RU2153660C1 - Method and device for vibration diagnosis of rotary mechanisms - Google Patents

Abstract

FIELD: measurement technology. SUBSTANCE: method consists in bringing several equivalent successive signals of measurements to common conditions followed by judgment on presence of defects by reformed data. To this end, signal proportional to vibrational acceleration is stored and is differentiated into several segments. Pairs of sequences are formed from levels of amplitudes and magnitudes of frequencies which are just spectra whose components are stored separately in respective similar zones and segments. Besides that, sequence of magnitudes of level of rotational speed is measured and stored at moments of differentiating the vibration signal and extremal magnitude is determined. Several cycles of standardizing the components are made; prior to identification, resultant spectrum formed by transposition of its components is shaped. EFFECT: enhanced efficiency and reliability of measurements. 8 cl, 1 dwg

Description

 The invention relates to measuring equipment, in particular to vibration diagnostics.

 The device is known and the method it implements according to the USSR copyright certificate N 1423934, A device for monitoring the condition of rolling friction units.

 There is another device and its corresponding method according to the patent of Russia N 2036455, Sensor diagnosis of rolling friction, which is taken as a prototype.

 Devices based on the conversion of information about the vibrations of a controlled object into amplitude-frequency characteristics with averaging, normalization, and comparison with threshold levels make it possible to distinguish between types and degrees of defects.

 The main disadvantage of the known technical solutions is the need to stabilize the rotation speed of the vibration source, depending on various reasons, including changes in temperature and viscosity of the lubricant or transients of untwisted masses.

 The aim of the invention is to increase reliability while reducing the time of diagnosis.

 The essence of the invention consists in fitting several equivalent short-term, consecutive measurement signals to conditions common to them, after which, according to adequately re-formed data, the presence and degree of defects are judged. For this purpose, an extremum is isolated by hardware, normalized, corrected, transposed and averaged components of the vibration spectrum, in which, thanks to the above measures, external influences on the real state of the diagnosed mechanism are eliminated.

 The method provides for partially sequentially parallel and partially cyclic actions.

 After receiving an electrical signal proportional to vibration acceleration, several segments are isolated from it. Each segment is converted into a pair of sequences, one of which is represented by amplitudes, and the other is represented by the frequencies corresponding to them, characterizing together the spectrum of an individual segment in the selected frequency range for analysis. The amplitude components of the spectra are stored separately from the frequency ones, but at the same time they retain their membership both in pairs of the same name and in zones adjacent to commensurate frequencies.

 Simultaneously with the allocation of segments, the values of the rotation speed of the diagnosed element are measured. Of these, an extreme level is selected that designates one of the limiting modes.

 The method is further presented by several cycles of bringing (fitting) the measurements to uniform conditions, which include: determining the correction for an individual segment, normalizing its frequencies, transposing the amplitudes by zones. The number of cycles is chosen to be less than or equal to the number of segments.

 At the end of the cycles, the amplitude levels in the zones are averaged and then the reduced real spectrum is formed. For the real spectrum, take the average levels of amplitudes and frequency values of the segment corresponding to the extreme measurement mode.

 The real spectrum is analyzed for the presence and types of defects.

 A device for implementing the diagnostic method is presented in the drawing and contains analog, computational and logical blocks and elements.

 The acceleration sensor 1 is connected to the buffer memory 2, which in turn is connected to the Fourier transducer 3. The transducer 3 is included in the drive 4.

 Along with the vibration acceleration sensor, the device is equipped with a tachometer 5. The tachometer is connected to the memory block modes 6, which is part of the recorder extremum modes 7.

 The control inputs of the blocks are connected by a synchronization block 8, which has a Start button 9 and a Stop button 10.

 Drive 4 has two address switches: segments 11 and channels (zones) 12 connected to separate inputs of two multi-channel memory 13 and memory 14. Memory 13 is intended for several sequences of frequency values (CCD), and memory 14 is for several sequences of amplitude levels ( PUA) spectra. The memory 13 and the memory 14 by the information inputs are connected to the corresponding outputs of the converter 3. The number of segment addresses in the memory 13 and the memory 14 is equal to the capacity of the memory 6.

 The extremum recorder of modes 7 contains a sequential chain of the mentioned memory block of modes 6, a logical block for allocating extremum 15 and memory of extremum 16. Block 7 has two outputs: one after memory 6 and the other after memory 16.

 Both outputs are connected to the determinant of amendment 17 of the auto-normalization unit of the segments 18. The output of the determinant is connected to one of the inputs of the corrector 19. The second input of the corrector is connected to the memory output of the CCD 13. The corrector is determined, in particular, by a divider, and the corrector is made by a multiplier.

 The output of the corrector 19 is connected to the inputs of the same line of adders 20, which is part of the frequency zone selector 21. Separate outputs of input unit 22 are connected to the second inputs of the said line. The number of adders in line 20 and memory cells with separate outputs in block 22 is selected with at least the number of zones (channels) in memory 13 and memory 14.

 Each adder of line 20 has an individual output that is connected to a line of threshold elements 23. The output of each threshold element is connected to one of the inputs of the same logical element of the AND-NOT line 24. Separately the outputs of each previous threshold element of the line 23 are connected to the second inputs of the elements AND-NOT The outputs of the line 24 are connected to the control inputs of the same name forwarding keys of the line 25.

 The information inputs of the line 25 are connected to the outputs of the memory channels of the PUA 14, and its outputs are connected to the same channels of the multi-channel adder 26 of the multi-channel averager 27. In parallel, the outputs of the line 25 are connected to the same threshold elements of the line 28, the outputs of which are connected to the inputs of the same meters of the line 29.

 The outputs of the line of counters 29 are connected to the inputs of the divider 30. The outputs of the adder 26 are connected to its other inputs.

 The synchronization block of cycle 8 has one of its outputs “Countdown” connected to the control inputs of memory 2 and memory 6. The output “Record” of block 8 is connected to the control input of switch II. The “Read” output of block 8 is connected to the control inputs: memory 6 of block 7, switches 11 and 12 of block 4, corrector 19 of the auto-normalization block, and logical element I 31. The second input of the element And is connected to the output of the determinant of correction 17.

 The output of the element And is connected to the control input of the channel selector 32. The information input of the switch is connected to the memory output 13 of the drive 4, and the output is connected to the frequency input of the spectrum former 33. The amplitude inputs of the spectrum former 33 are connected to the outputs of the divider 30. The “Diagnosis” output of the synchronizer 8 is connected in parallel with the control inputs of the divider 30 of the averager 27 and the identifier 34. The diagnosis process begins with the transmission of the signal from the vibration sensor 1 to the buffer memory 2.

 Simultaneously with the vibration sensor, a tachometer 5 starts to work, which is connected to the mode 6 memory. The memory is part of the extremum recorder 7. The memory 2 and memory 6 are both filled synchronously by the "Count" commands from the cycle 8 synchronizer. The commands alternate in time according to the rotations of the diagnosed rotor mechanism, after about 1.5-2.0 turns. The “Count” command during the measurement period in the buffer memory divides the vibration signal into several segments, and the rotor speed values are noted in the mode memory. Usually enough 4-6 samples. In the spectrum picture, relative to the spectral lines, half the gaps to the right and left form a zone. Each amplitude of the spectrum is characterized by its frequency. Thus, within the segment, the spectrum contains zones characterized by amplitude levels and frequency values.

 The vibration signal, divided into several segments, from the buffer memory is supplied to the Converter 3 by the command "Read". After the conversion, in particular Fourier, the spectrum of each segment is presented at one output of the converter by a sequence of amplitude levels (ACS), and at the other output, by a sequence of frequency values (CCP). The sequence of the PUA of the first segment goes to the first memory register 14. The sequence of the CCD of the same segment goes to the memory register of the same name 13. The segments are allocated to the addresses of the registers by the switch 11, whose outputs are connected in parallel to the memory 13 and memory 14. Each level of the ROM and the value of the CCD enters the register cells according to the sequence in the spectrum. The distribution of the spectrum parameters to the addresses of the cells simultaneously in the memory 13 and the memory 14 is performed by the switch 12 by the “Record” command. For simplicity, we accept that the same-name register cells of each memory form a corresponding channel. Thus, the switch 12 in the drive 4 PUA and CCD switches the channels, and the switch 11 segments.

 The start of the cycle and the forced stop are performed by the "Start" 9 and "Stop" 10 buttons.

 Since the rotation speed is not stabilized during measurements, memory 6 may contain different speeds for individual spectra. The difference in measurement modes distorts the assessment of defects.

 By the command "Extreme" from memory 6 to the logical block 7, a pair of speed values is alternately displayed. The principle of operation of block 7 consists in comparing all the values stored in memory 6 and isolating from inequalities, in particular, the largest. In this case, subtraction is used. In the first round of calculations in adder 15, the first term of all the differences represents the velocity value of the first segment, and the second terms the velocity values of the remaining segments. Then, in each circle, pairs of differences are compiled in a schematic way according to the given system. For the value of the last segment, the values of the previous segments are paired. The presence of a plus sign for all results in pairs of differences of one circle indicates the extreme, in particular the maximum, value of the first member of these pairs. The extreme value is stored in the memory 16 and is involved when the corrections are found by the auto-normalization unit 18. This completes the preparation of the reformation.

 At the first stage of the spectrum reformation cycle, the determinant of amendment 17, in particular as a divisor, receives the speed value of the first range, and an extreme value acts as a dividend. All results will always be less than one, but one.

 The second stage of the cycle — the normalization of spectra — is as follows. The corrector 19 enters the value of the first frequency of the first spectrum from the first memory channel 13 and corrects it, in particular, it is multiplied by the correction coming from determinant 17. If the speed value is not equal to the extremum, then the correction will change the frequency value. The change may be within the first zone of the spectrum, but often changes significantly.

 The third stage of the cycle - transposition - is realized by shifting the amplitude levels to other channels.

 The value of the normalized frequency is fed to the first inputs of the adders 20 line of the selector 21. The second inputs of the array 20 are supplied with the settings from the input unit 22. The settings determine the values of the width of the zones at the frequencies of the spectrum and are based on experiments. With increasing frequencies, the values of the settings increase (sometimes evenly). Each of the signals received from the autonorming unit 18 on the adders 20 is manifested by a pair of differences having either the same or opposite signs. The presence of opposite signs can only be in a single pair of neighboring adders belonging to a certain zone (channel), which determines the normalized frequency of the spectrum and, therefore, a new channel for transmitting the amplitude level.

 Signs with unit values are highlighted by a line of threshold elements 23. The number of threshold elements is selected not less than the number of zones and channels. At one of the outputs of each pair corresponding to a separate zone, an element of the NOT line of logical elements AND-NOT 24 is included. A single signal, changing its sign, is fed to the first input of element I. The signal comes directly from its threshold element to the second input of each AND element. Logical processing allows you to allocate a single control signal for each normalized frequency.

 The control signal is supplied to a specific key of the forwarding line 25.

 As a result of the address change, the amplitude level from the memory 14 will be directed to another region of the spectrum, for example, the amplitude of the first component will be shifted to the neighboring region or further.

 The fourth stage of the cycle is designed to increase the reliability of the amplitude levels of the spectrum.

 The amplitude level, passing into the other channel of the adders 26 of the multi-channel averager 27, is simultaneously converted by a line of threshold elements 28 connected to the adders 26 into single pulses entering the counter 29. This completes one conversion cycle in one of the channels.

 So, in the process of normalizing the contents of all channels for all segments in the channels of the adder 26, an unequal number of terms will be added up. The number of terms is entered per channel into the counter 29. Now, the amounts accumulated in the channels are divided by the corresponding numbers of terms in the divider 30. The averaging results are sent through 33 and then to identifier 34.

 An amendment equal to unity is monitored by the And element 31 and forms a control signal on the switch 32. Since it is connected in parallel with the outputs of the memory 13, access is made to the frequency values of the corresponding segment at 33.

 Thus, after normalization and reformation, the identifier concentrates the average amplitudes of the spectrum and the values of their frequencies, which together express the completeness of reliable data on the real spectrum.

 The presence of defects, the degree of their development or the fact of the absence of defects is displayed on indicator 35.

 The effectiveness of the method is manifested in the fact that there is no need to take mandatory measures to control or stabilize the speed of rotation during the measurement period. Typically, measurements are taken within 60-90 seconds and the speed, if no special measures are taken, can vary within 15%. Changes in speed distort the vibration signal.

 Signal distortion when converting to a spectrum affects its clarity. There is a blur effect in the form of the appearance of additional erroneous lines on a par with real ones. The distorted spectrum in the analysis leads to an erroneous diagnosis.

 Often, for the clarity of the spectrum, speed stabilizers are used or the measurement process is lengthened, choosing the moments when the rotation speed takes a predetermined value. One method requires expensive specialized stands, and the other is complicated in execution and associated with individual skills.

 The proposed method involves actions on the signal, the requirements for the measurement of which are significantly reduced and stabilization of speed is not required. Moreover, it has acquired sufficient versatility, as it allows measurements with increasing and decreasing speed or with its fluctuations.

 A device for implementing the method is carried out in a portable version, using elements that are usual for the state of the art, and does not require special staff skills or climatic and environmental conditions in use. Vibrodiagnostics can be carried out in any place equipped with a power source.

 The method and device are especially convenient for vibration diagnostics directly on large-sized products, since they do not require preliminary disassembling them into parts that can be mounted on a stand.

Claims (8)

 1. A method for diagnosing rotor mechanisms, which consists in simultaneously measuring vibration and an accompanying factor, which are converted into electrical signals, memorized, averaged, highlighted extremes and form functionally dependent signal components, some of which are normalized and redistributed based on the level, then part jointly identify the presence of signs of defects with subsequent indication of the results, characterized in that they memorize a signal proportional to the intensity of the vibration, for example, vibration acceleration, preliminarily delimiting the signal into several segments, form, from each segment, as the aforementioned constituent pairs of sequences from amplitude levels (AS) and from frequency values (RPS), which are spectra, elements of the AS and RPS are stored separately and in the corresponding zones of the same name in the segments , for a concomitant factor, a sequence of values of the rotation speed level (CCD) is taken, which is measured and stored at the moments of differentiation of the vibration signal, the extreme value is determined and PZUS, wherein the normalization is performed several cycles components from which to form the resultant identification spectrum formed transportation of its components.  2. The method according to claim 1, characterized in that in each cycle the components of the same name are normalized, taking the ratio of the next CCD to the extreme value as the normalization coefficient and keeping the normalization coefficient constant from the first to the last value in the CCD of the corresponding segment, the positions of the pair According to the normalized values of the CCD, after the cycles are completed, the values of the transport PUA of all segments are grouped over the same zones and the amplitudes in each zone averaged over nii are taken for the resulting ASA spectrum, and as a pair of the resulting CCD spectrum, the CCD of the segment corresponding to the extreme value of the CCD is taken.  3. The device for implementing the method according to claim 1, containing a vibration sensor, a defect identifier connected to the display unit, a block of vibration signal parameters, an extremum recorder, an auto-feeding unit, an accompanying signal source, an averaging unit, a control unit, a channel selector, switches, an input unit, characterized in that it is supplemented by a spectrum shaper, a buffer memory and a frequency zone selector, while the frequency zone selector is made with built-in threshold elements, the vibration signal parameter block is in the form e of a drive of sequences of values of amplitudes and frequencies of the constituent spectra (ACA and CCD), an autonorming unit - in the form of an autororming unit for spectral ranges, an accompanying parameter source - in the form of a tachometer, an extremum recorder - in the form of an extremum recorder of measurement modes, an averaging unit - in the form of a multi-channel averager and the control unit is in the form of a cycle synchronizer, the information input of the buffer memory is connected to the output of the vibration sensor, and the output is to the information input of the storage device ПУА and ПЗЧ, information the input of the recorder of the extremum of the measurement modes is connected to the output of the tachometer, and its first output is connected to the first information input of the autororming unit of the spectral ranges, the second information input of which is connected to the first output of the ACS and CCD, and the third information input is connected to the second output of the recorder of the extremum of measurement modes , the first input of the frequency zone selector is connected to the output of the input unit, the second input to the output of the normalized frequency of the auto-normalization unit of the spectral ranges, the third input to W by the next output of the ACS and CCD drive, and the output is with the information input of the multi-channel averager, the first input of the spectrum shaper is connected to the output of the multi-channel averager, the second input is with the output of the channel selector, and the output is with the information input of the defect identifier, the information input of the channel selector is connected to the first the output of the storage device ПУА and ПЗЧ, and the control input through the element And - with a single output of the unit for automatically spectral ranges, the output of the command "count" of the cycle synchronizer is connected in parallel flax with the first control inputs of the buffer memory and the recorder of the extremum of the measurement modes, the output of the “write” command with the first control input of the ACS and PZCh drive, the output of the “read” command with the second control inputs of the buffer memory, the extremum recorder of measurement modes and the memory of the ACP and PZCh as well as the control inputs of the auto-normalization block of the spectral ranges and the And element, the output of the “diagnosis” command is with the control inputs of the multi-channel averager and defect identifier, and the output of the “extremum” command is with tr tim control input registrar extremum measurement modes.  4. The device according to claim 3, characterized in that the storage device for the ACS and CCD components of the spectrum contains a Fourier transducer, two switches and two multichannel dual-address memory blocks PUA and CCD with control inputs for channel selection and segment selection, connected in parallel to the outputs of the respective switches, and information inputs separately - to the pair of outputs of the Fourier transform, while the information input of this drive is the information input of the Fourier transform, its first and second control inputs - volume the remote control inputs of the switches in pairs, and the first and second outputs are the outputs of the CCD memory block and the PUA memory block, respectively.  5. The device according to claim 3, characterized in that the recorder of the extremum of the measurement modes comprises a chain of consecutively connected mode memory block, an extremum logical allocation unit, in particular a maximum, and an extremum memory block, while the information input of this recorder is the information input of the memory block modes, its first and second control inputs are, respectively, the first and second control inputs of the mode memory block, the third control input is the combined third control input of the memory block modes and the control input of the extremum allocation block, and the first and second outputs are the output of the extremum memory block and the output of the mode memory block, respectively.  6. The device according to claim 3, characterized in that the spectral range auto-normalization unit comprises a chain of corrector determiners, in particular a divider, and a corrector, in particular a multiplier, connected in series, with the first and second information inputs of this block being the first input of the corrector and the second information input of the corrector, the third information input is the second input of the corrector, the control input is the control input of the corrector, the output of the normalized frequency is the output corrector, and a single output - the output of the determinant of the amendment.  7. The device according to claim 3, characterized in that the frequency zone selector contains a chain of series-connected adders, a line of threshold elements, a line of logical elements AND - NOT and a line of redirection keys, while the number of elements in the lines and, respectively, mutual connections is chosen not less the number of spectrum components, the first and second inputs of this selector are parallel inputs of the same names of the adders, the third input is the group information input of the forwarding keys, and the output home - group exit last.  8. The device according to claim 3, characterized in that the multi-channel averager contains a series-connected multi-channel adder and a divider, as well as series-connected line of threshold elements and a line of counters, the outputs of which are connected to the second inputs of the divider, while the information input of this averager is the combined inputs multi-channel adder and a line of threshold elements, the control input is the control input of the multi-channel divider, and the output is the output of the latter.