Compressor lubricating oil online monitoring vibration diagnosis method and system

Abstract

The invention relates to the technical field of lubricating oil monitoring, in particular to a compressor lubricating oil online monitoring vibration diagnosis method and system, and the method comprises the steps: obtaining vibration signal data of a compressor bearing, a connecting rod mechanism and a piston assembly and lubricating oil medium parameters, arranging the data according to a time sequence, and generating a lubrication working condition time sequence signal data set; according to the method, through time sequence matching of vibration signal data and lubricating oil state parameters, a lubricating working condition time sequence signal data set is established, and in combination with time window division, the bearing vibration frequency offset, the piston transverse impact amplitude and the connecting rod torsional vibration change rate are extracted; and calculating the viscosity drop rate, the temperature change range and the pollutant particle growth rate of the lubricating oil, screening the correlation trend of the bearing vibration frequency offset before and after the change of the lubricating state, and enhancing the lubricating state recognition capability.

Description

Technical Field

[0001] The present invention relates to the technical field of lubricating oil monitoring, and particularly to an online monitoring vibration diagnosis method and system for compressor lubricating oil. Background Art

[0002] The technical field of lubricating oil monitoring involves monitoring the lubricating oil during the use of mechanical equipment to ensure its good working condition and timely detect potential problems. This field involves the state monitoring, quality assessment of lubricating oil and related fault diagnosis technologies, which are mainly used to improve the operation efficiency of equipment, extend its service life, and prevent equipment failures caused by unqualified or insufficient lubricating oil. By continuously monitoring various performance parameters of the oil, such as temperature, pressure, viscosity and pollutant content, etc., accurate assessment of the equipment operation state can be achieved. The lubricating oil monitoring technology is widely applied to various mechanical equipment, including but not limited to compressors, engines, wind turbine generators, etc., and is particularly important in industrial fields with high requirements for equipment stability.

[0003] Among them, the online monitoring vibration diagnosis method for compressor lubricating oil refers to the real-time monitoring of the state of the lubricating oil during the operation of the compressor, and the analysis of the equipment operation status by combining the diagnosis of vibration signals. This patent theme aims at the problems of lubricating oil monitoring and vibration diagnosis in compressors, and proposes a technical method of comprehensively analyzing by online monitoring the state of lubricating oil and combining vibration data. Specifically, various physical parameters of the lubricating oil are collected by sensors, and the vibration signals are combined to identify and analyze the operation state of the compressor. This method can timely detect potential fault risks through real-time data collection and analysis, help maintenance personnel accurately judge the equipment health status, and avoid production interruption or major accidents caused by equipment failures.

[0004] The prior art relies on single vibration signal or lubrication parameter monitoring, fails to establish time series correlation, is difficult to predict the impact of lubrication anomalies in time, and can only passively diagnose. The short-term data analysis lacks dynamic tracking of viscosity, temperature, and pollutant concentration, affecting trend judgment, making it difficult to identify abnormal signs, and increasing the uncertainty of equipment operation. The matching evaluation of vibration signals and lubrication states is single, and the vibration characteristic offset before and after the change of lubrication states is not comprehensively considered, resulting in the risk of misjudgment, reducing the diagnostic accuracy, and affecting the scientific nature of maintenance strategies. Summary of the Invention

[0005] To solve the technical problems existing in the prior art, embodiments of the present invention provide a method and a device. The technical solutions are as follows:

[0006] An online monitoring vibration diagnosis method for compressor lubricating oil includes the following steps:

[0007] S1: Obtain the vibration signal data of the compressor bearings, connecting rod mechanism, and piston assembly, as well as the lubricating oil medium parameters, arrange them in chronological order, and generate a lubrication condition time series signal dataset;

[0008] S2: Divide the lubrication condition time series signal dataset by time window, calculate the change amount of the lubricating oil state parameters, screen the associated change trends of the vibration signal data before and after the change of the lubricating oil medium, and generate compressor lubrication influence characteristic values;

[0009] S3: Calculate the influence range of the lubricating oil on the vibration signal data according to the compressor lubrication influence characteristic values, select the prediction interval of the resonance frequency offset amount by the viscosity change rate under the lubrication state, and generate the lubrication state vibration prediction result;

[0010] S4: Calculate the vibration state deviation rate of the bearing vibration frequency offset amount according to the lubrication state vibration prediction result, screen the data beyond the prediction interval, determine the difference between the vibration signal and the lubrication state, and generate the compressor vibration characteristic matching result;

[0011] S5: Based on the compressor vibration characteristic matching result, judge whether the vibration state deviation rate is within the lubrication vibration matching data range, classify the fault types according to the judgment result, and generate the compressor lubrication vibration diagnosis result.

[0012] The improvement of the present invention is that the lubrication condition time series signal dataset includes vibration amplitude, vibration frequency, vibration phase, lubricating oil viscosity, lubricating oil temperature, and lubricating oil pollutant concentration. The compressor lubrication influence characteristic values include the associated change trends of the bearing vibration frequency offset amount and the lubricating oil viscosity decrease rate. The lubrication state vibration prediction result includes the predicted offset interval of the bearing resonance frequency offset amount, the influence range of the connecting rod journal torsion amplitude, and the influence range of the piston lateral impact strength. The compressor vibration characteristic matching result includes the matching data of the bearing vibration frequency offset amount and the lubrication vibration influence characteristic result. The compressor lubrication vibration diagnosis result is specifically the vibration change caused by the change of the lubricating oil state, bearing wear fault, connecting rod journal fatigue fault, and piston impact abnormal fault.

[0013] The improvement of the present invention is that the specific steps for obtaining the vibration signal data of the compressor bearings, connecting rod mechanism, and piston assembly, as well as the lubricating oil medium parameters, and arranging them in chronological order to generate a lubrication condition time series signal dataset are as follows:

[0014] S101: Obtain the vibration signals of the compressor bearings, connecting rod mechanism, and piston assembly, measure the amplitude, frequency, and phase of the vibration signals and perform time series calibration, archive and store the calibrated data, and generate a vibration signal dataset;

[0015] S102: Collect the viscosity, temperature, and pollutant concentration parameters of the lubricating oil. Keep the vibration signal dataset synchronized with the lubricating oil medium parameters in the time dimension, and store the collected lubricating oil medium parameters to generate a lubricating oil medium parameter set.

[0016] S103: Based on the vibration signal dataset and the lubricating oil medium parameter set, conduct data correlation analysis according to the time series, match the vibration signal data and the lubricating oil medium parameters within the corresponding time interval, and generate a lubrication condition time series signal dataset.

[0017] The improvement of the present invention is as follows. The specific steps of dividing the lubrication condition time series signal dataset by time window, calculating the change amount of the lubricating oil medium parameters, screening the correlation change trend of the vibration signal data before and after the change of the lubricating oil medium, and generating the compressor lubrication influence eigenvalue are as follows:

[0018] S201: Divide the lubrication condition time series signal dataset into data regions by time window, extract the bearing vibration signal, piston lateral impact signal, and connecting rod torsional vibration signal, calculate the vibration frequency offset, lateral impact amplitude, and torsional vibration change rate, and generate a vibration characteristic parameter set.

[0019] S202: Based on the time dimension of the lubrication condition time series signal dataset, call the vibration characteristic parameter set, extract the lubricating oil viscosity, temperature, and pollutant concentration, calculate the viscosity decrease rate, temperature change range, and pollutant particle growth rate, and generate a lubrication state parameter set.

[0020] S203: According to the vibration characteristic parameter set and the lubrication state parameter set, calculate the correlation coefficient between the bearing vibration frequency offset and the lubricating oil viscosity decrease rate, screen the correlation trend of the bearing vibration frequency offset before and after the change of the lubricating oil medium, and perform data fitting on the calculation results to generate the compressor lubrication influence eigenvalue.

[0021] The improvement of the present invention is that for calculating the correlation coefficient r between the bearing vibration frequency offset and the lubricating oil viscosity decrease rate, the formula is used:

[0022]

[0023] where M r represents the total number of correlation calculation time windows, represents the bearing vibration frequency offset in the m-th time window, represents the average value of the bearing vibration frequency offset in all time windows, represents the viscosity decrease rate in the m-th time window, represents the average value of the viscosity decrease rate in all time windows.

[0024] The improvement of the present invention is as follows. According to the lubrication influence characteristic value of the compressor, calculate the influence range of the lubricating oil on the vibration signal data, select the prediction interval of the resonance frequency offset amount by the viscosity change rate under the lubrication state, and the specific steps for generating the vibration prediction result of the lubrication state are as follows:

[0025] S301: Based on the lubrication influence characteristic value of the compressor, extract the temperature, viscosity, and pollutant particle concentration of the lubricating oil at the current moment, calculate the time series change trend of each parameter, and generate the lubrication parameter change trend value;

[0026] S302: According to the lubrication parameter change trend value, calculate the influence range of the lubricating oil parameters on the piston lateral impact strength, the connecting rod journal torsion amplitude, and the bearing resonance frequency offset amount, set the influence intensity threshold of each lubrication parameter on each vibration characteristic quantity, and construct a non - linear mapping relationship between the lubrication state and the vibration characteristics according to the distribution law of the influence intensity, and generate the lubrication state influence range;

[0027] S303: Based on the lubrication state influence range, extract the correlation data between the viscosity change rate of the lubricating oil and the bearing resonance frequency offset amount, calculate the predicted offset interval of the viscosity change rate, fit the vibration characteristic parameters according to the time series and adjust the matching time window between the lubrication parameter change trend value and the vibration characteristic parameters, extract the fluctuation range of the bearing resonance frequency offset, and generate the vibration prediction result of the lubrication state.

[0028] The improvement of the present invention is that for calculating the predicted offset interval of the viscosity change rate The formula is adopted:

[0029]

[0030] Wherein, represents the mean value of the viscosity change rate, represents the viscosity change rate within the w - th time window, and W η represents the number of samples in the prediction time window.

[0031] The improvement of the present invention is as follows. According to the vibration prediction result of the lubrication state, calculate the vibration state deviation rate of the bearing vibration frequency offset amount, screen the data beyond the prediction interval, determine the difference between the vibration signal and the lubrication state, and the specific steps for generating the compressor vibration characteristic matching result are as follows:

[0032] S401: Based on the vibration prediction result of the lubrication state, extract the current bearing vibration frequency offset amount, the connecting rod journal torsion amplitude, and the piston lateral impact strength, calculate the deviation value of each vibration characteristic value, and generate a set of vibration state deviation rates;

[0033] S402: Based on the set of vibration state deviation rates, screen out the vibration characteristic values that exceed the offset range of the vibration prediction result of the lubrication state, calibrate them as abnormal vibration signals, calculate the matching error between the abnormal vibration signals and the change trend of the lubricating oil medium, and generate a vibration signal matching error value;

[0034] S403: Based on the vibration signal matching error value, call all the matching data of the current bearing vibration frequency offset and the lubrication vibration influence characteristic results, analyze the distribution law of each vibration characteristic in the abnormal state, screen and classify the abnormal vibration modes, and generate a compressor vibration characteristic matching result.

[0035] The improvement of the present invention is that, based on the compressor vibration characteristic matching result, it is judged whether the vibration state deviation rate is within the lubrication vibration matching data range, and the fault types are classified according to the judgment result. The specific steps for generating the compressor lubrication vibration diagnosis result are as follows:

[0036] S501: Based on the compressor vibration characteristic matching result, extract the lubricating oil temperature, viscosity, and pollutant particle concentration, calculate the fluctuation range of each lubrication parameter, compare with the lubrication vibration matching data, screen the matching data and calculate the vibration state deviation rate, and generate a lubrication matching deviation value;

[0037] S502: According to the lubrication matching deviation value, screen out the abnormal vibration signals whose vibration state deviation rate exceeds the lubrication vibration matching data limit. If the data is within the matching interval, it is attributed to the change of the lubrication state. Otherwise, it is attributed to bearing wear, connecting rod journal fatigue, or abnormal piston impact, and generate an abnormal vibration classification result;

[0038] S503: Based on the abnormal vibration classification result, analyze the parameter offset direction and amplitude of each abnormal state, compare with the vibration characteristic distribution of each type of mechanical fault, screen the abnormal modes of the bearing, connecting rod, and piston components, set the vibration characteristic classification standard, and summarize the correlation between the parameter interval of each type of fault and the lubrication state, and generate a compressor lubrication vibration diagnosis result.

[0039] A compressor lubricating oil on-line monitoring vibration diagnosis system, the system includes:

[0040] The data acquisition module acquires the vibration signal data of the compressor bearing, connecting rod mechanism, and piston assembly, acquires the lubricating oil viscosity, temperature, and pollutant concentration parameters, arranges them in chronological order, and generates a lubrication condition time series signal data set;

[0041] Based on the lubrication condition time series signal dataset, the lubrication feature analysis module divides the data region according to time windows, extracts the bearing vibration frequency offset, the piston lateral impact amplitude, and the connecting rod torsional vibration change rate, calculates the lubricating oil viscosity decrease rate, the temperature change range, and the pollutant particle growth rate, screens the associated change trend of the bearing vibration frequency offset before and after the change of the lubricating oil medium, and generates the compressor lubrication influence eigenvalue;

[0042] Based on the compressor lubrication influence eigenvalue, the lubrication state prediction module inputs the current lubricating oil viscosity, temperature, and pollutant particle concentration, calculates the influence range of the lubricating oil on the vibration signal, determines the prediction interval of the resonance frequency offset by the viscosity change rate, and generates the lubrication state vibration prediction result;

[0043] Based on the lubrication state vibration prediction result, the vibration matching evaluation module calculates the vibration state deviation rate of the bearing vibration frequency offset, screens the data beyond the prediction interval, determines the difference between the vibration signal and the lubrication state, and generates the compressor vibration feature matching result;

[0044] Based on the compressor vibration feature matching result, the fault classification and diagnosis module analyzes the lubricating oil viscosity, temperature, and pollutant concentration, judges whether the vibration state deviation rate matches. If it matches, it is determined that it is caused by the change of the lubricating oil medium; otherwise, it is determined as a mechanical fault, and the compressor lubrication vibration diagnosis result is generated.

[0045] The beneficial effects brought by the technical solution provided by the embodiment of the present invention at least include:

[0046] By matching the time series of vibration signal data and lubricating oil state parameters, a lubrication condition time series signal dataset is established. Combining the time window division, the bearing vibration frequency offset, the piston lateral impact amplitude, and the connecting rod torsional vibration change rate are extracted, and the lubricating oil viscosity decrease rate, the temperature change range, and the pollutant particle growth rate are calculated. The associated trend of the bearing vibration frequency offset before and after the lubrication state change is screened to enhance the lubrication state recognition ability. Further calculate the influence of the lubricating oil on the component vibration, establish the prediction interval of the viscosity change rate and the bearing resonance frequency offset, and improve the prediction ability of the lubrication state change. By calculating the vibration state deviation rate, screening the data beyond the prediction interval, and comparing the trend with the historical matching data, the matching accuracy of the vibration signal and the lubrication state is improved. Finally, the fault types are classified by the offset direction and amplitude of the lubrication vibration matching data, realizing the accurate diagnosis of lubrication anomalies and mechanical faults, improving the accuracy of equipment health assessment, and reducing misjudgment and omission. Description of the Drawings

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0048] Figure 1 is the method flow chart of the present invention;

[0049] Figure 2 is the schematic diagram of the refinement process of step S1 of the present invention;

[0050] Figure 3 is the schematic diagram of the refinement process of step S2 of the present invention;

[0051] Figure 4 is the schematic diagram of the refinement process of step S3 of the present invention;

[0052] Figure 5 is the schematic diagram of the refinement process of step S4 of the present invention;

[0053] Figure 6 is the schematic diagram of the refinement process of step S5 of the present invention;

[0054] Figure 7 is the system module diagram of the present invention. Detailed implementation manners

[0055] The following will describe the technical solutions in the present invention in conjunction with the accompanying drawings.

[0056] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as "example" in the present invention should not be construed as more preferred or more advantageous than other embodiments or design solutions. Exactly, the use of the word "example" is intended to present concepts in a specific manner. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or either of the two can be selected.

[0057] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when not emphasizing their differences, the meanings they express are the same. "(of)", "corresponding" and "corresponding" can sometimes be used interchangeably. It should be noted that when not emphasizing their differences, the meanings they express are the same.

[0058] In the embodiments of the present invention, sometimes subscripts such as W1 may be written in a non-subscript form such as W1. When not emphasizing their differences, the meanings they express are the same.

[0059] To make the technical problems, technical solutions, and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments.

[0060] Please refer to Figure 1 , the present invention provides a technical solution: an online monitoring vibration diagnosis method for compressor lubricating oil liquid, including the following steps:

[0061] S1: Obtain the vibration signal data of the compressor bearings, connecting rod mechanisms, and piston assemblies, as well as the lubricating oil liquid medium parameters, and arrange them in chronological order to generate a lubrication condition time series signal data set;

[0062] S2: Divide the lubrication condition time series signal data set according to a time window, calculate the change amount of the lubricating oil liquid state parameters, screen the associated change trends of the vibration signal data before and after the change of the lubricating oil liquid medium, and generate a compressor lubrication influence eigenvalue;

[0063] S3: Calculate the influence range of the lubricating oil liquid on the vibration signal data according to the compressor lubrication influence eigenvalue, select the prediction interval of the resonance frequency offset amount by the viscosity change rate under the lubrication state, and generate a lubrication state vibration prediction result;

[0064] S4: Calculate the vibration state deviation rate of the bearing vibration frequency offset amount according to the lubrication state vibration prediction result, screen the data exceeding the prediction interval, determine the difference between the vibration signal and the lubrication state, and generate a compressor vibration characteristic matching result;

[0065] S5: Based on the compressor vibration characteristic matching result, judge whether the vibration state deviation rate is within the lubrication vibration matching data range, classify the fault types according to the judgment result, and generate a compressor lubrication vibration diagnosis result.

[0066] The lubrication condition time series signal data set includes vibration amplitude, vibration frequency, vibration phase, lubricating oil liquid viscosity, lubricating oil liquid temperature, and lubricating oil liquid pollutant concentration. The compressor lubrication influence eigenvalue includes the associated change trends of the bearing vibration frequency offset amount and the lubricating oil liquid viscosity decrease rate. The lubrication state vibration prediction result includes the predicted offset interval of the bearing resonance frequency offset amount, the influence range of the connecting rod journal torsion amplitude, and the influence range of the piston lateral impact strength. The compressor vibration characteristic matching result includes the matching data of the bearing vibration frequency offset amount and the lubrication vibration influence characteristic result. The compressor lubrication vibration diagnosis result is specifically the vibration change caused by the change of the lubricating oil liquid state, bearing wear fault, connecting rod journal fatigue fault, and piston impact abnormal fault.

[0067] Please refer to Figure 2, the specific steps to obtain the vibration signal data of the compressor bearings, connecting rod mechanism, and piston assembly, as well as the lubricating oil medium parameters, and arrange them in chronological order to generate a lubrication condition time series signal dataset are as follows:

[0068] S101: Obtain the vibration signals of the compressor bearings, connecting rod mechanism, and piston assembly, measure the amplitude, frequency, and phase of the vibration signals and perform time series calibration, archive and store the calibrated data, and generate a vibration signal dataset;

[0069] During the operation of the compressor, first install high-precision vibration sensors on the bearings, connecting rod mechanism, and piston assembly. These sensors capture the vibration signals of each component in real time and obtain their amplitude, frequency, and phase information. For example, select an acceleration sensor with a frequency range of 0 - 10 kHz and a sensitivity of 100 mV / g to ensure that small vibration changes during equipment operation can be accurately captured. The collected vibration signals are transmitted to a computer through a data acquisition sensor for digital processing. During the processing, a high-speed sampling rate (such as 20 kHz) is used to sample the signals to ensure that high-frequency vibration components are captured. Then, calibrate the collected time series data to eliminate environmental noise and system errors. For example, use a band-pass filter to filter out noise signals below 10 Hz and above 5 kHz to ensure the accuracy of the data. The calibrated data is stored in the database in chronological order to form a complete vibration signal dataset. In practical applications, a petrochemical enterprise monitored its reciprocating compressor and found abnormal vibrations in the piston assembly during operation. By collecting and analyzing vibration data using the above method, the fault location and cause were determined, providing a reliable basis for equipment maintenance.

[0070] S102: Collect the viscosity, temperature, and pollutant concentration parameters of the lubricating oil, synchronize the vibration signal dataset with the lubricating oil medium parameters in the time dimension, and store the collected lubricating oil medium parameters to generate a lubricating oil medium parameter set;

[0071] During the operation of the compressor, online sensors are used to monitor parameters such as the viscosity, temperature, and pollutant concentration of the lubricating oil in real time. For example, an online viscometer is installed to measure the dynamic viscosity of the lubricating oil in mPa·s; a thermocouple sensor is used to measure the oil temperature in °C; and a particle counter is employed to detect the pollutant concentration in the oil in particles / mL. To ensure the synchronization of the vibration signal dataset and the lubricating oil parameters in the time dimension, the data acquisition sensor needs to have a high-precision timestamp function to record various types of data according to a unified time reference. For example, the data acquisition frequency is set to 1 Hz, that is, the values of each parameter are recorded once per second. The collected lubricating oil parameter data is preliminarily processed, such as removing outliers and smoothing, and then stored in a database to form a lubricating oil medium parameter set. In practical applications, an abnormal vibration occurred during the operation of a compressor in a certain factory. Through monitoring, it was found that the viscosity of the lubricating oil decreased from the normal value of 68 mPa·s to 50 mPa·s, the temperature increased from 80 °C to 95 °C, and the pollutant concentration increased from 100 particles / mL to 500 particles / mL. These data changes indicate the deterioration of the lubricating oil performance, which may lead to increased equipment wear and requires timely replacement of the lubricating oil and inspection of related components.

[0072] S103: Based on the vibration signal dataset and the lubricating oil medium parameter set, conduct data correlation analysis according to the time series, match the vibration signal data and the lubricating oil medium parameters within the corresponding time interval, and generate a lubrication condition time series signal dataset;

[0073] After obtaining the vibration signal dataset and the lubricating oil medium parameter set, conduct correlation analysis on the two according to the time series. First, match the two datasets according to a unified timestamp to ensure that the vibration data and the lubricating oil parameters at the same time point can correspond. For example, for a certain time point t, the amplitude of the vibration signal is 0.5 g, the frequency is 150 Hz, the viscosity is 60 mPa·s, the temperature is 85 °C, and the pollutant concentration is 300 particles / mL. Then, adopt statistical analysis methods, such as correlation analysis, to evaluate the relationship between the vibration characteristics and the lubricating oil parameters. Calculate the Pearson correlation coefficient between the vibration amplitude and the viscosity. Suppose the obtained correlation coefficient is -0.85, indicating a strong negative correlation between the two, that is, a decrease in viscosity may lead to an increase in the vibration amplitude. By analyzing the data in different time periods, identify the influence mode of the change in lubricating oil parameters on the vibration characteristics. For example, it is found that when the oil temperature exceeds 90 °C, the vibration frequency increases significantly. Finally, store the analysis results as a lubrication condition time series signal dataset to provide data support for subsequent fault diagnosis and maintenance. In practical applications, an abnormal vibration occurred during the operation of a certain compressor. Through the above correlation analysis, it was found that there is a significant correlation between the increase in the pollutant concentration of the lubricating oil and the increase in the vibration amplitude, indicating that it is necessary to strengthen the lubricating oil filtration and replacement frequency to maintain the normal operation of the equipment.

[0074] See also Figure 3 , divide the lubrication condition time series signal data set by time window, calculate the change amount of lubricating oil medium parameters, screen the correlation change trend of vibration signal data before and after the lubricating oil medium changes, and generate the compressor lubrication influence characteristic value. The specific steps are as follows:

[0075] S201: Divide the lubrication condition time series signal data set into data regions according to time windows, extract bearing vibration signals, piston lateral impact signals and connecting rod torsional vibration signals, calculate vibration frequency offset, lateral impact amplitude and torsional vibration change rate, and generate a vibration characteristic parameter set;

[0076] The data in the lubrication condition time series signal data set are first divided according to time windows. The length of each time window is set to 100ms to ensure that the data has sufficient resolution. Then, the bearing vibration signal, piston lateral impact signal and connecting rod torsional vibration signal are extracted in each time window. The bearing vibration signal is obtained by the acceleration sensor installed on the bearing seat. The sensor sampling frequency is set to 10kHz to ensure sufficient sampling density to capture the vibration characteristics, and the piston lateral impact signal is collected by the impact sensor installed on the side wall of the piston. The data sampling interval is set to 1ms to ensure that the transient response during high-speed impact can be recorded. The connecting rod torsional vibration signal is obtained by the angular velocity sensor installed on the connecting rod shaft. The sampling frequency is 5kHz to ensure sufficient resolution to obtain angular velocity changes. For the bearing vibration signal, the data is firstly processed by time domain mean removal to eliminate the system The influence of zero drift is then removed, and the signal is short-time Fourier transformed to extract the main vibration frequency components. On this basis, the bearing vibration frequency offset is calculated. The offset is defined as the offset of the main frequency of the vibration signal relative to the reference main frequency in the current time window. The reference main frequency is set based on the stable vibration frequency of the equipment during normal operation. This value needs to be obtained through long-term measurement and the influence of short-term abnormal disturbances needs to be excluded. The setting principle is: statistics on the equipment operation data of the last 30 days, calculate the main frequency mean, and eliminate extreme data that is higher than the mean + 3 times the standard deviation or lower than the mean - 3 times the standard deviation to ensure the stability of the reference main frequency. For example, in the vibration main frequency measurement data set of a compressor during normal operation, its mean is 250Hz and the standard deviation is 5Hz. The reference main frequency setting range is between 235Hz and 265Hz. If the main frequency detected in the current window is 270Hz, the vibration frequency offset is calculated as follows:

[0077] X f =270-250=20Hz;

[0078] For the piston lateral impact signal, short-time energy calculation is first performed to analyze the impact amplitude in each time window. The optimized calculation formula is as follows:

[0079]

[0080] In the calculation of the impact signal energy, the principle of setting the energy threshold is based on the impact amplitude range under normal working conditions, and the setting method is as follows: collect the impact signal energy within 100 hours of the operation of the acquisition device, and calculate the average energy E mean and the standard deviation σ E , and set the impact anomaly threshold to E mean +4σ E . If during the historical operation of the device, E mean =3000 mJ and σ E =200 mJ, then the abnormal impact threshold is calculated as follows

[0081] E threshold =3000 + 4×200 = 3800 mJ;

[0082] If the energy value calculated within a certain time window is 5000 mJ, exceeding the abnormal threshold, it is determined as an abnormal impact.

[0083] S202: Based on the time dimension of the vibration characteristic parameter set called from the lubrication condition time series signal dataset, extract the lubricating oil viscosity, temperature, and pollutant concentration, calculate the viscosity decrease rate, temperature change range, and pollutant particle growth rate, and generate a lubrication state parameter set;

[0084] Based on the lubrication condition time series signal dataset, first call the time dimension information of the vibration characteristic parameter set to extract the lubricating oil viscosity, temperature, and pollutant concentration. The lubricating oil viscosity is measured by an online viscosity sensor, and the measurement range is set to 10 - 200 cP, and the sensor sampling interval is 1 s to ensure data stability. The lubricating oil temperature is obtained through a temperature sensor, and the measurement range is set to -40 to 150 °C. The pollutant concentration is measured by an oil particle counter, and the set detection particle diameter range is 4 - 100 μm to ensure coverage of the main wear particle sizes. For the viscosity decrease rate, the optimized calculation formula is as follows:

[0085]

[0086] For the setting of the viscosity decrease rate threshold, the setting method is as follows: select the data of the device running continuously for 500 hours, calculate the average viscosity η mean and the standard deviation σ η , and set the abnormal viscosity decrease rate threshold to For example, if the average viscosity change standard deviation then

[0087]

[0088] If the calculated value of R v is greater than -0.3 cP / min, it is determined that the viscosity decrease rate is abnormal.

[0089] S203: According to the vibration characteristic parameter set and the lubrication state parameter set, calculate the correlation coefficient between the bearing vibration frequency offset and the lubricating oil viscosity decrease rate, screen the correlation trend of the bearing vibration frequency offset before and after the change of the lubricating oil medium, perform data fitting on the calculation results, and generate the compressor lubrication influence characteristic value;

[0090] According to the vibration characteristic parameter set and the lubrication state parameter set, calculate the correlation coefficient between the bearing vibration frequency offset and the lubricating oil viscosity decrease rate. The optimized calculation formula is as follows:

[0091]

[0092] For the setting of the significant correlation threshold, the setting method is as follows: Based on the historical data of the vibration frequency offset and the viscosity decrease rate during normal operation of the equipment, calculate the mean value r mean and the standard deviation σ r of the two, and set the significant correlation threshold as:

[0093] r threshold = |r mean | + 2σ r ;

[0094] If the measured r mean = -0.6 and σ r = 0.1, then

[0095] r threshold = |-0.6| + 2×0.1 = 0.8;

[0096] If the calculated r value is greater than 0.8 or less than -0.8, it is determined to be significantly correlated; finally, screen the correlation trend of the bearing vibration frequency offset before and after the change of the lubricating oil medium, and use the data fitting method to process the calculation results, and finally form the compressor lubrication influence characteristic value as the input parameter for subsequent analysis.

[0097] Please refer to Figure 4 , and calculate the influence range of the lubricating oil on the vibration signal data according to the compressor lubrication influence characteristic value, select the prediction interval of the resonance frequency offset by the viscosity change rate under the lubrication state, and generate the specific steps of the lubrication state vibration prediction result as follows:

[0098] S301: Based on the compressor lubrication influence characteristic value, extract the lubricating oil temperature, viscosity, and pollutant particle concentration at the current moment, calculate the time series change trend of each parameter, and generate the lubrication parameter change trend value;

[0099] Based on the characteristic values affecting compressor lubrication, the temperature, viscosity, and pollutant particle concentration of the lubricating oil at the current moment are extracted. The data sampling interval is set to 1 s, and real-time measurements are respectively carried out through a temperature sensor, an on-line viscometer, and an oil pollutant particle sensor. The temperature range measured by the temperature sensor is set to -40 °C to 150 °C, the viscosity range measured by the on-line viscometer is set to 10 cP to 200 cP, and the particle concentration range measured by the pollutant particle sensor is set to 0 to 5000 particles / L. On this basis, the time-series change trend of each parameter is quantified by calculating the change rate and fluctuation amplitude over a past period. For example, within a certain period, the lubricating oil temperature is measured to drop from 85 °C to 80 °C, and the time interval is 10 minutes. Then the optimized calculation of the temperature change rate is as follows:

[0100]

[0101] When setting the temperature change threshold, statistical analysis needs to be carried out based on the long-term operation data of the equipment. The setting method is as follows: Select the temperature change data of the past 100 hours and calculate its mean value and standard deviation σ T , and define the abnormal threshold as If it is measured that σ T = 0.1 °C / min, then the abnormal threshold is calculated as follows

[0102]

[0103] If the calculated temperature change rate exceeds this range, it is necessary to further check whether there is an abnormal lubrication state. Similarly, the optimized calculations of the viscosity change rate and the pollutant particle growth rate are as follows:

[0104]

[0105] Similarly, set the viscosity change threshold and the pollutant particle concentration change threshold using the same method for calculation. For example, if the average viscosity change within the past 100 hours σ η = 0.3 cP / min, then

[0106]

[0107] If the viscosity decrease rate exceeds this range, it is determined that the lubricating oil is aging or the influence of pollutants is serious, and finally the change trend value of the lubrication parameter is calculated.

[0108] S302: According to the lubrication parameter change trend value, calculate the influence ranges of the lubricating oil parameters on the piston lateral impact strength, the connecting rod journal torsion amplitude, and the bearing resonance frequency offset. Set the influence intensity thresholds of each lubrication parameter on each vibration characteristic quantity. Based on the distribution law of the influence intensity, construct a non-linear mapping relationship between the lubrication state and the vibration characteristics, and generate the lubrication state influence range;

[0109] According to the lubrication parameter change trend value, calculate the influence ranges of the lubricating oil parameters on the piston lateral impact strength, the connecting rod journal torsion amplitude, and the bearing resonance frequency offset. Respectively set data acquisition time windows for each vibration characteristic quantity, set the window length to 1 second, and collect 10,000 sample points per second to ensure sufficient time-domain resolution. When calculating the influence range, extract the change rate of the lubrication parameter at the current moment and analyze its influence on the change rate of the vibration characteristic quantity. For example, when the lubricating oil temperature change rate is -0.5 °C / min, the measured piston lateral impact strength change rate is 10%, the connecting rod journal torsion amplitude change rate is 8%, and the bearing resonance frequency offset change rate is 12%. Optimize the calculation of the influence intensity as follows:

[0110]

[0111] When setting the influence intensity threshold, statistical analysis needs to be carried out based on the long-term operation data of the equipment, and the setting method is as follows: Calculate the mean value of the influence intensity within the past 100 hours and the standard deviation Set the abnormal influence range threshold to For example, if the mean value of the temperature influence intensity is -15% / °C and the standard deviation is 2% / °C, then

[0112]

[0113] If the calculated influence intensity is -20% / °C, it is determined that the change in the lubrication temperature has a greater impact on the piston lateral impact strength. Finally, construct a non-linear mapping relationship between the lubrication state and the vibration characteristics, and generate the lubrication state influence range.

[0114] S303: Based on the lubrication state influence range, extract the correlation data between the lubricating oil viscosity change rate and the bearing resonance frequency offset, calculate the predicted offset interval of the viscosity change rate, fit the vibration characteristic parameters according to the time series, and adjust the matching time window between the lubrication parameter change trend value and the vibration characteristic parameters. Extract the fluctuation range of the bearing resonance frequency offset, and generate the lubrication state vibration prediction result;

[0115] Based on the influence range of the lubrication state, extract the correlation data between the viscosity change rate of the lubricating oil and the bearing resonance frequency offset, calculate the predicted offset interval of the viscosity change rate, count the data of the past 50 hours, calculate the mean value and standard deviation of the viscosity change rate, and optimize the calculation of the predicted offset interval as follows:

[0116]

[0117] When setting the threshold of the predicted offset interval, it is necessary to calculate the predicted error range according to historical data, and the setting method is as follows: calculate the mean value of the viscosity change rate error within the past 50 hours and the standard deviation Define the predicted error threshold as If the measured error mean value is 0.2 cP / min and the standard deviation is 0.1 cP / min, then

[0118]

[0119] If the predicted error exceeds this range, it is necessary to adjust and fit the data, and finally generate the vibration prediction result of the lubrication state.

[0120] Please refer to Figure 5 , and calculate the vibration state deviation rate of the bearing vibration frequency offset according to the vibration prediction result of the lubrication state, screen the data exceeding the prediction interval, determine the difference between the vibration signal and the lubrication state, and the specific steps for generating the compressor vibration characteristic matching result are as follows:

[0121] S401: Based on the vibration prediction result of the lubrication state, extract the current bearing vibration frequency offset, the torsional amplitude of the connecting rod journal, and the piston lateral impact strength, calculate the deviation value of each vibration characteristic value, and generate a set of vibration state deviation rates;

[0122] Based on the vibration prediction result of the lubrication state, extract the current bearing vibration frequency offset, the torsional amplitude of the connecting rod journal, and the piston lateral impact strength. Set the data sampling interval to 1 s, and perform real-time measurement through vibration sensors installed on the bearing housing, connecting rod journal, and piston side wall respectively. The measurement range of the bearing vibration frequency offset is set to 0 to 100 Hz, and the measurement range of the torsional amplitude of the connecting rod journal is set to 0 to 10 rad / s 2 , and the measurement range of the piston lateral impact strength is set to 0 to 5000 mJ. On this basis, calculate the deviation value of each vibration characteristic value, and quantify it by calculating the change rate of the current measured value relative to the predicted value. For example, within a certain period of time, the measured bearing vibration frequency offset changes from 20 Hz to 30 Hz, and the torsional amplitude of the connecting rod journal changes from 2 rad / s 2 to 3 rad / s 2 , and the piston lateral impact strength changes from 3000 mJ to 4000 mJ, then optimize the calculation of the deviation rate as follows:

[0123]

[0124] Among them, D f represents the deviation rate of the bearing vibration frequency offset, f current represents the currently measured vibration frequency, f predicted represents the predicted vibration frequency, D θ represents the deviation rate of the connecting rod journal torsion amplitude, θ current represents the currently measured torsion amplitude, θ predicted represents the predicted torsion amplitude, D E represents the deviation rate of the piston lateral impact strength, E current represents the currently measured impact strength, E predicted represents the predicted impact strength. If the measured data are f current = 30 Hz, f predicted = 20 Hz, θ current = 3 rad / s 2 , θ predicted = 2 rad / s 2 , E current = 4000 mJ, E predicted = 3000 mJ, then the calculation is as follows:

[0125]

[0126] When setting the vibration deviation threshold, statistical analysis needs to be carried out based on the long-term operation data of the equipment, and the setting method is as follows: Calculate the mean value D mean and the standard deviation σ D of the deviation rate within the past 100 hours, and set the abnormal vibration deviation threshold as D threshold = D mean + 3σ D , for example, if the mean value of the bearing vibration deviation rate within the past 100 hours is 20% and the standard deviation is 5%, then

[0127] D threshold = 20 + 3(5) = 35%;

[0128] If the calculated vibration deviation rate exceeds 35%, it is determined that the vibration state is abnormal, and finally a vibration state deviation rate set is generated.

[0129] S402: According to the vibration state deviation rate set, screen out the vibration characteristic values that exceed the offset interval of the lubrication state vibration prediction result, calibrate them as abnormal vibration signals, calculate the matching error between the abnormal vibration signals and the change trend of the lubricating oil medium, and generate a vibration signal matching error value;

[0130] According to the vibration state deviation rate set, screen the vibration characteristic values that exceed the offset range of the lubrication state vibration prediction results. Calculate all the vibration state deviation rates and compare them with the predicted offset range. For example, set the predicted offset range of the bearing vibration frequency to 10, the predicted offset range of the crankpin torsional amplitude to 15, and the predicted offset range of the piston lateral impact strength to 5. If the calculated vibration state deviation rates are 50%, 50%, and 33.3% respectively, then these data all exceed the predicted range and are calibrated as abnormal vibration signals. Calculate the matching error between the abnormal vibration signal and the change trend of the lubricating oil medium, and the optimization calculation is as follows:

[0131]

[0132] When setting the matching error threshold, it needs to be calculated based on the historical data of the equipment, and the setting method is as follows: Calculate the mean value E of the matching error in the past 50 hours mean and the standard deviation σ E , and set the abnormal matching error threshold as E threshold =E mean +3σ E , for example, if the mean value of the matching error obtained from historical data analysis is 20% and the standard deviation is 4%, then

[0133] E threshold =20 + 3(4)=32%;

[0134] If the calculated matching error exceeds 32%, it is determined that there is an abnormal matching situation in the influence of the lubrication state on the vibration, and finally the vibration signal matching error value is generated.

[0135] S403: Based on the vibration signal matching error value, call all the matching data of the current bearing vibration frequency offset and the lubrication vibration influence characteristic results, analyze the distribution law of each vibration characteristic in the abnormal state, screen and classify the abnormal vibration modes, and generate the compressor vibration characteristic matching result;

[0136] Based on the vibration signal matching error value, call all the matching data of the current bearing vibration frequency offset and the lubrication vibration influence characteristic results, set the data sampling interval to 1 s, extract the bearing vibration frequency, crankpin torsional amplitude, and piston lateral impact strength data in the past 500 hours, and screen out all the vibration characteristic data whose matching error exceeds the abnormal threshold. Group all the vibration signal data that exceed the matching error threshold, and count the distribution law of each characteristic parameter in the abnormal state. For example, assume that in the bearing vibration frequency data in the past 500 hours, the mean value during normal operation is 40 Hz and the standard deviation is 5 Hz, then the normal vibration range is set as:

[0137] f normal=[40 - 3(5), 40 + 3(5)] = (25, 55] Hz;

[0138] If the bearing vibration frequency measured at a certain moment is 60 Hz, which has exceeded the normal range of 55 Hz, then the data at this point is marked as an abnormal vibration point. Then, all the measurement data within 10 minutes before and after this data point is analyzed and classified into the corresponding abnormal vibration modes. For example, the abnormal modes of the bearing vibration frequency are set as steady-state offset, fluctuating amplitude increase, and sudden increase and decrease. If the vibration frequency remains at 57 Hz - 60 Hz within 10 minutes, this data is classified as steady-state offset. If the vibration frequency shows a change of 50 Hz → 60 Hz → 50 Hz within 10 minutes, it is classified as fluctuating amplitude increase. If a sudden change of 40 Hz → 65 Hz → 40 Hz occurs, it is classified as sudden increase and decrease;

[0139] Similarly, for the torsional amplitude of the connecting rod journal, assuming the mean value during normal operation is 4 rad / s 2 , and the standard deviation is 0.5 rad / s 2 , then the normal range is set as:

[0140] θ normal =[4 - 3(0.5), 4 + 3(0.5)] = [2.5, 5.5] rad / s 2 ;

[0141] If the torsional amplitude of the connecting rod journal measured at a certain moment is 6 rad / s 2 , exceeding the normal range of 5.5 rad / s 2 , then the data at this point is marked as an abnormal torsional vibration point, and the change trend within 10 minutes before and after this data point is further analyzed. If the vibration amplitude is between 5.8 rad / s 2 - 6 rad / s 2 , it is classified as periodic change. If a change of 4 rad / s 2 → 6 rad / s 2 → 3 rad / s 2 occurs, it is classified as sudden instability;

[0142] For the piston lateral impact strength, assuming the mean value during normal operation is 2500 mJ and the standard deviation is 400 mJ, then the normal range is set as:

[0143] E normal =[2500 - 3(400), 2500 + 3(400)] = [1300, 3700] mJ;

[0144] If the measured impact strength at a certain moment is 4000 mJ, exceeding the normal range of 3700 mJ, then this data is calibrated as an abnormal impact signal and classified into the corresponding abnormal mode. If the impact strength within 10 minutes before and after this data point continuously ranges from 3900 mJ to 4000 mJ, it is classified as an increasing impact. If the impact strength changes between 4000 mJ and 2500 mJ, it is classified as a decaying impact;

[0145] When setting the classification threshold for abnormal vibration modes, it is necessary to statistically analyze the occurrence ratio of abnormal modes based on historical data, and calculate the mean value and standard deviation to set the reliable interval for abnormal classification. The setting method is as follows: Calculate the mean value M of all abnormal data points within the past 500 hours abnormal and the standard deviation σ abnormal , and set the recognition threshold for abnormal modes as

[0146] M threshold = M abnormal + 2σ abnormal ;

[0147] For example, if the mean vibration frequency of the sudden increase and decrease mode within the past 500 hours is 60 Hz and the standard deviation is 4 Hz, then

[0148] M threshold = 60 + 2(4) = 68 Hz;

[0149] If the measured vibration frequency of an abnormal point reaches 70 Hz, it indicates that this data point should be classified into a more extreme abnormal mode, such as a sudden jump, so as to refine the mode classification and finally generate the matching result of the compressor vibration characteristics.

[0150] Please refer to Figure 6 , based on the matching result of the compressor vibration characteristics, determine whether the vibration state deviation rate is within the lubrication vibration matching data range, and classify the fault type according to the judgment result. The specific steps for generating the compressor lubrication vibration diagnosis result are as follows:

[0151] S501: Based on the matching result of the compressor vibration characteristics, extract the lubricating oil temperature, viscosity, and pollutant particle concentration, calculate the fluctuation range of each lubrication parameter, compare with the lubrication vibration matching data, screen the matching data and calculate the vibration state deviation rate to generate the lubrication matching deviation value;

[0152] Based on the matching results of the compressor vibration characteristics, the lubricating oil temperature, viscosity, and pollutant particle concentration are extracted. The data sampling interval is set to 1 s, and sensors installed at the lubricating oil pipeline, oil tank, and filter outlet are used for real-time monitoring. The temperature measurement range is set from -40°C to 150°C, the viscosity measurement range is from 10 cP to 200 cP, and the pollutant particle concentration measurement range is from 0 to 5000 particles / L. On this basis, the fluctuation range of each lubrication parameter is calculated, the measurement data for the past 100 hours are extracted, and the maximum, minimum, and average values of each parameter are calculated. For example, if the measured temperature range is 85°C, 95°C, the viscosity range is 120 cP, 140 cP, and the pollutant particle concentration range is 2000 particles / L - 2500 particles / L, then the optimized calculation of the fluctuation range is as follows:

[0153] ΔT = T max - T min = 95 - 85 = 10°C;

[0154] Δη = η max - η min = 140 - 120 = 20 cP;

[0155] ΔP = P max - P min = 2500 - 2000 = 500 particles / L;

[0156] When setting the lubrication parameter fluctuation threshold, based on the operation data for the past 100 hours, the mean value M lub and the standard deviation σ lub are calculated, and the abnormal fluctuation threshold ΔT threshold = M lub ± 3σ lub is set. For example, if the mean value of the temperature fluctuation is 8°C and the standard deviation is 1°C, then

[0157] ΔT threshold = [8 - 3(1), 8 + 3(1)] = [5, 11]°C;

[0158] If the calculated lubrication parameter fluctuation range exceeds this range, it indicates that the lubrication state is abnormal. Then, by comparing the lubrication vibration matching data, the matching data are screened and the vibration state deviation rate is calculated. The optimized calculation is as follows:

[0159]

[0160] Among them, D lub-match represents the lubrication matching deviation rate, M lub represents the mean value of the lubrication parameters, M vib represents the mean value of the vibration characteristic matching data. When setting the lubrication matching deviation rate threshold, the mean value D mean of the historical data and the standard deviation σD and set a matching deviation threshold D threshold = D mean + 3σ D , for example, if the average lubrication matching deviation rate in the past 100 hours is 2.5% and the standard deviation is 1.0%, then

[0161] D threshold = 2.5 + 3(1.0) = 5.5%;

[0162] If the calculated lubrication matching deviation rate exceeds 5.5%, it is considered that there is an abnormal matching situation for this data point, and finally a lubrication matching deviation value is generated.

[0163] S502: According to the lubrication matching deviation value, screen out abnormal vibration signals whose vibration state deviation rate exceeds the lubrication vibration matching data limit. If the data is within the matching interval, it is attributed to the change in lubrication state; otherwise, it is attributed to bearing wear, connecting rod journal fatigue or abnormal piston impact, and an abnormal vibration classification result is generated;

[0164] According to the lubrication matching deviation value, screen out abnormal vibration signals whose vibration state deviation rate exceeds the lubrication vibration matching data limit. Set the lubrication matching interval to ±5. If the calculated deviation rate D lub-match is within this interval, it is determined that the vibration deviation is attributed to the change in lubrication state; otherwise, it is attributed to bearing wear, connecting rod journal fatigue or abnormal piston impact. Attribute classification is performed on the vibration signals that exceed the lubrication matching interval. For example, set the normal range of the bearing vibration frequency matching deviation rate to 0 - 5, the normal range of the connecting rod journal torsional amplitude matching deviation rate to 0 - 7, and the normal range of the piston impact strength matching deviation rate to 0 - 10. If the calculated matching deviation rates are 6%, 8%, and 12% respectively, all data exceeds the normal matching range, and further attribution judgment is required. Compare the historical data of the bearing, connecting rod, and piston, calculate the wear rate, fatigue rate, and impact strength increment, and the optimization calculation is as follows:

[0165]

[0166] When setting the wear rate threshold, calculate the historical data in the past 100 hours, calculate the average value W mean and the standard deviation σ W , and set the abnormal threshold W threshold = W mean + 3σ W , for example, if the average bearing wear rate in the past 100 hours is 0.02 Hz / h and the standard deviation is 0.005 Hz / h, then

[0167] W threshold = 0.02 + 3(0.005) = 0.035 Hz / h;

[0168] If the calculated wear rate exceeds 0.035 Hz / h, it is determined that the bearing has abnormal wear, and finally an abnormal vibration classification result is generated.

[0169] S503: Based on the abnormal vibration classification result, analyze the parameter deviation direction and amplitude of each abnormal state, compare the vibration characteristic distributions of each type of mechanical fault, screen the abnormal modes of the bearing, connecting rod, and piston components, set the vibration characteristic classification criteria, and summarize the correlation between the parameter intervals of each type of fault and the lubrication state to generate the compressor lubrication vibration diagnosis result;

[0170] Based on the abnormal vibration classification result, analyze the parameter deviation direction and amplitude of each abnormal state, compare the vibration characteristic distributions of each type of mechanical fault, and screen the abnormal modes of the bearing, connecting rod, and piston components. Statistically analyze the vibration data of the bearing, connecting rod, and piston in the past 500 hours, and calculate their vibration characteristic distributions. For example, set the abnormal mode of the bearing vibration frequency as steady-state deviation, fluctuating amplitude increase, sudden increase and sudden decrease, the abnormal mode of the connecting rod journal torsion as periodic change, sudden instability, and the abnormal mode of the piston impact as impact increase, impact attenuation. Classify the abnormal signals calculated within the matching error threshold, and set the vibration characteristic classification criteria. For example, the determination condition for the bearing steady-state deviation is that the vibration frequency is within a certain range for a long time. If the vibration frequency is stable at 55 Hz - 57 Hz within 10 minutes, it is classified as steady-state deviation. If the vibration frequency suddenly changes within a short time (50 Hz → 65 Hz → 50 Hz), it is classified as sudden increase and sudden decrease. Similarly, if the angular acceleration of the connecting rod journal is between 6 rad / s 2 -7 rad / s 2 for 10 consecutive minutes, it is classified as periodic change. If it changes between 4 rad / s 2 →7 rad / s 2 →3 rad / s 2 it is classified as sudden instability. If the piston impact intensity is continuously between 3900 mJ - 4000 mJ within 10 minutes, it is classified as impact increase. If the impact intensity changes between 4000 mJ → 2500 mJ, it is classified as impact attenuation. Then, summarize the correlation between the parameter intervals of each type of fault and the lubrication state, calculate the occurrence probability of the abnormal vibration signal under different lubrication states. For example, if the probability of the sudden increase and sudden decrease of the bearing vibration frequency increases from 5% to 15% when the lubricating viscosity is lower than 120 cP, it indicates that the lubricating viscosity has a significant impact on the bearing fault, and finally generate the compressor lubrication vibration diagnosis result.

[0171] Please refer to Figure 7 , a compressor lubricating oil liquid on-line monitoring vibration diagnosis system, the system includes:

[0172] The data acquisition module obtains the vibration signal data of the compressor bearings, connecting rod mechanism, and piston assembly, obtains the parameters of lubricating oil viscosity, temperature, and pollutant concentration, arranges them in chronological order, and generates a time-series signal dataset of lubrication conditions;

[0173] Based on the time-series signal dataset of lubrication conditions, the lubrication feature analysis module divides the data region according to a time window, extracts the bearing vibration frequency offset, piston lateral impact amplitude, and connecting rod torsional vibration change rate, calculates the lubricating oil viscosity decrease rate, temperature change range, and pollutant particle growth rate, screens the associated change trend of the bearing vibration frequency offset before and after the change of the lubricating oil medium, and generates the compressor lubrication influence characteristic value;

[0174] Based on the compressor lubrication influence characteristic value, the lubrication state prediction module inputs the current lubricating oil viscosity, temperature, and pollutant particle concentration, calculates the influence range of the lubricating oil on the vibration signal, determines the prediction interval of the resonance frequency offset due to the viscosity change rate, and generates the lubrication state vibration prediction result;

[0175] Based on the lubrication state vibration prediction result, the vibration matching evaluation module calculates the vibration state deviation rate of the bearing vibration frequency offset, screens the data outside the prediction interval, determines the difference between the vibration signal and the lubrication state, and generates the compressor vibration characteristic matching result;

[0176] Based on the compressor vibration characteristic matching result, the fault classification and diagnosis module analyzes the lubricating oil viscosity, temperature, and pollutant concentration, determines whether the vibration state deviation rate matches. If it matches, it is determined that it is caused by the change of the lubricating oil medium, otherwise it is determined as a mechanical fault, and the compressor lubrication vibration diagnosis result is generated.

[0177] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for online monitoring and vibration diagnosis of compressor lubricating oil, characterized in that: The following steps are involved: S1: Obtain vibration signal data of compressor bearings, connecting rod mechanisms, piston assemblies, and lubricating oil medium parameters, and arrange them in chronological order to generate a lubrication condition time series signal data set; S2: dividing the lubrication condition time series signal data set into time windows, calculating the change amount of the lubricating oil state parameter, screening the correlation change trend of the vibration signal data before and after the lubricating oil medium changes, and generating the compressor lubrication influence characteristic value; S3: calculating the influence range of the lubricating oil on the vibration signal data according to the compressor lubrication influence characteristic value, selecting a prediction interval of the viscosity change rate to the resonance frequency offset under the lubrication state, and generating a lubrication state vibration prediction result; S4: calculating the vibration state deviation rate of the bearing vibration frequency offset according to the lubrication state vibration prediction result, filtering the data beyond the prediction interval, determining the difference between the vibration signal and the lubrication state, and generating the compressor vibration feature matching result; S5: Based on the compressor vibration feature matching result, determine whether the vibration state deviation rate is within the lubrication vibration matching data range, classify the fault type according to the determination result, and generate the compressor lubrication vibration diagnosis result.

2. The compressor lubricating oil online monitoring vibration diagnosis method according to claim 1, characterized in that: The lubrication condition timing signal data set includes vibration amplitude, vibration frequency, vibration phase, lubricating oil viscosity, lubricating oil temperature, and lubricating oil contaminant concentration. The compressor lubrication influence characteristic value includes the associated change trend of the bearing vibration frequency offset and the lubricating oil viscosity decrease rate. The lubrication state vibration prediction result includes the bearing resonance frequency offset prediction offset interval, the connecting rod journal torsional amplitude influence range, and the piston lateral impact strength influence range. The compressor vibration feature matching result includes the matching data of the lubrication vibration influence feature result. The compressor lubrication vibration diagnosis result specifically includes the vibration change caused by the change of the lubricating oil state, the bearing wear failure, the connecting rod journal fatigue failure, and the piston impact abnormal failure.

3. The compressor lubricating oil online monitoring vibration diagnosis method according to claim 1, characterized in that: The specific steps to obtain the vibration signal data of the compressor bearing, connecting rod mechanism, piston assembly and lubricating oil medium parameters and arrange them in chronological order to generate the lubrication condition time series signal data set are as follows: S101: Obtain vibration signals of compressor bearings, connecting rod mechanisms, and piston assemblies, measure the amplitude, frequency, and phase of the vibration signals, and perform time series calibration, archive and store the calibrated data, and generate a vibration signal data set; S102: collecting the viscosity, temperature, and pollutant concentration parameters of the lubricating oil, synchronizing the vibration signal data set with the lubricating oil medium parameters in the time dimension, and storing the collected lubricating oil medium parameters to generate a lubricating oil medium parameter set; S103: Based on the vibration signal data set and the lubricating oil medium parameter set, data association analysis is performed according to the time series, the vibration signal data in the corresponding time interval is matched with the lubricating oil medium parameters, and a lubrication condition time series signal data set is generated.

4. The compressor lubricating oil online monitoring vibration diagnosis method according to claim 1, characterized in that: The specific steps of dividing the lubrication condition time series signal data set by time window, calculating the change amount of the lubricating oil medium parameter, screening the correlation change trend of the vibration signal data before and after the lubricating oil medium change, and generating the compressor lubrication influence characteristic value are as follows: S201: Divide the lubrication condition time series signal data set into data regions according to time windows, extract bearing vibration signals, piston lateral impact signals and connecting rod torsional vibration signals, calculate vibration frequency offset, lateral impact amplitude and torsional vibration change rate, and generate a vibration characteristic parameter set; S202: Based on the lubrication condition time series signal data set, the time dimension of the vibration characteristic parameter set is called to extract the viscosity, temperature, and pollutant concentration of the lubricating oil, calculate the viscosity drop rate, temperature variation range, and pollutant particle growth rate, and generate a lubrication state parameter set; S203: Calculate the correlation coefficient between the bearing vibration frequency offset and the lubricating oil viscosity decrease rate based on the vibration characteristic parameter set and the lubricating state parameter set, select the correlation trend of the bearing vibration frequency offset before and after the lubricating oil medium changes, perform data fitting on the calculation results, and generate the compressor lubrication influence characteristic value.

5. The compressor lubricating oil online monitoring vibration diagnosis method according to claim 4, characterized in that: To calculate the correlation coefficient r between the bearing vibration frequency deviation and the rate of decrease in lubricating oil viscosity, the formula is used: Among them, M r Represents the total number of correlation calculation time windows, represents the bearing vibration frequency offset in the mth time window, Represents the mean value of the bearing vibration frequency offset in all time windows, represents the viscosity decrease rate in the mth time window, Represents the mean of the viscosity decrease rate within all time windows.

6. The compressor lubricating oil online monitoring vibration diagnosis method according to claim 1, characterized in that: The specific steps of calculating the influence range of the lubricating oil on the vibration signal data according to the compressor lubrication influence characteristic value, selecting the prediction interval of the viscosity change rate to the resonance frequency offset under the lubrication state, and generating the lubrication state vibration prediction result are as follows: S301: Based on the compressor lubrication influence characteristic value, extract the lubricating oil temperature, viscosity, and pollutant particle concentration at the current moment, calculate the time series change trend of each parameter, and generate a lubrication parameter change trend value; S302: Calculate the influence range of the lubricating oil parameters on the lateral impact strength of the piston, the torsion amplitude of the connecting rod journal, and the offset of the bearing resonance frequency according to the lubrication parameter change trend value, set the influence intensity threshold of each lubrication parameter on each vibration characteristic quantity, and construct a nonlinear mapping relationship between the lubrication state and the vibration characteristic according to the distribution law of the influence intensity to generate the influence range of the lubrication state; S303: Based on the influence range of the lubrication state, the correlation data between the viscosity change rate of the lubricating oil and the bearing resonance frequency offset are extracted, the predicted offset interval of the viscosity change rate is calculated, the vibration characteristic parameters are fitted according to the time series and the matching time window between the lubrication parameter change trend value and the vibration characteristic parameters is adjusted, the fluctuation range of the bearing resonance frequency offset is extracted, and the lubrication state vibration prediction result is generated.

7. The compressor lubricating oil online monitoring vibration diagnosis method according to claim 6, characterized in that: For the calculation of the viscosity change rate, the predicted deviation interval Using the formula: in, represents the mean value of the viscosity change rate, represents the viscosity change rate in the wth time window, W η Represents the number of samples in the prediction time window.

8. The compressor lubricating oil online monitoring vibration diagnosis method according to claim 1, characterized in that: The specific steps of calculating the vibration state deviation rate of the bearing vibration frequency offset according to the lubrication state vibration prediction result, screening the data beyond the prediction interval, determining the difference between the vibration signal and the lubrication state, and generating the compressor vibration feature matching result are as follows: S401: Based on the lubrication state vibration prediction result, extract the current bearing vibration frequency offset, connecting rod journal torsional amplitude, and piston lateral impact strength, calculate the deviation value of each vibration characteristic value, and generate a vibration state deviation rate set; S402: According to the vibration state deviation rate set, the vibration characteristic values ​​exceeding the deviation interval of the lubrication state vibration prediction result are screened and calibrated as abnormal vibration signals, and the matching error between the abnormal vibration signal and the change trend of the lubricating oil medium is calculated to generate a vibration signal matching error value; S403: Based on the vibration signal matching error value, call all matching data of the current bearing vibration frequency offset and the lubrication vibration influence characteristic results, analyze the distribution law of each vibration characteristic under abnormal conditions, screen and classify abnormal vibration modes, and generate compressor vibration characteristic matching results.

9. The compressor lubricating oil online monitoring vibration diagnosis method according to claim 1, characterized in that: Based on the compressor vibration feature matching result, the specific steps of judging whether the vibration state deviation rate is within the lubrication vibration matching data range, classifying the fault type according to the judgment result, and generating the compressor lubrication vibration diagnosis result are as follows: S501: based on the compressor vibration feature matching result, extract the lubricating oil temperature, viscosity, and pollutant particle concentration, calculate the fluctuation range of each lubrication parameter, and compare the lubrication vibration matching data, filter the matching data and calculate the vibration state deviation rate, and generate the lubrication matching deviation value; S502: According to the lubrication matching deviation value, screen abnormal vibration signals whose vibration state deviation rate exceeds the lubrication vibration matching data limit; if the data is in the matching interval, it is attributed to the change of lubrication state; otherwise, it is attributed to bearing wear, connecting rod journal fatigue or piston impact abnormality, and generate abnormal vibration classification results; S503: Based on the abnormal vibration classification results, analyze the parameter offset direction and amplitude of each abnormal state, compare the vibration characteristic distribution of each type of mechanical fault, screen the abnormal modes of bearings, connecting rods, and piston components, set vibration characteristic classification standards, and summarize the correlation between the parameter range and lubrication status of each type of fault to generate the compressor lubrication vibration diagnosis results.

10. A compressor lubricating oil online monitoring vibration diagnosis system, characterized in that: According to the compressor lubricating oil online monitoring vibration diagnosis method according to any one of claims 1 to 9, the system comprises: The data acquisition module obtains the vibration signal data of the compressor bearing, connecting rod mechanism, and piston assembly, obtains the lubricating oil viscosity, temperature, and pollutant concentration parameters, and arranges them in chronological order to generate a lubrication condition time series signal data set; The lubrication characteristic analysis module is based on the lubrication condition time series signal data set, divides the data area according to the time window, extracts the bearing vibration frequency offset, the piston lateral impact amplitude, and the connecting rod torsional vibration change rate, calculates the lubricating oil viscosity drop rate, the temperature change range, and the pollutant particle growth rate, and screens the correlation change trend of the bearing vibration frequency offset before and after the lubricating oil medium changes, and generates the compressor lubrication influence characteristic value; The lubrication state prediction module inputs the current lubricating oil viscosity, temperature, and pollutant particle concentration based on the compressor lubrication influence characteristic value, calculates the influence range of the lubricating oil on the vibration signal, determines the prediction interval of the viscosity change rate to the resonance frequency offset, and generates a lubrication state vibration prediction result; The vibration matching evaluation module calculates the vibration state deviation rate of the bearing vibration frequency offset based on the lubrication state vibration prediction result, filters the data beyond the prediction interval, determines the difference between the vibration signal and the lubrication state, and generates the compressor vibration feature matching result; The fault classification diagnosis module analyzes the viscosity, temperature, and pollutant concentration of the lubricating oil based on the compressor vibration feature matching results, and determines whether the vibration state deviation rate matches. If it matches, it is determined that the fault is caused by a change in the lubricating oil medium; otherwise, it is determined to be a mechanical fault and the compressor lubrication vibration diagnosis result is generated.

Images