GB2415259A - Measuring warp in planar materials - Google Patents

Abstract

A method of measuring warp in planar materials (such as boards 4 travelling on a conveyor 2 during manufacture) measures the distance between the planar material 4 and a series of spaced sensors 3 and calculates a composite curvature for the planar material 4. The sensors 3 are provided above the conveyor 2 and measure the height of the boards 4passing below. Warp is calculated using successive trios of sensors (e.g. sensor-1, sensor-2, sensor-3) whereby distance measurements from sensors (e.g. sensor-1, sensor-3) to either side of an intermediate sensor (e.g. sensor-2) are used to calculate the warp height BE (approximately BD) for the section of board being sensed. Inaccuracies of curvature measurement due to stacking and overlapping on a conveyor which occurs in corrugated board production for example) are overcome by taking the position of overlap (board-edge) or gaps between boards into account.

Description

241 5259 Method for Detecting Warp in PLanar Materials This patent relates

to the provision of methods for detecting the existence of warp in planar materials such as board and particularly during the manufacture of the same. The invention also provides measurement techniques to improve the accuracy and effectiveness of the measurement of warp in flat board such as corrugated board, during, and after the production process. In particular the techniques arc utiliscd to remove measurement errors which can be caused by the effects of the material under scrutiny being in the form of multiple overlapping (shingled) sheets.

In the application, planar sheets of board (when manufactured) arc often not available for view as distinct single items. They are normally arranged as multiple sheets adjacent, overlapping or both. However, it is important to be able to measure whether the planar materials arc flat or not (i.e. have warp). To gain an accurate measurement of the warp in the board, it is necessary for the measurements to be confined to within the boundaries of that board alone and not to include measurements of a second board piece, or the underlying platform such as the conveyor.

its a result, in the measurement of warp of material, it is necessary to know where one piece of board ends and the next piece starts. Thus, conventionally, if distance measuremcuts arc made, it is not normally possible to distinguish whether two adjacent measurements being different is caused by warp, or the two measurements being different caused by the fact that the second measurement was on a second overlapping sheet, or on a gap between two sheets.

Warp can be defncd as follows: If the board is resting on a completely flat plane (the rcfercncc plane) then warp is the maximum deviation from a flat plane of any point on the board over a given distance. The deviation from the plane is measured as a distance perpendicular to the plane. Or alternatively, if two points on the board are joined by a straight dine, then the warp is defncd . . - A:: :. :e:: .: - a.. e.: as the maximum deviation from that straight line divided by the length of that straight line.

There are different forms of measuring apparatus known and one form is shown in figure 1 which is the subject of a co-pending patent application. This figure shows a cross section of the measuring apparatus. The warp measurement apparatus consists of a number (for example 50) of distance measuring sensors (3), arranged in a row and suspended from a beam (1), with each aiming vertically down. 'Lyle sensor row is in a straight line, and each sensor is equally spaced from its two adjacent sensors, typically 50mm apart.

F,ach sensor measures the distance to the point of material directly below it. In the diagram the dotted line (5) shows the measured distance from a sensor (3) to a point (6) on the board (4). 'late row of sensors is arranged to be above a conveyer (2), with the row being transverse to the flow of board material on the conveyer. In the diagram the conveyed board would be moving "out of the page".

At a point in time, a simultaneous measurement is taken on each of the sensors.

This group of measurements is stored in electronic memory. Collectively, this group of measurements represents a profile (in one dimension), at an instant in time, of the material passing below.

Repeated groups of measurements are taken at points in time when the material has moved a heed distance along the conveyer. The collection of repeated groups of measurements then forms a representation of the profile of the board in two dimensions, but it can be difficult to identify whether the measurements are truly representative of the board condition.

The aim of the present invention is to provide improvements in the manner in which data representing the distance of sensors from a planar material can be obtained and processed in order to provide a more reliable and indicative measurement of the boards condition with respect to the existence of warp. e.

- c:: ::: ::ë In a first aspect of the invention there is provided a method for detecting the presence of warp in a planar material, and the extent of the warp, if present, said method comprising the steps of measuring the distance of the planar material from a series of spaced sensors, calculating the height of the planar material from an intermediate sensor by using the height measurements from the sensors to either side of the intermediate sensor, repeating the step for each of the sensors, other than those at each end of the row of sensors and calculating a composite curvature figure for the planar material.

In a second aspect of the invention there is provided a method of detecting the presence of warp in a planet material and detecting that the readings obtained relate to a common planar material, said method comprising the steps of providing a row of sensors which detect the distance to the planar material at the same instant in time, said planar material moving with respect to said sensors, repeating the measurements at spaced time intervals to provide a series of sets <-if measurement data and wherein the data for each set of measurements is accumulated and the accumulated sums for each set compared, to identify whether or not the sets relate to the same planar material.

Thus if a number of measurement pairs are made along the line where two adjacent board pieces meet, then the accumulation of these reading pair differences will be distinctive, compared to other sets of readings where the accumulation will tend to cancel-out.

In a third aspect of the invention there is provided a method of detecting the leading edges of respective boards of planar material, said method including the steps of providing a linear array of sensors which are spaced across a conveyer system by which a series of said boards are moved past the sensors, said sensors provided to detect the position of the edge of the board as it passes thereunder and selected ones of said sensors readings used to provide a c. . : : :e:. : : :.: .... : predicted location of the board edge whereupon a comparison of actual and predicted board edge locations is performed.

It should be appreciated that these first, second and third aspects can be used individually or in any combination Specific embodiments of the invention will now be described with reference to the accompanying drawings wherein; Figure I illustrates one form of sensing apparatus; Igures 2 and 3 illustrate prior art measuring techniques; Figure 4 illustrates a first aspect of the invention; Figure 5 illustrates a second aspect of the convention; and Figure 6 illustrates a further aspect of the invention.

Referring firstly to Figures 2 and 3, there is shown a piece of warped board (1) resting -,n the flat surface of a conveyer (2). The warp is defined as the maximum height D, above the flat surface divided by the width of the board W. With a row of sensors as shown in F;ig.1, a simple way to measure the warp would be to read all the sensors, identify the sensor indicating the greatest height, and select that for the value of 19. 'this is not the method proposed by this patent, and is presented here for illustration purposes.

Fig.3 shows where this method fails. A piece of board which is not warped, could easily be resting on another piece of board so that it is positioned at an angle, rather than resting flat on the base plane. Hence the maximum height is : : :. : : : .. . . e . not indicative of warp and would give an erroneous indication using the previously described simple method.

In accordance with a hrst aspect of the invention and with reference to Figure 4 the following method overcomes the afore-mentioned limitation.

lhc measurement of warp (ie curvature) is achieved as follows; The diagram shows an end-on view of 4 sensors looking down on a curved piece of board (curve exaggerated for clarity). Curvature can be defined as length BD divided by length AC in the diagram. ie the maximum deviation from a straight line over a given distance.

Linc BD is perpendicular to line At,. The calculation can be greatly simplified by the tact that length of BD is approximately equal to the length of BE for small anglcs(of DBE). It is further simplified because the sensors are equally spaced. This means that the height of point B (hB) is the average of the height of point A (measured and denoted hA) and the height of point C (measured and denoted h(,).

lo hi] = (hA + hC) / 2. 'I'he height of point 1, (hE) is measured directly so the calculation of BE ( approx ED) is hence derived.

1317, = ((hA + hC) / 2) - hE Curvature is ED / AC and AC is known ( 2 x sensor pitch).

The above calculation is performed successively for sensors L1,2,31 then L2,3,41 then [3,4,5] ctc provided they are over board and not shingle or slit area. A composite curvature figure is then derived from an average of sets of calculations. The sign of the calculation will indicate up- warp or down-warp. If ce a..

r . e eee e * . the sensors are 50mm apart then the above sequence is warp over 100mm therefore this figure should be multiplied by 10 to give warp per metre.

The above procedure is repeated for multiple sheets of board, and further averaging performed.

Turning now to Figure 5 there is shown an aerial view of three pieces of shingled (overlapping) board and 4 sensors above the board taking height readings from points on the board directly below. The 3r board is resting on the 2"0 board, and the 2"6 board is resting on the 1St board. The horizontal movement of the board is represented in this diagram by movement from the bottom to the top of the page. Hence at an instant in time sensors sensor!, sensory, sensor3 and sensory would take a simultaneous reading at the marked points 1A, 2A, 3A, and 4A respectively. At a later instant the board has moved, and the sensors take simultaneous height measurements at points 1B, 2B, 3B and 4B respectively. And at a yet later instant the sensors take simultaneous height measurements at points 1C, 2(,, 3C and 4C respectively.

In theory it should be possible to recognise the shingle with a sensor because the height of the board should be greater just after the shingle passes under the sensor. But in practice, a combination of factors (such as thin board, vibration in the conveyer, warp in the board, and measurement inaccuracies) may mean that this is not always true.

This proposed technique utilises the fact that, whilst any individual sensor cannot be relied on to indicate the shingle position, on average most will show the increase in height over the shingle. Hence the proposed technique is to perform a summation of height increments across a number of sensors, where the inaccuracies and aberrations will tend to cancel out, and the overall height increment will be clear.

An example typical set of height readings illustrates the point e e ee. ec.

PA: 15mm 2A: 14mm 3A: 16mm 4A: 15mm 1B: 16mm 2B: 13mm 3B: 17mm 4B: 14mm 1C: 16mm 2B: 16mm 3C: 18mm 4C: 16mm late shingle lies between the "B" set of readings and the "C" set of leadings.

This means height differences between the "B" and "(," sets of readings should be greater than the height differences between the "A "and "B" sets of readings. However, if you look at sensor 1 readings by following column "1", this is not clear ( lB-lA=16-15=1 and 1(,-lB=16-16=0). Also if you look down column3 (3B-3=17-16=1 and 3B-3A=18-17=1), sensor 3 shows no clear difference.

If you employ the proposed technique of accumulating a group of sensor readings as follows: 1 \ + 2A + 3A + 4A = 15 + 14 + 16 + 15 = 60 1B + 2B + 3B + 4B = 16 + 13 + 17 + 14 = 60 1(,+2(,+3(,+4(,=16+ 16+ 18+ 16=66 Then the difference in height increments becomes clear.

Sum_of_row E3- sum_of_rowA = 60 - 60 = 0 Sum_of_row (,- sum_of_row_B = 66 - 60 = 6.

. .. . . e -. .-. e*e This result shows that the statistical accumulation of minor differences which may be individually indistinguishable reveals the discontinuity identifying the transition from one piece of board to the next, ie the shingle.

In some cases, such as on a stacker conveyer, upstream from the warpmeasunog apparatus, a combination of helical knife and brushes causes the shingle-eclge to be slightly delayed, from one board to the board adjacent. This means that the identifiable leading edge of a board may be a small distance (eg one centimetre) ahead of an adjacent board.

Scanning across the board, at a high scanning rate (eg Imm per scan) allows a discontinuity of a straight-line to be identified. At the point of the discontinuity, it is likely that the slit in the board is present. Statistical processing can be applied to multiple edge-identifcation, giving a reliable indication of slot position.

Figure 6 IS a representation of an aerial view of board travelling from the bottom of the page to the top.

Referring to the diagram, the position of the shingle (board edge) is recorded on all of the sensors. Sensors 1, 2, 3 and 4 of the group 1 to 50 are shown.

Some skew of the board leading-edge is allowed for and this is shown by the bc>ard-edge not being parallel to the x-axis. Board flows in the "y" direction shown. '{he x,y coordinates of points Point 1, Point 2,Point 3 and Point4 on the board edges are detected by sensors Sensor 1, Sensor 2, Sensor 3 and Sensor 4 respectively.

So taking Point 1 and Point 2, for the straight edge of the board it is possible to trigonometrically predict the position of the board edge at the third sensor (x3, y3). This trigonometric calculation is greatly simplified because the sensors are equally spaced (x2 - x1 = x3 - x2), so the predicted y-position of sensor 3 for a straight edge iS ce Be c e C c yl + (2 * dy2), where dy2 is equal to y2 - yl and the symbol "*" denotes multiplication.

his position is calculated, and then subtracted from the actual monitored position. This deviation is recorded. '['his process is repeated, starting with sensors 1,2,3 then 2,3,4 then 3,4,5 etc to 48,49,50.

In the diagram, Point 3 is shown in the predicted position and the deviation is very small. However this is not the case for the l'oint 4.

The expected position Ex4,Ey4 for point 4 is y2 + (2 * dy3) where dy3 is equal to y3 - y2.

\t the point "Point 4" the diagram shows the start of a second piece of board.

Between l'oint 2 and Point 3 is the position where the board has been slit, and the slightly delayed board causes the deviation from predicted position to actual position "dy4" to be relatively large.

So, of the 48 or so sets of calculations, some of the deviations will be larger.

this whole process is repeated for a number of shingles and averages taken to indicate the slit positions. 'I'he slit position is then recognised, relative to the sensors, and the sensor readings either side of the slits are disregarded from the warp calculations. Also the separate board sheets arc thus identified so that no attempt is made to calculate warp over the slit. la

Claims (1)

1. ( I) A method of detecting the presence of warp in a planar material, and the extent of the warp, if present, said method comprising the steps of measuring the distance of the planar material from a series of spaced sensors, calculating the height of the planar material from an intermediate sensor by using the height measurements from the sensors to either side of the intermediate sensor, repeating the step for each of the sensors, other than those at each end of the row of sensors and calculating a composite curvature figure for the planar material.

   (2) An apparatus utilising the method as claimed in claim (1).

   (3) A method or apparatus as claimed in claim 1 or claim 2, for detecting the presence of warp in a planar material and detecting that the readings obtained relate to a common planar material, said method comprising the steps of providing a row of sensors which detect the distance to the planar material at the same instant in time, said planar material moving with respect to said sensors, repeating the measurements at spaced time intervals to provide a series of sets of measurement data and wherein the data for each set of measurements is accumulated and the accumulated sums for each set compared, to identify whether or not the sets relate to the same planar material.

   (4) A method or apparatus as claimed in claim 1 or claim 2, wherein if a number of measurement pairs are made, and an accumulation of the measurement pairs is made, such that along the line where two adjacent board pieces meet, then the accumulation of these reading pair differences will be distinctive compared to other sets of readings where the accumulation will tend to cancel-out, and the consequent position of board- edge can be taken into account to provide a more accurate calculation of warp in the planar material. i\

   (5) A method or apparatus as claimed in claim 1 or claim 2, wherein there is provided a method of detecting the leading edges of respective boards of planar material, said method including the steps of providing a linear array of sensors which are spaced across a conveyer system by which a series of said boards are moved past the sensors, said sensors provided to detect the position of the edge of the board as it passes thereunder and selected ones of said sensors readings used to provide a predicted location of the board edge whereupon a comparison of actual and predicted board edge locations is performed and the position of slits (perpendicular board edges) derived and the consequent position of slits be taken into account to provide a more accurate calculation of warp in the planar material.