What Negative Torque Means on a Wide Heavy-Plate Straightener
On a seven-roll pre-straightening machine downstream of a wide heavy-plate mill, every roll is driven and every roll carries a share of the bending load applied to the plate. In normal operation the drive motor of each roll delivers positive torque: it pushes the plate forward and simultaneously resists the reaction from the bending moment. Negative torque is the condition in which the load torque on a particular roll reverses sign, so the roll is being driven by the plate rather than driving it. At that moment the roll-to-plate contact can break down, the roll skids on the surface, and the machine generates a characteristic squeal together with a visible vibration mark.
The importance of this phenomenon is not acoustic. When a roll alternately grips and slips, the torque transmitted through the drive train fluctuates, the roll surface is damaged by the resulting localised contact stress, and the flatness and residual stress state of the delivered plate become inconsistent from head to tail. In wide, thick plate, where the bending loads are large and the plate is stiff, the effect concentrates on a small number of rolls. Operators on such lines have reported the most severe vibration marks on the sixth roll of a seven-roll set, with the first and seventh rolls showing almost no marking and the middle rolls showing only slight marking, which points to the geometry of the bending schedule rather than to a general drive fault.
Kinematics of Seven-Roll Pre-Straightening
A seven-roll straightener applies a sequence of alternating bendings to the plate. The entry rolls impose a large plastic curvature of one sign, the middle rolls reverse the curvature, and the exit rolls reduce the curvature progressively until the plate leaves with a flatness error inside the acceptance limit. The amount of bending at each roll is set by the roll gap and by the intermesh, that is, the depth to which each upper roll is pressed below the pass line defined by the lower rolls. Because the plate is a continuous body, the curvature imposed at one roll constrains the deformation at its neighbours, and the resulting bending moment distribution is not symmetric across the roll set.
Three geometric parameters govern the load pattern. The roll pitch determines the length of the bending span and therefore the strain level for a given intermesh. The roll diameter determines the minimum curvature that can be imposed without exceeding the contact stress limit of the roll material. The plate thickness relative to the roll pitch determines whether the deformation is dominated by bending or by shear; in wide thick plate the bending component dominates, and the plate behaves as a stiff beam that transfers load several rolls away from the point of application.
This load transfer is the origin of the problem. When the entry rolls are heavily engaged to break the incoming crown and the residual curvature, the reaction on the following rolls can be large enough that, for a short interval, the plate's forward motion pulls a roll faster than the drive intends, producing the torque reversal.
Causes of Torque Reversal: Speed Mismatch, Bending Difference and Roll Wear
The classical explanation for negative torque on a group-driven straightener is the conflict between the nominal synchronous speed of the drive group and the actual surface speed demanded by the plate. When several rolls are connected to one motor through a common shaft or a fixed gear ratio, all rolls turn at the same angular velocity, so their surface speeds are identical only if their diameters are identical. In practice the rolls wear at different rates, are reground to slightly different diameters, and deflect differently under load. A roll whose effective diameter is 0.5 percent larger than its neighbour tries to move the plate faster than the neighbour allows, and the resulting internal shear in the plate generates a torque that can reverse the sign on one of the two rolls.
The second cause is the non-uniform bending schedule itself. Where the intermesh is set aggressively to correct a heavy cross-bow, the bending moment at the adjacent roll changes rapidly along the pass, and because the plate is continuous the change appears as an additional torque component on that roll. Measurements taken with strain-gauge bridges on the transmission shafts of a seven-roll pre-straightener, cross-checked against a finite element model of the plate-roll contact, have shown that the negative torque appears precisely where the imposed bending differs most sharply from the neighbouring pass.
The third contributor is the difference between the plate's entry profile and the machine's design assumption. A wide plate with a pronounced centre buckle or a wedge-shaped cross-section presents a non-uniform thickness to the roll gap, so the two ends of the roll see different reductions and the roll is loaded asymmetrically. That asymmetry is transmitted through the roll body into the drive as a fluctuating torque.
Consequences for Roll Surface, Plate Flatness and Residual Stress
When negative torque develops, the contact condition at the roll changes from rolling to mixed rolling and sliding. The friction shear stress at the surface rises sharply, the lubricating film of oxide scale breaks down, and the roll surface is subjected to a localised thermo-mechanical cycle that produces the familiar banded marks. Once a roll surface is marked, the marks are reproduced on every subsequent plate, so a single event can degrade several hundred tonnes of production before the roll is changed.
| Observable symptom | Probable mechanism | Process consequence |
|---|---|---|
| Vibration marks concentrated on one roll | Torque reversal and slip at that roll | Roll surface damage, repeated marking of subsequent plate |
| Alternating torque peaks on adjacent rolls | Bending schedule too aggressive at entry | Uneven flatness between head and tail |
| Audible squeal with no hardness change | Slip without plastic damage to the plate | Cosmetic defect, but indicates roll wear |
| Flatness good at the exit, poor after cutting | Residual stress imbalance from asymmetric bending | Distortion of laser-cut or plasma-cut parts |
| Roll diameter scatter above 0.3 percent | Unequal surface speed in a group drive | Chronic low-level torque fluctuation |
The flatness consequence is subtler than the visual defect. A straightener corrects flatness by imposing a controlled pattern of plastic strain through the thickness. If slip occurs at one roll, the strain pattern is interrupted, and the plate leaves with a longitudinal wave that is not visible on the straightener but appears when the plate is cut into blanks, because the internal stress balance has been disturbed. The residual stress distribution after straightening therefore becomes non-uniform along the length of the plate, which is far more damaging to a fabricator than a small flatness error measured at the mill.
Measurement: Strain-Gauge Torque Telemetry and Finite Element Cross-Check
Diagnosing negative torque requires measurement rather than inference. The standard approach is to bond strain-gauge bridges to the transmission shaft of each roll drive, arranged in a full-bridge configuration that cancels bending and axial strain and responds only to torsion, with the signal transmitted through a slip ring or a wireless telemetry unit. Calibration is carried out with a known static torque so that the recorded millivolt output can be converted to newton-metres.
The measured torque history is then compared with a finite element simulation of the same pass. The model represents the plate as an elastic-plastic continuum with a yield criterion and isotropic hardening matched to the actual grade, and represents the rolls as rigid or as elastic bodies with the measured diameters. The comparison is used in two directions: the measured torque validates the friction coefficient and the intermesh values assumed in the model, and the model explains why a particular roll shows reversal by revealing the bending moment distribution that the measurement alone cannot show.
Two practical points improve the quality of this work. First, the strain-gauge installation must be protected from coolant and scale dust, and the zero reference must be re-recorded before each pass, because thermal drift on a hot line biases the reading. Second, the sampling rate must be high enough to capture the slip events, which last a fraction of a second; a low-rate logger records only the mean torque and misses the reversal entirely.
Corrective Measures: Speed Compensation, Intermesh Revision and Roll Maintenance
The most direct remedy is to remove the speed conflict. Where the drives are group-coupled, converting the offending roll, or the pair of rolls either side of it, to an individually controlled variable-frequency drive allows the surface speed to be trimmed to match the plate rather than the shaft. In the published case of a seven-roll pre-straightener, increasing the speed of the second motor by 30 percent alleviated the negative torque on the sixth roll, eliminated the torque surges on the third and fifth rolls, improved the flatness of the straightened plate and produced a more uniform residual stress distribution after straightening.
The second remedy is to spread the bending schedule. Reducing the entry intermesh and moving part of the curvature correction to the middle rolls lowers the peak bending moment that drives the reversal, at the cost of requiring more passes or a larger machine for the same total correction. The optimum is found by running a set of trial plates at several intermesh patterns and recording both torque history and post-straightening flatness, rather than by adjusting the schedule on the basis of exit flatness alone.
The third remedy is roll maintenance discipline. Rolls should be measured for diameter and profile at every regrind, and the set should be matched so that diameter scatter within a drive group stays below about 0.3 percent. Roll bearing condition, coupling backlash and the condition of the roll cooling system all affect the friction coefficient and therefore the slip threshold; a dry, scaled roll surface slips at a much lower torque than a clean one.
Operating Envelope and Grade-Specific Considerations
Wide thick plate is not a single material. A 400 HBW abrasion-resistant plate per GB/T 24186 has a yield strength several times that of a structural Q355 plate of the same thickness, so the bending force required to reach plastic curvature is proportionally higher and the elastic springback after the pass is larger. The straightener schedule must therefore be written per grade and per thickness, and the schedule for a 500 HBW plate cannot simply be scaled from the schedule for a structural plate of equal thickness.
Temperature is the other variable in the envelope. Pre-straightening is carried out on plate that still retains heat from rolling, and the flow stress at 700 degrees Celsius is a fraction of the room-temperature value. A schedule that is safe at the top of the temperature window can overload the drives if the plate arrives cooler than expected, which happens when the transfer time between the mill and the straightener is extended. Tracking plate temperature at entry and interlocking the intermesh setting with the measured temperature is a reliable way to stay inside the drive torque limit.
Acceptance, Flatness Criteria and Documentation
Straightened plate is accepted against a flatness specification, commonly the flatness classes of EN 10029 or the equivalent clauses of GB/T 709 for hot-rolled plate, measured as maximum deviation over a defined gauge length with the plate on a level table. Surface condition is assessed against EN 10163 or the equivalent requirement in the purchase order, which is where the vibration marks from a slipping roll would be classified, usually as a surface defect that requires dressing or rejection depending on depth.
Good practice is to record, for each plate, the grade, thickness, entry temperature, intermesh pattern, measured torque history summary and exit flatness. That record lets a mill correlate a flatness excursion with a specific drive condition before the roll surface is damaged, and it provides the evidence trail a purchaser needs when a claim arises over cut-part distortion. Where the plate is subsequently cut into wear parts, an additional check of dimensional stability after cutting is more informative than a second flatness measurement on the full plate.
FAQ
Q: What causes negative torque on the sixth roll of a seven-roll pre-straightener?
The reversal is produced by a mismatch between the surface speed demanded by the plate and the speed imposed by the group drive, combined with a sharply changing bending amount at that pass. The plate then drives the roll instead of the roll driving the plate, and slip marks appear.
Q: How is roll torque measured on a hot straightening line?
Strain-gauge bridges bonded to the transmission shaft in a torsion-only full-bridge arrangement, with slip-ring or wireless telemetry, calibrated against a known static torque and zeroed before each pass to remove thermal drift.
Q: Does negative torque affect plate flatness?
Yes. Slip interrupts the intended pattern of plastic strain through the thickness, so flatness becomes inconsistent along the plate and the residual stress distribution after straightening is less uniform, which shows up as distortion after cutting.
Q: What roll diameter scatter is acceptable within one drive group?
Keeping diameter scatter below roughly 0.3 percent within a group-driven set avoids the systematic surface-speed difference that promotes torque reversal. Rolls should be measured at every regrind and matched before assembly.
Q: Why does a speed increase on one motor reduce torque on a different roll?
Because the rolls are mechanically coupled through the plate. Correcting the speed of the roll that is fighting the plate removes the internal shear that was being transmitted several rolls downstream, which relieves the reversal at the affected roll.
Q: Can the intermesh schedule be adjusted to avoid negative torque?
Yes. Spreading the curvature correction over more passes instead of concentrating it at entry lowers the peak bending moment, which raises the slip threshold. The revised schedule should be validated by torque measurement and by flatness and residual stress checks, not by exit flatness alone.




