| By David W. Manthey, Product Development Rebecca M. Pearce, Business Development CamSys, Inc., Troy, NY, |
Dr. Daeyong Lee, Professor of Mechanical
Engineering Rensselaer Polytechnic Institute, Troy, NY |
One of the most common analysis and evaluation methods used for the sheet metal stamping process is measuring the extent of deformation in critical areas on a stamped part. The extent of deformation is sometimes related to forming severity, where the level of deformation is categorized as safe, marginal or failure. In many press shops, strain measurement is used as a way to assess the formability of a stamping.
Fig. 1 -- Mylar tape used to measure major and
minor strains from deformed circles.
Surface strain data can be used effectively to diagnose production problems, and identify potential failure sites and contributing factors. These data also are used to verify predicted results from finite element analysis programs (FEA). Many techniques have been developed for measuring strain on sheet metal parts. To use these techniques, it is usually necessary to apply a pattern to the sheet surface before forming. Patterns of circles or squares are commonly used [1,2]. An example of a common manual strain measurement tool, the mylar tape, is shown in Fig. 1, and a description of strain measurement using circles is shown in Fig.2 [3].
Fig. 2 -- Measurement of surface strains
using circles.
During the stamping operation, the grid pattern deforms with the material, then strain can be measured from the deformed pattern by comparing it to the original size of the pattern. Most of the strain measurements done today are performed manually, where the maximum and minimum lengths across a deformed circle are measured. Types of tools used in manual measurement include traveling microscopes, mylar tape with graduations, and ruler and dividers. The strain data can then be plotted on a forming limit diagram (FLD). This diagram relates measured strains to the material formability limits. Analyzing the data on an FLD is the most common way to evaluate the formability of a stamped part [4,5].
Since manual strain measurement methods can be labor intensive and the results are significantly affected by user interpretation, considerable effort has been directed toward developing automated methods for strain measurement. One method that represents the latest developments in this technology is the Automated Strain Analysis and Measurement Environment (ASAME). This system has been described in previous papers [6,7].
Currently, ASAME is being used in either a production or research environment at Ford, General Motors, Motor Wheel, ALCOA, Fiat, Peugeot, Hyundai, Nissan, Toyota, Nippon Steel and many other leading companies throughout the world. Customer applications include production troubleshooting, verification of predicted results, die tryout and development of material forming limits.
To measure the surface geometry and strain on an area of a formed (stamped) part with ASAME, first a square or circle grid is applied to the surface of the sheet material before forming. Subsequently, ASAME is used to take several photographs of an area on the part. Advanced image processing software is used to identify and compute the three-dimensional coordinates of intersecting points on the grid pattern [6,7].
Surface geometry and strain can be measured using two different ASAME methods. Both ASAME methods require that a grid pattern is present on the surface of a part before deformation.
Fig. 3 -- Position based
measurement system consists of a CCD camera mounted on positioning equipment, and a
computer capable of accepting video input and positioning encoder data.
This measurement system consists of a CCD camera mounted on positioning equipment, and a computer capable of accepting video input and positioning encoder data, Fig. 3. The part to be measured is placed on a turntable, and then two or more photographs are taken from different angles. The photographs then are processed through the use of computer software to locate the 2-D grid coordinates. The 3-D coordinates are determined using the 2-D coordinates, camera position and internal camera parameters. The internal camera parameters are determined though a calibration procedure prior to measuring a part.
In general, the position-based system yields an accuracy of about ±2 percent strain using two photographs and ±1.5 percent strain using three photographs per measurement.
Fig. 4 -- Recently developed target-based
system for measuring printed grids does not require knowledge of camera position or
internal camera geometry.
A new measurement method (Method B) has been developed recently that does not require knowledge of the camera position or internal camera geometry prior to making a measurement. This method uses a Kodak DCS420c digital camera, a computer capable of reading PCMCIA hard drives, and a photogrammetry target. The photogrammetry target is an object of known dimensions with easily identifiable markings: a 25 mm cube with a 5 mm grid machined into each side, Fig.4.
The target cube is placed next to the part to be measured, and photographs then are taken from two or more locations using the digital camera. Each photograph must include the target as well as the area to be measured. The Kodak camera stores the photographs on a PCMCIA hard drive, which then is removed from the camera and inserted into the computer for processing by the user with the ASAME software to generate the surface geometry and strain data.
In general, the target-based system yields an accuracy of about ±2.5 percent strain using two photographs and ±1.5 percent strain using three photographs.
Fig. 5 -- New
target-based ASAME can be used for measuring coordinate data by projecting a grid onto the
surface of the part. Thus, grid does not have to be permanently applied to the surface.
The new target-based ASAME also can be used for measuring coordinate data by projecting a grid onto the surface of the part, Fig 5. In this way, a grid does not have to be permanently applied to the part surface. A square grid from a 35 mm slide is projected on the object, and the Kodak DCS is used to take two or more photo-graphs of an object from any orientation. The photographs then can transferred to the computer for processing as in Method B. [7]
Fig. 6 -- A sample photograph from
projected grid measurement of an automobile door panel.
To measure surface geometry, the position-based measurement system requires that a grid be permanently applied on the original surface of a part. In practice, a non-contact method of making measurements is desirable, where the material surface is not affected. The new target-based measurement system can be used to measure the surface geometry of a non-gridded part by projecting a grid of light from a slide projector onto the part area to be measured.
An area on an automobile panel was measured with the target-based system using both a printed grid and a projected grid. The grid was projected such that the squares on the part surface were a similar size to those of the printed grid. However, the two grid patterns were not aligned, nor were they identical in size. Each of the two measurements was made with five photographs. The target was not in exactly the same position for both measurements. A sample photograph from the projected grid measurement is shown in Fig. 6.
Fig. 7 -- B-car MBR front side INR
forming limit diagram: material SPRC35-E is shown on the left and material SPRC35-R is
shown on the right.
For many CMMs, there are at least two major issues to be examined: one is the accuracy and the other the time required to produce satisfactory data. As far as the accuracy is concerned, most of the commercial mechanical systems produce high accuracy. Many manufacturers state a linear accuracy range of 0.025 mm to 0.0125 mm [8]. One manufacturer claims an accuracy of 0.1 micrometer [9]. A laser-based system reports an accuracy of 0.005 mm to 0.0125 mm [9].
A summary of the comparison of the two
methods for coordinate measurement is given in Table I for several main issues.
| Issues | DCS Camera System | Conventional
CMMs Probe-Based and Laser-Based) |
| Linear Accuracy | ±0.05 mm to 0.3 mm1 | ± 0.1 um to 0.0125mm |
| Measurement Area | Based on Camera field-of-view2 | Limited by positioning hardware |
| Measurement Time | -20 minutes per area | Relatively long time
(dependent on number of probe contacts) |
| Portability | Completely portable | Not Portable |
| Cost | $50,000 - $60,000 | $100,000 - $500,000 |
| Additional Features | Surface strains can be measured if a grid was applied to the underformed blank | No strain measurement capability |
| 1Depends on the number of photographs used to make measurements and the size of the field of view. | ||
| 2 Areas larger than the field of view can be measured by combining smaller measurements. A larger target can be used to allow a larger area to be measured. | ||
Fig. 8a
-- Current material strain data overlaid on new material FLD. All points are within the
safety limit for the new material.
In 1994, Hyundai Motor Corporation published a paper entitled "A Problem Solving Plan in Automotive Panel Forming Using the Automated Strain Analysis and Measurement Environment" [10]. In one case study outlined in this paper, a high strength steel (SPRC35-R) was used to produce the MBR front side of the B-car. It was observed that tearing occurred frequently during the final stages of the flanging operation.
To verify a change in material to SPRC35-E, surface strain measurements were made with ASAME, and plotted against the forming limit diagram (FLD) for two materials, SPRC35-R and SPRC35-E, as shown in Fig. 7. The FLD is a graphical plot of the experimentally determined forming limit curve for a material, where major strain is plotted on the vertical axis and minor strain is plotted on the horizontal axis.
Measured strain data can be plotted against the forming limit curve to show how close a part is to failure. In this case, the FLD plot indicated that the R-material had about 13 percent Safety Strain and the E-material showed Safety Strain of approximately 21 percent. Since the E-material had the highest Safety Strain, the decision to change to the E-material was verified.
Fig. 8b
-- Strain data for new material's critical area plotted with the new material forming
limit. All points are predicted to be within the safety limit.
ASAME has been used by Phoenix Consulting at a stamping plant to generate quantitative data to support decisions on material change. Before a material change was made for a part, the most problematic areas on a part were measured with ASAME to confirm that the strain levels in those part areas were below the safety limit on the material's forming limit diagram. These same strain data then were automatically overlaid on the new material's forming limit diagram calculated from material properties of the proposed new material, as shown in Fig. 8a. The data points were in the safe zone of the new material's FLD indicating that strain levels should be adequately safe to allow use of the new material.
To support the decision to change material, a blank of the new material was likewise formed and evaluated using ASAME to confirm that strain levels remained safe. The measured strain data were plotted on the calculated FLD for the new material, as shown in Fig. 8b.
The (b) plot shows that data points remained in a similar shape and position as that of the old strain data, indicating that material flow and strain distribution were not significantly altered by any differences in properties between the two materials. Strain levels remained in the safe zone, so a decision was made to change material. In this case, ASAME provided a technical basis for decision making, reducing the risk of wasting time and money on an unsuccessful change.
A new vision-based surface measurement system has been used to determine the surface profile and surface strains of a part using a square or circle grid pattern. Surface profiles also can be measured by projecting a grid onto the part. This new portable system uses a Kodak DCS420c camera, a laptop computer, and a small photogrammetry target. Multiple digital photographs are taken, which then are used to locate the three-dimensional coordinates of the grid intersections.
The capabilities of the digital camera-based system have been compared to commercial CMM systems. While the accuracy of the digital camera-based system is substantially lower than that of mechanical systems, measurement times are much shorter, surface strains can be measured, and the vision-based system is completely portable. Applications of the surface geometry and strain measurement system include production troubleshooting and material selection. MF
| References |
|