SE467710B - Test arrangement which utilizes successive connection of sensors and vector transformation - Google Patents

Abstract

The present invention describes an arrangement for simulating a sensor and/or sensor movement by means of for example a number of sensors, for example eddy current- based surface sensors G, which are stationary relative to one another, and a selector/distributor 5. The sensors are for example mounted in a plane parallel to the surface 1 of the test object, and scan the test object for a quantity, for example cracks 2. The resultant imaginary/simulated sensor signal can be or is signal-processed 4 via for example automatic compensation and vector transformation. The principle involves the sensors being automatically connected in turn, for example in a continuously repeated sequence, and signals originating from different sensors in full or in part determining different time portions/sections of a resultant signal, by means of which a quantity or a combination of quantities can be tested and/or measured. <IMAGE>

Description

467 710 The advantage of this is, apart from the fact that no mechanically moving parts are required, that very high sensor speeds can be simulated, which, among other things, means that very short cracks can be detected.

The invention can be described as follows which is to be seen as one of many conceivable examples and embodiments. Figure 1 shows a number of sensors (G) placed above different parts of a test object (1) containing a crack (2). Via the selector / distributor (5) ¿the sensors are connected one by one in sequence, to the sensor supply part (3) which feeds the current sensor with a constant current (Ik) of at least one frequency / carrier frequency. The selector (5) goes / steps continuously and each, takes the time TS, and is called sequence. The voltage across the current sensor is measured and processed in the signal processing part (4), whose task is to detect, for example, rectify and amplify, etc., magnitude, in this case the crack (2).

Figure 2 shows the time course in figure 1. Figure 2A shows the sequence TS along the time axis, and the time Tn which is allocated to the respective sensor and which, for example, controls the incline of the selector etc. When the selector assumes position G1, the voltage across G1 will increase due to Ik, the impedance of the sensor and the inductive time constant of the sensor coil, as shown in Figure 2B.

Correspondingly, a decay process is obtained when the selector steps up to the next sensor, i.e. when the current I is interrupted in G Figure 2B thus shows the input voltage to block 4, where each carrier frequency pulse train refers to its special sensor.

The sensors are individually adjusted, eg balanced and compensated, so that the voltage drop across them is equal when, for example, the test object lacks defects / quantities. If, on the other hand, a crack (2) is located below the sensor G2, the impedance of the sensor will change and then the voltage across G2 will also change, which is shown by the dashed curves in Figure 2.

In Figure 2C, the inward and outward oscillation processes of the sensors have been gated away so that they do not give rise to disturbance phenomena, thus only the middle (Sb) amplitude-stable part in Figure 2B has been included.

Figure 2D shows the detected, for example, rectified signal in Figure 2C.

By using a suitable discharge time constant in the rectifier *, the signal has the appearance shown in Figure 2D, and where a possible crack under sensor G2 that appears to deviate significantly from the normal level at “faultless test object.

Figure 2E shows the filtered crack signal from 2D. un ... .1 467 / u As one easily realizes, the sensors must be electrically very similar to each other if the level in Figure 2D is to be acceptably even and stable when no magnitude, ie in this case crack, is present, otherwise the crack signal would be hidden due to disturbances caused by various oils. These differences then also often include the selector / distributor (In the case, for example, the signal in Figure 2E, despite measures taken to make the sensors equal to each other, exhibits interference, it may be advantageous to store the signal E for one or more sequences in a normal or reference cases, and use this stored signal directly or indirectly as a comparison signal when comparing with signals, we measured other times / sequences.The storage of signals from a sequence can take place with the help of both analog and digital circuits eg analogue disk register, computer etc.

The sensors can, if appropriate, be continuously connected to each sensor supply, and the selector (5) only used to select the input signal to block 4. The great advantage of the principle according to Figure 1 is, however, that the same sensor supply and signal processing can be used for all sensors. It should also be mentioned here that the selector (5) is of electronic type, for example consisting of analog C-MOS switches controlled on conventional V from, for example, crystal-controlled counters and the like.

In order to limit transient disturbances and the like when switching from one sensor to another, the switching can be done, for example, via controlled resistors and the like in order to bring about a soft / successive switching.

In connection with storing / remembering signal information, for example in a computer, opportunities arise to change the order of the different sensor signals, so that several so-called resulting signals are obtained, representing different surface scanning patterns. You can thus use the stored signals / information from several sensors to, according to your wishes, several different sensor paths, including a type of reorientation principle with extensive use.

Figures 1 and 2 show that it is possible to determine to which time pulse the average crack signal is derived, i.e. which sensor detects the crack. It follows that it is possible to determine where on the test object the crack / size is located.

In summary, it is thus possible to determine both the position of size and orientation / extent.

Properly applied and embodied, the invention includes in principle all the advantages offered by a rotating surface sensor according to, for example, Swedish Patent Document No. 8503894-1. 467 710 Regarding the sensors, the following can be mentioned. Figure 1 shows a number of sensors (G1-Gn), which together form a sensor arrangement. The sensor arrangement can have different appearances, for example as shown in Figures 3 and 4.

For the invention to work perfectly, the sensors should be as similar to each other as possible both mechanically and electrically. For this reason, the sensors in Figures 3 and 4 have been constructed as foil coils, i.e. they have been etched out in a similar manner as conventional electronic circuit boards are manufactured. The surface sensor coil, i.e. the sensor, can be etched on one side of the circuit board laminate, but of course you can also etch a coil on each side so that you get double-sided sensors. In the double-sided design, the coils can be given winding directions so that their fields cooperate, whereby the series connection takes place, for example, by connecting the coil windings via the center bushing. In Figs. 3 and 4, which can be regarded as double-sided, the connecting ears then constitute the beginning and end of the series-connected coils.

If you want to go one step further, you can stack / stack several circuit board laminates / sensors on top of each other and connect them in series (or connect them in parallel), thereby obtaining a multi-layer sensor coil with accompanying higher inductance.

For several reasons, it can be advantageous to place several sensors on a circuit board / laminate or other type of foil carrier. One reason may be the economic one and another that the sensors in this way become almost identical to each other as the foil thickness and foil width become the same for all sensors.

Figure 4 can be seen as a circuit board containing 16 surface sensors, for example intended for crack detection on flat blank surfaces and where the circuit board is placed in a plane parallel to the blank surface. The circuit board can of course, if appropriate, move relative to the subject surface and, for example, rotate around its own center. Figure 3 may also consist of a circuit board but may include either a rigid circuit board or a flexible one. In the latter case, you can bend the circuit board around, for example, a wire rod that moves in the direction of the arrow. In this way, a sensor arrangement is obtained which encloses the selection wire, and thereby the entire surface of the selection wire is scanned when it passes through the sensor arrangement. This can be seen as a repeat sensor built up of a surface sensor or a combination sensor. If, due to, for example, the limited inductance in the foil coil, you want to increase the impedance of the sensor, you can, for example, provide the coils with a ferrite core or, for example, glue ferrite disks to the side of the circuit board facing away from the test object or its surface.

In Figure 3, a clarifying sketch has been added which shows how the flexible sensor arrangement encloses the wire. This arrangement can then, for example, be placed in a metal pipe and cooled with water which, for example, forms a water gap between the metal pipe and the sensors / sensor arrangement.

In order that, for example, non-longitudinal cracks on the wire rod can fit undetected between the sensor coils, these should, for example, be placed in a zigzag pattern or the like, so that their field images overlap each other in the longitudinal direction of the wire rod.

A sensor pattern according to Figure 4 can, for example, be used on flat surfaces, such as slabzs surfaces. Here, too, it may be appropriate to supplement the scarring with an inner or outer ring of sensor so that a zig-zag pattern is obtained, or to completely fill the surface with sensor.

As described here, etching simple foil donors opens up many interesting possibilities. It is also possible, for example, to leave the foil sensor attached to a flexible foil carrier which can be glued or adhered to the test object, for example in corners and other hard-to-reach places, including a type of sensor patch.

You can also diffuse on a metal layer on a ceramic plate and then etch the foil spiral directly on the plate. In this way, for example, the lift-off distance can be reduced to a maximum during hot testing, especially if the ceramic plate on the foil side is water-cooled. Correspondingly, the ceramic plate can be replaced with a ceramic tube or the like when testing elongated, eg round test objects, through which the test object, for example the wire, may pass. The ceramic tube is suitably cooled with water on the outside, for example via a water gap.

Despite the fact that the sensors via etching technology etc. have been made almost identical to each other, they can still differ so much from each other that it interferes with the measurement of magnitude. For this reason, it is often necessary, for example directly on the sensor laminate next to the sensors, to mount trim components for individual balancing and compensation of the respective sensor coil. Relevant components are, for example, potent meters and rotary capacitors etc. mounted in parallel or in series with the sensor coil. Since the sensor has a Q-value, it may also be relevant to adjust the gain in the respective sensor branch so that the signal in the reference case has the same signal amplitude.

Each sensor should also have an individual height adjustment option so that the off signal in the reference case can be set to exactly the same value in all sensors.

Etching coils is not in itself new, but new is the way of using the coils as described here, ie utilizing the advantages that etched multiple coils bring.

The sensor arrangement can advantageously have a relative movement, for example so that the sensor arrangement is given a limited oscillation movement or the like.

This reduces the requirement for the donors' field images to overlap at rest. 467 710 In eddy current testing with surface sensors, as is well known, suppression of lift-off dependence is essential. The Swedish patent documents 7507857-6, 7813344-4, 8206678-8, 8302738-3, 8400698-0, 8400861-4 and others, describe different ways of mastering the LO problem.

The present invention is subject to the same in eddy current testing. type of problem as the sensor arrangement in practice is almost always slightly skewed relative to the surface of the test object. The signal processing block 4 in Figure 1 must therefore in some way suppress the LO dependence in eddy current testing. If the sensor arrangement is skewed, the LO will vary the sensor to the sensor so that block 4 generally perceives it as a sensor moving over the test object with varying LO distances.

The invention therefore advantageously comprises the principles described in the above-mentioned patent documents, and the terminology and content of which are otherwise applicable in this description.

Against this background, the invention can be briefly described as a method and / or device based on sensor multiplexing method in combination with vector transformation technique, or perhaps simpler, vector transformation in combination with simulated sensor movement based on the use of several sensors.

The invention also includes compensation as described, for example, in Swedish patent application 7813344-4, but then, as in the case of vector transformation, for example together with a fictitious / simulated sensor and sensor movement.

The compensation which via control servo strives to automatically balance from sensors originating signals to zero works in the present invention, for example via a multiplexing procedure, with several sensors in a certain sequence, for example so that the respective sensors are compensated in whole or in part.

In this way, the compensation part can also be common to all sensors, which is of course an advantage.

Both vector transformation techniques and compensation techniques are known per se by the patent documents cited herein. However, using these techniques for transformation and compensation of a fictitious or simulated donor is unique, and as far as is known is not described in the literature.

A consequence of this is that the invention also comprises the use of several carrier frequencies and / or carrier frequency components.

In Figure 5, as an example of how sensors can be balanced, individual trim components have been crossed out for the sensor GN. It can sometimes also be relevant to add trim resistors in series with the sensors to, for example, compensate for varying Ron in the selector / distributor. 467 'F10 Figure 5 shows a more detailed diagram of how the invention can be designed and function.

Here, the simulated sensor, i.e. a variable fictitious sensor that can be considered as a type of "average value sensor" representing several phase sensors (G1 - Gn) has been drawn as a variable sensor with the designation GM.

The subsequent electronics thus believe to work with the fictitious and variable sensor GM, which is characteristic of the invention! So some explanations. Pos 10 consists of a KL oscillator which can be crystal controlled by driving. Item 11, which is fed with clock pulses from item 10, contains counters, decoders, gate pulse Sb, sensor addresses ADR etc.

Pos 12 is an analog gate opened by the Sb pulse. Item 13 with associated electronic potentiometers, including the phase-controlled rectifiers, items 21 and 22, constitute simple zero-compensation servos and should not require further explanation. Pos 14 consists of a simple op amplifier.

Posts 15 and 16 consist of bandpass filters tuned to current carrier frequencies, ie in this example a high (H) and low (L) frequency, respectively.

In this way, the different carrier frequencies are separated so that you get a can for each carrier frequency. In other respects, see cited patent documents.

Posts 17 and 18 consist of phase-controlled rectifiers whose output signals are used as input signals to the transformation circuit, consisting of the weighting potentiometer 19 and subsequent op amplifiers of the summation type.

As stated earlier, the sensor arrangement according to Figure 4 can be near or, for example, movably mounted via the axis AX at point P. If in the latter case the sensor arrangement is allowed to move / rotate back and forth, ie oscillate with the angle 1 S, the field images from the sensors will overlap each other well, whereby, for example, the scanned surface becomes more homogeneous in nature. As an alternative to different zig-zag patterns, you can apparently oscillate the sensor arrangement, which in some applications is clearly more advantageous. A consequence of this is that you can connect the sensor, for example via a simple foil tape cable (BA) in the center of the sensor arrangement to the electronics, including the sensor supply.

Correspondingly, for example, water for cooling can be connected at point P via a suitable hose and the like.

When testing / measuring round test objects such as pipes etc., one can in the same way oscillate the curved / tubular sensor arrangement according to Fig. 3, and thereby obtain a very even magnetic field image, which is especially suitable for detecting so-called long axial faults and the like.

The combination of a high simulated sensor speed and a limited oscillation provides great metrological advantages compared with conventional mechanical solutions and arrangements. no 4 467 710 The tubular sensor arrangement according to figure 3 also enables measurement of, for example, the roundness of the wire rod, dimension in several directions, the position of cracks along the periphery of eg wire, etc., due to, among other things, the homogeneous and even field image.

Some definitions appear from the following.

TEST PROJECT refers to, for example, wire, rod, pipe, blanks, etc.

GREATNESS refers to, for example, dimension, shape, change, surface defect, crack, etc., ie size is a wide-ranging concept.

SENSORS refer to, for example, both sensors, sensors and combinations of these, ie everything that the person skilled in the art includes in the term sensor, eg eddy current-based surface sensor coil.

By TIDDEL is meant, for example, a part of the time a time sequence (TS) comprises, an example of tiddel is T1 in figure 2.

PULSE refers to, for example, a signal of a certain duration.

By CONSTANT CURRENT GENERATOR is meant, for example, a generator for generating constant current, eg alternating current, and with in this context large internal resistance, which means that the generator current becomes largely independent of the load resistance. å In this context, for example, a RESULT SIGNAL refers to a signal that is directly or indirectly composed of sub-signals received from different sensors at different times. The signals D and E in Figure 2 are examples of resulting signals.

LO refers to lift-off.

The concept of quantity can also include a combination of quantities.

It should also be pointed out that the functions specified in the invention can be realized with both software and hardware. It is, for example, possible to use a computer or the like for storing digital measured values, etc., which is why this is also included in the invention.

The concept of size can also include embers and the like as described, for example, in Swedish patent document 8603604-1.

The invention can be varied and applied in various ways within the scope of the appended claims.

In addition to the already mentioned patent documents, the Swedish patent documents 8500065-1 and 8505541-6 have points of contact with the present invention.

Likewise, the technique described in the Swedish patent 8206678-8 can advantageously be used within the framework of the signal processing part of the present invention, but then utilizing time delay of the fictitious and / or simulated sensor signal.

Claims (12)

9,467,710 PATENT CLAIMS Device for testing and / or measuring test objects such as wire rod with respect to a quantity, eg surface crack, or a combination of quantities, which device comprises sensors (G), eg eddy current-based surface sensor coils, arranged so that they each time completely or partially detects different parts of the test object and at least one selector (5), characterized in that sensors are automatically connected via selectors in a certain order, eg in a recurring selected sequence, and that from different sensors directly or indirectly derived signals are combined into a resultant signal, which can be considered, for example, as obtained from a fictitious simulated sensor (GM), which signal is processed via vector transformation technique into a resultant vector-transformed signal which is sensed with respect to said quantity. Device according to claim 1, characterized in that the sensors are arranged in planes, substantially parallel to the surface or part thereof of the test object, their field images overlapping each other seen in the longitudinal direction of the test object. Device according to claim 1, characterized in that the sensors are supplied with current, the frequency of which is suitable for eddy current testing / measurement. Device according to one or more of the preceding claims, characterized in that the respective sensor or signal originating from the sensor is arranged to be sensed / gated via gating technology, for example so that the sensor signal is only sensed during the middle part of the current time pulse ( Sb), where the signal is most stable. Device according to one or more of the preceding claims, characterized in that at least one signal originating from the sensor, directly or indirectly, is arranged to be compensated automatically. Device according to one or more of the preceding claims, characterized in that at least two different carrier frequencies are arranged to be used. 10 Device according to one or more of the preceding claims, characterized in that the sensors are individually adapted and / or adjusted automatically. Device according to one or more of the preceding claims, characterized in that the position of the sensors relative to each other and / or the test object is adjustable via automation. Device according to one or more of the preceding claims, characterized in that the sensors, in whole or in part, consist of foil coil. Device according to one or more of the preceding claims, characterized in that the sensors, for example foil coil, are mounted directly on the insulator or protection on a ceramic washer. 11. ll. Device according to one or more of the preceding claims, characterized in that the resulting vector-transformed signal, directly or indirectly, is arranged to be stored for use as a comparison signal in comparison with a signal obtained at another time. Device according to one or more of the preceding claims, characterized in that the sensor arrangement is arranged to oscillate.