JPH03123857A - Speed controller for probe - Google Patents

Abstract

PURPOSE:To improve flaw detection accuracy by driving the probe according to the sum of a speed command value corresponding to the deviation of a current position from the mean value of position command values at respective control points on the scanning line of the probe and an averaged speed command value. CONSTITUTION:A position command value arithmetic means 501 calculates the position command values at the respective control points on the scanning line of the probe 9 from position information on a body W to be inspected and a position command value averaging means 502 calculates the mean value of the position command values as a position command value at each control point. A speed command value arithmetic means 503 finds the speed command value to each control point from the position command value of the means 502 and a speed command value averaging means 504 finds the mean value of the speed command values at the respective control points with the speed command value. A control means 506 calculates a speed command value according to the deviation between the detected position of the probe 9 by the position detecting means 505 and the position command value averaged by the means 502 and a driving means 507 is driven according to the sum of the speed command value and the mean speed command value from the means 504. Therefore, errors that various arithmetic results contain become small and the flaw detection position accuracy of the probe is improved.

Description

Detailed Description of the Invention A. Industrial Application Field The present invention relates to the movement of an ultrasonic probe when scanning an object with a known arbitrary surface shape to detect flaws on the object. The present invention relates to a device for controlling speed.

B. Conventional technology Journal of Japan Society of Mechanical Engineers vol. 1.90. Nα826° p5-9 (Conventional Document 1) and Nondestructive Inspection Vol. 37 No. 2 p.
The water immersion automatic flaw detection method shown in 152-153 (Conventional Document 2) is known. This is for the purpose of precise ultrasonic flaw detection, in which an ultrasonic probe is scanned over the surface of a test object in water in which the test object is immersed. In this automatic water immersion flaw detection method, in order to accurately determine the size and location of flaws (defects), the distance between the ultrasonic probe and the test object is kept constant, and the direction of the central axis of the ultrasonic beam is must be aligned with the normal direction of the surface of the subject.

The device disclosed in Conventional Document 1 is capable of detecting flaws on a specimen with a flat surface, but when the surface is curved, it is difficult to align the direction of the central axis of the ultrasonic beam with the normal direction of the surface of the specimen. Due to the difficulty, it was not possible to detect defects on objects with curved surfaces.

On the other hand, the device disclosed in Conventional Document 2 first measures the shape of the entire surface of the object using a laser distance meter, and then scans the ultrasonic probe using the shape data. It is possible to detect defects on objects with surface shapes of

C0 Problems to be Solved by the Invention The device disclosed in Conventional Document 2 scans the surface of an object with a laser distance meter in the XY force directions and measures at N points along the X axis at a pitch of dx. This operation is repeated M times in the Y direction at a pitch of dy to obtain NXM pieces of measurement data. Based on this, a shape function Z"f (X, y) is determined. Then, using this shape function, the surface of the object is The point (xij, y
This is an open-loop control method in which a control point located at a constant distance in the normal direction at ij) is found and the probe is driven and controlled to this point.

Therefore, errors may be included in the data when calculating the shape function, and errors may occur when calculating the control points.
The probe may not face the normal direction accurately at each control point, making highly accurate flaw detection difficult, and the probe may not move smoothly.

A technical problem of the present invention is to accurately direct a probe to a measurement point on the surface of a subject.

90 Means for Solving Problems To be explained with reference to FIG. 1, which is a diagram corresponding to the claims, the present invention allows the probe 9 to be moved by the driving means 507 based on the positional information of the surface shape of the object W determined in advance. The present invention is applied to a speed control device for a probe 9 that detects flaws on the object W while moving it to a plurality of control points on an arbitrary scanning line. and,
The above technical problem can be achieved with the following configuration.

a position command value calculation means 501 which calculates a position command value for each control point of the probe 9 from the above position information obtained in advance; , a position command value averaging means 502 that sets the average value of the position command values of a plurality of control points including the point to be controlled as the position command value of the point to be controlled; and this averaging means 502
A speed command value calculation means 503 calculates a speed command value to each control point based on the position command value of each control point calculated in A speed command value averaging means 504 that sets the average value of speed command values to a plurality of control points including the point to be controlled as the speed command value of the point to be controlled, and a position detection means for detecting the position of the probe 9. 505, a speed command value is calculated according to the deviation between the averaged position command value and the position detected by the position detection means 505, and the speed command value is searched by the sum of this speed command value and the averaged speed command value. Control means 50 for controlling the moving speed of the feeler 9
6.

In particular, the apparatus of claim 2 includes a position command value calculation means 501;
Position command value averaging means 502 and speed command value calculation means 5
03 and the speed command value averaging means 504 operate while the probe 9 is performing flaw detection along the preceding scan line, and calculate the averaged position command value of each control point of the next flaw detection scan line. The averaged speed command value is calculated.

The invention according to claim 3 has a storage means that divides the surface of the subject into small regions in the X and Y directions and stores only one piece of position information for each small region, and stores a plurality of averaged position command values. The calculation is performed based on positional information within a small area of .

The position command value of each control point is calculated from the 81 action position information and averaged. The speed command value of each control point is calculated from the averaged position command value and averaged. A speed command value corresponding to the deviation between the averaged position command value and the current position is calculated, and the driving means 507 is driven by the sum of this speed command value and the averaged speed command value. As a result, errors included in various calculation results are reduced, the accuracy of the flaw detection position by the probe 9 at each control point is improved (for example, the probe is correctly oriented in the normal direction), and the probe can be moved smoothly. Can be moved.

Further, in the second aspect, the position command value and speed command value for the next flaw detection are calculated in parallel with the flaw detection operation, so that the waiting time is shortened.

Furthermore, in claim 3, since the position command value and speed command value for the flaw detection operation are calculated from the position information within the small area,
It is possible to perform flaw detection on a scanning line different from the position of the object surface whose shape was measured, and since the position command value and speed command value are averaged, the flaw detection point can be scanned with high accuracy.

In addition, in the above-mentioned section and section E which explain the present invention in detail, some of the symbols of the embodiments are used in order to make the present invention easier to understand, but the present invention is not limited to the embodiments. Embodiment 1 of Embodiment Part 1 - [Description of the entire apparatus] The entire flaw detection apparatus to which the present invention is applied will be described with reference to FIGS. 2 to 5.

As shown in FIG. 2, this flaw detection device consists of a gate-shaped traveling body 2 that runs, for example, in the X-axis direction by an A bracket 5 that is integrated with the Y-axis traveling body 4 is provided with a Z-axis arm 7 that moves up and down in the Z-axis direction by a Z-axis drive device 6. A wrist portion 8 of the robot is attached to the lower end of the Z-axis arm 7.

As shown in FIG. 3, the wrist portion 8 includes a bracket 8A fixed to the lower end of the Z-axis 7, a drive device 8B for α-axis rotation provided on the bracket 8A, and a drive device 8B for α-axis rotation provided on the bracket 8A.
The bracket 8C has a bracket 8C provided on the rotating shaft of the frame B, and a β-axis rotation drive device 8E that is attached to the bracket 8C and has a bracket 8D attached to the rotating shaft.
D has one ultrasonic probe 9 for flaw detection on the object W;
A pair of distance sensor units 10a and lob for detecting the position of the subject W in the Z-axis direction are installed in a positional relationship as shown in FIG. Distance sensor units 10a, 10b
can be composed of an ultrasonic probe, for example, and its detection signal is the fifth
The signal is input to the distance detection circuit 11 shown in the figure.

As shown in FIG. 5, the distance detection circuit 11 consists of a distance detection circuit 11A for the ultrasonic probe 10a and a distance detection circuit 11B for the ultrasonic probe 10b.
B is an ultrasonic probe 10a.

a transmitter 11a that transmits an ultrasound signal to the object W, and a receiver 11b that receives an ultrasound signal reflected from the subject W.
and a clock circuit 11c.

The clock circuit 11c measures the time interval between the transmission signal from the transmitter 11a and the ultrasound reflected signal from the surface of the subject W, and outputs the result to the control device 12.

Here, if the time interval is to, and the speed of sound in water is V, then the distance M between the ultrasonic probe 10a or 10b and the surface of the object W is determined by M=Vt, /2 - (1) .

Furthermore, in FIG. 2, the control device 12 is a CPU.

SWA is a microprocessor consisting of ROM, RAM, etc., and indicates the time interval between the detection signal ST from the ultrasonic probe 9 and the distance detection circuit 11.

SWB is input, and X, Y, Z, α.

Signals sx, sy, sz from a position or angle detector, such as a potentiometer (not shown), built into the β-axis drive device 1, 3, 6.8B, 8E.

Sα and Sβ are also input. 13X, 13Y, 13Z, 1
3α, 13β are purge devices for each axis 1, 3, 6° 8B, 8E
11 is an input device for inputting the flaw detection range, etc., and 12 is a recorder for recording the flaw detection results. Note that each shaft circumferential drive device includes, for example, an electric motor.

[Arithmetic processing of control device fii 12] (1) Main flowchart FIG. 6 is a main flowchart of the arithmetic processing executed by the control device 12.

First, initial processing such as memory is performed in step SIO, and then, in step S20, the ultrasonic probes 10a and 10b are
position to the control start position.

FIG. 7 shows a state in which the positioning operation is completed.

Then, the process moves to a shape measuring operation in step S30, in which the ultrasonic probe 10a is measured based on the surface shape data of the subject W that is known in advance from design specifications and the like.

10b is scanned in the X direction along the subject W while facing on the normal line of the subject W. This X-direction scanning is performed a plurality of times in the Y-direction with a predetermined pitch shift, and surface shape data at least within the distance range shown in FIG. 4 is collected prior to flaw detection. That is, the surface shape of the object W to be inspected is determined by signals SX, S from position or angle detectors built into each axis drive device.
Y, SZ.

Sα, Sβ and signals S from the ultrasound probes 10a, 10b
Calculations are performed for each scanning line based on WA and SWB.

The detailed procedure is shown in FIG.

Next, when this shape measuring operation is completed, the process advances to step S40 to proceed to a flaw detection operation procedure. Here, step S30
Based on the surface shape data of the object W obtained in Ultrasonic probe 9 at the timing when 9 is controlled to each control point.
The ultrasonic flaw detection signal is taken into the control device 12 from the ultrasonic flaw detection signal. Also, during this flaw detection operation, the ultrasonic probes 10a and 10b scan several lines (distance mr in FIG. 4 => ahead of the ultrasonic probe 9).
The distance between the ultrasonic probes 10a, 10b and the subject W is calculated, and the surface shape data of the preceding scanning line is collected together with the position data from the position or angle detector described above. When the ultrasonic probe 9 reaches the scanning line, position information of the control point of the ultrasonic probe 9 corresponding to the flaw detection point on the preceding scanning line is calculated from this collected data, and Perform flaw detection. The detailed procedure is in Chapter 13.
As shown in the figure.

When this flaw detection operation is completed, the ultrasonic probe 9 is moved to the end position in step S50, and the process ends.

Next, the shape measurement operation and flaw detection operation will be explained in detail.

■Flowchart of shape measurement operation FIG. 8 shows details of the shape measuring operation in step S30.

First, in step S31, command values for one scan of each collapsing device (hereinafter referred to as each axis command value) are taken in and stored in the memory. The command value for each axis for one scan is expressed by formula (2) ~
The data group is shown in (6).

Xr= (Xs+ X1t ・=・=v Xn,・・−
, Xmax) − (2) Yr=(Yst Yxy −・=
y Ynt--, Ymax)-(3) Zr= (Z3.
Z,, ..., Zn, --, Zmax)...
(4) αr= (αS9α,, -=, an, --,
amax)°= (5)βr= (βS, β0...
..., βn, ..., amax) ... (6)
These axis command values require less precision than the command values for flaw detection operations, which will be described later. Therefore, the positions of representative points on one scanning line are determined based on the design data of the object W, and interpolation is performed between them. Desired.

Next, in step S32, positioning is performed to the scanning start point. That is, Xr=Xs, Yr=Ys, Zr=Zs+αr=
Let αS, βr=βS, and position each axis.

After this positioning is completed, the variable N is set to 1 in step S33, and the process proceeds to step S34, where a scan start command is output, and the timer interrupt program shown in FIG. 9 operates at regular intervals.

In FIG. 9, first, in step 5341, it is determined whether or not there is a scanning stop command. Since the first command is a scan start command, the process moves to step 5342, and the Nth command value, that is, N=1
The command value of Xr = X S + Yr = Y3. Zr
=Zs, αr=(!S+ βr=βS. Then, the signals SX, S indicating the current position of each axis mentioned above are taken in.
Y, SZ, Sa, and Sβ are taken in from each detector and the current value X of each axis is obtained. t Yor Zot α. .

β. Find the command value and the current value X of each axis. l'1'o+
ZQTα. ,β. servo amplifiers 13X to 13β in FIG.
Output to. As a result, the distance sensor units 10a, 1
0b moves to the first commanded position. At this time, the ultrasonic probe 10a.

10b faces the normal direction of the surface of the measurement point on the subject W. Then, in step 5343, the ultrasonic probe 10a
, 10b and the surface of the subject W by a signal SWA from the distance detection circuit 11.

Import by SWB. Next, in step 5344, the signal SX from the detector of each axis drive device is detected.

Take in sy, sz, sα, Sβ, step 5345
The current value of each axis is X0IYOIZOI α. , β0, and determine the position of the point on the surface of the object W that is hit by the ultrasound beams of the ultrasound probes 10a and 10b (Xa, Ya, Z
a) Calculate (xb, yb, zb). here,
Xa''Zb is calculated from the following relational expression.

Xa=f1(Xot Yor Zoy αot
βat Qa) ++・(7)Ya=L (Xot
Yor Zav αo+ β. ,Qa)...(
8) Za=f3(Xot Yot Zoy α.,
β,,Qa)-(9)Xb=f1(Xo, Yo, Zo,
α. ,β. , Qb) ・(10) Yb=f2(X,
,Y,,Z,, α. ,β,,Qb)-(11)Z
b"fa (Xo, Yo, Za, αo, βat Qb
) =·(12) Next, in step 8346, the surface position of the subject W determined using equations (7) to (12) is stored.

An example of a method for storing this position will be explained using FIG. 10.

Figure 1o shows area division for position storage, and this storage area is set corresponding to the X-Y coordinate plane,
In the figure, the shaded area is the flaw detection range of the object W on the X-Y coordinate plane. The position memory area is a slightly larger area than the flaw detection area, and the area (X axis is xt
h1~xth,, y axis is the area surrounded by yth~y th2) divided into (P+1) in the X axis direction, and divided into (P+1) in the Y axis direction.
S+1) Divide into multiple small areas.

Then, it is checked to which small area in Fig. 10 the Xa and Ya obtained by equations (7) and (8) belong, and Xa, Ya, Za and storage completion are determined as the position data of the small area to which they belong. Remember the flag.

Similar processing is performed for the values obtained by equations (10) to (12).

Here, in sequential scanning in the X-axis direction in the shape measurement operation, the position data obtained in the current scan (for example, the scan when N = 2) is the same as that in the previous scan (for example, the scan when N = 1). If it is determined that the small area belongs to a small area that already stores position data, only one piece of position data needs to be stored in each small area, so (1) Store the new position data obtained this time. Update the memory contents as data.

(2) Old position data that has already been obtained is used as stored data and the stored contents are not updated.

(3) Update the storage contents by using the average value of the new data and the old data as new storage data.

You may also use a method such as

Next, the program hand j@ is step 5347 in FIG.
Then, 1 is added to the variable N, that is, N=2, and the process ends.

When the operation of the timer interrupt program in FIG. 9 is completed, the program procedure returns to step S35 in FIG.
It is checked whether one scan in the X direction is completed based on whether or not is (max+1). At this stage, N=2, so
Step S35 is repeated, and during this time the timer interrupt program shown in FIG. 9 is operated at a certain time interval, and the bracket 8D connecting the ultrasonic probes 9, 10a, and 10b is Scanned in the axial direction.

Along with this scanning, the x, y, and z position coordinates of the surface shape of the subject W are stored using the storage method described above.

When one scan in the X direction is completed and N=max+1, the program moves from step S35 to step 836, and outputs a scan stop command. With this stop command, the procedure of the timer interrupt program shown in FIG. 9 advances to step 5348, where Xmax, which is the last command value in one scan,
Ymax, Zmax.

Take in αmax and βmax and perform servo calculation,
The calculation results are output to servo amplifiers 13X to 13β.

When one scan in the X direction is completed in this manner, the process proceeds to step S37 in FIG. 8, where it is determined whether the shape measurement operation is completed. In this embodiment, an ultrasonic probe 10a,
10b, the surface shape data of the X-direction scanning line preceding the X-direction by a distance in the Y-direction is collected, and the scanning control points of the flaw detection ultrasonic probe 9 are calculated based on the results. Therefore, the ultrasonic probe 9 is sequentially driven and controlled to a plurality of control points on a certain scanning line according to the position data obtained by such a shape measurement operation to measure the object W.
In order to perform flaw detection, at least 10 ultrasonic probes are required.
Ultrasonic probe 9 is placed on the first scanning line of a and 10b.
It is necessary to repeat the scan for shape measurement until the point is reached, that is, until the ultrasonic probe 9 moves in the Y direction by L shown in FIG. Therefore, since one shape measurement scan is not completed, the program moves from step S37 to step S31, takes in each axis position command for the next one scan, and repeats steps 832 to 37.

Then, as shown in FIG. 12, when the ultrasonic beam of the ultrasonic probe 9 reaches the vicinity where the position of the surface of the subject W is memorized by the shape measurement operation shown in FIG. It is determined that the operation is completed, and the program moves from step S37 in FIG. 8 to the flaw detection operation in step S40 in FIG. 6.

By the operation of only shape measurement before the flaw detection operation explained above,
The positional coordinates of the surface of the subject W in the area indicated by the dashed line in FIG. 12 are stored using the storage method shown in FIG. 10.

■Flow chart of flaw detection operation Next, the flaw detection operation processing in step S40 in FIG. 6 will be explained.

FIG. 13 is a detailed flowchart of the flaw detection operation processing procedure S40, which is similar to the flowchart of the shape measurement operation in FIG. First, in step S41, 1
Calculate each axis position and speed command value for scanning. 14th
The calculation will be explained in detail with reference to the figure and FIG.

FIG. 14 is a flowchart of the calculation procedure for each axis position command value for one scan in step S41, and FIG. FIG. 3 is a diagram illustrating which data is used. The thick solid line in FIG. 15 is a scanning line, and the process in FIG. 14 is for calculating each axis command value at the control point position indicated by .

In FIG. 14, in step 5411, the normal vector of the surface of the subject W in the vicinity of the control point position is calculated. For example, in the case of the control point position . in FIG. 15, the result is as follows.

Neighboring areas (1,l), (1,2) containing the first .
, (2, 2), and (2, 1), if the data of the four regions are all in the very vicinity of This is undesirable because the line vector calculation error becomes large. Therefore, the outer regions (0, O), (0, 3), (3, 3
), (3,0) data will be used. Area (0
, 0) (7) Data Xl, Yl, Zl, area (0, 3
) data to X2. Y2. Z2, data of area (3, 3) to X3. Y3. Z3, data of area (3, O)
．． Y4. When Z4 is the control point position on the surface of the subject W, the nearby normal vector N (=Nx, Ny
, Nz) is determined by the following formula.

Nx=Σ(Yi−Yj)X (Zi+Zj) ・・・(
13) Ny=kai (Zi-Zj) X (Xi+Xj)・
...(14) Nz= Kai (Xi-Xj) X (Yi+
Yj) −(15) However, if i≠4, then j=i+1. i
If =4, then j=1.

Next, in step 5412, position calculation of the control point position is performed. The position of the control point fist in Fig. 15 is (Xk, Yk
, Zk), Xk and Yk are values given to set the storage area and are known. Therefore, the position calculation of the control point position in step 5412 is a calculation to obtain Zk.

Now, as shown in FIG. 16, a plane PL including control point positions in a certain coordinate system is expressed by the following plane equation.

NxX+NyY+NzZ+d=0・= (16) Therefore, the normal vector N (=
Once Nx, Ny, Nz) are found, the coordinates (Xm, Yn+) of the position around . in Fig. 15.

Zm), the coefficient d of the plane equation of the plane PL is given by the following equation.

d=-(NxXm+NyYm+NzZm)...(1
7) By using this coefficient d, the position Zk of the first control point is determined as follows.

Zk=-(d+NxXk+NyYk)/Nz-(18)
Next, it is used to calculate the coefficient d (Xm, Ym.

Zm) will be explained.

The calculation of the normal vector is performed in the area (0,0) in Figure 15.
, (0,3), (3,3), (3,O), but the stored data used to calculate the coefficient d has a truer value near the control position of . Close to. Therefore, the stored data in areas (1,1), (1,2), (2,2), and (2,1) are used. That is, the storage data of the area (1, 1) and the normal vector obtained by equations (13) to (15) are substituted into equation (17) for calculation, and the value is set as d.

In the same way, d2 is the value calculated from the stored data in area (1, 2), d8 is the value calculated from the stored data in area (2, 2), and d8 is the value calculated from the stored data in area (2゜1). d4, and calculate the coefficient d as an average value using the following equation.

d= (d1+d, +d3+d,)/4...(19)
By averaging the coefficients obtained from the four regions in this way, the coefficient d of the control point becomes close to the true value.

From the above explanation, the normal vector N of the control point position (=N
x g N y + N z ) and position (Xk, Yk
, Zk) was found.

The program procedure in FIG. 14 then moves to step 5413, where position command values for each axis are calculated and stored. Here, if the distance between the ultrasonic probe 9 and the surface of the subject W is set to 11o, the position command values of each axis (Xr, Y
r, Zr, ar, ar) are calculated from the following relational expression and stored.

Xr=f4(Xk, Yk, Zk, Nx, Ny,
NZ, L) ”・(20)Yr=f, (Xk,
Yk, Zk, Nx, Ny, Nz, u,) −
(21) Zr=f, (Xk, Yk, Zk, Nx
, Ny, Nz, Q,) −(22)ar=f7(
Nx, Ny, Nz) ・・−(2
3) βr=f, (Nx, Ny, Nz)
-(24) Next, proceed to step 5414, and 1
It is determined whether the calculation for the scan is completed. Of course, in the above explanation, only the control points indicated by . in FIG. 15 are calculated, so the process returns to step 5411 and
Perform normal vector calculation near the next ○ control point. In this way, by sequentially repeating the processing of steps 8411 to 414, calculations are performed up to the control point positions marked with . . . in FIG. 15, and the process moves to step 5415. At this time,
Each axis position command value for one scan is a data group expressed by equations (25) to (29).

Xr= (Xrs, Xr, -Xrn, =Xrm
ax) −(25)Yr= (Yrs, Yr,, °
=Yrn, -Yrmax) -(26)Zr= (Z
rst ZrH-...zrn, ...Zrmax)
”・(27) ar= (arB, ar,, °=ar
n, −αrmax) −(28)βr= (βr3.
βrll ”'βrn, ”'βrmax) =-(2
9) Here, the command value of each control point (Xrs, Yrsy
Zr5yarB, ar3), (X rl, Y r,
, Z rl, a rl, βrx)... contain some errors due to calculation errors and the like. Therefore, in step 5415, the position command values are averaged. The averaging is performed using, for example, a linear equation.

Xn= (Xrn-8+Xrn+Xrn+x) / 3
・= (30) Yn= (Yrn-1+ Yrn+ Y
rn, ) / 3- (31)Zn= (Zrn-0+
Zrn+ Zrn*z) / 3- (32) αn=
(αrn-1+αrn+αrn,)/3...(33
)βn=(βrn-4+βrn+βrn, 1)/3・
(34) This is the arithmetic mean of the command values of a total of three points at an arbitrary control point and the control points before and after it. Note that the start point and end point are not averaged.

When equations (30) to (34) are calculated for all control points, each axis position command value for one scan becomes a data group in the same format as expressed by equations (2) to (6).

By the way, in the flaw detection operation of the present invention, the ultrasonic probe 9
Position feedback control and speed feedforward control are used to improve trajectory accuracy. That is, as shown in FIG. 17, the position command value X1n
A deviation device 2 calculates the deviation between s and the detected current position Xdet.
1 and multiplied by the coefficient k using the coefficient unit 22 to obtain the speed command value
Set to err.

Furthermore, the sum of the speed command value X1ns and Xerr is calculated by the adder 23 and input to the servo amplifiers 13A to 13β.

At step 8416 in FIG. 14, speed command values for the above-mentioned speed feedforward are calculated and averaged. The speed command value is calculated using the following equations (35) to (39), for example.

arn= (arn+1- arr+) /ΔT-(3
8) βrn=(arn,,-arn)/ΔT・(3
9) Here, ΔT is the sampling time interval of the timer interrupt program. Equation (35) for all control points ~
When the calculation in (39) is performed, the speed command values for each axis for one scan become a data group expressed by the following equations (40) to (44).

mir = (rurs + mirwork, ... mirn, ... sir
max)...(43)ar=(Jars, fJr
l, -βrn, -βrmax) = - (44) Next, the speed command values are averaged in the same way as the position command values. The averaging is performed, for example, as shown in the following equations (45) to (49).

Valley=(mirn-□+arn+arn,)/3...
・(48) βn= (βrn-0+βrn+βrn, 1
)/3 (49) This is the arithmetic mean of the command values of an arbitrary control point and three points before and after the control point. Note that the start point and end point are not averaged.

When formulas (45) to (49) are calculated for all control points, the speed command value for each axis for one scan is calculated by formulas (50) to (54).
) is a data group represented by

mir = (mi rs + mi r worker, ... mi rn.

. . tx rmax) (53) The above-described averaging operation is performed in step 8416, the program in FIG. 14 is completed, and the process moves to step S42 in FIG. 13.

The processing of steps 342 to S45 in FIG. 13 is performed using equations (4o) to (44), equations (50) to
This is performed using a data group of command values represented by (54). Then, similarly to the shape measurement operation, step S42,
When step S43 is executed and a scan start command is output in step S44, the timer allocation program shown in FIG. 18 is started.

In step 5441 of the program in FIG. 18, it is determined whether or not scanning is to be stopped. If the answer is negative, in step 5442, the Nth each axis position command value is taken in, servo calculation, and outputted. This is performed by position feedback control and velocity feedforward control shown in FIG. After that, the process proceeds to step 5443, where the ultrasonic probe 9
The flaw detection results are taken in from the control point position (X
m, Y m) data to the recording device 15. Next, steps 8444 to 5448, which are similar to steps 8343 to S347 in FIG.
As shown in FIG. 9, in parallel with the flaw detection operation, the shape of the preceding scanning line (indicated by a broken line) is measured.

When one scan is completed by repeating such a procedure, the process proceeds from step S45 to 846, and a travel stop command is output. When it is determined in step 5441 that scanning is to be stopped, in step 5449 the Nmax-th axis position command value is taken in, servo calculation is performed, and output is performed. Then, the process moves to step S47, and it is checked whether the flaw detection operation is completed. That is, it is checked whether the entire range indicated by the thick solid line in FIG. Calculate the axis position command value.

After scanning the entire flaw detection range shown in Figure 1o,
The program starts from step S47 to step S5 in FIG.
0, each axis is positioned to a certain determined end position, and the control is completed.

As described above, in the embodiment described with reference to FIGS. 2 to 19, the measurement of the surface shape of the object W and the flaw detection operation are performed as follows.

(a) Leading distance sensor unit loa.

The surface shape of the object W is measured based on the detection results 10b and the like.

(b) The shape measurement data is stored as data for one of the small areas divided on the xy plane. However, only one piece of data is stored within the small area.

(c) Repeat steps (a) and (b) for the distance shown in FIG.

(d) Find the position command value of the control point of the probe 9 from the surface shape data of the object W stored in each small region in (a) to (c), and average the multiple position command values. This is the position command value for each control point.

(e) Calculate the speed command value to the control point from the averaged position command value, average the speed command value to each control point from the speed command value to multiple control points, and calculate the speed to each control point. Use as command value.

(f) Find a speed command value for the deviation from the deviation between the averaged position command value and the actual position, add it to the averaged speed command value, and control each axis drive device using this value.

Therefore, the error in the position command value used for 1-position feedback and the speed command value used for speed feedforward becomes small, and the ultrasonic probe 9
The ultrasonic beam is directed in the normal direction of the surface of the object W, and highly accurate flaw detection is performed. Furthermore, since velocity feedforward control is adopted, the time required for the ultrasonic probe to travel to each control point is shortened, and the trajectory accuracy is high.

In addition, the XY plane is divided into small regions, and only one position data is stored in each region, and the control point of the probe on any flaw detection scanning line is determined based on the position data of each small region. from.

b) Memory capacity can be reduced.

You can freely set a scan line different from the mouth and shape measurement scan line for flaw detection. As a result, the degree of freedom in flaw detection operation is increased, and the arrangement of the flaw detection probe and the distance measurement probe is not restricted.

In the configuration of the above embodiment, the control device 12 includes the position command value calculation means 501 and the position command value averaging means 50.
2, a speed command value calculation means 503, a speed command value averaging means 504, and a control means 506.
A position or angle detector built into each shaft drive device 07 constitutes the detection means 505, respectively.

Modified Embodiment> In the above embodiment, at the start of flaw detection scanning of an arbitrary line, the position command value and speed command value of each control point of the line scan are calculated. Therefore, the ultrasonic probes 9, 10a, 10b are
There is a loss time when the system is stopped. Therefore, in this modification, the command values for the position and speed of the next line scan are calculated in advance during each flaw detection operation to eliminate the lost time.

This will be explained with reference to FIGS. 20 to 22.

FIG. 20 shows the hardware configuration when there are two central processing units. Reference numeral 31 denotes a central processing unit, which is responsible for calculation processing of each axis command value for one scan in the flaw detection operation, and other processing is performed by the central processing unit 32. 33 is a memory for storing position data on the surface of the object to be inspected and a calculation start flag for each axis command value for one scan, and 34 is a memory for storing each axis command value and calculation completion flag calculated by the central processing unit W31. be.

FIG. 21 shows the flaw detection operation (13th test) in the central processing unit 32.
FIG. 22 is a detailed flowchart of the central processing unit 31 (instead of a diagram).

Next, the operation of this modified embodiment will be explained.

First, in step 5101, a scan counter CN is set to 1, and a command calculation start flag is set in the memory 33.

Then, the process waits until the computation completion flag is set in step 5102.

The process in FIG. 22 first waits until the calculation start flag is set in step 5201. and.

If the calculation start flag is set in step 5101 of FIG. 21, the process proceeds to step 5202, reads the count value of the scan counter, and resets the calculation start flag and calculation completion flag in the first step 5203. In addition,
As a matter of course, the operation completion flag is initially reset.

Then, the process moves to step 5204, in which each axis command value of the scan line corresponding to the value of CN, that is, the first scan line is calculated. This calculation is.

The calculations are exactly the same as those in steps 8411 to 8416 in FIG. Then, in step 5205, the obtained axis command values are written into the memory 34, and in step 82
In step 06, the computation completion flag is set in the memory 34, and the process returns to step 5201 to wait for a command to start the command value computation for the next scanning line.

By setting the computation completion flag, the program in FIG. 21 proceeds from step 5102 to step 5103, where each axis command value is fetched from the memory 34.

Then, in step 5104, the count value CN of the scan counter is incremented by 1 and a command value calculation start flag is set. Further, in steps 8105 to 5109, the first
Perform scanning line flaw detection operation. During this flaw detection operation, the central processing unit 31 calculates a command value for the second scanning line.

As described above, according to this embodiment, since flaw detection and command value calculation can be performed in parallel, the time from the start to the completion of flaw detection can be shortened.

In the above description, it was assumed that flaw detection was performed while directing the ultrasonic beam from the flaw detection probe in the normal direction of the surface of the object to be inspected. The present invention can also be applied to flaw detection by directing an ultrasonic beam in the surface wave critical angle direction.

In addition, the shape is measured in advance using a distance detection probe integrated with the flaw detection probe, and the position command value and speed command value of each control point of the flaw detection probe are calculated based on the results. However, the position information on the surface of the object used to calculate the command value at each control point of the flaw detection probe is the position information obtained by measuring with a separately installed distance detection probe. It may also be position information given from design data.

G9 Effects of the Invention According to the present invention, each position command value on each scanning line of a flaw detection probe is calculated based on position data obtained by shape measurement etc. at multiple control points on that line. At the same time, the speed command values of multiple control points are also calculated as the average value of the speed command values of each control point. The probe is oriented precisely in the normal direction of the surface, improving flaw detection accuracy and allowing the probe to operate smoothly.

[Brief explanation of drawings]

FIG. 1 is a complaint correspondence diagram. 2 to 19 explain one embodiment of the present invention, and FIG. 2 is an overall configuration diagram of the control system. FIG. 3 is a diagram showing details of the mounting structure of the flaw detection ultrasonic probe and the distance sensor unit. FIG. 4 is a diagram showing an example of the arrangement of a flaw detection ultrasonic probe and a distance sensor unit. FIG. 5 is a block diagram showing details of the distance detection circuit. FIG. 6 is the main flowchart. FIG. 7 is a perspective view showing the distance sensor unit and the subject at the initial position. FIG. 8 is a shape measurement flowchart. FIG. 9 is a flowchart showing details of the shape measurement flowchart. FIG. 10 is a diagram illustrating a flaw detection range and a storage area for position data. FIG. 11 is a perspective view showing the distance sensor unit during shape measurement operation. FIG. 12 is a perspective view showing the flaw detection probe and distance sensor unit at the start of flaw detection operation. FIGS. 13, 14, and 18 are flowcharts showing flaw detection operations. FIG. 15 is a diagram illustrating a flaw detection scanning line within a small area in which position data is stored. FIG. 16 is a diagram illustrating the coefficient d when calculating the position data of a control point from the position data of a small area. FIG. 17 is a circuit diagram illustrating position feedback control and velocity feedforward control. FIG. 19 is a diagram illustrating that the flaw detection operation and the shape measurement operation are performed in parallel. 20 to 22 explain a modified embodiment, and FIG. 20 is a block diagram showing the hardware thereof. FIG. 21 and FIG. 22 are flowcharts showing the processing procedure. 1: X @ drive device 3: Y-axis drive device 6: Z-axis drive device 8B: α-axis drive device 8E: β-axis drive device
9: Flaw detection probes 10a, 10b: Distance detection probe 11: Distance detection circuit 12: Control device 31.32: Central processing unit 33.34: Memory 501: Position command value calculation means 502: Position command Value averaging means 503: Speed command value calculating means 5o4: Speed command value averaging means 505: Detection means 506: Control means 507: Driving means

Claims (1)

[Claims] 1) A flaw detection method for detecting flaws on an object while moving a probe to a plurality of control points on an arbitrary scanning line using a driving means based on positional information of the surface shape of the object obtained in advance. A probe speed control device includes a position command value calculation means for calculating a position command value for each control point of the probe from the position information obtained in advance; position command value averaging means for determining, based on the position command value, the average value of the position command values of a plurality of control points including the point to be controlled as the position command value of the point to be controlled; a speed command value calculation means for calculating a speed command value for each control point based on the position command value of each control point, and a speed command value calculation means for calculating a speed command value for each control point based on the position command value of each control point, and speed command value averaging means for determining the average value of speed command values for a plurality of control points including as the speed command value for the point to be controlled; position detection means for detecting the position of the probe; and the averaging means for detecting the position of the probe. A speed command value is calculated according to the deviation between the position command value and the position detected by the position detection means, and the speed command value of the probe is calculated by the sum of this speed command value and the averaged speed command value. 1. A speed control device for a probe, comprising: control means for controlling movement speed. 2) The apparatus according to claim 1, wherein the position command value calculation means, the position command value averaging means, and the speed command value calculation means,
The speed command value averaging means operates while the probe is performing flaw detection along the preceding scan line, and averages the position command value and average speed command of each control point of the next flaw detection scan line. A speed control device for a probe, characterized in that it calculates a value. 3) The apparatus according to claim 1, further comprising a storage means for dividing the surface of the subject into small regions in the X and Y directions and storing only one piece of the position information for each small region, and storing an averaged position command value. is calculated based on position information within a plurality of small areas.