereby improving strip profile and making the profile control process more robust.
WRC Trials The low crown WRC Ref. Code 154 (Table 2), were concave in nature (typical of other WRC in use in the hot mill presently). Due to the low crown requirements placed upon the WideDWI product, they are rolled in rounds where the number of meters rolled is limited, before a roll change is required. This is to take into account possible wear and surface issues that may occur as the rolls slowly decay during rolling. The ground WRC are input into the mill model (see Fig.3), the accuracy of which is critical in maintaining uniform material deformation throughout the mill. Small incompatibilities between two work roll profiles can result in interstand shape (shape produced between finishing mill stands). Table 2. Original 154 Cambers F5 F6 F7 F8 F9 F10 F11 Code 154-0.090-0.060-0.030 +0.040 +0.095 +0.015-0.020 Fig. 3 Example of a ground WRC code 154 for stand 1: Target = 0.090, measured = 0.095. Step 1 involved modifying the WRC in only stands 1 and 2 in order to correct observed centre buckle in stands 5 and 6. This made the WRC in stands 1 and 2 parallel (flat). It was seen critical that once the interstand shape in the mill was at a minimum, then wholesale changes to the WRC in all stands could be made through maintaining the proportional crown (crown divided by strip thickness) through the mill. After several rollings, interstand shape was improved, and the next step (Step 2) to bring the bulk crown on the strip down could begin by making the WRC more positive (fatter). Step 2 required a small shape correction for edge work observed in Stand 6. Table 3 shows the changes to WRC made in Step 1 and 2; in addition, the roll shop code used to describe these trial cambers was changed to 300. Table 3. Steps 1 and 2 to modify low crown WRC. Code 300 F5 F6 F7 F8 F9 F10 F11 Step 1 0.000 0.000-0.030 +0.040 +0.095 +0.015-0.020 Step 2 +0.050 +0.030-0.010 +0.050 +0.110 +0.160 +0.020 Steps 3, 4 and 5 (Table 4) involved further progressive camber changes making the WRC more positive and convex in nature. Each WRC modification was made in sequence with detailed observations and mill conditions were recorded during each rolling. Operators were brought on board with the changes and actively encouraged to feedback their opinions on how well rollings went given their considerable experience. Table 4. Steps 3, 4 and 5 to modify low crown WRC. Code 300 F5 F6 F7 F8 F9 F10 F11 Step 3 0.150 0.095 0.030 0.075 0.110 0.160 0.020 Step 4 0.250 0.160 0.070 0.100 0.140 0.180 0.040 Step 5 0.350 0.225 0.110 0.125 0.140 0.180 0.040
In Fig. 4 the changes made to WRC to make them more positive is clearly visible, especially in the earlier stands. Generally, the gradient of the interstand proportional crown remains similar, along with the overall shape of the proportional crown through the mill. However, the effects of modifications to particular stands to correct for interstand shape can also be seen. This graph shows how we can change the level of the proportional crown at, or before a stand in order to remove observed shape; as in Steps 1, 2, 3 & 5. In Step 3 for example, it can be seen that in Stands 1 4, the proportional crown was raised to lower the crown exiting each stand, through doing more centre work although shape in these stands is a concern, the ratio of the strip width to gauge is small, and the hot strip will flow in the loaded roll gap both longitudinally and laterally [7], resulting in a significant change in profile. However, in Stands 5 7 the proportional crown was kept the same as Step 2. This had the effect of essentially doing less centre work and more edge work. This was due to the fact that centre work was observed in these stands due to the WRC doing too much work to lower the strip crown. In the latter stands the ratio of strip width to gauge is much lower and lateral flow of the material much smaller, hence leading to non-uniform reduction across the strip width producing a nonuniform longitudinal elongation, causing a change in strip flatness centre work in this case. 50 Proportional WR Crown 40 30 20 10 0 Original Step 1 Step 2 Step 3 Step 4 Step 5-10 1 2 3 4 5 6 7 Stand Fig. 4. Changes to proportional WRC for rolling WideDWI. Results and Discussion Fig. 5 displays the crown40 performance for WideDWI rolled before, during and after the WRC trials. The graph shows only rollings where mill conditions are comparable. Fig. 5 Run chart for WideDWI Crown Performance before, during and after the WRC trials
The performance on the original WRC set (unfilled boxes) was significantly off-target with a sizeable spread. The immediate impact of the Steps 1-3 is evident in bringing the crowns down to just above target. Step 4 showed slightly mixed results with the last two rollings on or below target. Step 5 gave the performance that was desired. It was also decided that after five changes to WRC a more long-term analysis of the improved crown performance was required and so this WRC set had been left in use. A significant improvement was that the target crown was obtained consistently and the deviation from target often caused by variable conditions, was considerably reduced. One of the key aims of the trials was to lower the overall bending level, thereby creating more flexibility when adverse mill conditions are experienced. These may include reduced push rates, WRB limited loading differential, poorly ground work rolls, worn back up rolls, large transfer bar profiles. From Fig. 6 there appears to be an improvement in reducing the overall bending loads applied across the first 6 stands. The first 6 stands have been considered since stand 7 operates essentially as a correction stand with reduced loading required. Changes to WRC s Steps 1-5 Fig. 6 Work roll bending levels during and after the camber trials Rundown Monitor As stated above, the main aims of the trials were to lower crown on the strip and ultimately reduce the rundown value that the canstock customer receives. Using the hot rolled coil rundown value as a guide should result in a correlation in with an improvement in the customer monitor. The hot rolled coil calculated rundown (Eq.4) could readily be used to assess which component of the rundown can be improved to lower the overall rundown. This has led to the development of an intranet based rundown monitor (Fig. 7) that shows on a coil-by-coil basis the rundown performance during rolling and is updated every 60 seconds (virtually live). This tool has allowed the mill teams/operators to see easily which component, crown or wedge, to attack to lower rundown This is supported by a handy guide as to what to do (generally) and links to supporting documents. A driving force behind this monitor was the need to develop a meaningful monitor that the operators could use to continuously improve performance during a rolling, but just as important was the concern that lowering our crown performance onto target, or as can be seen, below the target, may lead to (the commonly held belief) of increased strip wander [8]. Strip wander (or walking) can enlarge wedge and camber increasing the rundown and possibly causing work roll surface damage. The monitor allows operators to react more
quickly if wander occurs and what the possible cause of the wander could be (crown dipping to zero). Fig.7 Snapshot of PT HRP DWIwide rundown assessment intranet page. Rundown Correlation The final stage of the trials was to obtain feedback from the customer to ascertain whether the reduction in the hot rolled coil rundown correlated with the canstock customer assessment of rundown on the final tinplate product. Fig. 8 shows the correlation between the rundown on the hot rolled coil and that measured through weighing blanks (disks) of tinplate at the canstock customer. 5 60 4.5 Approximately a 2 Week Lag Between HRP Processing & Canstock Processing. Canstock Rundown Aim = 2.8µm HRP Rundown Aim = 30µm 55 4 50 Canstock Rundown. 3.5 3 2.5 45 40 35 2 1.5 1 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul HRC Rundown. 30 25 20 2004 2005 Hot Rolling Week Canstock Rundown HRC Rundown Fig. 8 Correlation between hot rolled coil (HRC) rundown and canstock customer rundown. The canstock customer randomly selects positions within the tinplate coil to punch out the centre and two edge disks for weighing, with an average performance value plotted for a sample size of 50. Although identifying individual coil performances or rolling performances
can be undertaken, Fig. 8 allows the bulk improvement that the customer has obtained to be clearly seen. The canstock customer requires that improvements made are wholesale. They can then re-engineer their process in order to maximise the improvements, through reduced edge discards (improved yield) for example. The improvements made in lowering the rundown value have led to the canstock customer re-calculating their control limits and aim. Conclusions Plant trials involving modifications to work roll cambers was undertaken to lower the crown on hot rolled coil, and subsequently improve the rundown on cold-reduced tinplate supplied to canstock customers. The following results were obtained: 1. After 5 step changes to the WRC the mean crown performance per rolling was brought down onto or below the target crown. This had led to significant step change improvement in rundown at the canstock customer. 2. The improvement in crown performance through changes to the WRC has also led to a lowering of the work roll bending force down from the limits, thereby creating a more flexible process that can react to variable mill conditions. 3. The new approach to calculating rundown and the use of the rundown web page have helped to target the key components to improving the rundown, whilst also simplifying the process and improving the feedback to operators. 4. Through the process involved in modifications to WRC, an insight has been gained into the strip crown and its interaction with the mill stands. This process and technology could be applied to other hot rolled coil customers with tight gauge tolerance requirements. References 1. BSEN 10051:1992 (Issue 2, 1998) Continuous hot-rolled uncoated plate, sheet and strip of non-alloy and alloy steels Tolerances on dimensions and shape. 2. T. H. Kim, W. H. Lee & S. M. Hwang, An integrated FE process model for the prediction of strip profile in flat rolling. ISIJ International, Vol. 43 (2003), No. 12, p 1947 1956. 3. M. Veryard. Internal Document: Set-up Profile and Flatness Strategy. 4. S. Ogawa, H. Matsumoto, S Hamauzu et al. On-Line Calculation Method of Strip Crown in Hot Strip Mills. Journal of the Japan Society for Technology of Plasticity, Vol.25. No.11 (1984), p1034-1041 (in Japanese). 5. How Food and Drink Cans Are Made: www.mpma.org.uk/design/overview.html 6. B. Beal. Rolling Mill Gauge Control Components (White Paper), Pfeiffer Engineering Ltd. www.pfeiffereng.com/gauge%20control%20components.pdf 7. Y. Hongo et al. Hot strip crown control utilizing crown-ratio hereditary distribution in the lateral direction. Journal of the Japan Society for Technology of Plasticity, Vol. 38, No.433 (1997), p165-170. 8. T. Shiraishi et al. Relationship between camber and wedge in flat rolling under restriction of lateral movement.. ISIJ International, Vol. 31 (1991), No. 6, p 583 587.