RU2821171C1 - Method of calibrating laser scanner for evaluating quality of weld surface - Google Patents

Abstract

FIELD: metrology.

SUBSTANCE: invention relates to metrology, specifically to calibration of laser devices built by laser triangulation and intended for evaluation of surface quality of welded seams. Calibration method consists in the fact that the reference surface is scanned and its rated dimensions are measured. For calibration, curvilinear convex spatial (volumetric) references are used, and the laser scanner measurement error is calculated from the ratio of the total area of deviations of the measured bulge sections of the reference to the total actual area of the sections of the reference bulge in the same points. Shape of the bulge of the references is limited by the values of the width and height of the welds specified in the normative and technical documents.

EFFECT: possibility of calibrating a laser scanner designed to control the quality of the surface of a weld seam, namely by the shape of a bulge of a spatial real weld seam.

3 cl, 2 dwg

Description

The invention relates to measuring technology, namely to the calibration of laser meters designed to control the quality of the surface of welds produced by surfacing, soldering or any known welding method, built using the laser triangulation method, in which radiation beams are directed from both sides perpendicular to the controlled surface, and the received optical signal is recorded by a multi-element receiver.

There is a known method for calibrating mobile 3D coordinate measuring instruments, which consists in monitoring rectangular coordinates in space along horizontal and vertical coordinates, characterized in that they create a reference system of 3D coordinates, unified with the calibrated mobile coordinate measuring instrument (MCSI), in which from one installation calibrated MCSI, while maintaining a single metrological base, calibrates three spherical coordinates, radius vector, horizontal and vertical angles and rectangular coordinates X, Y, Z in an unlimited number of controlled points declared by the manufacturer of the measurement space (RF Patent No. 2710900 SPK G01C 1/ 00, - 01/14/2020).

The disadvantage of this method is that not a material, but a virtual object is used as a standard - a “reference system of 3D coordinates”, which does not allow establishing the relationship between the readings of scanners used to control the surface quality of welds and the dimensions of the weld and determining the differences between the readings of the scanner and the readings of the measuring standard (a fixed device that gives correct readings).

Moreover, this method of calibrating mobile 3D coordinate measuring instruments is intended for laser scanners and trackers that are used to measure large-sized objects, for example, checking the accuracy of installation of units, measuring large-sized fastenings of equipment components, setting up assembly lines, etc. Welds have a curvilinear cross-section with small convex sizes: 0.5 mm-4.5 mm in height and 3 mm-25 mm in width, and for them such devices and the method of their calibration cannot be used.

There is also a known method for calibrating and tuning a laser sensor system and a device for its implementation, which contains a tuning sample. This sample is oriented in three-dimensional space relative to the camera-laser unit so that the light emitted by the laser is visible to the camera. Lasers and cameras are placed at a certain distance from each other so that the optical axes of the lasers and cameras are opposite at a certain angle, determining the properties of the laser from the light recorded by the camera, and the location of the laser relative to the camera. Moreover, the laser sensors are combined into one coordinate system and trays, which are tuning samples of the appropriate shape, are placed under each of the visible beams. Water, oil or any other liquid is used as a filler for the baths; calibration is carried out on the surface of the liquid onto which the laser beam falls, forming a visible mark on the surface of the liquid. During calibration, laser sensors are moved in three-dimensional space in height, on the sides, and rotated at certain angles, achieving visually and on the monitor the same values of the geometric parameters of the visible trace from laser beams on the liquid surface along the entire length of the bath, equal to the width of the adjustment sample. The level in all baths is maintained the same and horizontal. The technical result is ease of use due to the use of devices that form surfaces from a liquid rather than from solid samples (RF Patent No. 2541704, G01B 11/02 - 20.02.2015).

The disadvantage of this method is that baths with liquid are used as a standard (“tuning sample”), which during the calibration process are “moved in three-dimensional space in height, on the sides, and rotated at certain angles, achieving visually and on the monitor the same values of geometric parameters visible trace from laser beams." The use of liquid standards is associated with significant inconvenience for the calibration of laser meters intended for monitoring the surface quality of welds. The impossibility of using this method for our purpose lies in the fact that the shape of the surface of the welds has a curvilinear spatial (volumetric) shape in cross-section and physically cannot be horizontal (linear), which is the main condition of this method.

A laser thickness gauge and a method for calibrating it are also known (RF Patent No. 2542633 MPK G01B 11/02, - 02/20/2015). In accordance with this patent, calibration involves two photoelectric modules that form two narrow beams of radiation directed coaxially towards each other, which create light marks on opposite sides of the standard, and on two linear optically connected to the standard position-sensitive multi-element photodetectors included in the photoelectric modules image light marks, photodetectors simultaneously scan and measure the numbers of elements corresponding to the maximum and minimum values of the thickness of the standards, characterized in that a thickness standard is placed on the boundary of the measurement zone, which is discretely moved to another boundary and for each position of the standard the distances from the photovoltaic modules are measured to each side of the standard and the corresponding numbers of elements on multi-element photodetectors corresponding to these distances, and then determine the angular coefficients and displacements of the calibration lines for each photovoltaic module.

This calibration method uses a flat plate of varying thickness as a reference. To calibrate the base of the Ro thickness gauge, a standard is used, which consists of m sectors of different thicknesses. To measure Ro, a standard sector tm with a minimum thickness is installed, which sequentially occupies L positions in the measurement zone Δ=Rmax-Rmin. For further calibration of the Ro base, the standard is unfolded and a sector with a thickness of tm+1 is installed. At the end of L-cycles of measurements based on the displacement of the standard in the measurement zone Δ and M cycles of measurements on thicknesses of the available standards, the base of the thickness gauge is calculated as:

In this method, the laser thickness gauge is calibrated against a standard consisting of m sectors of different thicknesses; this is actually a set of gauge blocks of different thicknesses, combined in one product (standard), in which calibration is performed using one linear dimension.

Performing calibration using a flat standard that has one linear calibrated dimension in each sector is suitable for scanners that control the thickness of products, but is not correct for scanners designed to assess the quality of welds that have a curvilinear spatial (volume) shape in cross-section and physically cannot be horizontal (linear), which is the main condition of this method. This circumstance determines the main disadvantage of this calibration method - a flat plate of varying thickness, which is the standard, determines only one (linear) calibrated size Ro and does not allow establishing the relationship between the readings of scanners used to control the surface quality of spatial (volumetric) welds and the shape welds.

The closest in technical essence to the claimed solution is the method of calibrating laser scanners using gauge blocks or the linear method, published in the work “Linear calibration of a laser scanner”. Authors of the work: Pankov V.V. Ph.D., Pankov S.V. Ph.D., Bogorodsky I.G. Ph.D., Anikin D.A. (https://3dld.ru/?p=1536).

Using this method, calibration of a laser scanner consists of scanning gauge blocks, determining the absolute measurement error and the standard deviation of repeated measurements from the average values.

After scanning the CMD and automatically executing computational algorithms, the 3D coordinates of the CMD surface points are determined and stored.

Linear calibration of the laser scanner in height and width is performed by software by calculating the absolute measurement error and the standard deviation of the standardized size of the CMD.

The absolute measurement error of the CMD scanner in height and width (in millimeters) is determined by the formulas:

Where - true size of CMD, mm;

The uncertainty of measurement results is characterized by the standard deviation of repeated measurements from the average values of height and width according to the formulas:

The height h and width w of the KMD scanner are measured at least 20 times to determine the measurement error. The values of the height h and width w of the end block for each measurement are determined in accordance with the computational algorithms developed by the authors.

Average values of height and width are calculated using the formulas:

where K is the number of repeated measurements;

k - measurement number (k=1, …, K);

h _k is the height of the CM measured by the scanner at the kth measurement, mm;

w _k is the width of the CMD measured by the scanner at the kth measurement, mm.

When performing linear calibration, KMD or their analogues are used, which have only one calibrated (normalized) size. In this regard, calibration is performed sequentially: first, the scanner is calibrated linearly along the measurement height, and after rotating the gauge block around the longitudinal axis by 90° and re-scanning, the scanner is calibrated along the measurement width. The calculation of average values and standard deviations of the measured height and measured width is performed automatically by the program. The choice of gauge block size in accordance with GOST 9038-90 or analogues is determined by the calibrator, taking into account the capabilities of the laser scanner specified in its technical description, in the range of standardized values of gauge blocks from 2 mm to 20 mm.

The use of CMD for linear calibration shows a small error when measuring the height and a significant error when measuring the width of the CMD. This is due to the design features of the CMD, which have chamfers along the edges of the normalized plane, and the features of laser sensors that do not “see” the points of the normalized plane in the immediate vicinity of the vertical planes of the CMD.

The advantage of this method over the previous one is that with sequential calibration, the relationship is determined by two linear parameters, namely the height h and width w of the caliber (CMD), and in the previous method only by one linear parameter Ro.

The disadvantage of this method is that the measurement is a linear value of the standardized size between the plane-parallel measuring planes of the CMD, and the surface of the weld is curved and spatial (volumetric), therefore, sequential calibration of the scanner along the width and height of the CMD does not allow reliably establishing the relationship between the scanner readings used to control the surface quality of welds, and the shape of welds.

A significant disadvantage of calibrating laser scanners using this method is that when scanning a standard (CMD), triangulation laser sensors do not “see” the points of the normalized plane in the immediate vicinity of the vertical planes of the standard, and measuring the dimensions of the CMDS near such areas becomes impossible, therefore, during calibration Based on the width of the CMD, the result is not entirely reliable.

The methods described above for calibrating laser meters do not allow us to determine the relationship between the readings of laser scanners and the shape of the surface of a spatial (volumetric) weld. The forms of standards in analogues and prototypes are specified in the form of either electronic circuits, or horizontal (linear) surfaces of liquids, or in the form of flat (linear) plates, or KMD with one standardized size, do not allow ensuring the similarity of their shape - the shape of a spatial (volumetric) weld seam

In addition, such standards do not take into account the shapes and geometric dimensions of welds established by regulatory technical documentation at all, which is also a significant drawback of the above methods.

The objective of the present invention is to develop a method for calibrating a laser scanner designed to control the quality of the surface of a weld, namely the shape of the convexity, spatial, real weld.

The technical result is achieved due to the fact that in the proposed method, standards with a spatial (volumetric) convexity above the horizontal plane and having in cross-section the shape and dimensions of a circular segment with given values of height and chord length are used as standards required for calibrating the scanner. The shape of such standards approaches the shape of the surface of “ideal welds” prescribed in the regulatory technical documentation, in contrast to standards in the form of plates.

The height of the circular segment and the length of the chord are set in accordance with the requirements of the regulatory technical documentation, which prescribes the shapes and structural dimensions of the welds, and the convexity in the middle part of the standard is made according to the values of the width and height of the convex established by the regulatory technical documents according to the formulas: Radius of the circle of the convex section - according Pythagorean theorem:

Where

R is the radius of the circle of the convex section;

g is the height of the convex section;

e is the width of the convex section;

coordinate of the circle center Z _C :

convexity points of the standard - according to the circle formula in Cartesian coordinates:

Where

X - coordinate of the convexity point along the X axis;

Z - coordinate of the convexity point along the Z axis;

Z _c - coordinate of the circle center,

in this case, the overall dimensions of the standard convexity are within the limits:

in height - from 0.5 mm to 5 mm,

width from 3 mm to 30 mm,

length from 10 mm to 100 mm.

The measurement error of a laser scanner is calculated by the ratio of the total area of deviations of the measured cross-sections of the standard convexity to the actual cross-sectional area of the standard convexity using the formula:

Where

S is the total area of deviations of the measured sections of the standard convexity,

Where

- true height of the convexity in the measured sections of the standard;

- measured height of the convexity in the measured sections of the standard;

- absolute error in measuring the height of the convexity in the measured sections of the standard

- total, actual area of all M sections of the standard convexities:

Where

i=1, …, M - numbers of sections of the convexity of the standard;

j=1, …, N - numbers of points in the cross section of the standard convexity;

D is the distance between scanning points in the cross section of the standard convexity (step in N).

The appearance of the standards is shown in Fig. 1.

To determine the coordinates of the reference points specified in the normative technical documentation for the dimensions of welds, a calculation scheme was used. The calculation diagram for determining the shape and dimensions of the standard convexity is shown in Fig. 2.

The shape of the convexity in the middle part of the standard, in order to bring it even closer to the shape of the convexity of the real weld, can be made according to the values of the width and height of the convexity established by the regulatory and technical documents using the formulas:

Where

- a curve, for given values of a _k , e, g, ϕ, determines the shape of the standard

and _k - the capillary constant of molten steel or alloy is set in the range = 4.0-6.0 mm;

g is the height of the convex section. Established by regulatory and technical documents;

e is the width of the convex section. Established by regulatory and technical documents;

ϕ is the angle between the plane perpendicular to the reference axis in the Z0X coordinate system and the horizon line;

R ₀ - radius of curvature of the curve at point x equal to 0;

Integration of the curve in the range of values from x greater than or equal to 0 to x less than or equal to e, the cross-sectional area of the convexity of one standard was calculated:

The total, actual area of all M sections of the standard convexities:

The use of such forms of spatial standard when calibrating a laser scanner allows us to ensure maximum approximation of the proposed calibration method to real conditions for assessing the quality of welded joints and to obtain a result unattainable using known calibration methods.

The invention is illustrated by the following example.

Example

To calibrate the scanner of a device for assessing the quality of a weld seam (RF patent No. 2550673), standards with a convexity in the middle part were designed. The standards were made of steel on a high-precision electroerosive cutting machine Acctex A1600-SA with the dimensions of the width and height of the convexity established by the regulatory and technical documents for welds (GOST 16037-80, GOST 1154-80). Standards 1 and 2 with convexity according to the formulas:

The radius of the circle of the convex section was calculated using the Pythagorean theorem:

Where

R is the radius of the circle of the convex section;

g is the height of the convex section;

e is the width of the convex section;

coordinate of the circle center Z _C :

convexity points of the standard - according to the circle formula in Cartesian coordinates:

Where

X - coordinate of the convexity point along the X axis;

Z - coordinate of the convexity point along the Z axis;

Z _C - coordinate of the circle center.

For standards 3 and 4, the shape of the convexity in the middle part of the standard was made according to the values of the width and height of the convexity established by the regulatory and technical documents according to the formulas:

Where

- a curve, for given values of a _k , e, g, ϕ, determines the shape of the standard

and _k - the capillary constant of molten steel or alloy is set in the range = 4.0-6.0 mm;

g is the height of the convex section. Established by regulatory and technical documents;

e is the width of the convex section. Established by regulatory and technical documents;

ϕ is the angle between the plane perpendicular to the reference axis in the Z0X coordinate system and the horizon line;

R ₀ - radius of curvature of the curve at point x equal to 0;

Integration of the curve in the range of values from x greater than or equal to 0 to x; less than or equal to e, the cross-sectional area of the convexity of one standard was calculated:

The total, actual area of all M sections of the standard convexities:

Measurements of the manufactured standards were carried out on a three-coordinate measuring machine (CMM) GLOBAL modification PERFOMANCE having limits of permissible basic error of spatial measurements with a head PH10MQ-TP200=±1.7 microns, limits of permissible basic absolute error with head PH10MQ-TP200=1.9 microns. Resolution of the measuring system = 0.039 µm.

Measurements have shown that the absolute error in deviations of the coordinates of 50 heights of section profiles from the calculated values, in five fixed sections, is a negligible value = 0.05 mm. The data from the measurement results indicate that the manufactured standards in each cross section have a profile identical to the calculated one, and the coordinates of the heights in each section are equal to the calculated values at the same points.

Calibration of the laser scanner consisted of comparing the cross-sectional shapes of the manufactured standards (10×3 mm and 20×4 mm) with the results of the cross-sectional shapes measured (scanned) by the laser scanner, expressed in a digital value equal to the ratio of the total area of deviations of the scanned cross-sections of the standard from their true values by absolute value to their total area.

The calculation formulas used to calibrate the laser scanner are given below in the text; M is the number of sections of the standard;

N is the number of points in each section of the standard;

- true height of the convexity at the cross-sectional points of the standard;

- measured height of the convexity at the cross-sectional points of the standard, where

i=1, ...M - numbers of sections of the standard;

j=1, …N - numbers of points in the cross-section of the standard;

D is the distance between points in the cross-section of the standard (step in N).

The total area of all M sections of the true convex area of the standard:

During calibration, the scanner measured the heights of the convexity at the same points Z _t j that are given in the technical documentation for the manufactured standards.

The absolute error when measuring height at each point is equal to and the total area of deviations of the scanned sections from the total true areas of the standard sections, in absolute value, is equal to:

The overall measurement error during calibration was determined by the ratio of the total area of deviations of the scanned sections of the standard in absolute value to the true cross-sectional area of the standard specified in the technical documentation for the manufactured standard. The total measurement error expressed as a percentage is:

For comparison with the proposed method, calibration was performed using a prototype with a standard in the form of “end plane-parallel length measures” according to GOST 9038-90 (KMD).

The dimensions of the standards are given in table. 1.

Calibration was performed as follows:

We installed the LSP-U laser scanner (for technical specifications, see https://3dld.ru/wp-content/uploads/2023/03/Tech-inf-LSP-U.pdf) on a horizontal surface.

We prepared the LSP-U laser scanner for operation in accordance with the user instructions and loaded the software interface. A clearly visible bright red line from the laser sensor appeared on the surface of the table.

The standard or KMD, in the case of calibration according to the prototype, was placed so that the line of the laser beam was on its surface and completely covered it.

In the software interface window, the scan length and number of measurements were specified.

Scanning was performed by pressing the virtual arrow keys: “right” or “left”, displayed on the screen by the software interface. The LSP-U laser scanner moved relative to the standard and performed scanning.

After this procedure, a 3D replica (image) of the standard was displayed on the PC screen.

The measurement error of the LSP-U laser scanner was calculated automatically using software based on the ratio of the total area of deviations of the measured cross-sections of the convexity of the standards to the actual cross-sectional area of the convexity of the standards using the formula:

Where

S is the total area of deviations of the measured sections of the standard convexity,

Where

- true height of the convexity in the measured sections of the standard;

- measured height of the convexity in the measured sections of the standard;

- absolute error in measuring the height of the convexity in the measured sections of the standard

S ⁰ - total, actual area of all M sections of the standard convexities:

Where

r=1, …, M - numbers of sections of the convexity of the standard;

j=1, …, N - numbers of points in the cross section of the standard convexity;

D is the distance between scanning points in the cross section of the standard convexity (step in N);

When scanning a prototype CMD in accordance with GOST 9038-90 or analogues that have only one calibrated size h ₀ , to estimate the absolute error and standard deviation of the LSP-U laser scanner along the width of the CMD, it was necessary to rotate the CMD around its length axis by 90°. After that, the CMD scan was performed again to assess the absolute error and standard deviation of the LSP-U laser scanner when measuring along the width of the CMD. As a result, the results of a specified number of CMD measurements, their average values, and the standard deviation of CMD were displayed on the computer screen.

The calibration results for 20 measurements are given in table. 2

Thus, the above information indicates that the required set of conditions is met when using the claimed invention:

The method embodying the claimed invention when implemented allows quantitative calibration of the scanner using curvilinear convex standards;

For the claimed invention in the form as it is characterized in the claims, the possibility of its implementation has been confirmed using the means and methods described above in the application or known before the priority date;

The means embodying the claimed invention, when implemented, is capable of achieving the technical result envisaged by the applicant.

Consequently, the claimed invention meets the requirement of “industrial applicability” under current legislation.

Claims (51)

1. A method for calibrating a laser scanner designed to assess the surface quality of a weld, including scanning the surface of a standard, measuring its standardized dimensions, calculating the absolute error of these measurements and the standard deviation of each measurement from the standardized size of the standard, characterized in that convex spatial standards are used for calibration , and the measurement error of the laser scanner is calculated by the ratio of the total area of deviations of the measured sections of the standard convexity to the actual cross-sectional area of the standard convexity according to the formula: Where S is the total area of deviations of the measured sections of the standard convexity Where - true height of the convexity in the measured sections of the standard; Z _i,j - measured height of the convexity in the measured sections of the standard; - absolute error in measuring the height of the convexity in the measured sections of the standard; S ⁰ - total, actual area of all M sections of the standard convexities: Where i=1, …, M - numbers of sections of the convexity of the standard; j=1, …, N - numbers of points in the cross section of the standard convexity; D is the distance between scanning points in the cross section of the standard convexity (step in N). 2. The method according to claim 1, characterized in that the convexity in the middle part of the standard is made according to the values of the width and height of the convexity established by the regulatory and technical documents according to the formulas: The radius of the circle of the convex section - according to the Pythagorean theorem: Where R is the radius of the circle of the convex section; g is the height of the convex section; e is the width of the convex section; coordinate of the circle center Z _C : convexity points of the standard - according to the circle formula in Cartesian coordinates: Where X - coordinate of the convexity point along the X axis; Z - coordinate of the convexity point along the Z axis; Z _C - coordinate of the circle center, in this case, the overall dimensions of the standard convexity are within the limits: in height - from 0.5 mm to 5 mm, width from 3 mm to 30 mm, length from 10 mm to 100 mm. 3. The method according to claim 1, characterized in that the shape of the convexity in the middle part of the standard is made according to the values of the width and height of the convexity established by regulatory and technical documents according to the formulas: Where Z=Z(x) - a curve, for given values of a _k , e, g, ϕ, determines the shape of the standard and _k - the capillary constant of molten steel or alloy is set in the range = 4.0-6.0 mm; g is the height of the convex section, established by regulatory and technical documents; e is the width of the convex section, established by regulatory and technical documents; ϕ is the angle between the plane perpendicular to the reference axis in the Z0X coordinate system and the horizon line; R ₀ - radius of curvature of the curve Z=Z(x) at point x equal to 0; Integration of the curve in the range of values from x greater than or equal to 0 to x less than or equal to e, the cross-sectional area of the convexity of one standard was calculated: The total, actual area of all M sections of the standard convexities: S ⁰ =MS ^E in height - from 0.5 mm to 5 mm, width from 3 mm to 30 mm, length from 10 mm to 100 mm.