CN117109906A

CN117109906A - Oil online equipment fault analysis method and system based on visualization

Info

Publication number: CN117109906A
Application number: CN202311376604.1A
Authority: CN
Inventors: 叶新功; 赵之玉; 王统海; 王群英; 李永刚; 杜江南; 王玉强; 崔希国; 陈亚华; 胡卫争; 薛孝晗; 穆帅丽
Original assignee: Kasong Science And Technology Co ltd
Current assignee: Kasong Science And Technology Co ltd
Priority date: 2023-10-24
Filing date: 2023-10-24
Publication date: 2023-11-24
Anticipated expiration: 2043-10-24
Also published as: CN117109906B

Abstract

The invention relates to a visual-based oil online equipment fault analysis method and system, and belongs to the field of intelligent manufacturing. The invention determines infrared distribution characteristics based on infrared spectrum; determining impurity particle distribution characteristics according to particle distribution in an online oil sample image; determining whether an oil circuit component fault occurs according to the infrared distribution characteristics; when the oil circuit component fails, determining a reference failure covering oil circuit stroke according to impurity particle distribution characteristics, and determining a covering component related to the oil circuit failure according to infrared distribution characteristics; and providing reference information of equipment faults according to the reference fault coverage oil path and the coverage assembly. By the method, the fault analysis of the oil equipment can be realized.

Description

Oil online equipment fault analysis method and system based on visualization

Technical Field

The invention belongs to the field of intelligent manufacturing, and particularly relates to a visual-based oil online equipment fault analysis method and system.

Background

When the oil runs in the oil circuit, friction, material abrasion and oxidation are experienced, so that the property of the oil is changed.

In the prior art, oil is subjected to off-line testing through an oil pollution degree tester, and the off-line testing is a key device for evaluating the pollutant content in the liquid. The method can rapidly and accurately detect the pollutant level in the liquid through the sensor and the analysis technology. Oil contamination is one of the main causes of equipment failure and performance degradation, and is therefore critical to timely detection and control of oil contamination. The application of the oil on-line monitoring system can help the user to know the content of pollutants such as particles, moisture, oxidation products and the like in the oil, and timely take corresponding maintenance measures. Through accurate measurement of pollution degree, the user can optimize the liquid maintenance plan, improve the reliability of equipment, prolong equipment life-span to reduce cost of maintenance.

The oil monitoring technology plays an important role in ensuring the safe operation of equipment. By monitoring and evaluating the state and performance of lubricating oil or lubricating fluid, the abnormal condition of the oil can be found in time, and equipment faults and accidents can be prevented. The oil monitoring technology can provide key information such as viscosity, acid value, alkali value, water content and the like of oil, and the information can help operation and maintenance personnel to judge the health condition of equipment in time and take corresponding maintenance measures to avoid production interruption and safety accidents caused by equipment faults.

Taking lubricating oil as an example, chemical structure change under the working condition of high-temperature friction of the oil is an important reason for influencing the tribological performance of the oil, the consumption of an antioxidant and accumulation of oxidation products have remarkable influence on the lubricating performance of the oil, and the research on the structure change of the lubricating oil and the action rule of the antioxidant can provide guidance for the mechanism analysis of the decay of the tribological performance of the lubricating oil. Infrared spectroscopy is an important means for representing molecular structure information of lubricating oil, and is widely applied to oxidation and failure analysis of lubricating oil. There are devices for acquiring infrared spectrum of oil liquid by using infrared spectrum online test and devices for acquiring infrared spectrum of oil liquid by offline mode, but such analysis is often carried out on single type of information, and multiple types of sensors are needed to provide reference fault information.

Disclosure of Invention

At least one aspect and advantage of the present invention will be set forth in part in the description that follows, or may be obvious from the description, or may be learned by practice of the presently disclosed subject matter.

According to one embodiment of the invention, the oil online equipment fault analysis method based on visualization comprises the following steps:

determining an infrared distribution characteristic based on an infrared spectrum of the online oil sample;

determining impurity particle distribution characteristics according to particle distribution in an online oil sample image to be detected;

determining whether an oil way fault occurs according to the infrared distribution characteristics;

when the oil circuit component fails, determining a reference failure covering oil circuit stroke according to impurity particle distribution characteristics, and determining a reference component related to the oil circuit failure according to infrared distribution characteristics;

and providing reference information of equipment faults according to the reference fault coverage oil path and the coverage assembly.

According to one embodiment of the invention, the infrared spectrum comprises an infrared spectrum of hydraulic oil and an infrared spectrum of an oil outlet of the hydraulic oil, and the infrared distribution characteristic is based on the change trend of the infrared spectrum of the hydraulic oil and the infrared spectrum of the oil outlet.

According to one embodiment of the invention, determining whether an oil circuit assembly failure has occurred based on the infrared distribution characteristics includes:

Determining whether the change of the absorption peak is a continuous event based on the absorption peak corresponding to the infrared distribution characteristic and the corresponding absorption peak intensity;

when the change of the absorption peak is a continuous event, determining whether an abnormal peak exists according to the composition of the infrared light profile change of the absorption peak compared with the hydraulic oil;

and identifying abnormal peaks, and determining fault points in the oil path when the identification result comprises the characteristic absorption peak area.

According to one embodiment of the invention, the infrared spectrum is an oil outlet infrared spectrum at the outlet of the hydraulic oil, and the infrared distribution characteristic is a change trend determined based on the historical oil outlet infrared spectrum.

when the change of the absorption peak is a continuous event, determining whether an abnormal peak exists according to the composition of the change of the absorption peak;

According to one embodiment of the present invention, the impurity particle distribution characteristic determination reference fault coverage oil passage includes:

Acquiring the particle size and the number of impurity particles in an image, and dividing the impurity particles according to the particle size to obtain a particle size interval and the image proportion occupied by the particle size interval correspondingly;

normalizing the image proportion to obtain the distribution of impurity particle size in the image;

determining a reference source of impurities according to the distribution of the particle sizes of the impurities in the image;

a potential travel range for the impurity is determined based on a reference source of the impurity.

According to one embodiment of the invention, the overlay assembly is determined by:

determining the attribution and the content of impurity peaks according to the infrared distribution characteristics;

the wavelength or wave number of the region corresponding to the impurity peak is used for obtaining a first reference introduction source of the impurity;

acquiring a first reference introduction intensity of the impurity according to the absorption intensity corresponding to the impurity peak;

the overlay assembly is acquired based on the first reference inclusion source and the first reference inclusion intensity.

determining components within the reference fault coverage oil path travel based on the reference fault coverage oil path travel;

According to one embodiment of the invention, the visual-based oil online equipment fault analysis system comprises:

the infrared distribution characteristic acquisition unit is used for determining infrared distribution characteristics based on infrared spectrums of the online oil samples;

the impurity particle distribution characteristic acquisition unit is used for determining impurity particle distribution characteristics according to particle distribution in an online oil sample image to be detected;

the oil circuit fault acquisition unit is used for determining whether an oil circuit component fault occurs according to the infrared distribution characteristics;

the fault reference information acquisition unit is used for determining a reference fault coverage oil path travel according to impurity particle distribution characteristics when an oil path component fault occurs, and determining a coverage component related to the oil path fault according to infrared distribution characteristics;

The invention can realize non-invasive monitoring of the fault point of the oil way.

Drawings

FIG. 1 is a schematic flow chart of an oil online equipment fault analysis method based on visualization in an embodiment of the invention.

FIG. 2 is a chart showing an infrared absorption spectrum of inlet oil according to an embodiment of the present application;

FIG. 3 is a chart showing infrared absorption spectra of the exit oil according to an embodiment of the present application;

FIG. 4 is an image of particulates in an oil using a microscope in one embodiment of the present application.

Detailed Description

The disclosure will now be discussed with reference to several exemplary embodiments. It should be understood that these embodiments are discussed only to enable those of ordinary skill in the art to better understand and thus practice the present disclosure, and are not meant to imply any limitation on the scope of the present disclosure.

As used herein, the term "comprising" and variants thereof are to be interpreted as meaning "including but not limited to" open-ended terms. The term "based on" is to be interpreted as "based at least in part on". The terms "one embodiment" and "an embodiment" are to be interpreted as "at least one embodiment. The term "another embodiment" is to be interpreted as "at least one other embodiment". The terms "front", "rear", and the like refer to an azimuth or positional relationship based on that shown in the drawings. These terms are only used to better describe the present application and its embodiments and are not intended to limit the scope of the indicated devices, elements or components to the particular orientations or to configure and operate in the particular orientations. Also, some of the terms described above may be used to indicate other meanings in addition to orientation or positional relationships, for example, the term "upper" may also be used to indicate some sort of attachment or connection in some cases. The specific meaning of these terms in the present application will be understood by those of ordinary skill in the art according to the specific circumstances. Unless otherwise indicated, the meaning of "a plurality" is two or more.

Before further describing the invention, some detection objects related to online monitoring will be described. In the following content, monitoring is used for indicating that the corresponding state is obtained, and detection is performed on single substances or specific characteristics; features are mainly directed to the processed data, whereas the original spectrogram is indicated in images or spectra.

In the oil circuit system, the particulate matter is the pollutant with the greatest harm, and abrasive particle abrasion, metal crush scratch and metal fatigue can be caused. The particulate matter is generally of a certain hardness, many of which are very small in size and can pass through the gaps between the parts and circulate within the equipment causing wear. Common particulate matter are dust, gravel, fine metal particles produced during operation of the apparatus, rust, and the like.

The cleanliness of the oil liquid and whether the particle pollutants enter can be known by detecting the particle count in the oil liquid.

Moisture is a common contaminant in oil circuit systems, which can destroy the lubrication effect, deteriorate the oil, cause equipment wear, and cause metal corrosion. The moisture in lubricating oils is in three forms: dissolving water, emulsified water and free water, wherein the harm of the emulsified water is the greatest. In addition, some oils, such as lubricating oils, are hygroscopic and absorb moisture from the air and therefore contain a small amount of moisture. The oil, although containing water before reaching the water saturation point, does not show signs of water, such as emulsification, or turbidity, reduced transparency, etc.

When the oil contains emulsified water, the transparency of the oil is reduced, the oil is turbid, and the color is whitish or even turns into milky white. Emulsified water is a great hazard because they can freely flow, contaminate the oil in the entire oil circuit system, and in addition, moisture can deteriorate the lubricity of the oil. After the emulsified water reaches the pressure bearing areas where the equipment is operated, the areas are poor in lubrication, and friction is increased to wear.

Other types of oil are also objects to be detected by the oil circuit system, and when leakage occurs in the oil circuit system, different lubricating oils can be mixed.

In the following embodiments of the present invention, in the fault analysis process performed, it is assumed that the system is still in an operating state, and at least there is a flowing oil path, particularly, the oil speed and the oil pressure are all in a normal interval, and when there is a blockage or the oil pressure is too low, the fault point can be more quickly determined on site by other means, that is, the present invention is more suitable for determining the oil fault with a lower intervention degree.

determining infrared distribution characteristics based on infrared spectra of the online oil samples;

Determining whether an oil circuit component fault occurs according to the infrared distribution characteristics;

when the oil circuit component fails, determining a reference failure covering oil circuit stroke according to impurity particle distribution characteristics, and determining a covering component related to the oil circuit failure according to infrared distribution characteristics;

The invention collects and analyzes the data containing the equipment fault characteristics based on the equipment fault mechanism and the oil on-line analysis technology; determining impurity particle distribution characteristics according to particle distribution in an oil sample image to be detected; determining whether a fault occurs by adopting an optical imaging principle and infrared distribution characteristics; and determining equipment fault tracing analysis and fault positioning according to the impurity particle distribution characteristics. By the method, fault analysis and early warning of the equipment can be realized.

In one embodiment of the invention, the oil sample to be tested is diluted for testing. Impurities introduced into the oil include particulates, water, and other oils, which can cause air bubbles, water, and solids to interfere with the testing, and without processing such information, the above can interfere with the testing. In the present invention, tetrahydrofuran may be used as a solvent for eliminating the corresponding interference. Other solvents such as DMF, DMSO, etc. may be used, such as linear alkanes, as solvents to dilute the oil, regardless of the presence of water, for example. When the solvent is used, the solvent can be selected based on the rule of minimum solubility and phase number, and the solubility is used for providing the solvent for substances in oil liquid so as to enable the substances in the oil liquid to be analyzed, and the signal peaks are prevented from being masked during analysis; while the number of phases is at a minimum that in the presence of multiple phases, the number of phases actually observed is reduced by the introduction of a solvent. For example, the alkane solvent and water are mutually incompatible, and the solubility of the oil can be improved to a certain extent by introducing tetrahydrofuran, so that the uniformity of the sample can be improved.

In one embodiment of the invention, the oil sample to be tested is tested directly. At this time, although particulate matter, water and other oil materials are contained in the oil liquid, when infrared absorption of the oil liquid is obtained, errors possibly existing in a single image can be eliminated by superposing sampling data of a plurality of sampling times, and further, the errors can be eliminated by setting sampling frequencies of infrared images and microscopic images. For example, 6 images are acquired every 1s from 0s in 5s, and the characteristics of the 6 images are extracted respectively and processed to acquire microscopic image information of oil; or obtaining a sample of oil liquid at each interval of 1s, obtaining the characteristics of the infrared part of the oil liquid, extracting the characteristics of the 6 infrared maps, and homogenizing to obtain the infrared characteristics. Because the images of the oil liquid have little variation in a short interval time, such as 1s, the processed infrared spectrum can be provided by overlapping the infrared spectrograms, and the infrared distribution characteristic can be further obtained based on the processed infrared spectrum.

It will be appreciated that the infrared spectrum of the oil, as well as the acquisition of microscopic images, is performed in two different ways, either by solvent dilution or by direct testing.

In addition, the image of the oil sample to be measured is acquired by using an online particle microscope, the online particle microscope is used for acquiring micro particles, particles and fibers, such as size distribution, morphology and concentration, when the resolution is set, the resolution is preferably set between 10 micrometers and 50 micrometers, information is lost when the resolution is higher than 50 micrometers, the analysis difficulty is increased when the resolution is lower than 10 micrometers, and classification and attribution are too discrete in subsequent analysis, so that characteristics cannot be extracted.

In some embodiments of the invention, the infrared spectrum features correspond to the distribution and approximate proportion of groups in each region, for example, an absorption peak of water corresponds to the region of 3600-4000cm < -1 >, and an absorption peak also exists in the region of 1600cm < -1 >, and the approximate distribution and the change trend can be obtained by tracking the intensity of the feature peaks; because a plurality of impurity peaks may exist in one region, analysis cannot be performed in a quantitative manner when a plurality of processes which cause the change of the oil properties exist; and when the system is simpler, the quantitative change process can be determined by using infrared spectrum.

After the infrared distribution characteristics and the impurity particle distribution characteristics are determined, the image can be analyzed manually or automatically, and the abnormal position is determined.

For example, the mechanism and process of deterioration of the oil passage are determined by whether an infrared peak exists an oxidation peak and a water peak when the oil passage malfunction occurs, and whether an impurity peak exists. And determining the travel corresponding to the agglomeration of the particles according to the particle size, for example, the metal particles tend to come from moving parts, and the particles may come from interface sites.

From this, reference information for the device failure can be determined.

The principle of the oil liquid analysis by the on-line infrared spectrum in the invention is as follows:

molecular vibration-the molecules under the action of infrared light can cause specific vibration mode in the interior of the molecules, and different chemical components have specific vibration frequency. These vibration frequencies correspond to absorption peaks in the infrared spectrum, so that components in the liquid can be analyzed by detecting the positions and intensities of the absorption peaks.

Chemical fingerprints the shape and position of the absorption peaks of the infrared spectrum of different substances are unique, and the chemical composition in the liquid can be determined by standard spectral comparison with known substances.

The pollution detection, namely the infrared spectrum analysis can detect impurities, pollutants, degradation products and the like in the liquid.

Thus, oil information can be obtained by analyzing infrared spectra.

For example, in the presence of characteristic peaks corresponding to aldehydes, an oxidation process may occur, which is associated with high temperature or leakage, whereas high temperature oxidation and normal temperature oxidation may differ in their characteristics, the former having significantly higher strength than the latter, based on which it may be determined that the type of failure originates from local overheating, and based on which it may be determined that the oil circuit failure point based on the component where thermal runaway may occur. In the presence of continuous oxidation, mainly from the normal temperature oxidation process, moving parts can be eliminated, and the interface or the vulnerable pipeline is considered first.

In one embodiment of the invention, consistent oil is used in a certain period of time, and then the implementation spectrum of the oil and the basic infrared spectrum of the oil are tracked to determine the change characteristics corresponding to the real-time spectrum; the infrared distribution characteristics are determined by comparing the variance characteristics to differences of the original infrared spectrum to determine the infrared distribution characteristics, or by tracking data over a period of time.

According to one embodiment of the present invention, determining whether an oil circuit failure occurs based on the infrared distribution characteristics includes:

Further, whether the change of the absorption peak is a continuous event is determined based on the absorption peak corresponding to the infrared distribution characteristic and the corresponding absorption peak intensity, for example, whether the height of the peak is higher than a threshold value can be determined by calculating the absorption value of the peak intensity, so as to determine whether the event is a continuous event, and when the peak is lower than the threshold value, the event is regarded as an occasional event; for example, when the transmittance is 100%, the maximum characteristic absorption peak of the oil, such as 15% transmittance, is used as a standard, and when the absorption intensity is higher than that at the time, the absorption peak is regarded as having a continuous absorption peak according to 25% as a threshold;

when the change of the absorption peak is a continuous event, according to the composition of the infrared light profile change of the absorption peak compared with the hydraulic oil, corresponding base deduction is carried out to determine whether an abnormal peak exists or not;

The identification of abnormal peaks is mainly performed here for characteristic wave numbers, for example, free hydroxyl groups are concentrated at 3650-3590cm < -1 >, intermolecular hydrogen bonds are 3500-3300cm < -1 >, aldehydes are 2850-2710cm < -1 >, CO groups in carboxyl groups are 1740-1650cm < -1 >, and the above wave number ranges are not all infrared characteristic peaks of corresponding substances, but have the identified peak number ranges.

The characteristic peaks of the oil in the whole oil path are described, and the oil with mineral oil as a component is taken as an example, and mainly comprises straight-chain, branched alkane and cycloalkane, and may contain a small amount of alcohol, ester and other compounds according to the difference of production processes. Since the above-described substance partial characteristic peak and abnormal peak have overlapping ranges, the wave number range of the abnormal peak should be found by the persistence event, and the attribution of the abnormal peak should be further identified.

In this way, the attribution of the abnormal peak value can be identified, and the fault source is determined according to the power system and the pipeline system in the oil circuit system.

When the infrared spectrum is analyzed when the inlet oil does not have a continuously constant composition, an infrared change event determined by an infrared characteristic for a long time is considered, rather than an occasional event, and the continuous event is obtained by tracking the infrared characteristic of the oil in a certain time and comparing the infrared spectrum acquired in real time with the infrared change event to obtain the change trend.

In one embodiment of the invention, when the substantially uniform oil is used for a certain period of time, tracking the implementation spectrum of the oil and detecting the obtained basic infrared spectrum of the oil for a long period of time to determine the corresponding change characteristics of the real-time spectrum; the infrared distribution characteristics are determined by comparing the variance characteristics to differences of the original infrared spectrum to determine the infrared distribution characteristics, or by tracking data over a period of time.

In one embodiment of the invention, when the oil is replaced after the oil is basically consistent in use in a certain time period, the real-time spectrum of the oil is tracked and accumulated, and the basic infrared spectrum of the oil detected and obtained in a long time period determined based on the real-time infrared spectrum is generally basically stable in 10-15, so that the basic infrared spectrum can be used as the basic infrared spectrum, and the corresponding change characteristic of the real-time spectrum can be determined; the infrared distribution characteristics are determined by comparing the variance characteristics to differences of the base infrared spectrum to determine the infrared distribution characteristics, or tracking data over a period of time.

Here, first, a brief description will be given of the source of impurity particles in the oil passage system,

(1) Residual particles in the oil circuit system refer to pollutants left by abrasion in the initial stage of manufacturing and working, including sand grains which are not thoroughly cleaned, metal scraps, burrs, welding slag and the like which fall into the oil circuit system during assembly and repair;

(2) Particles generated by movement are products formed by mutual friction and abrasion of the surfaces of the friction pair in the relative movement process;

(3) Particles invaded from outside: the externally invaded particles mainly comprise dust, fine sand and other impurities, enter an oil way system through two ways, and are mainly related to interface sealing;

(4) Decay particles, friction polymer particles, fiber particles, carbon deposition particles and oil residue particles are generated due to oxidation, vulcanization, nitrification, additive consumption and other reasons in the use process of the oil.

Among the above-mentioned type 4 particles, for type 1, 3 and 4 particles, the particles are gradually agglomerated in the course of movement, and the variation of the particle size distribution is caused in the agglomeration course, so that fitting is possible by using historical data, wherein particles can be selectively added at different positions (1 position or a plurality of positions), a series of data sets are obtained by measuring the variation of the particles when the movement is ended, the added nodes of an oil circuit are used as input, the particle size distribution of the important particles is used as output, and the positive and negative propagation neural network is used for processing, so that the distribution of the end points which can be obtained when the particles are produced in different environments can be fitted in this way; the correlation of the impurity profile of the endpoint and the original added node can be obtained by approximation. In the method, only useful impurity particle information is considered, namely, impurities with larger particle sizes have larger reference significance, and impurities with lower particle sizes do not influence the normal operation of an oil way; therefore, when the forward and reverse propagation neural network is optimized, the distribution characteristic difference of the minimized large particle impurities can be selected as an objective function to optimize the weight of the neural network.

According to one embodiment of the invention, the covering assembly of the oil passage is determined by:

Here, the assignment and content of impurity peaks are carried out in the form of characteristic peaks, if overlapping peaks are concerned, the calculation can be repeated, for example, free hydroxyl groups are concentrated at 3650-3590cm-1, intermolecular hydrogen bonds at 3500-3300cm-1, aldehydes at 2850-2710cm-1, CO groups in carboxyl groups at 1740-1650cm-1, by which the distribution and content of hydroxyl groups, intermolecular hydrogen bonds, aldehydes and acids are determined; as previously mentioned, the amounts herein are approximate reference proportions and are not actual amounts.

When a large amount of oxidation exists, temperature and sealing factors need to be considered, namely, poor sealing causes air to enter an oil way system, and the temperature aggravates the oxidation process; in the case of small amounts of oxidation, however, the movement site can be excluded, so that a first reference source of introduction of impurities is determined on the basis of this, for example, if there is a potential for leakage, the site containing the joint should be used as the first reference source of introduction; when a large amount of oxidizing species is present, the motion site is introduced as part of the first reference source.

It has been pointed out above that the distribution and content of the correspondence of the impurity peaks is obtained based on the trend of variation, where the correspondence ratio can be determined in accordance with the corresponding peaks, the chemical impurity introduction process corresponding to the peak of the maximum ratio is obtained, such as oxidation to alcohol, acid or and aldehyde, and then the potential first reference introduction intensity is determined based thereon, which may have the correspondence relationship with the first reference introduction source. The determination of the covering assembly is done on the basis of this by means of other sensors.

And then, the covering component and the covering oil path travel can be compared to obtain the reference information of equipment faults, and particularly when the oxidation process caused by the moving part is accelerated, the moving part and the interface simultaneously contained in the covering component and the covering oil path travel are the reference points of faults.

Here, the possibility of failure of the component is first determined from the impurities, i.e., the component in the stroke is determined from the failure-covered oil path stroke, and then the attribution and content of the impurity peaks are determined in the form of characteristic peaks, by which the distribution and content of hydroxyl groups, intermolecular hydrogen bonds, aldehydes and acids are determined.

It has been pointed out above that the distribution and content of the correspondence of the impurity peaks is obtained based on the trend of variation, where the correspondence ratio can be determined in accordance with the corresponding peaks, the chemical impurity introduction process corresponding to the peak of the maximum ratio is obtained, such as oxidation to alcohol, acid or and aldehyde, and then the potential first reference introduction intensity is determined based thereon, which may have the correspondence relationship with the first reference introduction source. On the basis of this, the determination of the covering assembly is carried out by means of other sensors, so that reference information of the failure of the device is obtained.

Example 1

The test was performed using a single source of mineralized oil, the infrared spectrum and particle distribution of which were monitored on-line, and the partially used spectra and images were described with reference to FIGS. 2-4.

And constructing a model of the oil way according to the connecting component and the moving component contained in the oil way, adding particles with different particle diameters into pipelines with different lengths respectively, measuring to obtain an image, and counting the particles in the image according to the size to obtain the distribution of the particles with different particle diameters so as to obtain the correlation between the stroke and the distribution of the particles.

In this process, after an image is acquired, the input of the forward and reverse propagating neural network is constructed as follows:

acquiring an image of oil under a microscope, wherein as shown by referring to a particle diagram in the oil under one microscope in fig. 4, determining that a point corresponding to each pixel has a width of 3 micrometers, firstly removing the back according to red, green and blue (r, g, b) values of each pixel point, wherein the removing method is to calculate the sum of rgb, and when the sum is more than 500 or r+g, g+b or r+b is more than 270, regarding the sum as the back, and the pixels (corresponding to black points or areas on the graph) outside the back are particles, wherein the size of the particles is determined according to the smallest continuous length of the continuously distributed pixels in the transverse direction or the longitudinal direction, for example, the transverse width of a continuous pixel area is 5, and the longitudinal height is 6, and the size of the particles is 5, and corresponds to the size of 15 micrometers; in this way, a plurality of impurity particulate matter regions and corresponding counts can be obtained, and the counts and the image pixel ratio are the duty ratio of the corresponding particulate matters;

setting the step length as 2 pixel points, dividing the particles in the 15-45 micron interval according to the step length, and further obtaining the total number of the particles in each interval and the duty ratio of the particles in the image;

Then, classifying and normalizing each region according to the corresponding size, determining the overall proportion of impurity particles in the image according to the number of actual pixel points of the region, for example, adopting a linear normalization method, calculating 1630 points which are particles in one image, wherein 150 points which are pixels with the size of 5-7 pixels correspond to the region, and the content of 15-18 microns of particles in the image is 0.092; in this embodiment, the particles with an oversized size, for example, 20 pixels, are not counted, because the distribution of the impurity particles continuously distributed in the image is affected by the sharpness, and the judgment result is distorted according to the influence; when the data is generated, an exponential normalization mode can be adopted; in addition, when linear normalization is employed, the calculation of the proportion of particles below 15 microns is generally omitted; if the particulate matter proportion calculation of the part is considered, the exponential normalization can be used to avoid the problem of uneven distribution of the particulate matter interval at the low particulate matter content.

Based on which a series of values is obtained as input; the method is characterized in that the method comprises the steps of inputting the method to an impurity particle size distribution based on an oil path end point and constructing a forward and reverse propagation neural network based on the input oil path length, wherein the forward and reverse propagation neural network comprises a hidden layer, the first hidden layer is fully connected with an input layer, the input layer is the distribution of different particle sizes, and the output layer is the length of a corresponding adding point.

When monitoring the infrared spectrum, an image was acquired every 10s, please refer to fig. 3, which shows an infrared spectrum of the export oil acquired in real time, according to 2925 wavenumber cm ^-1 Superposing peak shapes as reference, taking the latest 10 maps, respectively superposing the maps and comparing the maps with an infrared spectrogram (figure 2) of inlet oil to obtain newly introduced peaks as infrared distribution characteristics by deducting according to wave numbers, and calculating 4000-1000 wave number cm ^-1 Wherein the new absorption peak ratio is 1000-4000 wavenumber cm ^-1 Peak 200 wave number cm in step ^-1 The ratio of 0.13, 0.17, 0.23, 0.19, 0, 0.07, 0, 0.12, 0.08, 0, and calculated regardless of the absorption peak of less than 2%, the rest is calculated according to the actual absorption value, and the sum is close to 1 but not equal to 1 due to rounding, and the data can be directly used for determining classification; 1600-1400 wave number cm was determined as described above ^-1 1800-1600 wavenumber cm ^-1 The absorption peak corresponding to the interval of (2) is a new region, which generally corresponds to the vibration of the ester group and is in the range of 1000-1200 wavenumber cm ^-1 Suction is also present in the sectionReceiving, based on which the existence of an oxidation process or leakage of the material containing the ester is indicated; and at 3200-3400 wavenumber cm ^-1 The presence of absorption in the interval also indicates the possibility of amine presence. Further analysis may be performed in conjunction with the oil circuit at this point.

This process involves a noise determination, in which the basis is referenced to 2925 wavenumber cm ^-1 The absorption default of the corresponding interval and the adjacent step interval is 0 so as to avoid introducing larger errors; in addition, the obtained data measured in different time periods can be adopted for comparison aiming at the change composition of other absorption peaks to calculate abnormal peak values, and the process of determining abnormal absorption comprises the following steps: performing according to standard deviation, deducting by using inlet oil and an infrared spectrogram acquired in the last time period to obtain peak duty ratio of each region determined according to step length, calculating the mean value of each non-zero step length interval, and then calculating the standard deviation to obtain an interval larger than 2.0 times of the standard deviation as abnormal absorption, wherein the threshold value obtained is 0.108;

the above process can be further shortened to 100 wave number cm ^-1 But this process causes a decrease in recognition sensitivity caused by the abnormal peak judgment process in the long-time recognition process;

after the judgment is completed, the prompt of the abnormal characteristic peak can be carried out on a user interface presenting the infrared spectrogram.

In the embodiment, the source corresponding to the impurity is determined based on the 10-35 micron impurity contained in each 5s acquired image combined with the forward and reverse propagation neural network, the components on the corresponding paths are selected and compared with the reference components determined according to infrared rays, and the covering component is determined to be used as the reference information of the fault.

Example 2

The method comprises the steps of constructing a forward and reverse propagation neural network based on impurity particle size distribution of an oil path end point and the length of an input oil path, wherein the forward and reverse propagation neural network comprises a hiding layer, the first hiding layer is fully connected with an input layer, the input layer is distributed with different particle sizes, and the output layer is the length of a corresponding adding point.

And then, a branch is led out from the tail end of the oil way, and the infrared spectrum and the particle distribution of the branch are monitored on line.

When the infrared spectrum is monitored, an image is acquired every 5s, the image is respectively overlapped and the infrared spectrum of the inlet oil is compared and buckled with the infrared spectrum of the inlet oil to obtain a newly introduced peak as an infrared distribution characteristic, an impurity peak in the newly introduced peak is extracted, a first reference introduction source is determined according to the comparison of the type and the proportion of the impurity peak and an oil path model, then a reference introduction source is determined according to the impurity peak with the maximum proportion, and a specific rule can comprise that when oxidation caused by movement is included, the reference assembly is correspondingly obtained according to the movement intensity of a movement assembly in the oil path model, namely, a component with high temperature is oxidized more quickly;

And then determining the source corresponding to the impurity based on the 10-50 micrometers of impurity contained in each 5s acquired image combined with the forward and reverse propagation neural network, selecting components on the corresponding paths, comparing the components with the reference components determined according to infrared, and determining the covering component as the reference information of the fault.

Example 3

The test was performed using mineralized oil of a single source, after which a branch was led out at the end of the oil line, which was mixed with tetrahydrofuran THF, and the infrared spectrum and particle distribution were monitored on line.

When monitoring infrared spectra, acquiring an image every 10s, respectively overlapping the image, comparing the image with an infrared spectrogram of inlet oil to obtain a newly introduced peak as an infrared distribution characteristic, extracting an impurity peak in the image, comparing the impurity peak with an oil path model according to the type and proportion of the impurity peak, determining a first reference introduction source according to the impurity peak with the maximum proportion, and determining a reference introduction source according to the impurity peak with the maximum proportion, wherein the specific rule can comprise that when oxidation caused by movement is included, the reference assembly is correspondingly obtained by the reference introduction source according to the movement intensity of a movement assembly in the oil path model, namely, the oxidization of a part with high temperature is faster;

And then determining the source corresponding to the impurity based on the 10-35 micron impurity contained in each 5s acquired image combined with the forward and reverse propagation neural network, selecting the components on the corresponding paths and comparing the components according to the infrared determined reference components, and determining the covering component as the reference information of the fault.

Example 4

A single source mineralized oil is used for testing, a branch is led out from the tail end of an oil path, and after the mineralized oil is mixed with dimethyl imine DMF, the infrared spectrum and the particle distribution of the mineralized oil are monitored on line.

And then determining the source corresponding to the impurity based on the 10-35 micron impurity contained in each 5s acquired image combined with the forward and reverse propagation neural network, selecting the components on the corresponding paths and comparing the components according to the infrared determined reference components, and determining the covering component as the reference information of the fault. This approach is not suitable for scenes containing the introduction of oxidizing species, as it contains NH2 and CO groups.

Example 5

A single source mineralized oil is used for testing, a branch is led out from the tail end of an oil path, and the infrared spectrum and the particle distribution of the mineralized oil are monitored on line.

When monitoring infrared spectra, acquiring an image every 5s, respectively overlapping the image, and comparing the image with an infrared spectrogram of inlet oil to obtain a newly introduced peak as an infrared distribution characteristic, extracting an impurity peak in the image, comparing the type and the proportion of the impurity peak with an oil path model to determine a first reference introduction source, and then determining a reference introduction source according to the impurity peak with the maximum proportion, wherein a specific rule can comprise that when oxidization caused by movement is included, the reference introduction source is correspondingly obtained with the reference introduction source according to the movement intensity of a movement assembly in the oil path model, namely, the oxidization of a part with high temperature is faster;

Example 6

And (3) testing by using recycled mineralized oil, leading out a branch at the tail end of the oil path, and monitoring the infrared spectrum and particle distribution of the oil path on line.

When monitoring infrared spectra, acquiring an image every 5s, respectively overlapping the image, comparing the image with a historical infrared spectrogram, deducting the background to obtain a newly introduced peak as an infrared distribution characteristic, extracting an impurity peak in the image, comparing the type and the proportion of the impurity peak with an oil path model to determine a first reference introduction source, and then determining a reference introduction source according to the impurity peak with the maximum proportion, wherein a specific rule can comprise that when oxidization caused by movement is included, the image is correspondingly obtained with the reference introduction source according to the movement intensity of a movement assembly in the oil path model to obtain a reference assembly, namely, oxidization of a part with high temperature is faster;

determining sources corresponding to the impurities based on the 10-50 micrometers of impurities contained in the images acquired every 5s and combining with a forward and reverse propagation neural network, selecting components on corresponding paths and comparing with reference components determined according to infrared rays, and determining a covering component as reference information of faults;

In this embodiment, the historical infrared spectrum is a spectrum obtained by overlapping infrared spectra of approximately 10 minutes.

Example 7

Determining a source corresponding to the impurity based on the 10-50 micrometers of the impurity contained in each 5s acquired image and combining a forward and reverse propagation neural network, and selecting a component on a corresponding oil path travel;

then analyzing based on an infrared spectrogram, acquiring an image every 5s, respectively overlapping the image, comparing the image with a historical infrared spectrogram, removing the back to obtain a newly introduced peak as an infrared distribution characteristic, extracting an impurity peak in the image, comparing the type and the proportion of the impurity peak with an oil path model to determine a first reference introduction source contained in a corresponding oil path stroke, and determining a reference introduction source according to the impurity peak with the maximum proportion, wherein a specific rule can comprise that when oxidization caused by movement is contained, the reference introduction source is corresponding to the reference introduction source according to the movement intensity of a moving component in the oil path model to obtain a covering component as fault reference information, namely that a component with high temperature oxidizes more quickly;

In this embodiment, the historical infrared spectrum is a spectrum obtained by overlapping infrared spectra of approximately 6 minutes.

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, specific working procedures of the apparatus and device described above may refer to corresponding procedures in the foregoing method embodiments, which are not described herein again.

The above description is only illustrative of the preferred embodiments of the present application and of the principles of the technology employed. It will be appreciated by persons skilled in the art that the scope of the application referred to in the present application is not limited to the specific combinations of the technical features described above, but also covers other technical features formed by any combination of the technical features described above or their equivalents without departing from the inventive concept. Such as the above-mentioned features and the technical features disclosed in the present application (but not limited to) having similar functions are replaced with each other.

Claims

1. The oil online equipment fault analysis method based on visualization is characterized by comprising the following steps of:

2. The visualization-based oil online equipment fault analysis method according to claim 1, wherein the infrared spectrum comprises an infrared spectrum of hydraulic oil and an infrared spectrum of an oil outlet of the hydraulic oil, and the infrared distribution is characterized by the change trend of the infrared spectrum and the infrared spectrum of the oil outlet of the hydraulic oil.

3. The visualization-based oil online equipment failure analysis method of claim 2, wherein determining whether an oil circuit assembly failure occurs according to the infrared distribution characteristics comprises:

4. The visualization-based oil online equipment fault analysis method according to claim 1, wherein the infrared spectrum is an oil outlet infrared spectrum at a hydraulic oil outlet, and the infrared distribution characteristic is a change trend determined based on a historical oil outlet infrared spectrum.

5. The visualization-based oil online equipment failure analysis method of claim 4, wherein determining whether an oil circuit assembly failure occurs according to the infrared distribution characteristics comprises:

6. The visualization-based oil online equipment fault analysis method as claimed in claim 1, wherein the impurity particle distribution feature determination reference fault coverage oil path travel comprises:

7. The visualization-based oil online equipment failure analysis method of claim 6, wherein the overlay assembly is determined by:

8. The visualization-based oil online equipment failure analysis method of claim 6, wherein the overlay assembly is determined by:

9. Oil on-line equipment fault analysis system based on visualization, which is characterized by comprising: