RU2771743C1 - Apparatus for monitoring the defect rate of wire insulation - Google Patents

Abstract

FIELD: testing.

SUBSTANCE: invention relates to the technology for electrical testing and is used to monitor the quality of insulation of wires in the process of manufacturing windings of electrical products therefrom. The technical result is achieved by using, as a defect sensor, elements of a winding machine secured on a conductive platform and connected to the grounded body of the winding machine by dielectric ties through an electric insulation gasket, wherein additionally introduced into the meter are a high-resistance amplifier, a delay line, a computer with a neural network, a router, a pickup generator, a wire length counter, and an inductor made in the form of a winding wound on a ferrite rod, said rod with the winding is inserted into the coil of the winding wire.

EFFECT: simplification of the apparatus for monitoring the defect rate of wire insulation.

1 cl, 2 dwg

Description

The invention relates to the technique of electrical testing and can be used to control the quality of wire insulation in the process of manufacturing windings of electrical products from them.

A device for monitoring the defectiveness of the insulation of wires, described in [1].

The device is designed to control high voltage spot damage, for wires with a core nominal diameter of more than 0.050 to 1.600 mm inclusive. At the same time, to control wire insulation defects on wire insulation from 0.050 to 0.25 mm, the high-voltage electrode (sensor) used is made in the form of two. The rollers in the device must be made of stainless steel and provide, each, contact with the wire for a length of (25 ± 2.5) mm.

When checking for point insulation of wires with a nominal wire core size of 0.250 to 1.600 mm, one high-voltage electrode in the form of a roller is used. The roller must be made of stainless steel and provide contact with the wire for a length

30 мм. 25

30 mm.

The disadvantage of the device is, firstly, the low versatility of the sensor, since for wires with a core diameter ranging from 0.050 to 0.25 mm, an electrode is used - a sensor made in the form of two rollers, and the controlled wire is pulled through 4 rollers , two of which are guides and the other two are sensor electrodes. For wires with a diameter ranging from 0.25 to 1.600 mm, this sensor is no longer applicable, and instead a single high-voltage electrode of a larger diameter is used.

Secondly, both when testing wires with a core diameter ranging from 0.050 to 0.25 mm, and when testing wires with a core diameter ranging from 0.25 to 1.600 mm, the wire is repeatedly bent. This leads to high mechanical loads on the wire insulation from the side of the rollers, which leads not only to a weakening of the mechanical and electrical strength of the insulation of the controlled wire, but also causes additional defects in the wire insulation. Therefore, using the prototype device, only selective control is carried out, at a constant and relatively low wire pulling speed equal to (275 ± 25) mm/s.

A device for monitoring the defectiveness of wire insulation [2]. The device contains a sensor-electrode, a high voltage source, the first output of which is connected to the first input of the defect pulse shaper, the output of the counter of the number of defects connected to the input, a speed pulse shaper and a speed sensor are introduced into it, and the second output of the high voltage source is connected to the sensor - electrode, the speed sensor is connected through the speed shaper to the second input of the defect pulse shaper, while the latter consists of a reference voltage source, a shaper of the leading and trailing edges and a differential amplifier, the first, second inputs and output of which are connected respectively to the first input of the defect pulse shaper, to to the output of the reference voltage source and to the first input of the shaper of the leading and trailing edges, the second input and output of which are respectively the second input and output of the defect pulse shaper.

The disadvantage of the device is the high control voltage, as well as the complexity of the design, due to the fact that the electrode-sensor and the speed sensor are made in the form of two separate functional blocks.

Closest to the claimed is a device for monitoring the defectiveness of the insulation winding wires, described in [3].

The prototype device contains a sensor-electrode, a high voltage source, a defect pulse shaper, consisting of a reference voltage source of the leading and trailing edge shaper and a differential amplifier, a defect counter, a speed sensor, a speed pulse shaper, while the first input of the high voltage source is connected to to the first input of the defect pulse shaper, the output of which is connected to the input of the defect counter, the second output of the voltage source is connected to the sensor - electrode, the speed sensor through the speed pulse shaper is connected to the second input of the defect pulse shaper, the first and second inputs of the differential amplifier are connected respectively to the first input of the shaper pulses of defects, to the output of the source of the shaper of the leading and trailing edges, the second input and output of which are, respectively, the second input and output of the defect pulse shaper, a pulse counter is additionally introduced into the device with speed, a key device, a defect length counter with an adjustable conversion factor, a trigger, an LED and a photodiode , while the sensor-electrode and the wire speed sensor are made in the form of a single functional unit, made in the form of two stainless steel rollers, having a U-shaped groove along generatrix, and the rollers are placed in a housing, which is made in the form of a channel, between the parallel walls of which a dielectric base is fixed to accommodate the sensor elements, also made in the form of a channel, the parallel walls of the specified base are fixed with fasteners to the parallel walls of the sensor housing, and the base of the said base , located perpendicular to the base of the sensor housing, two metal rocker arms, two springs, two sliding contacts, two outputs for connecting a power source, two guide bushings, a disk with through radial slots evenly made in it, one plane of which o is made in the form of a cylindrical glass, and the rocker arms are made in the form of metal plates, at one end of each of which cylindrical axles for bearings are rigidly fixed perpendicular to the plane of the plate, at the other end of each rocker plate, holes for the axles are made perpendicular to the plane of the rocker arms, which are rigidly fixed on the dielectric base for placing sensor elements, rotating rollers, pressed against each other by means of springs by generating surfaces at the point of contact lying on the vertical axis of symmetry of these rollers, a glass of said disk with radial slots is coaxially attached to the side surface of one of the rotating rollers, along the generatrix surfaces of the rollers grooves are made that lie against each other when the rollers come into contact and serve to fix and limit the movement of the wire in the transverse direction, bearings are pressed into the central part of the rollers, mounted on the above-mentioned cylindrical axles o fixed on the movable end of the rocker arms, the fixed ends of the rocker arms, with holes on them, are dressed on the axis, mechanically fixed on a dielectric base to accommodate the sensor elements, the rollers are pressed against each other by their generatrix surfaces with the help of two springs, one end of which is mechanically fixed to one of the rocker arms , and the other two ends of the springs are mechanically fixed to the dielectric base to accommodate the sensor elements, voltage is applied to the working surfaces of the rollers by sliding contacts made in the form of elastic leaf springs, one end of which is pressed against the axes of the rollers, the other end is electrically and mechanically connected to the end with the leads to connect the power source, guide bushings are fixed in the walls of the sensor housing, the longitudinal axes of symmetry of which coincide with the axis of the wire, the ultraviolet LED is located in the gap between the disk with through radial slots uniformly made in it, from the side of the plane, shaped in the form of a cylindrical cup and the surface of the roller to which the cup is mechanically attached, the ultraviolet photodiode is located on the opposite side of the mentioned disk with radial slots, while the output of the photodiode is connected to the input of the speed pulse shaper, one of the outputs of which is connected to the key device, the output of the key device connected to the input of the defect length counter with an adjustable conversion factor.

The disadvantage of the device is the impossibility of using the specified device for monitoring defects in the wire insulation introduced by the elements of the winding equipment (elements of the winding machine) into the wire insulation of the manufactured windings of electrical products, the manufacture of a special high-voltage defect sensor, the need to use high voltage during control, the need to strip the wire from insulation for galvanically connecting it to a common bus or power source.

The technical problem posed in the framework of this invention is to significantly simplify the device.

The solution of the stated technical problem is achieved by the fact that the device for monitoring the defectiveness of the wire insulation, containing a defect sensor, a counter for the number of defects, a counter for the length of defects, and a speed sensor connected to the input of the speed pulse shaper, while the elements of the winding machine, which are fixed, are used as a defect sensor. on the electrically conductive platform and through an electrically insulating gasket, they are connected by dielectric couplers to the grounded body of the winding machine, while the meter additionally includes a high-resistance amplifier, a delay line, a computer with a neural network, a router, a pickup generator and an inductor, which is made in the form of a winding wound on a ferrite rod, said rod with a winding is inserted inside the winding wire coil, the inductor winding leads are connected to the output of the pickup generator, the input of the pickup generator is connected to the output of the frequency multiplier , and the output of the pickup generator is connected to the input computer with a neural network, while the output of the pickup generator is also connected to the input of the wire length counter, the output of which is connected to the input of the computer with a neural network, while the electrically conductive platform is connected to the input of a high-resistance amplifier, the output of which is connected to a key device through a delay line, and to computer with a neural network, the output of the key device is connected to the input of the defect length counter and to the input of the defect pulse shaper, the output of which is connected to the input of the counter of the number of defects, the output of the computer with the neural network is connected to the input of the router, the output of which is connected to the inhibiting input of the key device.

In FIG. 1 shows a block diagram of the device of the claimed device.

In FIG. 2 shows signal diagrams that serve to explain the essence of the invention.

In FIG. 1 introduced the following designations: 1 - wire; 2 - coil; 3 - template; 4 - rewinder roller; 5, 6, 7, 8 - rollers of the winding machine; 9 – machine body; 10 – reel roller platform; 11 – machine roller platform; 12 – speed sensor; 13 – speed pulse shaper; 14 - pickup pulse generator; 15 - emitter;

16 - high-resistance amplifier; 17 - computer with neural network, 18 - router; 19 - low-frequency filter; 20 - counter of the length of input defects; 21 – defect counter; 22 - wire length counter; 23 - dielectric couplers 24 - key device; 25 - delay line.

The essence of the proposed device is as follows. When winding windings in the wire insulation, two types of defects can occur: input and technological defects. Input defects are those defects that are already present in the insulation of the wire when it is delivered. Technological defects are understood as damage to the wire insulation by elements 4, 5,6,7,8 (Fig. 1) of the winding equipment up to the wire strand. In a real winding machine, elements 4, 5,6,7,8 are disk-shaped rollers through which wire 1 passes when winding it on template 3. The distance L _i between each of the mentioned elements is determined and entered into the memory of the neural network 17. Measurement and the introduction of the mentioned L _i distances between the elements of the winding machine into the neural network is necessary for the correct determination of the number and extent of defects in the wire insulation and the elimination of errors in the control process.

Errors in the control process can be due to the fact that the same defect can be counted as many times as the number of times the wire strand in the defective place of the insulation comes into contact with various elements of the winding machine. To prevent this from happening, the neural network considers various options for a possible false count and develops algorithms for their elimination. Let's assume that there is some defect in the wire insulation in the state of delivery or some defect was formed during the winding process, and it passes through the element 4 of the winding machine, which is the rewinder roller (Fig. 1). The core of the wire in the place of this defect during the further movement of the wire, the movement of the wire during winding can come into contact only with rollers 5, 6, 7 and 8, which are elements of the winding machine. Let us calculate the possible distances S _j , after passing which the defect found in the element 4 of the winding equipment will reach the subsequent rollers and can cause a short circuit of the wire core with these elements. The path S ₁ to the roller 5 is equal to the distance L ₁ , ie S ₁ = L ₁ . The path S ₂ to roller 6 is equal to S ₂ = L ₁ + L ₂ ; S ₃ to roller 7 is equal to S ₃ = L ₁ + L ₂ + L ₃ ; the path S ₄ to roller 8 is equal to S ₄ = L ₁ + L ₂ + L ₃ + L _4.

If the defect caused the contact of the wire strand on roller 5, then when moving during the winding process, it can cause short circuits on rollers 6, 7 and 8, having passed the corresponding paths S ₅ \u003d L ₂ ; S ₆ \u003d L ₂ + L ₃ ; and S ₇ \u003d L ₂ + L ₃ + L _4.

If a defect has arisen or manifested itself on the roller 6, then it can later cause short circuits on the rollers 7 and 8, having passed the corresponding paths S ₈ = L ₃ ; S ₉ \u003d L ₃ + L ₄ .

If a defect has formed and manifested itself as a short circuit of the wire strand on roller 7, then it can cause short circuits only on roller 8, having passed the appropriate path

S ₁₀ = L ₄ .

The considered options for possible distances traversed by a defective section in the wire insulation from the element in which it appeared as a short circuit to subsequent elements in which it can cause a repeated short circuit and cause a false count of defects are introduced into the neural network in the process of its training.

_. When the wire 1 moves in the process of winding the windings, the speed sensor 12 (Fig. 1) generates speed pulses, the frequency of which is proportional to the speed V of the wire movement. These pulses are fed to the input of the shaper 13 speed pulses, which is essentially a frequency multiplier. In the shaper 13 speed pulses are formed by the voltage and steepness of the fronts. The generated speed pulses are fed to the input of the pickup pulse generator 14, and from its output to the input of the emitter coil 15. The emitter 15 emits these pulses and induces the corresponding pickup, in the form of a series of these pulses, in the coil 2 of the winding wire.

At the same time, the generated speed pulses in the pickup generator 14 enter the counter 22 of the length of the controlled wire.

To the counting input of the counter 22 of the length of the controlled wire, speed pulses are continuously received from the pickup generator 14, with a repetition period equal to the passage under the speed sensor, of a fixed certain lengthl _uh wires, for examplel _uh = 0.025 mm (Fig. 2 plots A and B).

Since the duration of one speed pulse corresponds to the passage of a strictly fixed elementary wire length l _e through the defect sensor, the value of which remains unchanged with a change in speed, the controlled wire length is determined by the value l _pr \u003d l _e × m _pr , where m _pr is the number of speed pulses passed into the counter 22 during the control.

. При этом за время одного периода сигнальных импульсов провод пройдет расстояние l _э , принятое за единицу измерения протяженности дефекта, равное по величине протяженности эквивалентного точечного повреждения (фиг. 2 эпюры А и Б)Let's designate the repetition period of speed pulses from the shaper (frequency multiplier)13 speed pulses through T ₁ . At the speed of the wire V ₁ , pulses appear at the output of the shaper 13, with a frequency f ₁ =

. In this case, during one period of signal pulses, the wire will pass the distance l _e , taken as a unit of measurement of the length of the defect, equal in magnitude to the length of the equivalent point damage (Fig. 2 diagrams A and B)

l _uh =V _one ×T _one (one)

When measuring wire pulling speeds by a factor of g, the frequency of equivalent point damage pulses changes proportionally to it by a factor of g, which leads to the invariance of the value determined by expression (1).

Indeed, the frequency of the speed pulses changes in proportion to the speed of the wire V _pr

V_пр (2),f= K ₁

V _pr (2),

where K ₁ - coefficient of proportionality, depending on the design of the speed sensor.

During one period of the voltage induced in the speed sensor, a section of wire with a length l _e passes through the sensor-electrode , equal to

, (3)l _e \u003d V _pr × T _e

, (3)

where T _e \u003d l / f - the period of pulses in the speed sensor.

As follows from expression (3), the value of l _e does not depend on the speed of the wire. Taking l _e as a unit of measurement, you can determine how long the wire l _pr passed through the wire length sensor for any time interval, if you count the number of speed pulses n, in the counter (Fig. 1) during the control time T _con of the specified wire segment.

L _pr \u003d n × l _e , (4)

where l_etc -the length of the piece of wire passed into the counter 22 of the length of the controlled wire; n is the number of speed impulses during the time T_con wire movement under control.

The pickup pulse generator 14 supplies these pulses to the input of the emitter 15, which induces EMF pulses in the coil 2 of the winding wire.

When winding wire sections with defect-free insulation pass through elements 4, 5, 6, 7, 8, if the elements of the winding machine do not damage the insulation, no changes, except for counting the number of speed pulses in the counter 22 of the length of the controlled wire, occur in the device signals.

Let at some time t_one the core of the wire in the defective place of the insulation came into contact with one of the elements 4, 5,6,7,8. Let the duration of the contact of the corresponding element with the core wire lasted until a certain time t₂. Since all of these elements are electrically isolated from the grounded body of the winding machine, then on the element with which the wire core came into contact, pulses of the EMF induced in the wire appear (Fig. 2 diagram B). These pulses are fed to the input of a high-resistance amplifier 16 (Fig. 1), are amplified, and from the output of the said amplifier are fed to the input of the neural network 17. If this was the first contact of the wire strand with any of the elements of the winding machine, then the neural network does not generate any signals, and the speed pulses delayed in the delay line 25 for the time of polling and decision-making by the neural network 17 pass through the key device 24 to the input of the defect length counter 20 and to the input of the defect pulse shaper 19. In shaper 19, a defect pulse is generated, the duration of which is T_i equals T_i= t₂-t_one (Fig.2 diagram D) the time of contact of the wire core at the defect site with one of the elements of the winding machine. This pulse is input to the counter 21 of the number of defects and counted. The length of the defect is determined by the number of counted speed pulses passed into the length counter 20 .

If this was not the first contact of the wire strand with any of the elements of the winding wire, for example, at the time t ₃ , but the first contact with some element of the winding equipment occurred at time t ₁ , then the neural network 17 counts the number N of speed pulses for time interval t ₃ - t ₁ and calculates the value L _k of the path that the wire has passed during the time interval t ₃ - t ₁ . The neural network compares the value L _k with each of the possible distances S _i between the elements of the winding equipment, after which a previously detected defect can cause a re-closing in subsequent elements of the device. The equality of L _k = S _i to one of the variants of the indicated distances, calculated and stored in the neural network, indicates that the defect that manifested itself at time t ₃ was previously registered by counters 21 and 20 for the number and length of defects, when it is in time t ₁ passed through some of the elements of the winding equipment. Therefore, in order to exclude a possible error that occurs when the same defect passes through various elements of winding equipment, and repeated short circuits of the wire core to these elements, i.e. false counts of the number and extent of defects, in case of repeated contacts of a previously registered defect, with subsequent elements of winding equipment, must be excluded. This exception occurs as follows. When the equality L _k \u003d S _i is fulfilled, the neural network issues a prohibiting signal, the duration T _zap , which is determined by the contact time t ₄ - t ₃ , (Fig. 2 diagram D), and usually by some value ∆ exceeding the specified time, i.e. e. T _zap \u003d (t ₄ - t ₃ ) + ∆. The value of ∆ is usually determined experimentally from pragmatic considerations. The generated prohibition pulse arrives at the key device 24 and prohibits for the time T _zap the passage of the signal delayed by the line 25 from the high-resistance amplifier 16 to the counters 21 and 20 for the number and length of defects, these counters are turned off for the time T _zap and the specified defect and its length in these counters are not counted. After the time T has _elapsed , the key device returns to its original state, allowing the signal to pass to the counters 20 and 21 from the output of the high-resistance amplifier 16.

If L _k is not equal to any of the above-mentioned values S _j of the calculated options, then the neural circuit does not generate a prohibiting signal, and the key device passes the amplified EMF pulses delayed in the delay line to the input of counters 20 and 21.

Example of a specific implementation

A control was carried out using the claimed device for the defectiveness of the winding wires, the block diagram of which is shown in Fig. 1. When winding, the winding wire passed through the winder roller 4 and rollers 5,6,7,8, which are elements of the winding machine. These rollers were isolated from the grounded body of the winding machine 9. To do this, they were fixed on metal platforms 10 and 11, and these platforms were attached to the machine body 9 through electrical insulating gaskets made of Teflon with dielectric ties 23 made of caprolactam.

An inductive sensor 12 was used as a speed sensor. The speed sensor 12 and a rotating disk 3, on which a template of the manufactured random winding of the stator of the electric motor was fixed, are shown at the bottom of Fig. 1 to explain the operation of the inductive speed sensor 12. When winding the windings on the template, the steel disk 3, on which this template is located, starts to rotate. When the toothed part of the disk 3 passes near the coil of the speed sensor 12, due to a change in the magnitude of the magnetic flux, electrical impulses are induced, the frequency of which is determined by the size and number of teeth on the disk 3 and the speed of rotation of this disk. With a uniform disk rotation speed, the frequency of the induced pulses remains constant. When the disk rotation speed changes, the frequency of the mentioned pulses also changes, in proportion to the change in the disk rotation speed. The pulses obtained at the output of the sensor 12 are fed to the input of the speed pulse shaper 13, which is a frequency multiplier. The introduction of a frequency multiplier makes it possible to increase the accuracy in determining the extent of defects. The generated speed pulses are fed to the pickup generator 14, and from its output pass to the coil 15 of the emitter. The electromagnetic pulses of the emitter 15 induce a pulsed EMF in the wire coil 2, the pulse frequency of which is proportional to the speed. During one induced in the coil 2 speed pulse, passes the elementary path l _e . Due to the introduction of a frequency multiplier into the device, it was possible to provide the value

Before the control, the distances between the rollers of the reeler and the winder were measured 4.5, 6.7.8. They were equal respectively L ₁ = 620 mm, L ₂ =500 mm, L ₃ =400 mm and L ₄ =400 mm.

According to the measured distances L _i , variants S _i of possible distances that a defective section can pass after its detection, due to the occurrence of contact of the wire strand, in this section, with some of the elements of the winding equipment, were determined. These S _i are given below:

S ₁ \u003d L ₁ \u003d 620 mm; S ₂ \u003d L ₁ + L ₂ \u003d 620 + 500 \u003d 1120mm; S ₃ \u003d L ₁ + L ₂ + L ₃ \u003d 620 + 500 + 400 \u003d 1520 mm; S ₄ \u003d L ₁ + L ₂ + L ₃ + L ₄ \u003d 620 + 500 + 400 + 400 \u003d 1920 mm. All values of L _i and S _i were entered into the neural network 17. To test the performance of the proposed method, two defects were artificially applied in the wire insulation, the boundaries of which were at a distance of 300 mm, with lengths l ₁ =5 mm and l ₂ =10 mm. First, a piece of wire with artificial defects applied to it was pulled through all the elements of the winding equipment with the neural network turned off. At the same time, both defects caused a short circuit with all elements of the winding equipment, and without a trained neural network, the counter for the number of defects showed 10, which indicated that both defects caused the wire core to short circuit to the corresponding elements of 4,5,6,7,8 machine elements according to once.

After that, the neural circuit was connected and the said piece of wire with two defects deposited in its insulation was pulled at a speed of V ≈ 50 mm/s through the same elements of the winding machine, but using the circuit shown in Fig. 1, which implements the proposed method.

Let us consider how the registration of the number and extent of defects took place.

S_i , то на ключевое устройство 24 из нейронной сети 17 запрещающего сигнала не поступало и импульсы, задержанные в линии задержки 25 прошли в счетчик 20 протяженности дефектов и на вход формирователя импульса дефекта 19, выходе которого появился импульс дефекта длительностью Т₁=0,1. В счетчик 20 прошло 20 импульсов скорости, которые соответствовали протяженности дефекта l₁= l_э×n₁=0,25×20= 5 мм. Счетчик 21 посчитал 1 дефект. At the moment of time from the beginning of the wire pulling t ₁ =12.4 s, the first short circuit was detected. It lasted until t ₂ =12.5 s. At the output of the amplifier 16 during the time T ₁ = t ₂ - t ₁ = 12.5-12.4 =0.1 s, 20 speed pulses appeared. Since up to the time t ₁ no short circuit was found in the circuit, then L _j , _k =0. Neural network 17 compared the value of L _j , _k =0 with each of the options S _i above. Since L _j , _k was not equal to any of the above options for distances, i.e. L _k

S _i , then the key device 24 from the neural network 17 did not receive a prohibiting signal and the pulses delayed in the delay line 25 went to the counter 20 of the length of defects and to the input of the defect pulse shaper 19, the output of which appeared a defect pulse with a duration of T ₁ = 0.1 . 20 speed pulses passed into the counter 20, which corresponded to the length of the defect l ₁ = l _e ×n ₁ = 0.25 × 20 = 5 mm. Counter 21 counted 1 defect.

The second and subsequent closures of the wire strand in one of the applied defects with various elements of winding equipment and the analysis of the analysis options that occurred in the trained neural network are shown in Table 1 below.

At the time t ₃ =18.5 s, the second short circuit of the wire core occurred in one of the defects with some element of the winding machine. From that moment, a series of 40 speed pulses appeared at the output of the high-resistance amplifier 16, which continued until the time t ₄ =18.7 s. From time t ₁ to time t ₃ , the counter 22 of the length of the controlled wire received 1200 pulses, which corresponded to the distance L _k =300 mm, which the wire passed after the first contact of the wire strand in the defective section of insulation, detected at time t ₁ , s some element of the contact winding machine.

Table 1. Analysis of options for closing the wire core to the elements of the machine

contact number,j Contact time, s Contact duration, s Distance L _jk traversed by the wire from the moment of contact j to the moment of contact k, mm Neural network output Forbidding signal of the neural network, (yes, no). one 12.5 0.1 L _0.1 =0

S_{i L01}

_Si No 2 18.5 0.2 L _1.2 \u003d 300 L ₁₂

_Si No 3 24.9 0.1 L _2.3 \u003d 320
L _1.3 \u003d 620 L ₁₃ = S ₁
L ₂₃

_S2 Yes 4 30.9 0.2 L _3.4 \u003d 320
L _2.4 \u003d 620
L _1.4 \u003d 920 L ₃₄

_Si
L ₂₄ = S _i
L ₁₄

_Si Yes 5 34.9 0.1 L _4.5 \u003d 320
L _3.5 \u003d 620
L _2.5 \u003d 920
L _1.5 \u003d 1120 L ₃₄

_Si
L ₂₄ = S _i
L ₁₄

_Si
L ₁₅

_Si Yes 6 40.9 0.2 L _5.6 \u003d 320
L _4.6 \u003d 620
L _3.6 = 920
L _2.6 \u003d 1120
L _1.6 \u003d 1420 L _5.6

_Si
L _4.6 = S _i
L _3.6

_Si
L _2.6 = S _i
L _1.6

_Si Yes 7 42.9 0.1 L _6.7 \u003d 320
L _4.6 \u003d 620
L _3.6 = 920
L _2.6 \u003d 1120
L _1.6 \u003d 1420
L _1.7 \u003d 1520 L _6.7

_Si
L _4.6 = S _i
L _3.6

_Si
L _2.6 = S _i
L _1.6

_Si
L _1.7 = S _i Yes eight 48.9 0.2 L _7.8 = 320
L _6.8 = 620
L _5.8 = 920
L _4.8 = 1120
L _3.8 \u003d 1420
L _2.8 \u003d 1520
L _1.9 \u003d 1820 L _7.8

_Si
L _6.8 = S _i
L _5.8

_Si
L _4.8 = S _i
L _3.8

_Si
L _2.8 = S _i
L _1.9

_Si Yes nine 50.9 0.1 L _8.9 \u003d 320
L _7.9 = 620
L _6.9 = 920
L _5.9 \u003d 1120
L _4.9 \u003d 1420
L _3.9 \u003d 1520
L _2.9 \u003d 1820
L _1.9 \u003d 1920 L _8.9

_Si
L _7.9 = S _i
L _6.9

_Si
L _5.9 = S _i
L _4.9

_Si
L _3.9 = S _i
L _2.9

_Si
L _1.9 = S _i Yes ten 56.9 0.2 L _9.10 = 320
L _8.10 = 620
L _7.10 = 920
L _6.10 = 1120
L _5.10 = 1420
L _4.10 = 1520
L _3.10 \u003d 1820
L _2.10 =1920
L _1.10 = 2220 L _9.10

_Si
L _8.10 = S _i
L _7.10

_Si
L _6.10 = S _i
L _5.10

_Si
L _4.10 = S _i
L _3.10

_Si
L _2.10 = S _i
L _1.10

_Si Yes

S_i , то на ключевое устройство 24 из нейронной сети 17 никакого сигнала не поступало, и импульсы скорости с выхода высокоомного усилителя 16, задержанные в линии задержки 25, прошли в формирователь импульса дефекта 19, и на входы счетчиков 20 и 21. Счетчик количества дефектов 21 зарегистрировал второй дефект, а счетчик 20 протяженности дефектов, в котором было подсчитано 40 импульсов скорости, показал что протяженность обнаруженного второго дефекта Since L _k =300 mm was not equal to any of the above options for distances, i.e. L _k

S _i , then no signal was received from the neural network 17 to the key device 24, and the speed pulses from the output of the high-resistance amplifier 16, delayed in the delay line 25, passed to the defect pulse shaper 19, and to the inputs of the counters 20 and 21. Counter of the number of defects 21 registered the second defect, and the counter 20 of the length of defects, in which 40 velocity pulses were counted, showed that the length of the detected second defect

l ₂ \u003d l _e ×n ₂ \u003d 0.25 × 40 \u003d 10 mm.

A similar procedure for analyzing and developing solutions in a neural network for each next short circuit of a wire strand in one of the defects is shown in Table 1.

As a result of the control of the wire with two artificial defects applied in its insulation using the device of Fig.1 implementing the inventive method, 2 defects were registered, one of which had a length of 5 mm, and the other 10 mm.

Thus, compared with the prototype, the claimed device is significantly simplified due to the fact that it allows you to control the actual process of winding the windings of electrical products with galvanic isolation of elements and for its implementation there is no need for special manufacturing of the defect sensor, the need to use high voltage to power the defect sensor is eliminated. , there is no need to clean one of the ends of the wire from insulation and connect it to a common bus.

Sources used:

1. GOST R IEC 60851-5-2008. Winding wires. Test methods. Part 5. Electrical properties.

2. Smirnov G.V. Device for quality control of enamel insulation of winding wires. Zh. Reliability and quality control, 1987, No. 10, p.51.

3. Author's certificate of the USSR No. 364885, class. G01N27/68.

4. RF patent 2737515 ( according to the application 2020107811 dated 21.02.20.) Method for monitoring the defectiveness of the insulation of winding wires / Smirnov G.V. Published 01.12.2020 Bull. No. 34, 15 p . (Prototype)

Claims (1)

A wire insulation defectiveness control device containing a defect sensor, a defect number counter, a defect length counter and a speed sensor connected to the input of the speed pulse shaper, characterized in that winding machine elements are used as a defect sensor, which are fixed on an electrically conductive platform and through an electrically insulating gasket connected by dielectric couplers to the grounded body of the winding machine, while the meter additionally introduced a high-resistance amplifier, a delay line, a computer with a neural network, a router, a pickup generator and an inductor, which is made in the form of a winding wound on a ferrite rod, the said rod with a winding is inserted inside winding wire coils, the outputs of the winding of the inductor are connected to the output of the pickup generator, the input of the pickup generator is connected to the output of the frequency multiplier, and the output of the pickup generator is connected to the input of a computer with a neural network, while the generator output the pickup torus is also connected to the input of the wire length counter, the output of which is connected to the input of the computer with a neural network, while the electrically conductive platform is connected to the input of a high-resistance amplifier, the output of which is connected to a key device through a delay line, and to a computer with a neural network, the output of the key device connected to the input of the counter for the length of defects and to the input of the defect pulse shaper, the output of which is connected to the input of the counter of the number of defects, the output of the computer with the neural network is connected to the input of the router, the output of which is connected to the inhibiting input of the key device.

Images