Online Roughness Measurement in a Coil Line

Size: px

Start display at page:

Download "Online Roughness Measurement in a Coil Line"

Thomas Sherman
5 years ago
Views:

1 Online Roughness Measurement in a Coil Line Wolfgang Bilstein Amepa GmbH Karl-Carstens-Str Aachen, Germany Tel.: Fax: wolfgang.bilstein@amepa.de Kevin Bertermann Amepa America Inc Engle Rd. #27C Cleveland, OH Tel.: +1 (440) Fax: +1 (440) info@amepa.com Key words: Roughness, Online Roughness Measurement, EBT, EDT, Pretex INTRODUCTION Rising demands of the automotive industry in quality of sheet surface make it necessary to control continuously the roughness of strip steel. Moreover, the optimization of the deep drawing process requires sharp tolerances in the tribological behavior of the material to achieve higher production rates. High-strength steel and aluminum sheet can only be deformed without defect under optimized conditions. Homogeneity of the surface roughness is an important factor to reach the best product. The on-line measurement of the roughness parameters enables not only a complete documentation as expected by ISO 9000 but also a more economic production. The immediate availability of roughness data allows the direct correction of process parameters if the quality of the surface roughness is out of its tolerance range. This avoids a cost and time intensive second milling of the product. The system presented here for the on-line measurement of surface roughness is based on approved measurement technique. The new aspect is the adaptation of this technique for on-line measurements under the rough conditions of a cold rolling mill. Today the roughness is measured on samples taken from head or tail of the coil. This measurement does not give a representative value for the roughness distribution over the length of the coil because rolling force and speed are reduced when the welding passes the skin pass. The optical on-line roughness measurement provides a measurement over a significant part of each coil and could be used for a more complete documentation of this important product property. The on-line measurement could not only be used for sharper production tolerances but it could also be used as a tool for a more precise trigger for the change of the texturing rolls. Here we will present the development, laboratory results and on-line measurements of the industrialized version by Amepa based on the original development by the Center for Research in Metallurgy (CRM) in Liège, Belgium /1, 5/.

2 MEASURING PRINCIPLE The presented roughness measurement system uses the two-dimensional triangulation, also known as lightsection measurement. The CRM has refined this principle for the on-line roughness measurement on strip steel. Successful tests were done at Arcelor Flat Carbon Steel, Gent /1/. For the roughness measurement a laser line is projected under a certain angle onto the surface. Due to the surface topography this line is distorted if it is observed under a different angle than the projection. This distortion allows a direct calculation of the surface profile and hence the surface roughness. Figure 1 shows the correlation between distortion of the line and the topography of the surface. A more detailed description can be found in /1/. Figure 1 Geometrical correlation between topography and line distortion /1/ As an example the projected laser line is shown in Figure 2 on an EBT texturized galvanized surface. The background illumination is not necessary for the measurement but gives the user the opportunity to validate the result. This example shows, that even the fine structures of the zinc layer can be resolved by the distortion of the laser line. Figure 2 Galvanized EBT textured surface with distorted line; Image area 530µm x 200µm.

On-line roughness measurement systems that use a laser spot for the roughness measurement depend on the movement of the strip to achieve data in the second dimension.

3 On-line roughness measurement systems that use a laser spot for the roughness measurement depend on the movement of the strip to achieve data in the second dimension. Therefore these devices are only able to measure parallel to the rolling direction. The light section principle used in the Amepa SRM - allows an arbitrary orientation of the line and so roughness can be measured perpendicular or under 45 to the rolling direction as SEP 1940 and EN /3/ demand. The other advantage of the line projection is the fixed correlation between all points of the line. When the exposure time is short enough the system is not affected by vibrations or the velocity of the strip. SYSTEM DESIGN The sensor consists of a laser system for line projection and of a microscope with a high resolution CCD-camera for the observation of the distorted line. An additional laser is used for background illumination enabling a direct inspection of the microscopic image of the strip surface which can be done normally only in laboratory. The used diode lasers are equipped with special driver electronics which enable not only the control of the pulse energy but also the variation of the pulse duration which could be decreased down to 8 ns. This means in combination with the used optical setup that even at speeds up to 1200 m/min no reduction of image quality could occur. The assumptions in the above mentioned estimation are very conservative, because they are valid for the microscopic background image. We can expect that two times higher velocities should not affect the result of the measurement itself. The other advantage of this electronics is the possibility to set the illumination of the laser diodes to the optimum point for each sheet surface with very different scattering properties from nearly black to highly reflective. Excentricity of Roll (blue) Trigger Level (red) Captured Image Figure 3 Schematics of the SRM system

4 A very high resolution and magnification of laser line and image is necessary to measure surface structures in the micrometer range. The small depth of field of the required microscope objectives leads to high demands in the positioning accuracy of the sensor. To match this task, the sensor is equipped with a precise distance sensor which triggers the lasers and the image acquisition process when the strip surface is within the focus of the microscope. The sensor is placed above a roll where the strip is held in a defined position. To make sure that the strip is well aligned, an angle of contact of 5 degree is sufficient. Would the roll be perfect, the system would focus once on the sheet surface and could acquire images continuously. But in fact all rolls have an excentricity. As an example the distance between sensor and strip is shown in Figure 3 over two revolutions. In this case the amplitude is 200µm with an diameter of the roll of 800 mm. In Figure 3 also the trigger level is marked as red line which is the distance between strip and sensor where the strip is in focus of the microscope. The lasers are triggered and an image is taken when the measured distance matches the trigger level. Due to the noise in the distance signal, which is probably caused by the surface roughness as well, a trigger-unit inside the sensor does some data processing, to calculate averaged values without loosing the dynamics. Afterwards the image processing software the roughness parameters can be calculated as described in the previous section. For the necessary adjustments of the optical system fully motorized stages are used. This means the settings for installation or corrections can be changed whenever necessary. No production stop is required. An installation at nearly inaccessible locations is possible as well. These properties guarantee a high availability and a flexible operation. LABORATORY RESULTS The laboratory results can be divided in three sections. The most important are the comparative measurements with the stylus on different textures and surface qualities over a wide range of roughness. Optical roughness measurements on roughness standards used for the calibration of stylus systems are also presented. These two groups allow a comparison of the optical measured statistic roughness values with the results of a stylus and with the values of the calibration standards. For a qualitative measurement the optical sensor was also used in a scanning mode which is only possible in laboratory to get a three dimensional image of the surface. The plausibility of the measurement principle can be shown better by the 3D-scan than with the analysis of the statistical values. Roughness Standards The stylus measurement requires a measurement length of 12.5 mm. Therefore we used the average value of 24 optical measurements with a laser line length of 530µm each. The results of the optical measurements on roughness standards are listed in table 1. The standards were chosen to cover the roughness range important for strip steel production. Table 1 Optical measurement on roughness calibration standards /5/ Roughness Standard nominal Ra [µm] optical measured Ra [µm] Halle KNT 2058/ Halle KNT 2058/ Mitutoyo Specimen No

5 Sheet Samples The results of laboratory measurements with the stylus and the optical system show a very good correlation (Figure 4). The investigated samples cover a wide range of roughness, but also stochastic and deterministic textures (EDT, Pretex, EBT) and different surface qualities (cold rolled, annealed, galvanized) as well. This collection of 70 different samples was measured with the same optical setup and identical software settings. This means that there are no texture specific settings necessary. The only parameter which has to be adjusted is the laser intensity to avoid over- or underexposed images due to the very different reflectivity of the analyzed surfaces. Only for roughness values smaller below 0.8µm the deviation is higher than 10%. For most samples the deviation was less than 5% which is in the same order than the deviation of stylus measurements done by different operators with different stylus systems /4/. Figure 4 Comparison of the optical roughness measurement (SRM) with stylus on different sheet samples with different textures and surface qualities 3D-Scan of Texturized Sheet Surfaces To prove the plausibility of the detected lines we used this system to scan over sheet surfaces. Using a step width of 1 µm complete three-dimensional data sets of different surfaces were acquired. One example of such a data set is presented in Figure 5. The lateral resolution is 1µm and the resolution of the height profile is approximately 0.1µm which is of course less than can be obtained with interferometric systems /2/. Due to the high reflectivity of galvanized surfaces they are very interesting for a system which analyses the scattered light. In Fig. 3.2 the 3d-measurement on a galvanized EBT sample is shown. The color-coded 3d-scan shows, that 100% of the strip surface is detected. The clearly visible fine structure gives a good impression of the resolution of the system.

Figure 5 SRM on-line sensor moved in steps of 1 µm over a galvanized EBT surface to generate a 3d-data set of an area of 1mm x 0.6 mm. The unfiltered raw data are shown.

6 Figure 5 SRM on-line sensor moved in steps of 1 µm over a galvanized EBT surface to generate a 3d-data set of an area of 1mm x 0.6 mm. The unfiltered raw data are shown. The EBT and the fine structures of the zinc layer are visible. ONLINE MEASUREMENTS For on-line measurements the system was installed at Arcelor Flat Carbon Steel, Gent. These online measurements were overseen by the Rolling Annealing Finishing (RAF) group of Arcelor Research. The Amepa system was tested on different production lines. In the beginning the sensor was installed behind the skin-pass at the exit of the continuous annealing line (CAPL). To test also its performance on polluted surfaces the sensor was installed at the entrance of the CAPL. Since spring 2006 the system is installed at the exit of the galvanizing line at Sidmar. Exit of Continuous Annealing Line For the first on-line test the prototype of the roughness measurement system was installed at the exit of the continuous annealing line behind the skin pass mill. In Figure 6 the mechanical off-line and optical on-line roughness measurements of one production day are shown. For the mechanical roughness three stylus measurements were done at the end of each coil at the strip position where the optical sensor was installed. The optical measurements were saved in the Sidmar database and the mean value of the last 10% of the coil was calculated for the comparison with the stylus results.

7 Ra [µm] Amepa Ra [µm] y = x R 2 = Stylus Ra [µm] Figure 6 On-line results of one production shift at the continuous annealing line at Sidmar, Gent (green +/- 10%, blue +/-15%, red +/-20%) /5/. Entrance of Continuous Annealing Line To test whether this system could also be used on polluted surfaces, the sensor was installed at the entrance of the continuous annealing line where the surface of the coils is still contaminated from the milling process. The result of one of these tests is shown in Figure 7, where a coil was produced in the tandem mill with increasing rolling forces. Afterwards the roughness was measured in the inspection line with the stylus and than this coil was sent through the continuous annealing line where the optical on-line measurement was done Stylus Measurement Optical Measurement Rolling force [t] Figure 7 EBT surface, rolled with different rolling forces in the tandem mill: comparison of optical and mechanical roughness measurement on polluted surface at the entrance of the continuous annealing line. Exit of Galvanizing Line

8 Ra Amepa [µm] The first online test on galvanized sheet were performed at the galvanizing line (Sidgal) at Arcelor Flat Carbon Steel, Gent. The results of a step test are shown in Figure 7. A prepared coil was rolled with different elongations which were increased in steps of 0.5% starting at 0.3%. Afterwards the roughness was measured with a stylus on the inspection line. The shown results were only corrected for the delay of the optical system which depends on the measuring frequency and the averaging process. The correlation between stylus and optical roughness measurement was very good, but the slope of 0.72 was unexpected. The reason for this slope was the measurement of the optical roughness under 90 to the rolling direction instead of 45. Due to the orientation of the honeycomb pattern of the EBT structure it happened regularly that the laser line hits completely flat regions. Therefore the mean roughness is underestimated for higher roughness values. 2,0 1,5 y = 0,7199x + 0,4784 R 2 = 0,9239 1,0 0,5 0,0 0 0,5 1 1,5 2 Ra Stylus [µm] Figure 7 Step test on a galvanized EBT surface, rolled with different elongation which was increased in steps: comparison of optical and mechanical roughness To overcome the detected problems of the first prototype a new system (shown in Figure 8) was installed end of 2006 at the galvanizing line. The most important change was the new optical configuration to measure under 45 to the rolling direction. Furthermore the measuring length was increased by a factor 2 and the measuring frequency is increased to 10 Hz. This leads to a significant better dynamic response of the system to changes of roughness.

9 Figure 8 SRM Sensor at galvanizing line CONCLUSION Amepa has developed, based on the CRM prototype, an on-line roughness measurement system. Its results show a good agreement with the stylus measurements in laboratory and in on-line trials as well. It delivers roughness information over the complete length of the coil. This opens the possibility of process optimization during cold rolling and a complete documentation of this product property. The here described measuring principle enables a roughness measurement perpendicular or under 45 to the rolling direction (EN 10049; SEP 1940) unlike optical systems which calculate the roughness out of a series of punctual measurements measuring always in rolling direction. The Amepa system works independent of the investigated surface, of its roughness, coating or texture, therefore no calibration is required. For a verification of the systems performance it is possible to use the same roughness standards used for stylus systems as well. The first permanent installed system, at Arcelor Flat Carbon Steel, Gent (Belgium), was installed in June 2006 at the exit of the continuous annealing line. The second system will be installed at Aceralia, Aviles (Spain), in spring 2007 on the galvanizing line. REFERENCES 1. G. Moreas et al., Advanced sensor for on-line topography in continuous lines, 26th Journées Sidérurgiques Internationales, Paris, A.W. Koch et al., Optische Messtechnik an technischen Oberflächen, Renningen-Malmsheim: expert- Verlag, H. Sorg, Praxis der Rauheitsmessung und Oberflächenbeurteilung, Munich; Wien: Hanser, F. Schwingenschlögel, Abschlußbericht Ringversuch Oberflächenkenngrößen, Fachhochschule Ulm, W. Bilstein et al., Two systems for on-line oilfilm and surface roughness measurement for strip steel production, ATS Steel Rolling Conference, Paris, 2006

Advanced sensor for on-line topography in continuous lines

Advanced sensor for on-line topography in continuous lines Moréas Geneviève (CRM) Van De Velde Frederik (Arcelor, Sidmar) Bilstein Wolfgang (Amepa) INTRODUCTION Complex metal forming, high quality in term