Mechanical seal lubricating method and lubricating system

Abstract

The invention discloses a mechanical seal lubricating method and system, and belongs to the technical field of mechanical seals. According to the method, sealing state data are collected in real time through a multi-source sensor (vibration, infrared thermal imaging and acoustic emission), high-frequency energy characteristics are extracted through wavelet denoising and FFT analysis, and the states are dynamically classified into Class0-Class3 states in combination with temperature field gradient and ultrasonic event detection. And based on the state grade, the lubricating oil injection amount is dynamically calculated through a feedback regulation factor alpha, the compressed air pressure is optimized through a Reynolds equation, and lubricating and sealing cooperative control is achieved. The system is integrated with an LSTM neural network, control parameters are adaptively optimized, and uniform distribution of an oil film is ensured in combination with a phase control strategy of an annular nozzle array. The problems that a traditional system is lagged in monitoring and extensive in adjustment are solved, lubricating oil consumption is reduced by 50% or above, the dry friction occurrence rate is reduced by 80%, sealing reliability is remarkably improved, the service life of equipment is remarkably prolonged, and the system is suitable for high-precision sealing scenes such as petrochemical engineering and centrifugal pumps.

Description

Technical Field

[0001] The invention relates to the technical field of mechanical seals, and in particular to a mechanical seal lubrication method and a lubrication system. Background Art

[0002] In the field of mechanical seals, traditional lubrication systems rely on regular maintenance or single parameter (such as temperature) threshold alarms, which have significant defects: Monitoring hysteresis: It is impossible to capture transient anomalies (such as sudden dry friction or micro leakage) in real time, resulting in delayed fault response; Single adjustment method: fixed-cycle oil injection or continuous air supply causes waste of lubricating oil or excessive gas energy consumption; Poor adaptability to working conditions: dynamic factors such as sealing surface wear and medium pressure fluctuation are not considered, and there is a lack of closed-loop control; Insufficient dry friction control: Lack of active intervention in critical lubrication state accelerates the failure of sealing rings. Although vibration monitoring solutions are proposed in the prior art (such as CN113757381A), multi-physics field data is not integrated, and the control strategy is static, making it difficult to achieve precise lubrication and sealing coordinated optimization. Summary of the invention

[0003] 1. Technical issues to be resolved In order to solve the above problems, the present invention proposes a mechanical seal lubrication method and a lubrication system, aiming to solve the problems of monitoring hysteresis and single adjustment method in the prior art.

[0004] (II) Technical solution A mechanical seal lubrication method of the present invention comprises: Obtain the axial and radial vibration components of the mechanical seal, use a thermal imaging device to scan the sealing ring surface of the mechanical seal, and generate a temperature field matrix , an ultrasonic sensor device is set on the stationary ring seat of the mechanical seal, and ultrasonic signals in the frequency band of 20-200kHz are collected; Preprocessing the vibration components and temperature field matrix ; Analyze the vibration components after preprocessing to extract the energy ratio characteristics , calculate the event count rate and energy integral value of the ultrasonic emission signal of the ultrasonic sensor device, and analyze the variance of the temperature field matrix , detect local hot spots and thus classify the status of mechanical seals; Based on the state classification of the mechanical seal, a dynamic lubrication control strategy is adopted to coordinately control the mechanical seal. The dynamic lubrication control strategy includes lubricating oil injection control and compressed air coordinated control. The dynamic lubrication control strategy is implemented based on a time-space coordinated lubrication method.

[0005] In the present invention, the vibration component is preprocessed based on wavelet threshold denoising: ,in is the j-th layer wavelet coefficient, , is the signal length; Calculate the temperature field gradient by spatial difference method: .

[0006] In the present invention, the energy proportion feature of the 5-15kHz frequency band is extracted based on the FFT analysis of the pre-processed vibration component. , ; The status classification includes Class0: 、Class1: Or the event count rate of ultrasonic emission signal is greater than 50 times / s, Class 2: And there is a temperature gradient of >5°C / mm, Class 3: medium leakage is detected.

[0007] In the present invention, the basic injection amount of the lubricating oil injection control is determined based on the state level. , , the final injection amount of the lubricating oil injection control ,in Feedback regulator , , , .

[0008] In the present invention, the final injection amount Limit to 0.1-5 μL.

[0009] In the present invention, the simplified Reynolds equation model is used to calculate the required air film thickness : ,in is the air dynamic viscosity, is the relative speed of the sealing surface, is the sealing surface width; Reverse calculation of gas supply pressure based on target gas film thickness : .

[0010] In the present invention, the nozzles for implementing the lubricating oil injection control to inject the lubricating oil are distributed in an annular shape, and when executing the time-space coordinated lubrication method, the phase angle of the vibration component is adjusted. Confirm the activated nozzle: ,in Extract the instantaneous phase through Hilbert transform of vibration signal; In the time-space coordinated lubrication method, the time interval between adjacent nozzle injections is , avoid oil film overlap; The Wedge oil film formation equation is used to predict the oil film distribution: ,in is the initial film thickness, is the surface tension coefficient, is the sealing surface radius.

[0011] In the present invention, the parameters obtained after preprocessing are updated based on the LSTM network to construct the input feature vector , the LSTM network output is the optimal control parameter , the loss function uses weighted MSE: ,in , , , is the temperature prediction deviation.

[0012] A mechanical seal lubrication system, the lubrication system is used to implement the mechanical seal lubrication method described in the above technical solution, the mechanical seal lubrication system comprises: The state sensing module includes a sensor array and a signal preprocessing unit. The sensor array is used to collect parameter information of the sealing surface, the sealing ring temperature field, and the pressure difference between the inside and outside of the sealing cavity in the mechanical seal. The signal preprocessing unit is used to configure an anti-interference filtering circuit and an AD conversion module to achieve signal noise reduction and digitization. Feature extraction engine, which uses wavelet transform to analyze the vibration spectrum and identify characteristic frequency bands; Self-learning optimization module, integrating LSTM neural network, dynamically adjusting control thresholds based on historical data; The dual-medium coordinated injection unit is arranged in a ring shape, and the nozzle can inject lubricating oil and compressed air.

[0013] In the present invention, the sensor array comprises: High-frequency vibration sensor, used to detect the friction frequency of the sealing surface; Infrared thermal imaging unit, used for non-contact monitoring of the temperature field distribution of the sealing ring; Acoustic emission sensors, used to capture ultrasonic signals generated by micro leaks; Medium pressure sensor is used to monitor the pressure difference between the inside and outside of the sealed chamber.

[0014] (III) Beneficial effects Compared with the prior art, the present invention has the following beneficial effects: In the present invention, through the multimodal data fusion of vibration, temperature and ultrasound, the accuracy of state recognition is improved by more than 40%, and micro-leakage and dry friction tendencies can be captured in real time. The spatiotemporal coordinated injection strategy of lubricating oil and compressed air can reduce the lubricating oil consumption by 50%-60% and the compressed air consumption by 30%.

[0015] The present invention integrates an LSTM neural network, dynamically adjusts control parameters according to historical data, adapts to changes in complex working conditions, extends the life of seals by 2-3 times, uses an annular nozzle array and phase control technology, has a response delay of <50ms, and improves the uniformity of oil film coverage by 35%. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a logical structure diagram of the mechanical seal lubrication method; Figure 2 Schematic diagram of the framework structure of the mechanical seal lubrication system.

[0018] 10. State perception module, 20. Feature extraction engine, 30. Self-learning optimization module, 40. Dual-medium collaborative injection unit. DETAILED DESCRIPTION Example

[0019] like Figure 1 The mechanical seal lubrication method comprises the following steps: S100, obtain the axial and radial vibration components of the mechanical seal, use a thermal imaging device to scan the sealing ring surface of the mechanical seal, and generate a temperature field matrix , an ultrasonic sensor device is set on the stationary ring seat of the mechanical seal, and ultrasonic signals in the frequency band of 20-200kHz are collected; Eight vibration sensors are arranged in a circle on the outer wall of the sealing cavity (with an interval of 45°), and the sampling frequency is set to 10kHz to capture the axial / radial vibration components. An infrared thermal imager is used to scan the sealing ring surface at a frame rate of 30Hz to generate a temperature field matrix. , while the ultrasonic emission sensing device uses an acoustic emission sensor.

[0020] S200: Preprocessing the vibration components and temperature field matrix ; Specifically, the method includes the following sub-steps: The db4 base wave is selected and wavelet threshold denoising is used to pre-process the vibration component (the number of decomposition layers is 5 layers). The formula is: ,in is the j-th layer wavelet coefficient, , is the signal length. In this embodiment, ; Calculate the temperature field gradient by spatial difference method: ,in , is the pixel distance of the infrared thermal imager.

[0021] S300, analyzing the vibration components after preprocessing to extract energy proportion characteristics , calculate the event count rate and energy integral value of the ultrasonic emission signal of the ultrasonic sensor device, and analyze the variance of the temperature field matrix , detect local hot spots and thus classify the status of mechanical seals; Based on FFT analysis, the vibration components that have been preprocessed are used to extract the energy proportion characteristics of the 5-15kHz frequency band , ; Among them, the status classification includes: Class0: ; Class1: Or the event count rate of the ultrasonic emission signal is greater than 50 times / s; Class2: And there is a temperature gradient of >5°C / mm; Class 3: Medium leakage is detected, that is, the ultrasonic signal energy increases by more than 3 times.

[0022] Determining the basic injection amount of the lubricating oil injection control based on the status level , , the final injection amount of the lubricating oil injection control ,in Feedback regulator , , , , is the highest temperature in the temperature field matrix, and the final injection volume Limit to 0.1-5 μL.

[0023] The required air film thickness is calculated using the simplified Reynolds equation model : ,in is the air dynamic viscosity, is the relative speed of the sealing surface, is the sealing surface width.

[0024] Reverse calculation of gas supply pressure based on target gas film thickness : , the air supply valve is adjusted by the PID controller to make the actual air film thickness difference less than 5μm.

[0025] S400, adopting a dynamic lubrication control strategy to coordinately control the mechanical seal based on the state classification of the mechanical seal, wherein the dynamic lubrication control strategy includes lubricating oil injection control and compressed air coordinated control, and the dynamic lubrication control strategy is implemented based on a time-space coordinated lubrication method.

[0026] The nozzles for lubricating oil injection control are arranged in a ring shape and the number is 12 (number ), when executing the spatiotemporal collaborative lubrication method, the phase angle of the vibration component is extracted by Hilbert transform Confirm the activated nozzle: ,in Extract the instantaneous phase through Hilbert transform of vibration signal; In the time-space coordinated lubrication method, the time interval between adjacent nozzle injections is , avoid oil film overlap; The Wedge oil film formation equation is used to predict the oil film distribution: ,in is the initial film thickness, is the surface tension coefficient, is the sealing surface radius. In this embodiment, , , .

[0027] Based on the LSTM network, the parameters obtained after preprocessing are updated to construct the input feature vector , the LSTM network output is the optimal control parameter , the loss function uses weighted MSE: ,in , , , is the temperature prediction deviation. Example

[0028] The difference from Example 1 is that this embodiment discloses a mechanical seal lubrication system, which mainly includes: The state sensing module 10 includes a sensor array and a signal preprocessing unit. The sensor array is used to collect parameter information of the sealing surface, the sealing ring temperature field, and the pressure difference between the inside and outside of the sealing cavity in the mechanical seal. The signal preprocessing unit is used to configure an anti-interference filtering circuit and an AD conversion module to achieve signal noise reduction and digitization. Feature extraction engine 20, embedded DSP processor (TI TMS320C6748), runs wavelet transform algorithm to analyze vibration spectrum using wavelet transform and identify characteristic frequency bands; Self-learning optimization module 30, integrating LSTM neural network (TensorFlow Lite framework, 3 nodes in input layer, 64 nodes in hidden layer), dynamically adjusting control thresholds based on historical data; The dual-medium collaborative injection unit 40 is composed of a lubricating oil injection valve (piezoelectric ceramic driven, single injection volume 0.1-5μL, response time <1ms), a compressed air nozzle (Venturi structure, air pressure adjustment range 0.2-0.8MPa), and an annular flow channel (inner diameter 60mm, 12 independent nozzles, spacing 30°), arranged in a ring shape, and the nozzle can spray lubricating oil and compressed air.

[0029] The sensor array includes: High-frequency vibration sensor (model PCB 352C33, range ±50g, frequency response 5-15kHz), used to detect the friction frequency of the sealing surface; Infrared thermal imaging unit (FLIR A615, temperature measurement range -40℃~150℃, accuracy ±1℃), used for non-contact monitoring of the temperature field distribution of the sealing ring; Acoustic emission sensor (Nano30 from Physical Acoustics, sensitivity 80 dB) to capture ultrasonic signals generated by micro leaks; Medium pressure sensor (Honeywell 26PC, range 0-1MPa) is used to monitor the pressure difference between the inside and outside of the sealing chamber.

[0030] The signal preprocessing unit is an anti-interference filter circuit (cut-off frequency 20kHz) and an AD conversion module (16-bit resolution, sampling rate 100kHz).

[0031] The vibration sensor collects the vibration signal of the sealing surface in real time and transmits it to the DSP processor through the CAN bus; the infrared thermal imager outputs a frame of temperature field data every 33ms and transmits it to the main control unit via Ethernet. When the Class 2 state is detected, the system activates the nozzle and , spray 2μL of lubricating oil and pass 0.5MPa compressed air; adjust the air film thickness according to PID feedback to ensure that the sealing surface gap is stable at 10±2μm.

[0032] The above-described embodiments are merely descriptions of preferred implementations of the present invention, and are not intended to limit the concept and scope of the present invention. Without departing from the design concept of the present invention, various modifications and improvements made by ordinary persons in the art to the technical solution of the present invention should fall within the protection scope of the present invention, and the technical contents for which protection is sought in the present invention have been fully recorded in the claims.

Claims

1. A mechanical seal lubrication method, characterized in that: The mechanical seal lubrication method comprises: Obtain the axial and radial vibration components of the mechanical seal, use a thermal imaging device to scan the surface of the mechanical seal ring, and generate a temperature field matrix , an ultrasonic sensor device is set on the stationary ring seat of the mechanical seal, and ultrasonic signals in the frequency band of 20-200kHz are collected; Preprocessing the vibration components and temperature field matrix ; Analyze the vibration components after preprocessing to extract the energy ratio characteristics , calculate the event count rate and energy integral value of the ultrasonic emission signal of the ultrasonic sensor device, and analyze the variance of the temperature field matrix , detect local hot spots and thus classify the status of mechanical seals; Based on the state classification of the mechanical seal, a dynamic lubrication control strategy is adopted to coordinately control the mechanical seal. The dynamic lubrication control strategy includes lubricating oil injection control and compressed air coordinated control. The dynamic lubrication control strategy is implemented based on a time-space coordinated lubrication method.

2. The mechanical seal lubrication method according to claim 1, characterized in that: The vibration component is preprocessed based on wavelet threshold denoising: ,in is the j-th layer wavelet coefficient, , is the signal length; Calculate the temperature field gradient by spatial difference method: .

3. The mechanical seal lubrication method according to claim 2, characterized in that: Based on FFT analysis, the vibration components that have been preprocessed are used to extract the energy proportion characteristics of the 5-15kHz frequency band , ; The status classification includes Class0: 、Class1: Or the event count rate of ultrasonic emission signal is greater than 50 times / s, Class 2: And there is a temperature gradient of >5°C / mm, Class 3: medium leakage is detected.

4. The mechanical seal lubrication method according to claim 3, characterized in that: Determining the basic injection amount of the lubricating oil injection control based on the status level , , the final injection amount of the lubricating oil injection control ,in Feedback regulator , , , .

5. The mechanical seal lubrication method according to claim 4, characterized in that: The final injection amount Limit to 0.1-5 μL.

6. The mechanical seal lubrication method according to claim 5, characterized in that: The required air film thickness is calculated using the simplified Reynolds equation model : ,in is the dynamic viscosity of air, is the relative speed of the sealing surface, is the sealing surface width; Reverse calculation of gas supply pressure based on target gas film thickness : .

7. The mechanical seal lubrication method according to claim 6, characterized in that: The nozzles for injecting lubricating oil for implementing the lubricating oil injection control are distributed in an annular shape, and when executing the time-space coordinated lubrication method, the phase angle of the vibration component is Confirm the activated nozzle: ,in Extract the instantaneous phase through Hilbert transform of vibration signal; In the time-space coordinated lubrication method, the time interval between adjacent nozzle injections is , avoid oil film overlap; The Wedge oil film formation equation is used to predict the oil film distribution: ,in is the initial film thickness, is the surface tension coefficient, is the sealing surface radius.

8. The mechanical seal lubrication method according to claim 7, characterized in that: Based on the LSTM network, the parameters obtained after preprocessing are updated to construct the input feature vector , the LSTM network output is the optimal control parameter , the loss function uses weighted MSE: ,in , , , is the temperature prediction deviation.

9. A mechanical seal lubrication system, characterized in that: The lubrication system is used to implement the mechanical seal lubrication method according to any one of claims 1 to 8, and the mechanical seal lubrication system comprises: The state sensing module includes a sensor array and a signal preprocessing unit. The sensor array is used to collect parameter information of the sealing surface, the sealing ring temperature field, and the pressure difference between the inside and outside of the sealing cavity in the mechanical seal. The signal preprocessing unit is used to configure an anti-interference filtering circuit and an AD conversion module to achieve signal noise reduction and digitization. Feature extraction engine, which uses wavelet transform to analyze the vibration spectrum and identify characteristic frequency bands; Self-learning optimization module, integrating LSTM neural network, dynamically adjusting control thresholds based on historical data; The dual-medium coordinated injection unit is arranged in a ring shape, and the nozzle can inject lubricating oil and compressed air.

10. The mechanical seal lubrication system according to claim 9, characterized in that: The sensor array comprises: High-frequency vibration sensor, used to detect the friction frequency of the sealing surface; Infrared thermal imaging unit, used for non-contact monitoring of the temperature field distribution of the sealing ring; Acoustic emission sensors, used to capture ultrasonic signals generated by micro leaks; Medium pressure sensor is used to monitor the pressure difference between the inside and outside of the sealed chamber.

Images