The idea in one paragraph

US9778013B2 covers an inertial angle-measurement system for bending machines. A gyroscope coupled to the workpiece measures angular velocity; a processing module turns that into rotation angle; a processing unit calculates the bend angle. A second independent apparatus claim uses two sensor units, one on each portion of the workpiece. The control can then drive an additional bending movement from the measured angle.

The important design choice is the coupling. The sensor follows the sheet instead of watching it from across the machine. That removes the camera's line-of-sight problem and creates a different engineering problem: the sensor must remain in known contact with a moving flange without slipping or changing the bend.

The document

Granted publication
US9778013B2, 19 claims
Application
US14/490,743
Title
Apparatus and method for measuring a bending angle of a workpiece
Inventor
Enzo Gesuita
Assignee on the register
Salvagnini Italia S.p.A.
Priority / US filing
7 August 2014 / 19 September 2014
Grant publication
3 October 2017

Why tool position is not the angle

A press control can know ram position to a very fine resolution and still miss the finished angle. Thickness, yield strength, rolling direction and tool condition change the relationship between tool displacement and sheet rotation. Springback changes it again when the force is removed. A depth command is an input; the angle of the workpiece is the result.

Optical systems attack the result from outside: project light or read a profile, then calculate the sheet's attitude from an image. This patent attacks it from the sheet itself. The sensor measures motion of a portion of the workpiece during bending, so material variation appears directly in the measured response rather than as an error in a material model.

Claim 1 is a four-part measurement chain

ElementWhat the granted claim requiresWhat it contributes
Motion sensorAt least one sensor unit containing a gyroscopeMeasures angular velocity
CouplingA device connecting the sensor unit to a portion of the workpieceMakes the sensor follow the sheet
Processing moduleCalculation of rotation angle from angular velocity about an axisTurns rate of rotation into accumulated rotation
Processing unitCalculation of bending angle from the rotation angleProduces the value the machine control can use

In physical terms, angular velocity is how fast the flange is rotating; integrating it over the motion gives how far it has rotated. The granted claim writes that as a processing module and a processing unit rather than as a formula. Claim 11 further allows the gyroscope to include a register and the processing module that records the angular-velocity data and calculates rotation.

One sensor measures motion; two sensors measure both portions

Claim 12 is a second independent apparatus claim. It places gyroscope units on both the first and second portions of the workpiece and calculates their respective rotations before calculating the angle between them. That is a different measurement architecture from claim 1, not merely an optional extra sensor. Claims 16 and 10 connect the two-sensor and one-sensor versions respectively to a bending-machine control.

The description shows removable attachment as one coupling route and a machine-mounted support that keeps the sensor against the sheet as another. Claims 6 and 7 cover those alternatives. Adhesive and magnetic attachment appear in the description; the support assembly is the more production-like arrangement because the machine, rather than an operator, manages contact.

The accelerometer has a different job

A gyroscope measures rotation rate. An accelerometer can infer attitude from the direction of gravity when the part is substantially stationary. Claim 2 adds an accelerometer measuring along at least two orthogonal axes; claims 18 and 19 cover using that acceleration to calculate angle and taking the reading when angular velocity is close to zero.

That timing condition is essential. During rapid flange motion the accelerometer sees motion acceleration as well as gravity. The granted method therefore uses the gyroscope to identify the near-stationary condition and then activates the accelerometer measurement. The two sensors are not redundant: one follows motion, the other checks attitude when motion has stopped.

Optical and inertial measurement fail in different places

QuestionOptical measurementThis patent's inertial measurement
What is observed?Light, a profile or an image of the sheetAngular velocity and, when stationary, gravity components
What must reach the sheet?A clear optical path and a usable reflected or projected patternA mechanical coupling or support that maintains contact
Typical environmental sensitivityDust, oil, ambient light, reflection and occlusionSlip, contact loss, vibration and sensor alignment
Small-flange constraintEnough visible surface for the optical methodEnough room and access for the sensor and its coupling
Where verification belongsCalibration against known angles on real surface conditionsCoupling stability plus calibration against known angles through the motion cycle

The patent argues that its approach avoids optical sensitivity to the factory environment and can serve small workpieces. It also states that tests gave precision comparable to or better than optical systems, but the published document supplies no numerical accuracy result in that passage. The useful engineering conclusion is the changed failure mode, not an unsupported accuracy ranking.

Where the closed loop closes

Claims 10 and 16 put the measured angle into a control unit that operates the bending tools. The description gives the consequence: if the measured angle is smaller than required, the control commands an additional tool movement. The loop therefore closes on workpiece rotation, not on a second reading of ram position.

A production specification still has to name the angle that counts. A reading while the tool holds the sheet and the final angle after unloading respond to different states. Closed-loop correction during bending does not remove the need to measure the finished part after springback and define that result in the acceptance record.

What to test before buying the phrase “closed-loop angle control”

  • Show the coupling on the smallest flange in the part family. Record how the sensor approaches, maintains contact and releases without marking or pulling the sheet.
  • Run a surface and material matrix. Include the actual alloy, thickness, grain orientation and surface finish; the point of the measurement is to respond to variation, so the test must introduce it.
  • Disturb contact deliberately within a safe test plan. A slipped sensor should be detected as an invalid measurement, not accepted as a plausible angle.
  • Check the correction ceiling. Ask how many additional movements the control may make and what stops a correction sequence that is not converging.
  • Measure the unloaded part independently. Compare the control's reported angle with the acceptance measurement after release, and define the allowable difference before the trial.

For a measured example of how strongly material and model choice affect springback, see the three-point bending paper reading. For the machine mechanism that gives difficult flanges room to move, read the swiveling-blade patent analysis.

Cite it

Primary source
US9778013B2, claims 1–19
One-sensor architecture
Gyroscope + workpiece coupling + rotation calculation + bend-angle calculation, claim 1
Two-sensor architecture
Gyroscopes coupled to both workpiece portions, claim 12
Control connection
Bending-tool control using the measured angle, claims 10 and 16
Stationary acceleration reading
Near-zero angular velocity, claim 19