Dry gas seal multi-source data fusion dynamic test and fault simulation experiment table

Abstract

The invention provides a dry gas seal multi-source data fusion dynamic test and fault simulation experiment table, which comprises a power and transmission device, a dry gas seal control system, an electric heat tracing, a gas circuit branch, an experiment tool and a control cabinet, and is characterized in that the power and transmission device comprises a motor, a gear box and a spindle box which are connected with one another, and the control cabinet controls the motor to rotate; the dry gas seal control system, the electric heat tracing, the gas path branch and the experiment tool are connected in sequence, a water path is arranged in the experiment tool, two balance weight discs which are equidistant from the two edges of the experiment tool are arranged in the experiment tool, a loading structure is further arranged in the experiment tool, and the experiment tool is used for fixedly installing a to-be-tested seal and an accompanying seal. The experiment tool, the spindle box, the gear box and the motor are located on the base. According to the technical scheme, the problems that in the prior art, electric, mechanical and dry gas seal composite fault working conditions cannot be simulated, and multi-source data cannot be subjected to fusion dynamic test analysis are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of dry gas seals, and particularly to a dynamic test and fault simulation test bench for multi-source data fusion of dry gas seals. Background Art

[0002] Dry gas seals are the standard seal forms for key refining units. Their structures are precise and complex. Due to factors such as manufacturing and equipment errors, phenomena such as static and dynamic ring friction are likely to occur during their operation. At the same time, phenomena such as temperature rise and scratching also occur. These factors will accumulate continuously. When the friction exceeds a certain number of times or time, it will lead to seal failure, thus causing leakage of key equipment and resulting in major safety accidents. The leakage failure of dry gas seals in refining enterprises has become a "technical bottleneck" for the long-term, safe, stable, and efficient operation of refining enterprise devices. A single fault shutdown often causes economic losses in the tens of millions and huge safety and environmental protection risks.

[0003] At present, the relevant research on the state monitoring of dry gas seals at home and abroad is in its infancy, and the typical fault mechanisms, fault modes, and failure laws of dry gas seals are still not clear. The dry gas seal test bench is the basis for deeply exploring and studying the above mechanisms and laws.

[0004] The existing dry gas seal test platforms mainly focus on studying the sealing performance of dry gas seals, pay more attention to process parameters such as leakage rate, and are basically difficult to effectively simulate typical dry gas seal faults such as dry gas with liquid and static and dynamic rubbing, and lack the synchronous acquisition and fusion analysis of multi-parameters such as vibration, acoustic emission, temperature, pressure, flow rate, and film thickness.

[0005] Therefore, there is a need for a test bench that can simulate electrical, mechanical, and dry gas seal composite fault conditions and perform dynamic test analysis of multi-source data fusion. Summary of the Invention

[0006] The main object of the present invention is to provide a dynamic test and fault simulation test bench for multi-source data fusion of dry gas seals, so as to solve the problems in the prior art that electrical, mechanical, and dry gas seal composite fault conditions cannot be simulated and multi-source data cannot be fused for dynamic test analysis.

[0007] To achieve the above object, the present invention provides a dynamic test and fault simulation experimental bench for dry gas seals, including: a power and transmission device, a dry gas seal control system, electric tracing heat, a gas circuit branch, an experimental tooling, and a control cabinet. The power and transmission device includes: a motor, a gearbox, and a main shaft box that are connected to each other. The control cabinet controls the rotation of the motor. The dry gas seal control system, electric tracing heat, gas circuit branch, and experimental tooling are connected in sequence. There is a water circuit inside the experimental tooling. There are two counterweight disks inside the experimental tooling that are equidistant from the two edges of the experimental tooling. There is also a loading structure inside the experimental tooling. The experimental tooling is used to fixedly install the seal under test and the backing seal. The experimental tooling, main shaft box, gearbox, and motor are located on the base.

[0008] Further, the loading structure includes: a moving ring, a stationary ring, a stationary ring seat, a piezoelectric crystal, and a spring. There is a support inside the experimental tooling. The two counterweight disks are respectively located at the left and right ends of the support and are equidistant from the left and right ends of the support.

[0009] The moving ring and the stationary ring are arranged opposite to each other and have a certain gap. There are three piezoelectric crystals arranged between the stationary ring and the stationary ring seat. One end of each spring is connected to the stationary ring, and the other end is connected to the piezoelectric crystal.

[0010] When the three groups of piezoelectric crystals compress the spring under the action of the piezoelectric effect, the stationary ring is driven to move towards the moving ring. Conversely, when the three groups of piezoelectric crystals stretch the spring, the stationary ring is driven to move away from the moving ring.

[0011] Further, there are 12 bolt holes arranged at intervals of 30 degrees on the counterweight disk. When the counterweights are installed at the same angle on the two counterweight disks, force imbalance can be simulated. When the counterweights are installed at opposite angles on the two counterweight disks, moment imbalance can be simulated.

[0012] Further, the three piezoelectric crystals are arranged on the stationary ring seat at intervals of 120 degrees.

[0013] Further, the piezoelectric crystal is a high-voltage-resistant YCOB piezoelectric crystal.

[0014] Further, the dry gas seal control system includes an air compressor, a pressure stabilizing tank, and a liquid separation tank that are connected in sequence. The air compressor compresses the air and transports it to the pressure stabilizing tank, and then the liquid separation tank injects it into the seal cavity of the experimental tooling through electric tracing heat and the gas circuit branch.

[0015] Further, the water circuit includes: a circulating water pump, a circulating water tank, and a refrigerating machine. The water circuit is used to take away the heat inside the experimental tooling.

[0016] Further, the main shaft box is connected to a lubricating oil pump.

[0017] Further, it also includes: a test signal channel and a control signal channel. The signals collected by the test signal channel include: motor current, voltage, spindle speed, torque, spindle box temperature, temperature, pressure, and flow rate of the gas path branch, the thickness signal of the to-be-tested sealing gas film, the acoustic emission signal of the to-be-tested sealing stationary ring, the acoustic emission signal and vibration signal of the experimental tooling end cover.

[0018] Further, the vibration signals include: the radial signal of the spindle box, the radial signals of the gear box and the non-driving end of the experimental tooling, the axial signal of the test bench spindle, and the radial signal of the driving end of the tooling.

[0019] Further, the acoustic emission signals are collected using the high-frequency channels of the data acquisition card, and the remaining signals are collected using the low-frequency channels of the data acquisition card. The high-frequency channels and the low-frequency channels share a set of hardware clock signals.

[0020] Further, it also includes an electric control system, which is used to control the motor speed, torque, the pressure and flow rate of the gas path branch, and the current direction in the piezoelectric crystal.

[0021] Further, the acoustic emission signals include: the inner-side acoustic emission signal and the outer-side acoustic emission signal, which are processed using the empirical wavelet transform method to segment the frequency domain of the acoustic emission signals. The segmentation boundary is shown in formula (1):

[0022] Λ n =[ω n-1 ,ω n ,n=1,2,…,N(1); where ω n represents the boundary of each part of the frequency band;

[0023] Based mainly on the frequency domain distribution of the inner-side acoustic emission signal, the signal is decomposed, and the decomposed signal is shown in formula (2):

[0024]

[0025] and are the approximation coefficient and the detail coefficient respectively, is the empirical wavelet function;;F -1 represents the inverse transform from the frequency domain to the time domain.

[0026] Based on entropy and kurtosis, the sensitive frequency band of the dry gas seal operating state is selected for band-pass filtering, and the sub-signals f 内 (t) and f 外 (t) are extracted from the inner-side acoustic emission signal and the outer-side acoustic emission signal respectively.

[0027] f 外 (t)=Af 外 (t - 1)+w p(w)~N(0,Q) (3);

[0028] f 内 f(t) = Hf 外 (t) + vp(v) ~ N(0, R) (4);

[0029] Wherein, A represents the state transition matrix, which is used to describe the transition relationship between the dry gas seal state signals at the previous moment and the current moment; H represents the observation matrix, which is used to describe the mapping relationship between the true state value of the dry gas seal and the externally measured signal; p(w) ~ N(0, Q) represents that p(w) conforms to the Gaussian distribution with a mean of 0 and a variance of Q; p(v) ~ N(0, R) represents that p(v) conforms to the Gaussian distribution with a mean of 0 and a variance of R.

[0030] The present invention has the following beneficial effects:

[0031] 1. The present invention can realize the comprehensive simulation and analysis of motor faults (stator winding faults, rotor bar breaking faults, motor bearing faults), gearbox faults (tooth surface pitting, tooth surface wear), shafting faults (misalignment, imbalance), and dry gas seal faults (dry gas with liquid, impure dry gas, static-dynamic rubbing).

[0032] 2. The present invention adopts a different sampling frequency signal synchronization and registration strategy combining hard synchronization and soft synchronization, and uses an integrated data acquisition card for high-frequency and low-frequency channels to realize the synchronous acquisition, transmission, and storage of all signals. All high-frequency and low-frequency acquisition channels share a set of hardware clock signals, achieving clock synchronization between multiple acquisition channels.

[0033] 3. The present invention constructs a dry gas seal gas flow field change model, a high-pressure-resistant YCOB piezoelectric crystal driver, and an electro-coupling model under transient and steady-state working conditions, and dynamically calculates and corrects the key undetermined parameters of the high-pressure-resistant YCOB piezoelectric crystal driver and the electro-coupling model such as the hysteresis factor by introducing a neural network in a data-driven manner, realizing the online precise regulation of the real-time operating posture of the dry gas seal.

[0034] In summary, the test bench provided by the present invention can realize the simulation and analysis of multi-system compound faults of electricity-mechanics-dry gas seal, which is more comprehensive in fault simulation than the prior art. The test bench provided by the present invention can realize the synchronous registration of high-frequency and low-frequency signals combining hard synchronization and soft synchronization, and the analysis signal results are more accurate. The test bench provided by the present invention dynamically calculates and corrects the key undetermined parameters of the high-pressure-resistant YCOB piezoelectric crystal driver-electro-coupling model such as the hysteresis factor by introducing a neural network in a data-driven manner, and the regulation of the real-time operating posture of the dry gas seal is more precise. Brief Description of the Drawings

[0035] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. In the drawings:

[0036] Figure 1 Shows a schematic diagram of the overall structure of a dynamic test and fault simulation experimental platform for multi-source data fusion of dry gas seals according to the present invention.

[0037] Figure 2 Shows Figure 1 a schematic diagram of the internal structure of the experimental tooling of the experimental platform.

[0038] Figure 3 Shows Figure 1 a schematic diagram of the structure of the counterweight disk of the experimental platform.

[0039] Figure 4 Shows Figure 1 a schematic diagram of the structure of the stationary ring and the stationary ring seat of the experimental platform.

[0040] Figure 5 Shows Figure 1 a flowchart of high-frequency and low-frequency signal synchronous acquisition of the experimental platform.

[0041] Figure 6 Shows Figure 1 a schematic diagram of real-time operation monitoring of the dry gas seal of the experimental platform.

[0042] Among them, the reference numerals in the above drawings are:

[0043] 1. Dry gas seal control system; 11. Air compressor, 12. Pressure stabilizing tank; 13. Liquid separation tank.

[0044] 2. Electric tracing

[0045] 3. Gas path branch

[0046] 4. Experimental tooling; 40. Support body; 41. Circulating water pump; 42. Circulating water tank; 43. Refrigerator; 44. Rotating ring; 45. Stationary ring; 46. Stationary ring seat; 47. Piezoelectric crystal; 48 Spring; 49. Counterweight disk; 491. Bolt hole.

[0047] 5. Control cabinet

[0048] 6. Motor

[0049] 7. Gearbox

[0050] 8. Headstock

[0051] 9. Lubricating oil pump. Specific embodiments

[0052] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative work shall fall within the protection scope of the present invention.

[0053] Embodiment 1

[0054] As Figure 1 shown, a dynamic test and fault simulation experimental bench for dry gas seal multi-source data fusion includes: a power and transmission device, a dry gas seal control system 1, an electric tracing 2, a gas circuit branch 3, an experimental tooling 4, and a control cabinet 5. The power and transmission device includes: a motor 6, a gearbox 7, and a main spindle box 8 that are connected to each other. The control cabinet 5 controls the rotation of the motor 6. The dry gas seal control system 1, the electric tracing 2, the gas circuit branch 3, and the experimental tooling 4 are connected in sequence. There is a water circuit inside the experimental tooling 4. There are two counterweight disks 49 inside the experimental tooling 4 that are equidistant from the two edges of the experimental tooling 4. There is also a loading structure inside the experimental tooling. The experimental tooling 4 is used to fixedly install the seal to be tested and the backing seal. The experimental tooling 4, the main spindle box 8, the gearbox 7, and the motor 6 are located on the base.

[0055] The motor 6 can simulate stator winding faults, rotor bar breaking faults, and motor 6 bearing faults. The gearbox 7 can realize faults such as tooth surface pitting and tooth surface wear, and realize the simulation and analysis of electrical, mechanical, and dry gas seal compound faults.

[0056] Specifically, as Figure 2 shown, the loading structure includes: a moving ring 44, a stationary ring 45, a stationary ring seat 46, a piezoelectric crystal 47, and a spring 48. There is a support 40 inside the experimental tooling 4. The two counterweight disks 49 are respectively located at the left and right ends of the support 40 and are equidistant from the left and right ends of the support 40.

[0057] The moving ring 44 and the stationary ring 45 are arranged opposite to each other and have a certain gap. There are three piezoelectric crystals 47 arranged between the stationary ring 45 and the stationary ring seat 46. One end of each spring 48 is connected to the stationary ring 45, and the other end is connected to the piezoelectric crystal 47.

[0058] When the three groups of piezoelectric crystals 47 compress the spring 48 under the piezoelectric effect, the static ring 45 is driven to move towards the dynamic ring 44. Conversely, when the three groups of piezoelectric crystals 47 stretch the spring 48, the static ring 45 is driven to move away from the dynamic ring 44. A high-pressure-resistant YCOB piezoelectric crystal 47 is installed at the spring 48 between the static ring seat 46 to be measured for sealing and the static ring 45. When the dry gas seal operates, a gap of several micrometers will be generated between the dynamic ring 44 and the static ring 45 due to the hydrodynamic pressure effect of the gas. At this time, using the inverse piezoelectric effect of the piezoelectric crystal 47, a positive current is passed through the high-pressure-resistant YCOB piezoelectric crystal 47, and the piezoelectric crystal 47 will compress the spring 48, thereby driving the static ring 45 of the dry gas seal to move towards the dynamic ring 44 of the dry gas seal. Conversely, the static ring 45 will move in the opposite direction under the action of the spring 48, so as to realize the dynamic adjustment of the gap between the dynamic ring 44 and the static ring 45 of the dry gas seal. The three groups of high-pressure-resistant YCOB piezoelectric crystals 47 are arranged at intervals of 120 degrees in the radial plane, and together they form a dynamic micro-motion loading structure for the axial force of the dry gas seal. When the three groups of piezoelectric crystals 47 elongate the same length, the static ring 45 undergoes a pure axial displacement at this time. When the elongation lengths of the three groups of piezoelectric crystals 47 are inconsistent, the static ring 45 will undergo axial and angular displacements. By controlling the reduction of the gap between the static ring 45 and the dynamic ring 44, the simulation of the rubbing fault between the static and dynamic rings 45 of the dry gas seal can be realized. By controlling the increase of the gap between the static ring 45 and the dynamic ring 44, the simulation of faults such as excessive leakage of the dry gas seal can be realized.

[0059] Specifically, as Figure 3 shown, 12 bolt holes 491 are arranged on the counterweight disk 49 at intervals of 30 degrees. When the counterweights are installed at the same angle on the two counterweight disks 49, the force imbalance can be simulated. When the counterweights are installed at opposite angles on the two counterweight disks 49, the couple imbalance can be simulated. Different weights of counterweight bolts can be installed on the counterweight disk 49 to realize the simulation of the imbalance fault of the dry gas seal.

[0060] Specifically, as Figure 4 shown, the three piezoelectric crystals 47 are arranged on the static ring seat 46 at intervals of 120 degrees. The three loading structures use the high-pressure-resistant YCOB piezoelectric crystal 47 as the main power source. By controlling the voltage of the high-pressure-resistant YCOB piezoelectric crystals 47 at different positions, the movement of the top of the piezoelectric crystal 47 is realized, and the axial movement and angular swing of the static ring 45 of the dry gas seal are realized, so as to realize the dynamic controllable loading under the conditions of pressure and speed of the dry gas seal.

[0061] Specifically, the piezoelectric crystal 47 is a high-pressure-resistant YCOB piezoelectric crystal 47.

[0062] Specifically, the dry gas seal control system includes an air compressor 11, a pressure stabilizing tank 12, and a liquid separation tank 13 that are connected in sequence. The air compressor 11 compresses air and transports it to the pressure stabilizing tank 12, and then the liquid separation tank 13 injects it into the sealed cavity of the experimental tooling 4 through the electric tracing heating 2 and the gas path branch 3. Liquid can be injected after the liquid separation tank 13 to simulate the dry gas with liquid failure, and minute solid impurities can be injected to simulate the dry gas impurity failure. By individually adjusting Figure 1 the pressure and flow rate of each gas path branch 3 shown in the figure to simulate typical failures such as unstable process parameters, blocked leakage gas pipelines, and reduced intake of sealing gas.

[0063] Specifically, the water circuit includes a circulating water pump 41, a circulating water tank 42, and a refrigerator 43. The water circuit is used to remove the heat inside the experimental tooling 4. The experimental tooling 4 is mainly used to install and fix the seal under test and the accompanying seal, and is equipped with a cooling water circulation winding to remove the heat generated during the high-speed rotation of the dry gas seal through the refrigerator 43.

[0064] Specifically, the main spindle box 8 is connected to the lubricating oil pump 9.

[0065] Specifically, it also includes a test signal channel and a control signal channel. The signals collected by the test signal channel include: the current and voltage of the motor 6, the spindle speed and torque, the temperature of the main spindle box 8, the temperature, pressure, and flow rate of the gas path branch 3, the dry gas film thickness signal of the seal under test, the acoustic emission signal of the stationary ring 45 of the seal under test, the acoustic emission signal and vibration signal of the end cover of the experimental tooling 4.

[0066] There are four paths for the acoustic emission signals. One path of the stationary ring 45 of the dry gas seal under test realizes the internal measurement of the acoustic emission signal, and two radial paths (arranged perpendicular to each other at 90 degrees) and one axial path of the end cover of the experimental tooling 4 corresponding to the dry gas seal realize the external measurement of the acoustic emission signal.

[0067] Specifically, the vibration signals include: the radial signal of the main spindle box 8, the radial signals of the gear box 7 and the non-driving end of the experimental tooling 4, the axial signal of the main spindle of the test bench, and the radial signal of the driving end of the tooling.

[0068] Specifically, acoustic emission signals are collected using the high-frequency channel of the data acquisition card, and the remaining signals are collected using the low-frequency channel of the data acquisition card. The high-frequency channel and the low-frequency channel share a set of hardware clock signals. The data acquisition card is an integrated high-frequency and low-frequency data acquisition card. A synchronous registration strategy for different sampling frequency signals combining hard synchronization and soft synchronization is adopted. By using the integrated high-frequency and low-frequency channel data acquisition card, synchronous acquisition, transmission, and storage of all signals are realized. Acoustic emission signals are collected using the high-frequency channel of the data acquisition card, and the remaining signals are collected using the low-frequency channel of the data acquisition card. All high-frequency and low-frequency acquisition channels share a set of hardware clock signals, achieving clock synchronization between multiple acquisition channels. For the collected signals, least-squares soft multiplication registration is performed based on the high-frequency channel, that is, registration is performed using the least-squares method.

[0069] Specifically, the multi-source data in the multi-source data fusion dynamic test includes the current of motor 6, the voltage of motor 6, the spindle speed, the spindle torque, the vibration of motor 6, the vibration of gearbox 7, the vibration of headstock 8, the temperature of headstock 8, the pressure of each gas path, the temperature of each gas path, the flow rate of each gas path, the thickness of the dry gas seal gas film to be measured (the gap between the rotating ring 44 and the stationary ring 45), the acoustic emission of the stationary ring 45 of the dry gas seal to be measured (inner measurement), the acoustic emission of the tooling end cover (outer measurement), and the tooling vibration signal. Among them, there are four acoustic emission signals. One path of the stationary ring 45 of the dry gas seal to be measured realizes the inner measurement of the acoustic emission signal, and two radial paths (arranged perpendicular to each other at 90 degrees) at the corresponding position of the tooling end cover for the dry gas seal and one axial path of the tooling end cover realize the outer measurement of the acoustic emission signal. There are nine vibration signals, two radial paths (arranged perpendicular to each other at 90 degrees) of the headstock 8, two of the gearbox 7, two radial paths (arranged perpendicular to each other at 90 degrees) at the non-driving end of the tooling, one axial path, and two radial paths (arranged perpendicular to each other at 90 degrees) at the driving end of the tooling, realizing the vibration measurement of the entire test bench.

[0070] As Figure 5 shown, a synchronous registration strategy for different sampling frequency signals combining hard synchronization and soft synchronization is adopted. By using the integrated high-frequency and low-frequency channel data acquisition card, synchronous acquisition, transmission, and storage of all signals are realized. Acoustic emission signals are collected using the high-frequency channel of the data acquisition card, and the remaining signals are collected using the low-frequency channel of the data acquisition card. All high-frequency and low-frequency acquisition channels share a set of hardware clock signals, achieving clock synchronization between multiple acquisition channels. For the collected signals, least-squares soft registration is performed based on the high-frequency channel.

[0071] Specifically, it further includes an electric control system, which is used to control the speed and torque of motor 6, the pressure and flow rate of the gas path branch 3, and the current direction in the piezoelectric crystal 47. According to signals such as the gas film thickness, speed signal, main seal gas pressure, temperature, and flow rate, the dynamic calculation and correction of dynamic model parameters are realized using a neural network, achieving precise control of the real-time operating posture of the dry gas seal.

[0072] Specifically, the neural network inputs are signals such as air film thickness, rotational speed of the main shaft of the test bench, pressure, temperature, and flow rate of each gas path, and the neural network output is the piezoelectric crystal control current signal. Specifically, the acoustic emission signals include: inner acoustic emission signals and outer acoustic emission signals, which are processed using the empirical wavelet transform method to segment the frequency domain of the acoustic emission signals, and the segmentation boundary is shown in formula (1):

[0073] Λ n =[ω n-1 ,ω n ,n=1,2,…,N(1); where ω n represents the boundary of each part of the frequency band;

[0074] Based mainly on the frequency domain distribution of the inner acoustic emission signals, the signals are decomposed, and the decomposed signals are shown in formula (2):

[0075]

[0076] and are the approximation coefficient and the detail coefficient respectively, is the empirical wavelet function;;F -1 represents the inverse transform from the frequency domain to the time domain.

[0077] Taking entropy and kurtosis as the basis, the sensitive frequency band of the dry gas seal operating state is selected for band-pass filtering, and the sub-signals f 内 (t) and f 外 (t) are extracted from the inner acoustic emission signals and the outer acoustic emission signals respectively;

[0078] f 外 (t)=Af 外 (t - 1)+w p(w)~N(0,Q) (3);

[0079] f 内 (t)=Hf 外 (t)+v p(v)~N(0,R) (4);

[0080] Among them, A represents the state transition matrix, which is used to describe the transition relationship between the dry gas seal state signals at the previous moment and the current moment, H represents the observation matrix, which is used to describe the mapping relationship between the true state value of the dry gas seal and the outer signal, p(w)~N(0,Q) represents that p(w) conforms to the Gaussian distribution with a mean of 0 and a variance of Q, and p(v)~N(0,R) represents that p(v) conforms to the Gaussian distribution with a mean of 0 and a variance of R.

[0081] Such as Figure 5As shown, the Kalman filter algorithm is introduced to further analyze the externally measured acoustic emission signal and the internally measured acoustic emission signal, and extract the signal features that can reflect the opening and closing state between the dynamic ring 44 and the static ring 45 of the dry gas seal.

[0082] The electric control system can realize the stepless adjustment of the motor 6 speed and torque, and the separate stepless adjustment of the pressure and flow rate of each gas path. By constructing a dry gas seal gas flow field change model, a high-pressure-resistant YCOB piezoelectric crystal 47 drive machine and an electro-coupling model under transient and steady-state working conditions, and introducing a neural network to dynamically calculate and correct the key undetermined parameters of the high-pressure-resistant YCOB piezoelectric crystal 47 drive machine in a data-driven manner, the online precise control of the real-time operating posture of the dry gas seal is realized. In view of the problems of insufficient utilization of multi-source data fusion, difficulty in realizing online dynamic loading, and insufficient simulation of typical faults in the existing dry gas seal test bench, the present invention proposes a dry gas seal multi-source data fusion dynamic test and fault simulation test bench, which can realize the simulation of electrical-mechanical-dry gas seal composite fault conditions and the dynamic test analysis of multi-source data fusion, and provide guidance for the dry gas seal condition monitoring and early warning technology.

[0083] The present invention adopts the method of multi-source data fusion analysis and intelligent control, which can be used to analyze the basic scientific problems such as the gas film disturbance characteristics, the mechanism of dry gas film condensation failure and the end face wear mechanism, the conditions for the occurrence of typical faults of dry gas seals and the corresponding fault characteristics, the process parameters of dry gas seals, the variation law of acoustic emission signals with the gas film state and the key influencing factors, and the influence law of liquid-carrying dry gas on the stability of the gas film between end faces, etc., provide support for related technologies such as dry gas seal condition monitoring and early warning, and have a relatively broad application prospect.

[0084] Embodiment 2

[0085] As Figure 1 shown, a dry gas seal multi-source data fusion dynamic test and fault simulation test bench includes: a power and transmission device, a dry gas seal control system 1, an electric tracing 2, a gas path branch 3, an experimental tooling 4 and a control cabinet 5. The power and transmission device includes: a motor 6, a gearbox 7 and a main shaft box 8 that are connected to each other. The control cabinet 5 controls the rotation of the motor 6. The dry gas seal control system 1, the electric tracing 2, the gas path branch 3 and the experimental tooling 4 are connected in sequence. The experimental tooling 4 has a water path inside. There are two counterweight disks 49 inside the experimental tooling 4 that are equidistant from the two edges of the experimental tooling 4. There is also a loading structure inside the experimental tooling. The experimental tooling 4 is used to fixedly install the seal to be tested and the lining seal. The experimental tooling 4, the main shaft box 8, the gearbox 7 and the motor 6 are located on the base.

[0086] The motor 6 can simulate stator winding faults, rotor bar breaking faults, and motor 6 bearing faults. The gearbox 7 can realize faults such as tooth surface pitting and tooth surface wear, and realize the simulation and analysis of electrical, mechanical, and dry gas seal composite faults.

[0087] Specifically, as Figure 2 shown, the loading structure includes a moving ring 44, a stationary ring 45, a stationary ring seat 46, a piezoelectric crystal 47, and a spring 48. There is a support 40 inside the experimental tooling 4. Two counterweight disks 49 are respectively located at the left and right ends of the support 40 and are equidistant from the left and right ends of the support 40.

[0088] The moving ring 44 and the stationary ring 45 are arranged oppositely with a certain gap. Three piezoelectric crystals 47 are arranged between the stationary ring 45 and the stationary ring seat 46. One end of each spring 48 is connected to the stationary ring 45, and the other end is connected to the piezoelectric crystal 47.

[0089] When the three groups of piezoelectric crystals 47 compress the spring 48 under the action of the piezoelectric effect, the stationary ring 45 is driven to move towards the moving ring 44. Conversely, when the three groups of piezoelectric crystals 47 stretch the spring 48, the stationary ring 45 is driven to move away from the moving ring 44. A high-pressure-resistant YCOB piezoelectric crystal 47 is installed at the spring 48 between the stationary ring seat 46 of the seal under test and the stationary ring 45. When the dry gas seal is operating, a gap of several micrometers will be generated between the moving ring 44 and the stationary ring 45 due to the hydrodynamic pressure effect of the gas. At this time, using the inverse piezoelectric effect of the piezoelectric crystal 47, a positive current is passed through the high-pressure-resistant YCOB piezoelectric crystal 47, and the piezoelectric crystal 47 will compress the spring 48, thereby driving the stationary ring 45 of the dry gas seal to move towards the moving ring 44 of the dry gas seal. Conversely, the stationary ring 45 will move in the opposite direction under the action of the spring 48, so as to realize the dynamic adjustment of the gap between the moving ring 44 and the stationary ring 45 of the dry gas seal. The three groups of high-pressure-resistant YCOB piezoelectric crystals 47 are arranged at intervals of 120 degrees in the radial plane and jointly form a dynamic micro-motion loading structure for the axial force of the dry gas seal. When the three groups of piezoelectric crystals 47 elongate by the same length, the stationary ring 45 undergoes a pure axial displacement at this time. When the elongation lengths of the three groups of piezoelectric crystals 47 are inconsistent, the stationary ring 45 will undergo axial and angular displacements. By controlling the reduction of the gap between the stationary ring 45 and the moving ring 44, the simulation of the rubbing fault between the stationary and moving rings 45 of the dry gas seal can be realized. By controlling the increase of the gap between the stationary ring 45 and the moving ring 44, the simulation of faults such as excessive leakage of the dry gas seal can be realized.

[0090] Specifically, the acoustic emission signals are collected using the high-frequency channels of the data acquisition card, and the remaining signals are collected using the low-frequency channels of the data acquisition card. The high-frequency channels and the low-frequency channels share a set of hardware clock signals. The data acquisition card is an integrated high-frequency and low-frequency data acquisition card. A synchronous registration strategy for different sampling frequency signals that combines hard synchronization and soft synchronization is adopted. Using the integrated high-frequency and low-frequency channel data acquisition card, synchronous acquisition, transmission, and storage of all signals are realized. The acoustic emission signals are collected using the high-frequency channels of the data acquisition card, and the remaining signals are collected using the low-frequency channels of the data acquisition card. All high-frequency and low-frequency acquisition channels share a set of hardware clock signals, achieving clock synchronization between multiple acquisition channels. For the collected signals, least-squares soft registration is performed with the high-frequency channel as the reference, that is, registration is performed using the least-squares method.

[0091] Specifically, the multi-source data in the multi-source data fusion dynamic test includes the current of motor 6, the voltage of motor 6, the spindle speed, the spindle torque, the vibration of motor 6, the vibration of gearbox 7, the vibration of headstock 8, the temperature of headstock 8, the pressure of each gas path, the temperature of each gas path, the flow rate of each gas path, the thickness of the gas film of the dry gas seal to be measured (the gap between the rotating ring 44 and the stationary ring 45), the acoustic emission of the stationary ring 45 of the dry gas seal to be measured (inner measurement), the acoustic emission of the tooling end cover (outer measurement), and the vibration signal of the tooling. Among them, there are four channels of acoustic emission signals. One channel on the stationary ring 45 of the dry gas seal to be measured realizes the inner measurement of the acoustic emission signal. There are two radial channels (arranged perpendicular to each other at 90 degrees) at the corresponding position of the tooling end cover for the dry gas seal, and one axial channel on the tooling end cover realizes the outer measurement of the acoustic emission signal. There are nine channels of vibration signals, including two radial channels (arranged perpendicular to each other at 90 degrees) on the headstock 8, two channels on the gearbox 7, two radial channels (arranged perpendicular to each other at 90 degrees) at the non-driving end of the tooling, one axial channel, and two radial channels (arranged perpendicular to each other at 90 degrees) at the driving end of the tooling, realizing the vibration measurement of the entire test bench.

[0092] As Figure 5 shown, a synchronous registration strategy for different sampling frequency signals that combines hard synchronization and soft synchronization is adopted. Using the integrated high-frequency and low-frequency channel data acquisition card, synchronous acquisition, transmission, and storage of all signals are realized. The acoustic emission signals are collected using the high-frequency channels of the data acquisition card, and the remaining signals are collected using the low-frequency channels of the data acquisition card. All high-frequency and low-frequency acquisition channels share a set of hardware clock signals, achieving clock synchronization between multiple acquisition channels. For the collected signals, least-squares soft registration is performed with the high-frequency channel as the reference.

[0093] Specifically, it also includes an electric control system, which is used to control the speed and torque of motor 6, the pressure and flow rate of the gas path branch 3, and the current direction in the piezoelectric crystal 47. According to signals such as the gas film thickness, speed signal, main seal gas pressure, temperature, and flow rate, the dynamic calculation and correction of the dynamic model parameters are realized using a neural network, achieving precise control of the real-time operating attitude of the dry gas seal.

[0094] Specifically, the neural network inputs are signals such as air film thickness, spindle speed of the test bench, pressure, temperature, and flow rate of each air circuit, and the neural network output is the control current signal of the piezoelectric crystal. Specifically, the acoustic emission signals include: inner acoustic emission signals and outer acoustic emission signals, which are processed using the empirical wavelet transform method to segment the frequency domain of the acoustic emission signals, and the segmentation boundary is shown in formula (1):

[0095] Λ n =[ω n-1 ,ω n ,n=1,2,…,N(1); where ω n represents the boundary of each part of the frequency band;

[0096] Based mainly on the frequency domain distribution of the inner acoustic emission signals, the signals are decomposed, and the decomposed signals are shown in formula (2):

[0097]

[0098] and are the approximation coefficient and the detail coefficient respectively, is the empirical wavelet function;; F -1 represents the inverse transform from the frequency domain to the time domain.

[0099] Based on entropy and kurtosis, the sensitive frequency band of the dry gas seal operating state is selected for band-pass filtering, and the sub-signals f 内 (t) and f 外 (t) are extracted from the inner acoustic emission signals and the outer acoustic emission signals respectively;.

[0100] f 外 (t)=Af 外 (t - 1)+w p(w)~N(0,Q) (3);

[0101] f 内 (t)=Hf 外 (t)+v p(v)~N(0,R) (4);

[0102] Among them, A represents the state transition matrix, which is used to describe the transition relationship between the dry gas seal state signals at the previous moment and the current moment, H represents the observation matrix, which is used to describe the mapping relationship between the true state value of the dry gas seal and the outer signal, p(w)~N(0,Q) represents that p(w) conforms to the Gaussian distribution with a mean of 0 and a variance of Q, and p(v)~N(0,R) represents that p(v) conforms to the Gaussian distribution with a mean of 0 and a variance of R.

[0103] Such as Figure 5As shown, the Kalman filtering algorithm is introduced to further analyze the externally measured acoustic emission signal and the internally measured acoustic emission signal, and signal features that can reflect the opening and closing state between the dynamic ring 44 and the static ring 45 of the dry gas seal are extracted.

[0104] Embodiment III

[0105] As Figure 1 shown, a dry gas seal multi-source data fusion dynamic test and fault simulation test bench includes: a power and transmission device, a dry gas seal control system 1, an electric tracing 2, a gas circuit branch 3, an experimental tooling 4, and a control cabinet 5. The power and transmission device includes: a motor 6, a gearbox 7, and a main spindle box 8 that are connected to each other. The control cabinet 5 controls the rotation of the motor 6. The dry gas seal control system 1, the electric tracing 2, the gas circuit branch 3, and the experimental tooling 4 are connected in sequence. There is a water circuit inside the experimental tooling 4. There are two counterweight disks 49 inside the experimental tooling 4 that are equidistant from the two edges of the experimental tooling 4. There is also a loading structure inside the experimental tooling. The experimental tooling 4 is used to fixedly install the seal to be tested and the accompanying seal. The experimental tooling 4, the main spindle box 8, the gearbox 7, and the motor 6 are located on the base.

[0106] The motor 6 can simulate stator winding faults, rotor bar breaking faults, and motor 6 bearing faults. The gearbox 7 can realize faults such as tooth surface pitting and tooth surface wear, and realize the simulation and analysis of electrical, mechanical, and dry gas seal composite faults.

[0107] Specifically, as Figure 3 shown, there are 12 bolt holes 491 on the counterweight disk 49 at intervals of 30 degrees. When the counterweights are installed at the same angle on the two counterweight disks 49, force imbalance can be simulated. When the counterweights are installed at opposite angles on the two counterweight disks 49, couple imbalance can be simulated. Different weights of counterweight bolts can be installed on the counterweight disk 49 to simulate the imbalance fault of the dry gas seal.

[0108] Specifically, as Figure 4 shown, three piezoelectric crystals 47 are arranged on the static ring seat 46 at intervals of 120 degrees. The three loading structures use the high-voltage-resistant YCOB piezoelectric crystal 47 as the main power source, and the movement of the top of the piezoelectric crystal 47 is realized by controlling the voltage of the high-voltage-resistant YCOB piezoelectric crystal 47 at different positions, so as to realize the axial movement and angular swing of the static ring 45 of the dry gas seal, and thus realize the dynamic controllable loading under the conditions of pressure and speed of the dry gas seal.

[0109] Specifically, the piezoelectric crystal 47 is a high-voltage-resistant YCOB piezoelectric crystal 47.

[0110] Specifically, the dry gas seal control system includes an air compressor 11, a pressure stabilizing tank 12, and a liquid separation tank 13 connected in sequence. The air compressor 11 compresses air and transports it to the pressure stabilizing tank 12, and then the liquid separation tank 13 injects it into the sealed cavity of the experimental tooling 4 through the electric tracing 2 and the gas path branch 3. Liquid can be injected after the liquid separation tank 13 to simulate the dry gas with liquid failure, and minute solid impurities can be injected to simulate the dry gas impurity failure. By separately adjusting Figure 1 the pressure and flow rate of each gas path branch 3 shown can simulate typical failures such as unstable process parameters, blocked leakage gas pipelines, and reduced intake of sealing gas.

[0111] Specifically, the water circuit includes a circulating water pump 41, a circulating water tank 42, and a refrigerating machine 43. The water circuit is used to remove the heat inside the experimental tooling 4. The experimental tooling 4 is mainly used to install and fix the seal to be tested and the accompanying seal to be tested, and is equipped with a cooling water circulation winding to remove the heat generated during the high-speed rotation of the dry gas seal through the refrigerating machine 43.

[0112] Specifically, the main spindle box 8 is connected to the lubricating oil pump 9.

[0113] Specifically, it also includes a test signal channel and a control signal channel. The signals collected by the test signal channel include: the current and voltage of the motor 6, the spindle speed and torque, the temperature of the main spindle box 8, the temperature, pressure, and flow rate of the gas path branch 3, the gas film thickness signal of the seal to be tested, the acoustic emission signal of the stationary ring 45 of the seal to be tested, the acoustic emission signal and vibration signal of the end cover of the experimental tooling 4.

[0114] There are four paths for the acoustic emission signal. Among them, one path of the stationary ring 45 of the dry gas seal to be tested realizes the internal measurement of the acoustic emission signal, and two radial paths (arranged vertically at 90 degrees) and one axial path of the end cover of the experimental tooling 4 corresponding to the dry gas seal realize the external measurement of the acoustic emission signal.

[0115] Specifically, the vibration signals include: the radial signal of the main spindle box 8, the radial signals of the gearbox 7 and the non-driving end of the experimental tooling 4, the axial signal of the main spindle of the test bench, and the radial signal of the driving end of the tooling.

[0116] Example 4

[0117] As Figure 1A dynamic test and fault simulation experimental bench for dry gas seal multi-source data fusion, as shown in the figure, includes: a power and transmission device, a dry gas seal control system 1, an electric tracing 2, a gas circuit branch 3, an experimental tooling 4, and a control cabinet 5. The power and transmission device includes: a motor 6, a gearbox 7, and a main shaft box 8 that are connected to each other. The control cabinet 5 controls the rotation of the motor 6. The dry gas seal control system 1, the electric tracing 2, the gas circuit branch 3, and the experimental tooling 4 are connected in sequence. The experimental tooling 4 has a water circuit inside. There are two counterweight disks 49 inside the experimental tooling 4 that are equidistant from the two edges of the experimental tooling 4. There is also a loading structure inside the experimental tooling. The experimental tooling 4 is used to fixedly install the seal under test and the mating seal. The experimental tooling 4, the main shaft box 8, the gearbox 7, and the motor 6 are located on a base.

[0118] The motor 6 can simulate stator winding faults, rotor bar breaking faults, and motor 6 bearing faults. The gearbox 7 can achieve faults such as tooth surface pitting and tooth surface wear, realizing the simulation and analysis of electrical, mechanical, and dry gas seal compound faults.

[0119] Specifically, as Figure 2 shown, the loading structure includes: a moving ring 44, a stationary ring 45, a stationary ring seat 46, a piezoelectric crystal 47, and a spring 48. There is a support body 40 inside the experimental tooling 4. The two counterweight disks 49 are respectively located at the left and right ends of the support body 40 and are equidistant from the left and right ends of the support body 40.

[0120] The moving ring 44 and the stationary ring 45 are arranged opposite to each other with a certain gap. There are three piezoelectric crystals 47 arranged between the stationary ring 45 and the stationary ring seat 46. One end of each spring 48 is connected to the stationary ring 45, and the other end is connected to the piezoelectric crystal 47.

[0121] When the three groups of piezoelectric crystals 47 compress the spring 48 under the piezoelectric effect, the stationary ring 45 is driven to move towards the rotating ring 44. Conversely, when the three groups of piezoelectric crystals 47 stretch the spring 48, the stationary ring 45 is driven to move away from the rotating ring 44. A high-pressure-resistant YCOB piezoelectric crystal 47 is installed at the spring 48 between the measured sealed stationary ring seat 46 and the stationary ring 45. When the dry gas seal operates, a gap of several micrometers will be generated between the rotating ring 44 and the stationary ring 45 due to the hydrodynamic pressure effect of the gas. At this time, using the inverse piezoelectric effect of the piezoelectric crystal 47, a positive current is applied to the high-pressure-resistant YCOB piezoelectric crystal 47, and the piezoelectric crystal 47 will compress the spring 48, thereby driving the dry gas seal stationary ring 45 to move towards the dry gas seal rotating ring 44. Conversely, the stationary ring 45 will move in the opposite direction under the action of the spring 48, so as to realize the dynamic adjustment of the gap between the dry gas seal rotating ring 44 and the stationary ring 45. The three groups of high-pressure-resistant YCOB piezoelectric crystals 47 are arranged at intervals of 120 degrees in the radial plane, and together they form a dynamic micro-motion loading structure for the axial force of the dry gas seal. When the three groups of piezoelectric crystals 47 elongate the same length, the stationary ring 45 undergoes a pure axial displacement at this time. When the elongation lengths of the three groups of piezoelectric crystals 47 are inconsistent, the stationary ring 45 will undergo axial and angular displacements. By controlling the reduction of the gap between the stationary ring 45 and the rotating ring 44, the simulation of the rubbing fault between the stationary and rotating rings 45 of the dry gas seal can be realized. By controlling the increase of the gap between the stationary ring 45 and the rotating ring 44, the simulation of faults such as excessive leakage of the dry gas seal can be realized.

[0122] Specifically, as Figure 3 shown, 12 bolt holes 491 are arranged on the counterweight disk 49 at intervals of 30 degrees. When the counterweights are installed at the same angle on the two counterweight disks 49, the force imbalance can be simulated. When the counterweights are installed at opposite angles on the two counterweight disks 49, the couple imbalance can be simulated. Different weights of counterweight bolts can be installed on the counterweight disk 49 to realize the simulation of the imbalance fault of the dry gas seal.

[0123] Specifically, as Figure 4 shown, the three piezoelectric crystals 47 are arranged on the stationary ring seat 46 at intervals of 120 degrees. The three loading structures use the high-pressure-resistant YCOB piezoelectric crystal 47 as the main power source. By controlling the voltage of the high-pressure-resistant YCOB piezoelectric crystal 47 at different positions, the movement of the top of the piezoelectric crystal 47 is realized, and the axial movement and angular swing of the dry gas seal stationary ring 45 are realized, so as to realize the dynamically controllable loading under the conditions of pressure and speed of the dry gas seal.

[0124] Specifically, the piezoelectric crystal 47 is a high-pressure-resistant YCOB piezoelectric crystal 47.

[0125] Specifically, the dry gas seal control system includes an air compressor 11, a pressure stabilizing tank 12, and a liquid separation tank 13 that are connected in sequence. The air compressor 11 compresses air and transports it to the pressure stabilizing tank 12, and then the liquid separation tank 13 injects it into the sealed cavity of the experimental tooling 4 through the electric tracing 2 and the gas path branch 3. Liquid can be injected after the liquid separation tank 13 to simulate the dry gas with liquid failure, and minute solid impurities can be injected to simulate the dry gas impurity failure. By separately adjusting Figure 1 the pressure and flow rate of each gas path branch 3 shown in the figure to simulate typical failures such as unstable process parameters, blocked leakage gas pipelines, and reduced sealed gas intake volume.

[0126] Specifically, the water circuit includes a circulating water pump 41, a circulating water tank 42, and a refrigerator 43. The water circuit is used to remove the heat inside the experimental tooling 4. The experimental tooling 4 is mainly used to install and fix the seal under test and the accompanying seal, and is equipped with a cooling water circulation winding to remove the heat generated during the high-speed rotation of the dry gas seal through the refrigerator 43.

[0127] Specifically, the main shaft box 8 is connected to the lubricating oil pump 9.

[0128] Specifically, it also includes a test signal channel and a control signal channel. The signals collected by the test signal channel include: the current and voltage of the motor 6, the spindle speed and torque, the temperature of the main shaft box 8, the temperature, pressure, and flow rate of the gas path branch 3, the gas film thickness signal of the seal under test, the acoustic emission signal of the stationary ring 45 of the seal under test, the acoustic emission signal and vibration signal of the end cover of the experimental tooling 4.

[0129] The acoustic emission signal includes four channels. Among them, one channel of the stationary ring 45 of the dry gas seal under test realizes the internal measurement of the acoustic emission signal, and two radial channels (arranged perpendicular to each other at 90 degrees) at the corresponding positions of the dry gas seal on the end cover of the experimental tooling 4 and one axial channel on the end cover of the experimental tooling 4 realize the external measurement of the acoustic emission signal.

[0130] Specifically, the vibration signal includes: the radial signal of the main shaft box 8, the radial signals of the gear box 7 and the non-driving end of the experimental tooling 4, the axial signal of the main shaft of the test bench, and the radial signal of the driving end of the tooling.

[0131] Specifically, the multi-source data in the multi-source data fusion dynamic test includes the current of motor 6, the voltage of motor 6, the spindle speed, the spindle torque, the vibration of motor 6, the vibration of gearbox 7, the vibration of headstock 8, the temperature of headstock 8, the pressure of each gas path, the temperature of each gas path, the flow rate of each gas path, the thickness of the dry gas seal gas film to be measured (the gap between the rotating ring 44 and the stationary ring 45), the acoustic emission of the stationary ring 45 of the dry gas seal to be measured (inner measurement), the acoustic emission of the tooling end cover (outer measurement), and the vibration signal of the tooling. Among them, there are four acoustic emission signals. One path of the stationary ring 45 of the dry gas seal to be measured realizes the inner measurement of the acoustic emission signal, and there are two radial paths (arranged perpendicular to each other at 90 degrees) at the corresponding position of the dry gas seal of the tooling end cover, and one axial path of the tooling end cover realizes the outer measurement of the acoustic emission signal. There are nine vibration signals, two radial paths (arranged perpendicular to each other at 90 degrees) of the headstock 8, two of the gearbox 7, two radial paths (arranged perpendicular to each other at 90 degrees) at the non-driving end of the tooling, one axial path, and two radial paths (arranged perpendicular to each other at 90 degrees) at the driving end of the tooling, realizing the vibration measurement of the entire test bench.

[0132] As Figure 5 shown, a different sampling frequency signal synchronization and registration strategy combining hard synchronization and soft synchronization is adopted. An integrated data acquisition card for high-frequency and low-frequency channels is used to realize the synchronous acquisition, transmission, and storage of all signals. The acoustic emission signal is acquired through the high-frequency channel of the data acquisition card, and the remaining signals are acquired through the low-frequency channel of the data acquisition card. All high-frequency and low-frequency acquisition channels share a set of hardware clock signals to achieve clock synchronization between multiple acquisition channels, and the least squares soft registration is performed on the acquired signals based on the high-frequency channel.

[0133] Specifically, it also includes an electric control system, which is used to control the speed and torque of motor 6, the pressure and flow rate of the gas path branch 3, and the current direction in the piezoelectric crystal 47. According to signals such as the gas film thickness, speed signal, main seal gas pressure, temperature, and flow rate, a neural network is used to realize the dynamic calculation and correction of dynamic model parameters, and the precise control of the real-time operating posture of the dry gas seal is achieved.

[0134] Specifically, the neural network input is signals such as the gas film thickness, the spindle speed of the test bench, the pressure, temperature, and flow rate of each gas path, and the neural network output is the piezoelectric crystal control current signal. As Figure 5 shown, the Kalman filter algorithm is introduced to further analyze the outer measurement acoustic emission signal and the inner measurement acoustic emission signal, and signal characteristics that can reflect the opening and closing state between the rotating ring 44 and the stationary ring 45 of the dry gas seal are extracted.

[0135] The electronic control system can achieve stepless adjustment of the motor 6 speed and torque, and separate stepless adjustment of the pressure and flow rate of each gas path. By constructing a dry gas seal gas flow field change model, a high-pressure-resistant YCOB piezoelectric crystal 47 drive machine, and an electro-coupling model under transient and steady-state working conditions, and introducing a neural network to dynamically calculate and correct the key undetermined parameters of the high-pressure-resistant YCOB piezoelectric crystal 47 drive machine with a data-driven method, the online precise control of the real-time operating attitude of the dry gas seal is realized. In view of the problems of insufficient utilization of multi-source data fusion, difficulty in realizing online dynamic loading, and insufficient simulation of typical faults in existing dry gas seal test benches, the present invention proposes a dry gas seal multi-source data fusion dynamic test and fault simulation test bench, which can realize the simulation of electrical-mechanical-dry gas seal composite fault conditions and multi-source data fusion dynamic test analysis, and provide guidance for dry gas seal condition monitoring and early warning technology.

[0136] Embodiment 4

[0137] As Figure 1 shown, a dry gas seal multi-source data fusion dynamic test and fault simulation test bench includes: a power and transmission device, a dry gas seal control system 1, an electric tracing 2, a gas path branch 3, an experimental tooling 4, and a control cabinet 5. The power and transmission device includes: a motor 6, a gearbox 7, and a main shaft box 8 that are connected to each other. The control cabinet 5 controls the rotation of the motor 6. The dry gas seal control system 1, the electric tracing 2, the gas path branch 3, and the experimental tooling 4 are connected in sequence. The experimental tooling 4 has a water path inside. There are two counterweight disks 49 inside the experimental tooling 4 that are equidistant from the two edges of the experimental tooling 4. There is also a loading structure inside the experimental tooling. The experimental tooling 4 is used to fixedly install the seal to be tested and the lining seal. The experimental tooling 4, the main shaft box 8, the gearbox 7, and the motor 6 are located on the base.

[0138] The motor 6 can simulate stator winding faults, rotor bar breaking faults, and motor 6 bearing faults. The gearbox 7 can achieve faults such as tooth surface pitting and tooth surface wear, realizing the simulation and analysis of electrical, mechanical, and dry gas seal composite faults.

[0139] Specifically, as Figure 2 shown, the loading structure includes: a moving ring 44, a stationary ring 45, a stationary ring seat 46, a piezoelectric crystal 47, and a spring 48. There is a support 40 inside the experimental tooling 4. The two counterweight disks 49 are respectively located at the left and right ends of the support 40, and the distances from the left and right ends of the support 40 are equal.

[0140] The moving ring 44 and the stationary ring 45 are arranged oppositely and have a certain gap. There are three piezoelectric crystals 47 arranged between the stationary ring 45 and the stationary ring seat 46. One end of each spring 48 is connected to the stationary ring 45, and the other end is connected to the piezoelectric crystal 47.

[0141] When the three groups of piezoelectric crystals 47 compress the spring 48 under the piezoelectric effect, the stationary ring 45 is driven to move towards the rotating ring 44. Conversely, when the three groups of piezoelectric crystals 47 stretch the spring 48, the stationary ring 45 is driven to move away from the rotating ring 44. A high-pressure-resistant YCOB piezoelectric crystal 47 is installed at the spring 48 between the measured sealing stationary ring seat 46 and the stationary ring 45. When the dry gas seal operates, there will be a gap of several micrometers between the rotating ring 44 and the stationary ring 45 due to the hydrodynamic pressure effect of the gas. At this time, using the inverse piezoelectric effect of the piezoelectric crystal 47, a positive current is applied to the high-pressure-resistant YCOB piezoelectric crystal 47. The piezoelectric crystal 47 will compress the spring 48, thereby driving the dry gas seal stationary ring 45 to move towards the dry gas seal rotating ring 44. Conversely, the stationary ring 45 will move in the opposite direction under the action of the spring 48, so as to realize the dynamic adjustment of the gap between the dry gas seal rotating ring 44 and the stationary ring 45. The three groups of high-pressure-resistant YCOB piezoelectric crystals 47 are arranged at intervals of 120 degrees in the radial plane, jointly forming a dynamic micro-motion loading structure for the axial force of the dry gas seal. When the three groups of piezoelectric crystals 47 elongate the same length, at this time, the stationary ring 45 undergoes a pure axial displacement. When the elongation lengths of the three groups of piezoelectric crystals 47 are inconsistent, the stationary ring 45 will undergo axial and angular displacements. By controlling the reduction of the gap between the stationary ring 45 and the rotating ring 44, the simulation of the rubbing fault between the stationary and rotating rings 45 of the dry gas seal can be realized. By controlling the increase of the gap between the stationary ring 45 and the rotating ring 44, the simulation of faults such as excessive leakage of the dry gas seal can be realized.

[0142] Specifically, as Figure 3 shown, 12 bolt holes 491 are arranged on the counterweight disc 49 at intervals of 30 degrees. When the counterweights are installed at the same angle on the two counterweight discs 49, the force imbalance can be simulated. When the counterweights are installed at opposite angles on the two counterweight discs 49, the couple imbalance can be simulated. Different weights of counterweight bolts can be installed on the counterweight disc 49 to realize the simulation of the imbalance fault of the dry gas seal.

[0143] Specifically, as Figure 4 shown, the three piezoelectric crystals 47 are arranged on the stationary ring seat 46 at intervals of 120 degrees. The three loading structures use the high-pressure-resistant YCOB piezoelectric crystal 47 as the main power source. By controlling the voltage of the high-pressure-resistant YCOB piezoelectric crystals 47 at different positions, the movement of the top of the piezoelectric crystal 47 is realized, and the axial movement and angular swing of the dry gas seal stationary ring 45 are realized, so as to realize the dynamic controllable loading under the conditions of pressure and speed of the dry gas seal.

[0144] Specifically, the piezoelectric crystal 47 is a high-pressure-resistant YCOB piezoelectric crystal 47.

[0145] Specifically, the dry gas seal control system includes an air compressor 11, a pressure stabilizing tank 12, and a liquid separation tank 13 that are connected in sequence. The air compressor 11 compresses air and transports it to the pressure stabilizing tank 12, and then the liquid separation tank 13 injects it into the sealing cavity of the experimental tooling 4 through the electric tracing 2 and the gas path branch 3. Liquid can be injected after the liquid separation tank 13 to simulate the dry gas with liquid failure, and minute solid impurities can be injected to simulate the dry gas impurity failure. By separately adjusting Figure 1 the pressure and flow rate of each gas path branch 3 shown in the figure, typical failures such as unstable process parameters, blocked leakage gas pipelines, and reduced sealing gas intake can be simulated.

[0146] Specifically, the water circuit includes a circulating water pump 41, a circulating water tank 42, and a refrigerating machine 43. The water circuit is used to remove the heat inside the experimental tooling 4. The experimental tooling 4 is mainly used to install and fix the seal under test and the accompanying seal, and is equipped with a cooling water circulation winding to remove the heat generated during the high-speed rotation of the dry gas seal through the refrigerating machine 43.

[0147] Specifically, the main shaft box 8 is connected to the lubricating oil pump 9.

[0148] Specifically, it also includes a test signal channel and a control signal channel. The signals collected by the test signal channel include: the current and voltage of the motor 6, the main shaft speed and torque, the temperature of the main shaft box 8, the temperature, pressure, and flow rate of the gas path branch 3, the gas film thickness signal of the seal under test, the acoustic emission signal of the stationary ring 45 of the seal under test, the acoustic emission signal and vibration signal of the end cover of the experimental tooling 4.

[0149] There are four acoustic emission signals. Among them, one path of the stationary ring 45 of the dry gas seal under test realizes the internal measurement of the acoustic emission signal, and two radial paths (arranged perpendicular to 90 degrees) at the corresponding position of the dry gas seal on the end cover of the experimental tooling 4 and one axial path on the end cover of the experimental tooling 4 realize the external measurement of the acoustic emission signal.

[0150] Specifically, the vibration signals include: the radial signal of the main shaft box 8, the radial signals of the gear box 7 and the non-driving end of the experimental tooling 4, the axial signal of the main shaft of the experimental bench, and the radial signal of the driving end of the tooling.

[0151] Specifically, the acoustic emission signals are collected using the high-frequency channel of the data acquisition card, and the remaining signals are collected using the low-frequency channel of the data acquisition card. The high-frequency channel and the low-frequency channel share a set of hardware clock signals. The data acquisition card is an integrated high-frequency and low-frequency data acquisition card. A synchronization registration strategy for different sampling frequency signals combining hard synchronization and soft synchronization is adopted. Using the integrated high-frequency and low-frequency channel data acquisition card, the synchronous acquisition, transmission, and storage of all signals are realized. The acoustic emission signals are collected using the high-frequency channel of the data acquisition card, and the remaining signals are collected using the low-frequency channel of the data acquisition card. All high-frequency and low-frequency acquisition channels share a set of hardware clock signals, achieving clock synchronization between multiple acquisition channels. For the collected signals, the least squares soft multiplication registration is performed based on the high-frequency channel, that is, the registration is performed using the least squares method.

[0152] Specifically, the multi-source data in the multi-source data fusion dynamic test includes the current of motor 6, the voltage of motor 6, the spindle speed, the spindle torque, the vibration of motor 6, the vibration of gearbox 7, the vibration of headstock 8, the temperature of headstock 8, the pressure of each gas path, the temperature of each gas path, the flow rate of each gas path, the thickness of the gas film of the dry gas seal to be measured (the gap between the rotating ring 44 and the stationary ring 45), the acoustic emission of the stationary ring 45 of the dry gas seal to be measured (inner measurement), the acoustic emission of the tooling end cover (outer measurement), and the tooling vibration signal. Among them, there are four channels of acoustic emission signals. One channel of the stationary ring 45 of the dry gas seal to be measured realizes the inner measurement of the acoustic emission signal. There are two radial channels (arranged perpendicular to each other at 90 degrees) at the corresponding position of the tooling end cover for the dry gas seal, and one axial channel of the tooling end cover realizes the outer measurement of the acoustic emission signal. There are nine channels of vibration signals, including two radial channels (arranged perpendicular to each other at 90 degrees) of the headstock 8, two channels of the gearbox 7, two radial channels (arranged perpendicular to each other at 90 degrees) at the non-driving end of the tooling, one axial channel, and two radial channels (arranged perpendicular to each other at 90 degrees) at the driving end of the tooling, realizing the vibration measurement of the entire test bench.

[0153] As Figure 5 shown, a synchronization registration strategy for different sampling frequency signals combining hard synchronization and soft synchronization is adopted. Using the integrated high-frequency and low-frequency channel data acquisition card, the synchronous acquisition, transmission, and storage of all signals are realized. The acoustic emission signals are collected using the high-frequency channel of the data acquisition card, and the remaining signals are collected using the low-frequency channel of the data acquisition card. All high-frequency and low-frequency acquisition channels share a set of hardware clock signals, achieving clock synchronization between multiple acquisition channels. For the collected signals, the least squares soft registration is performed based on the high-frequency channel.

[0154] Of course, the above description is not a limitation of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions, or substitutions made by those skilled in the art within the scope of the essence of the present invention should also fall within the protection scope of the present invention.

Claims

1. A dynamic test and fault simulation experimental bench for dry gas seal multi-source data fusion, characterized in that, Including: A power and transmission device, a dry gas seal control system, electric tracing heat, a gas path branch, an experimental tooling, and a control cabinet. The power and transmission device includes a motor, a gearbox, and a main spindle box that are connected to each other. The control cabinet controls the rotation of the motor. The dry gas seal control system, the electric tracing heat, the gas path branch, and the experimental tooling are connected in sequence. There is a water path inside the experimental tooling. There are two counterweight disks inside the experimental tooling that are equidistant from the two edges of the experimental tooling. There is also a loading structure inside the experimental tooling. The experimental tooling is used for fixedly installing a seal under test and a mating seal. The experimental tooling, the main spindle box, the gearbox, and the motor are located on a base.

2. The dynamic test and fault simulation experimental bench for dry gas seal multi-source data fusion according to claim 1, characterized in that The loading structure includes a dynamic ring, a static ring, a static ring seat, a piezoelectric crystal, and a spring. There is a support inside the experimental tooling. The two counterweight disks are respectively located at the left and right ends of the support and are equidistant from the left and right ends of the support. The dynamic ring and the static ring are arranged opposite to each other with a certain gap. There are three piezoelectric crystals arranged between the static ring and the static ring seat. One end of each spring is connected to the static ring, and the other end is connected to the piezoelectric crystal. When the three groups of piezoelectric crystals compress the spring under the action of the piezoelectric effect, the static ring is driven to move towards the dynamic ring. Conversely, when the three groups of piezoelectric crystals stretch the spring, the static ring is driven to move away from the dynamic ring.

3. A dynamic test and fault simulation experimental platform for dry gas seals with multi-source data fusion according to claim 2, characterized in that, There are 12 bolt holes arranged at intervals of 30 degrees on the counterweight disk.

4. A dynamic test and fault simulation experimental platform for dry gas seal multi-source data fusion according to claim 2, characterized in that, The three piezoelectric crystals are arranged at intervals of 120 degrees on the static ring seat.

5. The dynamic test and fault simulation test bench for dry gas seal multi-source data fusion according to claim 2, characterized in that The piezoelectric crystal is a high-voltage-resistant YCOB piezoelectric crystal.

6. A dynamic test and fault simulation experimental platform for dry gas seal multi-source data fusion, according to claim 1, characterized in that The dry gas seal control system includes an air compressor, a pressure stabilizing tank, and a liquid separation tank that are connected in sequence. The air compressor compresses air and transports it to the pressure stabilizing tank, and then the liquid separation tank injects it into the seal cavity of the experimental tooling through the electric tracing heat and the gas path branch.

7. A dynamic test and fault simulation experimental platform for dry gas seal multi-source data fusion, according to claim 1, characterized in that The water path includes a circulating water pump, a circulating water tank, and a refrigerating machine. The water path is used to take away the heat inside the experimental tooling.

8. A dynamic test and fault simulation experimental platform for dry gas seal multi-source data fusion, characterized in that, The main spindle box is connected to the lubricating oil pump.

9. A dynamic test and fault simulation experimental bench for dry gas seal multi-source data fusion, characterized in that, Also including: A test signal channel and a control signal channel. The signals collected by the test signal channel include: motor current, voltage, main spindle speed, torque, main spindle box temperature, temperature, pressure, and flow rate of the gas path branch, gas film thickness signal of the seal under test, acoustic emission signal of the static ring of the seal under test, acoustic emission signal and vibration signal of the end cover of the experimental tooling.

10. A dynamic test and fault simulation experimental platform for dry gas seal multi-source data fusion, characterized in that, The vibration signal includes: radial signal of the main spindle box, radial signals of the non-driving ends of the gearbox and the experimental tooling, axial signal of the main spindle of the test bench, and radial signal of the driving end of the tooling.

11. A dynamic test and fault simulation experimental platform for dry gas seal multi-source data fusion, according to claim 9, characterized in that The acoustic emission signal is collected using the high-frequency channel of a data acquisition card, and the other signals are collected using the low-frequency channel of the data acquisition card. The high-frequency channel and the low-frequency channel share a set of hardware clock signals.

12. A dynamic test and fault simulation experimental bench for dry gas seal multi-source data fusion, according to claim 1, characterized in that Also including an electric control system. The electric control system is used to control the motor speed, torque, pressure, and flow rate of the gas path branch, and the current direction in the piezoelectric crystal.

13. The dynamic test and fault simulation experimental platform for dry gas seal multi-source data fusion according to claim 1, characterized in that, The acoustic emission signal includes an internal acoustic emission signal and an external acoustic emission signal, and is processed using the empirical wavelet transform method. The frequency domain of the acoustic emission signal is segmented, and the segmentation boundary is shown in formula (1): Λ n = [ω n-1 , ω n , n = 1, 2, …, N (1); where ω n represents the boundary of each part of the frequency band; Based on the frequency-domain distribution of the internal measurement acoustic emission signal as the main basis, the signal is decomposed, and the decomposed signal is shown in formula (2): and are the approximation coefficient and the detail coefficient respectively, is the empirical wavelet function; F -1 represents the inverse transform from the frequency domain to the time domain; Band - pass filtering is performed on the basis of entropy and kurtosis to select the sensitive frequency band of the dry gas seal operating state, and sub - signals f 内 (t) and f 外 (t) are respectively extracted from the inner - side acoustic emission signal and the outer - side acoustic emission signal; f 外 f(t) = Af 外 (t - 1)+ w p(w) ~ N(0, Q) (3); f 内 f(t) = Hf 外 (t) + vp(v) ~ N(0, R) (4); Among them, A represents the state transition matrix, which is used to describe the transition relationship between the dry gas seal state signals at the previous moment and the current moment. H represents the observation matrix, which is used to describe the mapping relationship between the true state value of the dry gas seal and the external measurement signal. p(w)~N(0,Q) means that p(w) follows a Gaussian distribution with a mean of 0 and a variance of Q, and p(v)~N(0,R) means that p(v) follows a Gaussian distribution with a mean of 0 and a variance of R.

Images