WO2019134751A1

WO2019134751A1 - Rolling weight deflectometer

Info

Publication number: WO2019134751A1
Application number: PCT/EP2018/050281
Authority: WO
Inventors: Jack LARSEN; Mathias Bækbo ANDERSEN; Benjamin Maurice CUMMING; Albert Navarro COMES
Original assignee: Dynatest International A/S
Priority date: 2018-01-05
Filing date: 2018-01-05
Publication date: 2019-07-11

Abstract

A rolling weight deflectometer comprises a beam (3), a number of laser range sensors (6) placed on said beam (3) and a hoisting mechanism for lifting said beam (3) to an elevated position above a test profile (9) and lowering said beam (3) to a measuring position. A horizontal displacement mechanism adapted to displace said beam (3) horizontally along the test profile (3) in the elevated position above the test profile (9). The rolling weight deflectometer further comprises a data processing device adapted to receiving and process data from at least some among said number of laser range sensors (6) so as to record and compare recorded features of said test profile (3) and storage means for storing one or more correction values determined on the basis of said comparison of recorded and compared features.

Description

Rolling weight deflectometer

The present invention relates to a rolling weight deflectometer, more specifically to the calibration of the measuring system of a rolling weight de- flectometer.

Rolling weight deflectometers as e.g. disclosed in WO2013/185759 are used to survey the conditions of pavements by measuring the minute elastic deformation of the pavement forming a temporary deflection basin around a heavily loaded wheel as the wheel rolls along the pavement. Rolling weight deflectometers are also often referred to as rolling wheel deflectome- ters. Both terms yield the same abbreviation RWD which is also commonly used. In the present description as well as in the interpretation of the prior art all three are taken as synonyms.

The measurements are performed using a number of sensors, such as laser range sensors, mounted one after the other in the direction of motion of the rolling weight deflectometer on a rigid beam with one sensor as close to the loaded wheel of the rolling weight as possible and the others suitably spaced therefrom. By identifying the same points of the pavement in loaded and unloaded conditions the depth and/or shape of the deflection basin may be calculated, allowing knowledge of the condition of the pavement to be gained.

As explained in WO2013/185759 incorporated herein by reference the depth of the deflection basin formed around the wheel of the rolling weight is quite small e.g. in the range of 100 micrometers to 2000 micrometers whereas the overall length of the rigid beam is several meters. This, in turn, means that it does not take much bending or warping of the beam before the sensors are potentially misaligned and/or misplaced. In effect each sensor has six degrees of freedom, i.e. pitch, roll, yaw misalignment and misplace- ment along the X-, Y-, Z-axes. If for instance the beam sags in the middle the ends will bend upwards giving the sensors different pitch. If the sensors do not look vertically down on the pavement, they will measure a larger distance to the pavement than the shortest vertical distance. Furthermore, the sensors themselves may change over time. However, such misalignment, misplace- ments and other error sources can be compensated for in the calculations comparing the measurements, provided that the sensors are calibrated or recalibrated against a known surface so that the measurement errors are de- termined. One problem in doing so is that the calibration needs to be per- formed quite often, essentially before each measurement cycle, because overnight cooling and heating cycles etc. may minutely deform the beam, whereas, as long as the test run is performed, the beam is at least regarding temperature kept in a controlled environment and will not change noticeably.

Calibration of rolling weight deflectometers is dealt with in WO2011/009460. Here the beam is moved between various positions and individual test runs over a predetermined stretch of pavement is made. These repeated runs for calibration inter alia takes a substantial accumulated time which is not desirable.

Another known method is to provide a horizontal measuring surface using a large water filled tray placed under the rolling weight deflectometer. This, however, is cumbersome because the large tray needs to be placed under the rolling weight deflectometer and then filled with water. Alternatively, the rolling weight deflectometer, which is typically accommodated in a large trailer, such as a semitrailer, needs to be manoeuvred over the water tray without running over and potentially damaging the tray in the process.

Based on this prior art it is the object of the present invention to pro- vide a fast efficient and reproducible calibration of a rolling weight deflectome- ter.

According to a first aspect of the present invention this object is achieved by a rolling weight deflectometer comprising a beam, a number of laser sensors placed on said beam, a horizontal displacement mechanism adapted to displace said beam horizontally along the test profile above the test profile, a data processing device adapted to receiving and process data from at least some among said number of laser sensors so as to record and compare recorded features of said test profile and storage means for storing one or more correction values determined on the basis of said comparison of recorded and compared features.

According to a second aspect of the present invention this object is achieved by a method for calibrating the measuring system of a rolling weight deflectometer comprising providing a rolling weight deflectometer according to the first aspect of the invention, displacing said beam horizontally along the test profile above the test profile, receiving and processing data from at least some among said number of laser sensors with data processing device so as to record and compare recorded features of said test profile, determining and storing one or more correction values based on said comparison of recorded and compared features.

Both of these aspects allow the entire calibration process to be car- ried out in situ within the essentially closed and air-conditioned trailer accom- modating the measuring equipment of the rolling weight deflectometer, in turn, yielding reproducible calibration results. It moreover renders itself to an automated or semi-automated calibration procedure. Furthermore, by inte- grating the calibration system in the rolling weight deflectometer itself the cal- ibration becomes independent of the actual surface on which the rolling weight deflectometer is resting, thus avoiding any potential influences that the weight of the loaded wheel of the rolling weight deflectometer could have on the surface.

According to a first preferred embodiment according to the first as- pect of the invention, the test profile is accommodated within the rolling wheel deflectometer. This allows the calibration to be performed on a specially adapted test profile at any suitable time when the trailer is stationary, rather than having first to find an area of pavement with suitable surface features for the image recognition by the image recognition software of the data pro- cessing means.

According to a further preferred embodiment according to the first aspect of the invention, the rolling wheel deflectometer further comprises a hoisting mechanism for lifting said beam to an elevated position above a test profile and lowering said beam to a measuring position, wherein the horizon- tal displacement mechanism is adapted to displace said beam horizontally along the test profile in the elevated position. This has the advantage that the space of the semitrailer in use located above the saddle of the tractor may be put to use for the calibration, thus making the resulting rolling wheel deflec- tometer shorter than if a fully towed trailer was used.

According to another preferred embodiment of the first aspect of the invention, said test profile comprises a number of peaks and valleys. It has been found that a test profile using a number of peaks and valleys is advan- tageous, because it provides measurement over a large part of the measuring range of the laser sensors. This, in turn, means that even minute misalign- ments in pitch, roll and yaw yield large deviations in the resulting error, as compared to e.g. a flat surface. In other words, a large signal/noise ratio is achieved, the signal being the error. This not only makes alignment errors readily detectable in recordings from different laser sensors, but also the range variations between peaks and valleys allow for the detection and com- pensation of difference in linearity and other deviations between the individual sensors, rather than just the lack of alignment. It is, however, not excluded that a flat planar surface provided by a liquid and detectable by the range sensors is used.

According to a further preferred embodiment, the test profile corn- prises a surface texture. Having a surface texture, in particular a surface tex- ture which at least to some degree resembles the surface texture of the pavements of the actual measurements, facilitates the identification of the same points on the test profile by different laser sensors. As explained in WO2013/185759 image recognition be it on light images or virtual range im- ages works best when image comprises random and incoherent features.

According to another preferred embodiment, the horizontal displace- ment mechanism comprises a displacement sensor. Knowing the distance between the laser sensors along the beam, the displacement sensor helps in identifying the target area in which the same features of the test profile rec- orded with different laser sensors are to be searched for. In normal operation this information is provided by an odometer connected to the wheel of the rolling weight deflectometer, but since the calibration is made independently thereof with the wheel stationary this information is not available.

According to yet another preferred embodiment, the horizontal dis- placement mechanism comprises an electric stepper motor driving a threaded spindle. An electric stepper motor drive is efficient for slow and controlled movement of the beam along the test profile, and may directly yield a position that may aid the image recognition software in identifying corresponding are- as in images from different laser sensors.

The invention will now be described in greater detail based on exem- plary non-limiting embodiments and with reference to the drawings, on which:

Fig. 1 is a side view of a semitrailer accommodating the rolling weight deflectometer according to the invention,

Fig. 2 is a top view of the semitrailer of Fig. 1 ,

Fig. 3 is a cross-sectional view taken along the line Ill-Ill of Fig. 1 showing the test profile in a storage position,

Fig. 4 is a cross-sectional view corresponding to Fig. 3 but showing the test profile in the calibration position,

Fig. 5 is a perspective view in cross-section taken along the line V-V of Fig. 2 showing the rolling weight deflectometer in the operating position with the beam lowered,

Fig. 6 is a side view of the cross-section of Fig. 5,

Fig. 7 is a side view corresponding to Fig. 6 of a first calibration posi- tion with the beam raised, and

Fig. 8 is a side view corresponding to Fig. 6 of a second calibration position with the beam raised and translated towards the front of the semi- trailer.

Turning first to Fig. 1 a side view of a semitrailer 1 accommodating the rolling weight deflectometer. The semitrailer 1 is preferably an inloader. That is to say, a trailer without no fixed bottom and no through axles, but in- dependently suspended wheels 2 on either side. This has two major ad- vantages, one is that it is there is in principle free view downward towards the pavement to be surveyed, because there is neither a fully fixed bottom nor any axles to obscure the view. The other is that it is quite easy to provide the loading of the rolling weight deflectometer, as the semitrailer 1 may simply be pneumatically or hydraulically lowered to a pick-up position, then reversed under the loads, and the loads lifted up by the semitrailer 1 itself. This itself is a well-known procedure for inloaders. Though inloaders in general are without bottoms in order to be able to pick up various different loads that need to be transported, the inloader of the present invention is provided with a bottom serving as a floor for personnel, and providing a closed compartment that can be airconditioned. Though the preferred embodiment utilises an inloader sem- itrailer 1 the skilled person will understand that this is merely an example, and that the invention could alternatively be incorporated in an automotive vehicle or in other, different types of trailers.

Turning now to Fig. 5, a cross-section of the semitrailer 1 taken along the line V-V of Fig. 2. Flere the semitrailer is shown in its operation configura- tion as it would be during surveying, except of course that the semitrailer 1 would be attached to a tractor which does not form part of the invention and therefore is not shown.

The rolling weight deflectometer comprises a beam 3. To make the beam 3 light and rigid, it preferably comprises at least to parts, such as a tube 4 of fibre reinforced composite material, e.g. carbon fibre reinforced compo- site material, and a supporting lattice girder 5 of preferably stainless steel to prevent the tube 4 from sagging.

On the beam 3, a number of downward facing sensors, in particular laser range sensors 6, are mounted. In the illustrated embodiment, a large number of laser range sensors 6 are used, in casu twelve, but less will do. If the beam 3 is also fitted with a gyro 7 and/or other inertial sensors the neces- sary number of range sensors 6 is three, and without a gyro 7 four range sen- sors are needed for the Harr algorithm as explained in e.g. WO2013/185759.

The range sensors 6 are preferably laser range sensors shining a la- ser beam 8, e.g. in a fan-shaped line scan as shown, towards the pavement (not shown). Suitable slits for the laser range sensors 6 to shine through are provided in the bottom of the semitrailer 1. These slits are preferably closable when not in use. Since, as will be understood by the skilled person the beam 3 with the laser range sensors 6 is moved with respect to the test profile the invention is not limited to laser range sensors but also Doppler laser sensors could be used. This is also independent of whether the test profile is the pavement below the rolling wheel deflectometer itself, or an on-board test profile 9 accommodated within the airconditioned compartment of the semi- trailer 1.

As indicated above, there is however the problem that it is virtually impossible to prevent the beam 3 to change shape over time. Though it may be kept stable during measurements, because the climate conditions within the semitrailer 1 can largely be kept constant during operation, the conditions will change whenever the climate is not controlled, and changing tempera- tures will influence the beam 3 in an irreversible manner, meaning that if the beam is cooled down and reheated, it will not resume its exact prior shape. This, in turn, means that there will be some uncertainty on the exact orienta- tion of the laser range sensors 6 mounted on the beam 3, unless recalibrated. Calibration before each use is therefore desirable. Moreover, the laser range sensors 6 and the electronics thereof may change over time as components age.

To make this calibration feasible the rolling weight deflectometer ac- cording to the invention is provided with its own built-in calibration system, allowing for horizontal displacement of the beam 3 above the test profile. The test profile could be the pavement below the rolling wheel deflectometer, but preferably the calibration system comprises its own test profile 9, so that a suitable area of pavement need not be found before calibration can take place. The test profile 9 comprises a series of peaks 10 and valleys 11. In a simple version the test profile is a length of roofing felt arranged in a several metres long U-shaped aluminium profile. The width of the U-shaped profile is in the range of 20 to 30 centimetres, preferably approximately 25 centimetres, to cover the width of the scans by the laser range sensors 6, which typically scan in a fan-shape 8. Roofing felt has the advantage that it has a coarse tex- ture allowing for easy recognition of the same points along the length thereof captured by different range sensors 6 and compared by suitable image pro- cessing software. In other words, roofing felt exhibits some of the features of the pavement surface which the image processing software of the rolling weight deflectometer is already adapted to identify.

To perform the calibration, the test profile 9 needs to be placed below the laser range sensors 6, and the beam 3 with the laser range sensors 6 need to be displaced along the test profile 9 so as to allow several of the laser range sensors 6 to pass over and capture images or virtual images of the same areas for comparison. Having identified corresponding points or areas, the distances measured by different laser range sensors 6 can be compared and correction values corresponding to identified deviations stored. The im- age processing, the comparing and the storing of correction values is readily performed using a computer or similar data processing means.

To place the test profile below beam 3 with the laser range sensors 6 the beam 3 is lifted up along a pair of rails 12, 12’; 13, 13’ to the position shown in Figs. 3 and 7. Each rail is segmented into two rail segments 12, 12’ and 13, 13’ to provide a gap for accommodating the test profile 9 when not in use. To guide the beam 3, the beam 3 is mounted on a pair of carriages with wheels adapted to engage the rails 12, 12’; 13, 13’. Each carriage is large enough to bridge the gap between the respective rail segments 12, 12’; 13, 13’ so as to always be in engagement with at least one segment 12, 12’; 13, 13’. The carriages are moved up and down using a hoisting mechanism with belts attached to the carriages and driven by a hydraulic motor 15. A hydrau- lic motor 15 is preferred because of the good lifting power, and hydraulic sys- tems are already provided in the semitrailer for other purposes. However, other lifting mechanisms such as electric motors could evidently also be used. The carriages comprise hydraulic clamping means allowing the carriages to clamp firmly onto both the lower rails segments 12, 13 in the operating posi- tion, and onto the upper rail segments 12’, 13’ during non-operation, e.g. dur- ing maintenance where easy access is needed to the bottom of beam 3, i.e. to the bottom of the tube 4 of fibre reinforced composite material, and during calibration. During transport, the beam 3 is preferably clamped to the lower rail segments 12, 13 which, in turn, are firmly attached to the body of the sem- itrailer 1.

To prevent mechanical strain on the beam 3, which might deform it, the beam is suspended in only two points, of which at least one is a ball sus- pension allowing the beam 3 to move slightly with respect to the carriages so as not to introduce any strain. Similarly brackets for mounting parts on the trailer body are all made with long holes so than any deformation of the trailer body, does not influence the actual measuring parts of the rolling weight de- flectometer including the calibration mechanism.

When the beam 3 is in the position shown in Fig. 3, the test profile 9 may now be moved through the gap to the position below the beam 3, as shown in Fig. 4. This can be done manually or as a part of an automated pro- cess. In effect, the test profile is made similar to a drawer that may be drawn out below the beam 3 when the beam is in the upper position.

With the test profile 9 in place below the beam 3 with the laser range sensors 6, the beam 3 may now be moved in a translatory motion along the test profile 9, e.g. from the initial position of Fig. 7 to an end position shown in Fig. 8, and back again. During this the carriages are clamped firmly onto the upper rail segments 12’, 13’ which are displaced along with the beam 3 during the translatory motion. The test profile 9 is in a fixed position and will hardly change during the short time the calibration takes place. The translatory mo- tion of the beam 3 is preferably effected by a stepper motor 14 turning a threaded spindle, so as to allow a slow and smooth horizontal displacement motion. Since the distance between the laser range sensors 6 is essentially known, a position sensor is used to track the motion of the beam 3, so as to allow the image processing algorithms to have a qualified idea of where to look for similarities in images or virtual images from different laser range sen- sors 6. The stepper motor 14 could itself provide the position sensor, but it is currently preferred to also provide an encoder identical to the one used as the odometer on the loaded wheel 2 of the rolling wheel deflectometer. Virtual images in this context indicates that the images are not light intensity images where pixels represent light intensities, but instead the pixel values represent distances.

As can be seen the length of the test profile 9 is longer than the length of the beam 3. This allows the scan for all laser range sensors 6 to be performed in one go without having to perform any translatory movement of the test profile 9. This can efficiently be achieved by locating the gaps be- tween the rail segments 12, 12’ and 13, 13’, respectively, at a height allowing the test profile 9 to extend into the room above the saddle of the tractor. The length with which the test profile extends into the room above the saddle should be sufficient for allowing the laser range sensors 6 to scan the same suitable combination of peak-valley-peak or valley-peak-valley by at least three laser range sensors 6. The distances between the peaks and/or valleys are chosen accordingly. Since the peaks 10 and valleys 11 are not used for the image recognition itself, but only to provide an increased signal/noise ra- tio, the peaks 10 need not be sharp but can include plateaus and correspond- ingly valley floors may be flat.

Scanning both valleys 11 and peaks 10 allows the detection of later- ally skewed laser range sensors as well as for detecting linearity differences between sensors. It is however not excluded that the profile could be entirely flat, or even include a canal filled with coloured water or another suitable liq uid providing a surface detectable by the laser range sensors 6.

Having compared and identified the same points or areas in images from different laser range sensors 6 using a suitable data processing means, such as a computer with suitable software receiving the data, errors or devia- tions in the measurements from the different laser range sensors 6 may be detected and compensated for. Ideally there should be no deviations if all la- ser range sensors 6 are perfectly aligned with the same pitch, roll and yaw, and no deviations from the intended positions on the X-, Y-, and Z-axes, i.e. longitudinally, laterally and vertically). This is however not a simple feat be- cause during calibration, just as during actual pavement measurements, the beams 8 for the laser range sensors 6 can move both longitudinally and verti- cally during the measurement. A difference in the readings from two laser range sensors 6 may therefore not be due to any error from alignment, dis- placement or linearity error, but could instead be due to vertical movement of the beam. Therefore, the same calculations as those for actual pavement de- flection measurements are performed during calibration measurement. Cur- rently, this deflection measurement is the so-called ’’curvature”. However, since the calibration surface is undeformed during calibration, it is known that the deflection measurement should be zero during calibration. If the deflection measurement does not yield zero, it indicates the existence of at least one of the above mentioned alignment/placement/linearity errors. These can now be identified in an iterative process where the deflection-measure is minimized to its known value, namely zero, while adjusting the calibration parameters, e.g., the degrees of freedom (DOF) of each laser range sensor 6. As mentioned there are six degrees of freedom for each laser range sensor 6, viz. pitch, roll and yaw angles as well as the X, Y, and Z positions. This iterative process for all six degrees of freedom is somewhat cumbersome, in particular for a large number of laser range sensors 6. However, it is often sufficient to determine only a few of the degrees of freedom to achieve successful calibration, as they are not all equally important. Based on this iterative adjustment, a set of calibration parameters may then be determined and stored, e.g. in the digital memory of the computer, and used for actual deflection measurements.

The present invention thus integrates the calibration system entirely into the rolling weight deflectometer itself, thereby allowing for quick and fre- quent calibration of the measuring system, while at the same time excluding external error sources. Because the calibration is quick, the test profile - be it the on-board test profile 9 or the pavement below, both of which are in a fixed position with respect to the semitrailer 1 , will not change during the calibra- tion.

Claims

P A T E N T C L A I M S

1. A rolling weight deflectometer comprising

a beam,

a number of laser sensors placed on said beam,

a horizontal displacement mechanism adapted to displace said beam horizontally along a test profile above the test profile,

a data processing device adapted to receiving and process data from at least some among said number of laser sensors so as to record and com- pare recorded features of said test profile and

storage means for storing one or more correction values determined on the basis of said comparison of recorded and compared features.

2. A rolling wheel deflectometer according to claim 1 , wherein the test profile is accommodated within the rolling wheel deflectometer.

3. A rolling wheel deflectometer according to any one of the preced- ing claims, further comprising a hoisting mechanism for lifting said beam to an elevated position above a test profile and lowering said beam to a measuring position, wherein the horizontal displacement mechanism is adapted to dis- place said beam horizontally along the test profile in the elevated position.

4. A rolling weight deflectometer according to claim 3, wherein said test profile comprises a number of peaks and valleys.

5. A rolling weight deflectometer according to any one of the preced- ing claims, wherein said test profile comprises a surface texture.

6. A rolling weight deflectometer according to claim 1 , wherein said test profile comprises a planar surface provided by a liquid and detectable by said laser sensors.

7. A rolling weight deflectometer according to any one of the preced- ing claims, wherein the horizontal displacement mechanism comprises a dis- placement sensor.

8. A rolling weight deflectometer according to any one of the preced- ing claims, wherein the horizontal displacement mechanism comprises an electric motor, preferably a stepper motor.

9. A method for calibrating the measuring system of a rolling weight deflectometer comprising

providing a rolling weight deflectometer according to any one of the preceding claims,

displacing said beam horizontally along the test profile above the test profile,

receiving and processing data from at least some among said num- ber of laser sensors with data processing device, so as to record and com- pare recorded features of said test profile,

determining and storing one or more correction values based on said comparison of recorded and compared features.

10. A method according to claim 9, further comprising

placing said beam in an elevated position above the test profile, and displacing said beam horizontally along the test profile in the elevat- ed position above the test profile.