JP2983103B2 - Control method of mobile robot having welding bead sensor - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a control method for moving a mobile robot, such as a trackless ultrasonic inspection and traveling apparatus, along a welding line, and more particularly, to a method for linearly and densely intersecting a traveling direction of a bogie. The present invention relates to a control method of a mobile robot having a plurality of welding bead sensors that scans a number of coils arranged one by one to detect a bead position by eddy current.

[0002]

2. Description of the Related Art Generally, nondestructive inspection of welds of tanks, nuclear reactors, ships, bridges, and the like is important in quality control and quality evaluation of these welds, and without laying a running track. It is desired to provide an ultrasonic testing and traveling apparatus that travels on a weld bead along a welding line (that is, following the welding line).

The problem here is a technique for preventing the flaw detection traveling device from coming off the welding bead (tracing bead) that is going to travel along. When the traveling device is sucked and run on the surface of a container with a certain curvature, such as a spherical tank, the traveling posture of the traveling device is slightly changed from the direction of the tracing welding line by the action of gravity on the side of the sphere or the upper part of the sphere. There is a danger that this will accumulate and that the traveling equipment will fall off the bead. Therefore, it is necessary to control the traveling device that should follow the welding line so that it does not deviate from above the copying bead, and as a precondition, always grasp the position of the following bead with respect to the traveling device. Requires a weld bead sensor.

[0004] As for the weld bead sensor, a mechanical scanning type sensor based on the principle of eddy current flaw detection is conventionally known. That is, a set of an excitation coil and a detection coil are supported on a plate for detecting a weld bead, and the plate is mechanically reciprocated in a direction crossing the weld bead, thereby mechanically scanning the coil to detect eddy currents. In this case, the welding bead position is detected. However, the mechanical scan type welding bead sensor requires a mechanism for mechanically reciprocating the coil carrying plate, so that the sensor is inevitably heavy and large, and the configuration of the flaw detection traveling device is also complicated. Therefore, as a weld bead sensor capable of scanning eddy current flaw detection without a mechanical scanning mechanism, the inventors of the present invention sequentially shifted a large number of coils linearly and densely arranged one by one in a direction crossing the bogie traveling direction. We have proposed a welding bead sensor that selectively performs eddy current testing at the bead position.

If two sets of such mechanical scan or electric scan type weld bead sensors are mounted, two points at which the welding line intersects with the welding bead sensor can be determined, so that a straight line equation of the weld bead can be obtained. Therefore, it is considered that the mobile robot can travel along the welding line.

[0006]

However, simply moving the robot based on the equation of the welding line obtained from the above-mentioned two points requires that a mobile robot such as a flaw detection traveling device actually travel along the welding line. Can not. This is because the mobile robot has only orthogonal coordinates with its own bogie center as the origin, and the origin of the coordinate system moves because the bogie coordinate system shifts as the robot moves. That is, when viewed in a coordinate system at each movement position of the robot, the same welding line is recognized as being different each time in the equation.

An object of the present invention is to solve the above problem and to provide a control method for causing a mobile robot having a welding bead sensor to actually travel along a welding line.

[0008]

In order to achieve the above-mentioned object, a moving control method for a mobile robot according to the present invention is characterized in that a plurality of coils arranged in a direction intersecting with a traveling direction of a carriage are scanned one by one for each bead position. In a mobile robot having a plurality of welding bead sensors that detect eddy currents in a trolley, (1) a known bogie center line from two points on the welding line obtained from the bead sensor at every sampling interval when scanning of the bead sensor is started and the bead position data is collected. Distance l ₁ , l
₂ , the movement amount ΔV _L of the robot and the posture angle Va of the robot
Are input as one set, and the collected distance data l ₁ and l ₂ are converted into values at the current time when the collection is completed,
Using this together with the previous data, the equation of the welding line is obtained by the least squares method. (2) In the next movement of one sampling interval, the previously obtained equation is used as the robot movement amount ΔV
_The coordinate angle is converted by _{L and} used as an estimated value of the welding line equation. From the obtained relationship between the robot and the welding line, the attitude angle Va is controlled so that the distance between the welding line and the bogie center becomes zero, This controls the robot to follow the welding line.

[0009]

When the bead sensor 15 is activated to scan at point O in FIG. 14, it is point A that scanning is actually performed by the bead sensor and data collection is performed, and detection data of the bead position is output. Is point B after a further delay. During this time, the bogie has already moved to point B. That is, while moving, the robot takes in sensor data at every sampling interval (for example, every 5 mm) shown between points O and B in FIG.

Therefore, when controlling the mobile robot so as to follow the welding line using the bead sensor, the scanning of the bead sensor is started and every two sampling intervals at which the bead position data has been collected, the two points on the welding line obtained from the bead sensor are determined. the distance l _1, l ₂ from the point P1, P2 to a known bogie center line, inputs the posture angle Va movement amount [Delta] V _L and the robot of the robot as a set. This movement amount ΔV
_{The L} and the posture angle Va are obtained from, for example, the planimeter 48 and the vertical sensor 56 shown in FIG.

However, when the data of l ₁ and l ₂ are collected at point A, the robot has already moved to point B,
L ₁ at the value of l _1, l ₂ between A and B coordinate conversion, taken was complete value at the current time, i.e. point B (coordinates),
l ₂ position. By collecting some data l ₁ and l ₂ at each sampling interval and obtaining the equation of the welding line by the least square method, the relationship between the robot and the welding line can be understood.

Therefore, in the next movement at one sampling interval, the previously obtained equation is coordinate-transformed by the robot movement amount ΔV _{L and} used as an estimated value of the welding line equation.
If the attitude angle Va is controlled such that the distance between the welding line and the center of the bogie becomes zero based on the obtained relationship between the robot and the welding line, the robot follows the welding line even while the bead sensor is capturing data. Can be controlled as follows.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the accompanying drawings.

The mobile robot shown in FIG. 1 includes a permanent magnet wheel type flaw detection traveling device 1, a control device 3 connected to the flaw detection traveling device via a cable 2, and a printer 4 for recording the flaw detection results in a graph by dots. It is composed of The flaw detection traveling device 1 here has a combined mode in addition to the normal flaw detection traveling mode. That is, while checking (coarse flaw detection) the presence or absence of a defective spot to be subjected to precision flaw detection, the spot is moved linearly along the welding line on the welding bead 6 of the spherical tank wall 5, and the welding bead location requiring precision flaw detection is required. When it is detected, it stops at the position and scans for precise flaw detection.After that flaw detection, it moves linearly again while searching for a defective point along the welding line, stops at the required position and performs precision flaw detection in the same manner. This is an operation mode in which operations such as scanning are repeated to inspect all the weld beads of the spherical tank.

First, the configuration of the flaw detection traveling device and the welding bead sensor, which are the premise of this control, will be described.

The flaw detection traveling device flaw traveling device 1 is pivotally with self-propelled carriage 7 having a permanent magnet wheel 8 which strongly magnetically attracted to the inspected wall serving tank wall 5, the shaft 9 provided thereon (FIG. 3) The inspection arm 10 is supported and tiltable in the direction of arrow A in FIG. The permanent magnet wheel 8 is driven by a motor 35 provided separately and coaxially therewith. A guide plate 1 is provided on the distal end side of the inspection arm 10.
1 extends in a direction intersecting the traveling direction and is provided so as to be pivotable within a predetermined angle range (θ in FIG. 2). Probe arm 12 carrying ultrasonic probe 14 for fine search (FIG. 2)
The apparatus is movable (scanned) along the guide plate 11, that is, in a direction intersecting the traveling direction of the carriage. Further, the inspection arm 10 is provided with one welding bead sensor 15 on each of the front side and the rear side of the probe arm 12.

The flaw detection traveling device 1 travels to a required position with the inspection arm 10 raised to the position indicated by the broken line in FIG. Then, after coming on the weld bead 6 to be inspected, the inspection arm 10 is tilted forward (in the direction of the arrow A + in FIG. 3) about the axis 9 so that the tip side is at a predetermined lower position with respect to the spherical tank wall 5 (see FIG. 3 (solid line position). In order to determine the lower position, the inspection arm 10 includes the probe arm 12
An arm support roller 16 (FIG. 3) is provided between the arm and the welding bead sensor 15.

The ultrasonic probe 13 for coarse search has a weld bead 6
The ultrasonic probe 14 for coarse search for searching only for the presence of a defective portion (whether there is a flaw) in the welding bead, and the fine probe for flaw detection for flaw detection up to the actual position and size of the flaw in the weld bead 6, The welding beads 6 are arranged at positions before and after each other in the bogie traveling direction, and can detect flaws at different positions in the depth direction of the weld bead 6 by the oblique method. In this embodiment, four coarse ultrasonic probes 13 are provided on each side of the probe arm 12, two in total, and four fine ultrasonic probes 14 are provided on both sides of the probe arm 12. , One for each.

The two types of ultrasonic probes 13 and 14 are different from each other in the width of the radiated ultrasonic beam, and the fine probe ultrasonic probe 14 emits a narrow ultrasonic beam close to the main beam. On the other hand, the coarse probe ultrasonic probe 13 radiates a wide ultrasonic beam 152 diverging from the main beam 151 as shown in FIG. As shown in FIG. 11, the arrangement of a total of four coarse search ultrasonic probes 13 is such that the ultrasonic beam 152 from the inner coarse search ultrasonic probe (first probe) 131 is formed. Approximately 1 /
2 in a non-reflective manner, and the ultrasonic beam 152 from the outer coarse probe ultrasonic probe (second probe) 132 covers about one half of the surface side of the weld bead 6 in the depth direction area. , The light is reflected once by the spherical tank wall 5 and then covered.

Each of the above-described coarse and fine ultrasonic probes 13 and 14 can be individually lowered to the spherical tank wall 5 by dedicated pressing mechanisms 17 and 18. Reference numeral 19 denotes a pressing cylinder which constitutes the pressing mechanisms 17 and 18, and further includes a linear guide bearing 20.
Is also used.

4 and 5 (a) and 5 (b), the probe arm 12 is suspended on the lower surface of the guide plate 11 so as to be slidable in the S direction in FIG. Arm 1
The grooved rollers 21 are provided at four positions (see FIG. 2) facing the side edge of the second plate 2, while the probe arm 12 is engaged with the grooves of these grooved rollers 21. A guide plate 22 having an engaging edge 23 is fixed. On the other hand, the guide plate 1
The guide plate 11 is attached to a base member 24 provided on the inspection arm 10 via a large-diameter bearing 25 in order to pivotally mount the probe arm 1 on the inspection arm 10 and constitute a scanning drive system for the probe arm 12. A rotating shaft 26 is provided at the center thereof and penetrates between both members, and is rotatable with respect to the guide plate 11 and the base member 24 by a bearing 27. A sprocket 28 is fixed to the rotating shaft 26 in the inspection arm 10, and a gear 29 is fixed in the U-shaped cross section of the probe arm 12. I support it. The former sprocket 28
As shown in FIG. 4, it is rotated by a chain 31 wound around a sprocket 30 driven by a motor 36. The latter gear 29 meshes with a rack 32 fixed to the inner side wall of the probe arm 12, and accordingly, the motor (S
The guide plate 11 can be moved in the left-right direction (the S-axis direction in FIG. 2) with respect to the inspection arm 10 by the (axis motor) 36. The movement position in the S-axis direction is detected by a linear potentiometer (S-axis sensor) 37 provided in the inspection arm 10 so as to detect the movement of the chain 31.

In FIG. 6, on the lower surface of the probe arm 12, four pressing cylinders 19 forming the above-described coarse search pressing mechanism 17 and two pressing cylinders 19 forming the fine search pressing mechanism 18 are provided. Is attached to the tip of the piston rod 19a.
3 or mounting members 33, 33 'of the ultrasonic probe 14 for fine search
Are swingably mounted by shafts 34 and 34 '. The coarse and fine ultrasonic probes 13 and 14 are pivotally supported by a gimbal support mechanism including the mounting members 33 and 33 'and the shafts 34 and 34' so that they can be in close contact with the inspection surface. It has become. The reason why the number of the coarse probe ultrasonic probes 13 is larger than that of the fine probe ultrasonic probe 14 is that the probe arm 12 is not moved in the S direction (the probe is scanned). This is to make it possible to detect a portion having a flaw to be subjected to precision flaw detection. Since both the probes 13 and 14 perform flaw detection by the oblique method, the probe body is obliquely attached to a shoe that closely contacts the surface to be inspected.

In FIG. 7, the guide plate 11 and the probe arm 12 suspended on the lower surface are configured so that the whole thereof can be turned within the range of θ in FIG. The movement direction (scanning direction) of the camera 12 can be corrected. That is, in FIG. 7, an operation disk 52 driven through a gear by a motor (θ-axis motor) 51 is provided in the inspection arm 10, and the operation disk 5
2, two ends of two rods 53 are pivotally connected to the rods 53.
The other end of 3 is pivotally attached to a pin 54 of the guide plate 11.
Therefore, when the operation disk 52 is rotated by the θ-axis motor 51, one of the two rods 53 projects and the other retreats, and the guide plate 11 turns right or left within the range of θ.
The turning angle θ is detected by a potentiometer (θ-axis sensor) 55 attached concentrically and integrally with the operation disk 52.

The ultrasonic flaw detecting and traveling apparatus 1 moves on the welding bead 6 along the welding line while being magnetically attracted to the spherical tank wall 5 by the permanent magnet wheel 8. At this time, the probe arm 12 is held at a substantially constant height position by the arm support roller 16, and is lowered until the coarse search ultrasonic probe 13 comes into contact with the tank wall 5. The ultrasonic flaw detection traveling device 1 travels (moves forward) at a constant speed. At this time, the scanning of the coarse detection ultrasonic probe 13 becomes a linear scanning pattern 154 as shown in FIG. The contactor 13 detects whether or not there is a defective portion to be flaw-detected in the weld bead 6 on the front side of the self-propelled carriage 7. In the case of the coarse flaw detection, the probe arm 12 is controlled (S-axis control) so as to be always on the welding line.

When a defective portion requiring precision flaw detection is detected, the ultrasonic flaw detection traveling device 1 moves the coarse flaw detection ultrasonic probe 13 when the coarse flaw detection ultrasonic probe 13 detects the location. And move forward a little, if necessary, to move the ultrasonic probe 14
Stops at a position in front of the defective portion. And
The ultrasonic probe 14 for fine search is lowered, and precise ultrasonic flaw detection is performed by the rectangular scanning pattern 155 shown in FIG. In this precise flaw detection, the S-axis movement of the probe arm 12 by the S-axis motor 36 (that is, B1 → B2, B3 →
B4, B5, B6... In the direction S in FIG. 2) and the forward movement of the ultrasonic testing and traveling apparatus 1 (that is, B2 → B3, B4 →).
B5, B6 → B7...), The inspection volume is flaw-scanned at a specified pitch, thereby flaw-detecting the weld bead 6 up to the actual position of the defect and the size of the flaw. Thereafter, if necessary, the vehicle moves a little backward, and the ultrasonic inspection traveling apparatus 1 further moves along the welding bead 6 while performing the coarse inspection based on the linear scanning pattern 154,
When a defective portion requiring precision flaw detection is detected, the apparatus stops at that position and performs ultrasonic precision flaw detection using the rectangular scanning pattern 155 in the same manner. Repeat this operation to inspect all weld beads in the spherical tank.

As described above, in the case of the composite mode operation, it is confirmed whether or not there is a portion to be subjected to precision flaw detection by the coarse probe ultrasonic probe 13 by linear scanning, and the fine probe ultrasonic probe 14 is used only at the relevant portion. , The time required for flaw detection is drastically reduced as compared with the case where only flaw detection by rectangular scanning or rough and precision flaw detection by rectangular scanning is performed over the entire length of the weld bead.

The flaw detection result is sent from the flaw detection traveling device 1 to the control device 3 via the cable 2 and output by the printer 4 in real time. The printer 4 records the flaw detection results as dots A, B, and C using dots. A scope has peak value (percent), B scope has depth position (mm), and C scope has planar position (mm).
Is printed out according to the moving distance of the cart.

The welding Bidosensa welding Bidosensa 15, as shown in FIG. 8, "1" ~
It has a coil board 40 in which a total of 41 coils 41 numbered "41" are closely arranged in a planar and staggered manner. Electrically, as shown in FIG.
A switching circuit 44 is provided between an excitation circuit (oscillation circuit) 42 serving as an excitation AC power supply 43 (FIG. 10), and the switching circuit 44 is switched by a timing generation circuit 45 to obtain a switching circuit shown in FIG.
0 are sequentially formed, and the output is guided to a detection circuit 47 through an amplifier 46. Here, three coils 41 are simultaneously selected as one unit for the excitation coil L01, the excitation coil L02, and the detection coil L03, and at the next timing, one of the three coils 41 on one end side is cut off. At the same time, one coil 41 on the other end side is selected and newly added, so that three coils are sequentially selected and switched. Therefore, even the same one coil 41 functions as the excitation coil L01, the excitation coil L02, and the detection coil L03. Double excitation coil L
01 and L02 are differentially applied, and the detection coil L
The output of 03 is normally zero.

When all the coils 41 are sequentially switched as described above, the area of the lower container wall surface is scanned based on the principle of the eddy current flaw detection method.
A voltage corresponding to a portion in the scan area is output from each detection coil L03. The detected analog signal for one scan (a waveform having positive and negative peaks centering on the position of the welding bead 6) is converted into a digital signal by an A / D converter before or after being sent to the control device 3. The data processing is performed to recognize the position of the welding bead 6 existing in one scan area, and to determine the amount of deviation between the position and the position at which the traveling device 1 has traveled to a local point. Therefore, it is possible to maintain a state in which the ultrasonic beam is applied to the weld bead 6 at right angles.

Control Method Next, a control method according to the present invention will be described.

First, FIG. 13 is a block diagram showing prerequisite electrical elements. Reference numeral 15 denotes the above-described welding bead sensor, which gives bead center position data (the center of the welding line) as an output. Here, the front one of the two front and rear probe arms 12 is referred to as a “front bead sensor”, and the rear one is referred to as a “rear bead sensor”. The S-axis motor 36 moves the probe arm 12 in the left-right direction with respect to the inspection arm 10 (FIG. 2).
The S-axis sensor 37 is a linear potentiometer that detects the movement position in the S-axis direction (see FIG. 4). 48 is a planimeter (distance sensor) attached to the antinode of the self-propelled carriage 7 (see FIG. 3),
When the flaw detection traveling device 1 has advanced along the bead 6 as shown in FIG.
8 and evaluated by the control device 3. However, since a cumulative error may occur in the distance measurement value, the weld bead sensor 15 is provided with a cross bead sensor unit 49 (FIG. 8), and the cross bead sensor unit 49
When b (FIG. 1) is detected, the distance measured by the planimeter is automatically corrected. The θ-axis motor 51 is a motor for turning the probe arm 12 right or left within the range of θ,
The θ-axis sensor 55 is a potentiometer that detects the turning angle θ (see FIG. 7).

Reference numeral 56 denotes a vertical sensor indicated by a dotted line in FIG. The vertical sensor 56 has a pendulum (not shown) that attempts to become vertical by gravity, and has a structure that detects the inclination angle (posture angle Va) of the bogie with respect to the vertical line from the inclination angle α of the pendulum. is there. The pendulum of the vertical sensor 56 is directly connected to a motor (α-axis motor) 57 and its inclination angle α
(FIG. 1) can be variably designated. The α-axis motor 57 has a potentiometer (α-axis sensor)
58 are directly connected to each other so that the instructed tilt angle α of the pendulum can be detected. Reference numeral 59 denotes a servo mechanism for controlling the forward and backward movements of the wheels 8 separately for the left and right wheels.

Here, an outline of the control will be described. A general mobile robot only needs to specify its own position based on global coordinates, but in the case of a flaw detection traveling device, the origin of the coordinate system also moves. FIG. 14 shows a case where the bead sensor 15 is used to control the movement of the flaw detection traveling device 1 along a welding line, and even if the bead sensor 15 is started to be sampled at the point O, a predetermined coil scan is actually performed by the sensor. It is point A that data collection is performed, and that detection data is output at point B with a further delay, indicating that the bogie has already moved to point B during this time. That is, while moving, the robot takes in sensor data at every sampling interval (for example, every 5 mm) shown between points O and B in FIG.

Therefore, between point O and point B, it is assumed that the welding line is a straight line, and the control is performed with an estimated value. First, as information, the distances l ₁ and l ₂ from the two points P1 and P2 on the welding line (weld bead 6) obtained from the bead sensor to the known bogie center line, and the timing obtained from the planimeter 48 when this data is obtained. “Moving Δ
_VL ”and“ attitude angle Va ”obtained from the vertical sensor 56.
Are considered as one set. When the data of l ₁ and l ₂ are collected at the point A, the robot has already moved to the point B, so the coordinates of the values of l ₁ and l ₂ are converted between A and B, and
Are the positions of l ₁ and l ₂ at the point (coordinate system). Some data for each sampling interval are collected and the equation of the welding line is obtained by the least square method. The robot is controlled to follow the welding line according to the obtained relationship between the robot and the welding line. Also, between O and B, where the sensor takes in the data, the previously obtained equation is coordinate-transformed by the robot movement and used as an estimated value of the welding line equation.

More specifically, the control method will be described with reference to FIGS. FIG. 15 shows a computer 60 of the control device 3.
The tasks assigned to the computer 60 are "trajectory calculation" for performing linear approximation of the welding line, "trolley calculation" for determining the next moving direction, and And a task such as "arm operation" for controlling at right angles. The tasks of “trajectory calculation” and “trolley calculation” are shown in FIGS. 16 and 17.
The symbols and coordinate system used below are shown in FIG. In the figure, x _v -y _v coordinate system is a trolley coordinate system, y _v denotes the carriage traveling direction. The xy coordinate system is an absolute coordinate system, and x′− obtained by rotating the xy coordinate system by π / 4.
The y 'new absolute coordinate system is also used.

(1) Trajectory Calculation Processing (FIG. 16) First, based on the data l ₁ and l ₂ from the bead sensor 15, an equation y = ax + b that approximates the welding line in a straight line is obtained.

(A) Coordinates on the trolley (step to) Assuming that the bead sensor 15 starts sampling at the point O in FIG. 14 and moves the bead sensor to the point B via the point A in FIG. Reference numeral 15 performs predetermined coil scanning and outputs detection data at point B. That is, while moving, the robot activates the sensor task at every sampling interval (for example, every 5 mm) shown between the points O and B in FIG. 14 and takes in the sensor data.

In the trajectory calculation task, first, distance data l ₁ and l ₂ (FIGS. 14 and 19) to the bogie center line y _v are read for two points P 1 and P 2 on the welding line (step). x _v -y _v calculated coordinates in trolley coordinate system (step), the P1, P2 is converted into 2 points of the welding line in the x-y absolute coordinate system, further x'-y 'the new absolute coordinate system Convert (step).

In the above coordinate conversion, the distances r ₁ and r ₂ (FIGS. 18 and 19) from the center of the bogie to the front bead sensor and the rear bead sensor are known. Also, from the output of the vertical sensor 56, the bogie posture angle Va shown in FIG.
(I.e. the angle formed y _v-axis of trolley coordinate system with x-axis of the absolute coordinate system) is also known. Deviation angle θ between bogie coordinate system and absolute coordinate system
_v is θ _v = {Va− (π / 2)}. Therefore,
x _v -y _v dolly coordinate system _{-θ v = - {Va- (π} /
2)} only rotated by P1, P2 in advance to convert the x-y absolute coordinate system, when this substitutes the value of l _1, l ₂ and r _1, r _2, the x _v -y _v trolley coordinates Two points P1,
P2 is converted into two points on the weld line in the xy absolute coordinate system. Next, an x'-y 'new absolute coordinate system obtained by rotating the xy absolute coordinate system by π / 4 will be considered in obtaining the welding line equation. This is because the weld line is actually vertical or horizontal, so if the weld line is vertical, the weld line equation y = a
This is because the value of a may be ∞ for x + b. Therefore, P1, P in the xy absolute coordinate system obtained above
2 is converted to the x'-y 'new absolute coordinate system to solve the problem that the value of a becomes ∞. Hereinafter, the values of the x'-y 'new absolute coordinate system are distinguished by adding "".

(B) Carriage movement correction (step to) The robot activates a sensor task at one sampling interval (between points O and B in FIG. 14) while moving, and fetches sensor data. Therefore, the amount of movement of the carriage is predicted, and correction is made for this.

(I) Incremental movement amount of bogie The traveling distance of the bogie path during one sampling period (the current sensor data collection and the previous sensor data collection) is determined by the planimeter 4.
The output of 8 is ΔV _L (step). FIG.
In (a), (n−
1) The carriage posture angle at time and n time is Va (n-1), V
a (n), assuming that the carriage movement during one sampling is an arc, ΔV _L = RΔθ (however, Δθ = θ ₂ −θ _1, which is very small), and θ ₂ = Va (n) −π / 2. , Θ ₁ =
From Va (n−1) −π / 2, the increment x,
Δx and Δy as y components are obtained by the following equations.

Δx = x ₂ −x ₁ = Rcos θ ₂ −Rcos θ ₁ Δy = y ₂ −y ₁ = Rsin θ ₂ −Rsin θ _{1 where} Δx and Δy are xy at (n−1) time and (n) time.
The displacement is in the absolute coordinate system (step). This displacement is also converted to an x'-y 'new absolute coordinate system (step).

(Ii) Shift of P1 and P2 (Step) In order to obtain the locus of the welding line, the coordinates P1 and P2 of the welding line at each sampling must be converted to the current x'-y 'new absolute coordinate system. Must. In other words, in the sampling at the time (n), the (n-1) pairs of P1 and P2 represented by the x'-y 'new absolute coordinate system at the time (n-1) are calculated as shown in FIG.
By shifting by ΔQ ′ as in (b), n sets of P1 and P2 coordinate data in the x′-y ′ new absolute coordinate system at (n) time are obtained (step).

_N P1 (i) = _n−1 P1 (i) −ΔQ ′ _n P2 (i) = _n−1 P2 (i) −ΔQ ′ (i = 1 to n−
1) where _n P1 (i): P1 represented by x'-y 'coordinate system at n times _n P2 (i): represented by the x'-y' coordinate system at n time P2 _n-1 P1 (i): P1 _n-1 P2 (i) expressed in the x'-y 'coordinate system at the time n-1; expressed in the x'-y' coordinate system at the time n-1 Given P2 (c) Equation of welding line by least-squares approximation (step) From the n sets of data of P1 and P2, determine the locus y = ax + b of the welding line in the x'-y 'new absolute coordinate system at time n. Obtained by the method of least squares (step).

The trajectory y = ax + b obtained above is a trajectory at the time of the sensor taking-in at the time n.
There is a difference of about 100 ms between the sensor fetch timing and the calculation timing due to the relationship between the bead sensor. Therefore, in order to eliminate this deviation, the equation y = ax + b of the welding line obtained above is shifted in the same manner as in the case of FIG. 20B, and corrected to the timing at the time of calculation (step).

(2) Dolly Calculation Process (FIG. 17) This process is for moving the dolly while following the welding line obtained in the locus calculation process. In FIG. 21 showing the relationship between the bogie and the welding line, first, the distance l _VB between the welding line and the bogie center and the angle θ _VB between the welding line and the bogie traveling direction (y _V axis) are determined (step). Thus, if the truck is controlled so that θ _VB = 0 and l _VB = 0, the truck follows the welding coin. Since control uses only Va (posture angle), θ _VB
Not controlled with respect to, if l _VB → 0 by adjusting the Va in accordance with the deviation of l _VB, theta _VB is also zero. As for the tracking of the attitude angle, the attitude angle tracking control is performed in the soft servo, and if Va is set, the bogie attitude angle follows.

(A) Judgment of bogie center (step) FIG. 22 (a) shows a case where Va is substantially zero, FIG. 22 (b) shows a case where Va is substantially π / 2, and FIG. 22 (c) shows a posture angle Va. Shows the relationship between the bogie center and the welding line when is approximately π, and indicates that the welding line 1 and the welding line 2 are on the right and left sides of the bogie center, respectively. From the relationship shown in the figure, x′−y
′ In the new absolute coordinate system, when the coordinate value x _mv of the welding line detection point x _m is x _mv > 0, the center of the bogie is on the left side of the welding line, x _mv
When <0, it is determined that the bogie center is on the right side of the welding line.

(B) When the bogie center is on the left side of the welding line (x _mv > 0) It is determined whether the bogie is moving forward or backward (step), and when the bogie center is on the left side of the welding line (x _mv > 0) ), When the bogie is moving forward (step), as shown in FIGS. 23A to 23C, in each of the three modes in which the attitude angle Va is substantially zero, approximately π / 2, and approximately π, respectively. θ _VB is θ _VB <0
The posture angle Va is set so as to be as follows (step). When the bogie center is on the left side of the welding line (x _mv > 0) and the bogie is moving backward (step), 3 in FIG.
In the embodiment, Va is set such that θ _VB > 0 (step).

(C) When the bogie center is on the right side of the welding line (x _mv <0) Similarly, it is determined whether the bogie is moving forward or backward (step), and when the bogie center is on the right side of the welding line (x _mv > 0), when the truck is moving forward (step), Va is almost zero,
Va is set so that θ _VB > 0 in each of the three modes of π / 2 and π (step). When the bogie center is on the right side of the welding line (x _mv <0) and the bogie is moving backward (step), in the same three aspects,
Va is set so that θ _VB <0 (step).

With the Va thus set, the pendulum of the vertical sensor 56 is moved by the α-axis motor 57, and the servo mechanism unit 5 is moved so that the tilted angle of inclination becomes zero.
9 corrects the posture of the cart. In this case, the left and right wheels 8 are independently controlled. For example, one of the left wheel (two front and rear) and the right wheel (two front and rear) is moved forward,
By turning the other side backward, the vehicle turns right or left.

(3) Arm Calculation Process (FIG. 17) This process is a process of calculating the target position of the contact arm in accordance with the movement of the bogie, and obtaining the movement amount of each axis of the arm from the target position.

(A) Next sampling position First, the arm center position (θ-axis rotation center, S-axis movement center) R at the next sampling is determined. In the trolley calculation processing, the trolley movement amount ΔV ₁ during one sampling and the next posture angle Va (i
+1) is obtained, the position at the time i is Q1,
The position at (i + 1) time is Q2, and the x and y components Δx and Δy of the movement amount from Q1 to Q2 are obtained and converted to the x′-y ′ new absolute coordinate system.

(B) Setting of Flaw Detection Point The flaw detection point at the time (i + 1) is a point A where a perpendicular line is drawn from the rotation center R of the probe arm 12 at Q2 to the welding line.
(See FIG. 24).

(C) Setting of Arm Target Position The probe arm 12 is controlled on the θ axis and the S axis. Control is performed so that the probe arm 12 and the welding line are perpendicular to each other. Here, as in the case of the trolley arithmetic processing, the cases are considered separately according to the trolley posture. The position of the S-axis arm center R on the welding line is determined by x ′ based on the relationship shown in FIGS.
−y ′ coordinates x _r ′, y of the arm center R in the new absolute coordinate system
'In, x _{r' r1} positive and negative, as well as from the magnitude relation between + b 'equation ax _r1 weld line _{and' y r1, x r '·} (y r1
When '-ax _r1 ' -b) <0, the left side of the weld line, x _r '
_{(Y r1 '-ax r1'} -b)> found as the right, the weld line at 0.

Next, a target value of the θ axis is obtained. That is, (i
+1) From target attitude angle Va (i + 1) at time (i + 1)
Find θ _VB of time.

(D) Control of θ axis and S axis For the θ axis, the θ _VB (i +
1) Control may be performed in the opposite direction by the amount of minutes. However, the method of controlling the S-axis differs between the coarse flaw detection and the precision flaw detection. In the case of the coarse flaw detection, control is performed such that the S-axis is always on the welding line. On the other hand, in the case of precision flaw detection, movement is divided into a carriage and an arm, movement in the traveling direction is carried by a carriage, and scanning of the flaw detection range is carried out by moving the S axis. That is, as shown in FIG.
The carriage is moved in the traveling direction as B2 → B3, B4 → B5, B6 → B7, and the probe arm 12 is moved to B1 → B2, B3.
It is moved in a direction crossing the traveling direction, such as → B4, B5 → B6.

(I) At the time of moving the trolley At the time of moving the trolley, the trolley performs the tracing operation as the welding line follows, but the probe arm 12 cannot perform the tracing operation. This is because the S axis is at the maximum or minimum position. Therefore, only the θ axis is controlled so that the S axis is not moved.

(Ii) At the time of moving the arm Since the control of the θ axis is performed at the time of moving the bogie, here, S
Only the control of the S-axis is performed by the shaft motor 36, and when the S-axis reaches its maximum or minimum target value, the mode is switched to the bogie movement mode. In the structure of the probe arm, when the θ axis moves, the S axis interferes. Therefore, when giving the target value of the S-axis, a correction for the θ-axis movement is added.

Although the preferred embodiment of the present invention has been described above, the present control method can be applied to a known normal operation mode other than the above-described composite operation mode. Here, the known normal operation mode includes, for example, the following.

(1) As shown in FIG. 26 (a), an ultrasonic beam 162 having a narrowed width with respect to the main beam is emitted.
Of the ultrasonic beam 1
In this mode, a fixed volume of ultrasonic flaw detection is performed at a time by applying the beam 62 to the welding bead 6 and shaking the beam 162, that is, moving the ultrasonic probe 161 in the direction of the arrow 163. Therefore, when performing the entire inspection over the entire length of the weld bead 6, the position of the ultrasonic probe 61 is gradually moved while gradually moving the flaw detection traveling device so as to partially overlap before and after the above-mentioned fixed volume ( (Scanning), and as a result, the ultrasonic probe 61 is scanned with a rectangular scanning pattern 64 as shown in FIG. However,
A considerable amount of inspection time is required to complete the inspection of the entire weld bead, and even if the weld bead has no defects or only one spot, the same time must be spent for the detailed inspection. There are drawbacks.

(2) In a state where the sensitivity of the ultrasonic probe 61 in FIG. 26A is lowered (the beam width of the ultrasonic beam 162 is not changed), scanning is performed roughly according to the scanning pattern 164. In this mode, flaw detection is performed, and precision flaw detection is performed at necessary locations. However, since only the sensitivity of one type of ultrasonic probe is switched, the same rectangular scanning pattern as in the case of the precision inspection is used for the rough inspection for the inspection over the entire length of the weld bead 6. 164 have to be taken, and the scanning length is the same as in the case of only the flaw detection. Accordingly, it takes a considerable inspection time to complete the inspection of the entire length of the weld bead 6.

Further, in the above embodiment, the ultrasonic flaw detecting and traveling apparatus has been described as an example. However, the present invention is not limited to this, and the present invention is not limited to this. The present invention can be applied not only to a destructive inspection robot but also to a mobile robot for other purposes that travels along a welding line.

[0063]

As described above, according to the present invention, even if the coordinate system of the bogie is shifted due to the movement of the mobile robot,
The data l ₁ , l _{2 of the} weld bead sensor by the movement amount
Is corrected to the value at the current time at each sampling interval. Therefore, the equation of the welding line in the current coordinate system is correctly calculated, and based on this equation, it can be estimated and used as the trajectory of the movement at the next one sampling interval. It can be moved according to.

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing an ultrasonic flaw detection traveling device as a mobile robot to which the control method of the present invention is applied.

FIG. 2 is a plan view of the flaw detection traveling device of FIG.

FIG. 3 is a front view of the flaw detection traveling device of FIG. 1;

FIG. 4 is a diagram showing an S-axis drive system of the probe arm of FIG. 1;

FIG. 5 is a partial cross-sectional view showing a structure for attaching a probe arm to the inspection arm of FIG. 4;

FIG. 6 is an end view showing a structure for mounting an ultrasonic probe on a probe arm.

FIG. 7 is a diagram showing a θ-axis drive system of a probe arm.

FIG. 8 is a plan view of a coil board of the weld bead sensor.

FIG. 9 is a schematic diagram of an electric circuit configuration of a weld bead sensor.

FIG. 10 is a conceptual diagram of a detection circuit of a weld bead sensor.

FIG. 11 is an explanatory diagram of coarse detection and precision flaw detection in the flaw detection traveling device of FIG. 1;

FIG. 12 is an explanatory diagram of a scanning pattern of the coarse detection and the fine flaw detection.

FIG. 13 is a block diagram showing a control system of FIG. 1;

FIG. 14 is a transition state diagram of the mobile robot for explaining the control method of the present invention.

FIG. 15 is an explanatory diagram of a control task according to the embodiment of the control method of the present invention.

FIG. 16 is a flowchart showing tasks of the miracle calculation processing of FIG. 15;

FIG. 17 is a flowchart showing a task of the bogie calculation processing in FIG. 15;

FIG. 18 is an explanatory diagram of a coordinate system and codes used for calculation.

FIG. 19 is an explanatory diagram of coordinates on a trolley used for calculation.

FIG. 20 is an explanatory diagram of bogie movement correction used for calculation.

FIG. 21 is an explanatory diagram showing a relationship between a bogie and a welding line in calculation.

FIG. 22 is an explanatory diagram of the determination of the center of the bogie in the calculation.

FIG. 23 is an explanatory diagram of the determination of which side of the bogie center is on the welding line.

FIG. 24 is an explanatory diagram of setting of flaw detection points to be subjected to ultrasonic flaw detection.

FIG. 25 is an explanatory diagram of setting a target position of a contact arm.

FIG. 26 is an explanatory diagram of a flaw detection and a scanning pattern in a conventional flaw detection traveling device.

[Explanation of symbols]

 Reference Signs List 1 flaw detection traveling device 2 cable 3 control device 4 printer 5 spherical tank wall (wall to be inspected) 6 welding bead 7 self-propelled trolley 8 permanent magnet wheel 9 axis 10 inspection arm 11 guide plate 12 probe arm 13 ultrasonic for coarse search Probe 14 Precision ultrasonic probe 15 Weld bead sensor 16 Arm support roller 17, 18 Pressing mechanism 19 Pressing cylinder 20 Linear guide bearing 21 Slotted roller 22 Guide plate 23 Engagement rim 24 Base member 25 Bearing 26 Rotation shaft 27 Bearing 28 Sprocket 29 Gear 30 Sprocket 31 Chain 32 Rack 33,33 'Mounting member 34,34' Shaft 35 Motor (wheel) 36 Motor (S-axis) 40 Coil board 41 Coil 42 Excitation circuit (Oscillation circuit) 43 Excitation AC Source 44 switching circuit 45 timing Raw circuit 46 Amplifier 47 Detection circuit 48 Planimeter (distance sensor) 49 Cross bead sensor unit 51 Motor (θ axis) 52 Operating disk 53 Rod 54 Pin 55 Potentiometer (θ axis sensor) 56 Vertical sensor 57 Motor (α axis) 58 Potentiometer (Α-axis sensor) 59 Servo mechanism 60 Computer 151 Main beam 152 Ultrasonic 154 Scanning pattern (linear scanning) 155 Scanning pattern (rectangular scanning) 161 Ultrasonic probe 162 Ultrasonic beam 163 Arrow 164 Scanning pattern

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Yasuhito Tanaka 3-1-1-15 Toyosu, Koto-ku, Tokyo Ishikawa Shima-Harima Heavy Industries Co., Ltd. Tojin Technical Center (72) Inventor Kayoko Hayakawa Toyosu 3-chome, Koto-ku, Tokyo No. 1-15 Ishikawa Shima Harima Heavy Industries, Ltd. Toji Technical Center (56) References JP-A-63-30756 (JP, A) JP-A-3-104787 (JP, A) JP-A 5-133942 (JP) , A) Fully open 1986-193368 (JP, U) (58) Fields investigated (Int. Cl. ⁶ , DB name) G01N 29/00-29/28

Claims (1)

(57) [Claims] In a mobile robot having a plurality of welding bead sensors for scanning a plurality of coils arranged one by one in a direction intersecting with a traveling direction of a carriage and detecting a bead position by eddy current, scanning of the bead sensor is started and bead position data is activated. For each sampling interval at which sampling of the robot is completed, the distances l ₁ and l ₂ from the two points on the welding line obtained from the bead sensor to the known bogie center line, the movement amount ΔV _L of the robot and the posture angle Va of the robot are set to 1 Input as a set, convert the collected distance data l ₁ and l ₂ to the value at the current time when the collection was completed, and use this together with the data up to the previous time to find the equation of the welding line by the least squares method. In the movement at one sampling interval, the previously obtained equation is coordinate-transformed by the robot movement amount ΔV _{L and} used as an estimated value of the welding line equation. And a robot having a welding bead sensor characterized by controlling the attitude angle Va so that the distance between the welding line and the center of the bogie is zero based on the relationship between the welding line and the welding line, and controlling the robot to follow the welding line. Robot control method.