Method and device for fault diagnosis and early warning of air compressor

Abstract

The application discloses a fault diagnosis early warning method and device for an air compressor, and relates to the technical field of fault diagnosis. The method comprises the following steps: acquiring a cutting head of a laser cutting device; dynamically monitoring a laser cutting focusing mirror to obtain focusing mirror monitoring information; dynamically monitoring a capacitance connection induction stylus to obtain a capacitance value time sequence; analyzing the focusing mirror monitoring information and the capacitance value time sequence to obtain a real-time cutting head index; determining whether the real-time cutting head index meets a predetermined index threshold; if not, issuing an early warning instruction, and performing fault early warning on the target air compressor based on the early warning instruction. The application solves the technical problem that, in the prior art, when an air compressor is matched with a laser cutting device, the quality of compressed air cannot be effectively monitored, and cutting head failures occur frequently. The application achieves real-time monitoring and accurate fault early warning of the air compressor matched with the laser cutting device, improves equipment maintenance efficiency, and reduces the risk of failure.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault diagnosis, in particular to a fault diagnosis and early warning method and device for an air compressor. BACKGROUND

[0002] In the industrial manufacturing field where laser cutting technology is widely applied, the running stability of the laser cutting equipment is crucial, and the quality of the compressed air provided by the air compressor has a significant impact on it. The existing technology focuses more on the cutting function of the laser cutting equipment itself, and lacks effective monitoring and associated fault early warning of the compressed air quality of the air compressor, which causes impure and water-containing impurities in the compressed air to enter the cutting head, causing faults such as burning of the inner wall of the laser cutting focusing lens barrel, damage to the lens performance, and erosion and short circuit of the capacitor connection sensing probe, which not only reduces the cutting quality and equipment life, but also makes it difficult to quickly locate the root cause after the fault occurs, resulting in long equipment downtime, low production efficiency, high maintenance cost, and inability to prevent faults in advance, increasing the risk of equipment operation.

[0003] The existing technology has the technical problem that when the air compressor is matched with the laser cutting equipment, the quality of the compressed air cannot be effectively monitored, resulting in frequent failures of the cutting head. SUMMARY

[0004] The present application provides a fault diagnosis and early warning method and device for an air compressor, which is used to solve the technical problem that the quality of the compressed air cannot be effectively monitored when the air compressor is matched with the laser cutting equipment, resulting in frequent failures of the cutting head in the prior art.

[0005] In view of the above problems, the present application provides a fault diagnosis and early warning method and device for an air compressor.

[0006] In a first aspect of the present application, a fault diagnosis and early warning method for an air compressor is provided, which comprises:

[0007] acquiring a cutting head of the laser cutting equipment, the cutting head comprising a laser cutting focusing lens and a capacitor connection sensing probe; dynamically monitoring the laser cutting focusing lens by an optical analyzer to obtain focusing lens monitoring information; dynamically monitoring the capacitor connection sensing probe by an LCR meter to obtain a capacitor value time sequence; calling a cutting head state evaluation model to analyze the focusing lens monitoring information and the capacitor value time sequence to obtain a real-time cutting head index; determining whether the real-time cutting head index meets a predetermined index threshold; if not, issuing a warning instruction, and based on the warning instruction, performing fault early warning on the target air compressor.

[0008] In a second aspect of the present application, a fault diagnosis and early warning device for an air compressor is provided, which comprises:

[0009] The cutting head acquisition module is configured to acquire a cutting head of the laser cutting device, the cutting head comprising a laser cutting focusing mirror and a capacitance connection sensing stylus; the focusing mirror monitoring information acquisition module is configured to dynamically monitor the laser cutting focusing mirror by using an optical analyzer to obtain focusing mirror monitoring information; the capacitance value time sequence acquisition module is configured to dynamically monitor the capacitance connection sensing stylus by using an LCR meter to obtain a capacitance value time sequence; the real-time cutting head index acquisition module is configured to call a cutting head state evaluation model to analyze the focusing mirror monitoring information and the capacitance value time sequence to obtain a real-time cutting head index; the predetermined index threshold judgment module is configured to judge whether the real-time cutting head index meets a predetermined index threshold; and the fault early warning module is configured to issue a warning instruction if the real-time cutting head index does not meet the predetermined index threshold, and perform fault early warning on the target air compressor based on the warning instruction.

[0010] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0011] The cutting head acquisition module is configured to acquire a cutting head of the laser cutting device, the cutting head comprising a laser cutting focusing mirror and a capacitance connection sensing stylus; the focusing mirror monitoring information acquisition module is configured to dynamically monitor the laser cutting focusing mirror by using an optical analyzer to obtain focusing mirror monitoring information; the capacitance value time sequence acquisition module is configured to dynamically monitor the capacitance connection sensing stylus by using an LCR meter to obtain a capacitance value time sequence; the real-time cutting head index acquisition module is configured to call a cutting head state evaluation model to analyze the focusing mirror monitoring information and the capacitance value time sequence to obtain a real-time cutting head index; the predetermined index threshold judgment module is configured to judge whether the real-time cutting head index meets a predetermined index threshold; and the fault early warning module is configured to issue a warning instruction if the real-time cutting head index does not meet the predetermined index threshold, and perform fault early warning on the target air compressor based on the warning instruction. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0013] Figure 1 The flowchart of the fault diagnosis and early warning method for the air compressor provided by the embodiments of the present application is shown.

[0014] Figure 2 The structural diagram of the fault diagnosis and early warning device for the air compressor provided by the embodiments of the present application is shown.

[0015] Cutting head acquisition module 10, focusing mirror monitoring information acquisition module 20, capacitance value time sequence acquisition module 30, real-time cutting head index acquisition module 40, predetermined index threshold judgment module 50, fault early warning module 60. DETAILED DESCRIPTION

[0016] The present application provides a fault diagnosis and early warning method and device for an air compressor, which is used to solve the technical problem that the quality of compressed air cannot be effectively monitored when the air compressor is matched with a laser cutting device, resulting in frequent cutting head failures.

[0017] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0018] Embodiment one, as shown, the present application provides a fault diagnosis and early warning method for an air compressor, the method comprising: Figure 1

[0019] Step S100: acquiring a cutting head of the laser cutting device, the cutting head comprising a laser cutting focusing mirror and a capacitance connection sensing stylus.

[0020] Specifically, in the actual scenario of laser cutting device operation, it is necessary to accurately acquire the cutting head to carry out subsequent fault diagnosis and early warning work. As a key component of the laser cutting device, the laser cutting focusing mirror in the cutting head plays an important role in accurately focusing the laser beam, and its performance state directly affects the cutting precision and quality. The capacitance connection sensing stylus is responsible for sensitively sensing the change of relevant parameters during cutting and providing important feedback information. By using professional tools and operation processes, the cutting head is removed from the laser cutting device, ensuring that no damage is caused during the acquisition process, thereby laying a foundation for subsequent detailed monitoring and analysis of the components of the cutting head and starting the first step of air compressor fault diagnosis and early warning.

[0021] Step S200: dynamically monitoring the laser cutting focusing mirror by using an optical analyzer to obtain focusing mirror monitoring information.

[0022] ​Specifically, when dynamically monitoring the laser cutting focusing mirror, first, the light analyzer is effectively connected with the laser cutting focusing mirror. As an advanced optical detection equipment, the light analyzer has the characteristics of high sensitivity and high precision. During the monitoring process, it emits a light beam of specific frequency and intensity to the focusing mirror, and the light beam reflects and refracts on the surface of the focusing mirror. The light analyzer captures various characteristics of the reflected and refracted light in real time through its internal precise optical sensors, including light intensity, phase, polarization state, etc. These information forms a series of data sequences over time, thus constituting the focusing mirror monitoring information. For example, changes in light intensity reflect the cleanliness of the focusing mirror surface. If dust or oil stains adhere, the light intensity will fluctuate accordingly; changes in phase and polarization state may indicate whether there is stress or damage in the internal structure of the focusing mirror. The entire monitoring process is continuously carried out under the normal operation state of the laser cutting equipment to obtain comprehensive and accurate focusing mirror state data, providing reliable basis for subsequent analysis and fault diagnosis.

[0023] Step S300: dynamically monitoring the capacitance connection sensing stylus by LCR table to obtain capacitance value time sequence.

[0024] Specifically, when dynamically monitoring the capacitance connection sensing stylus, first, the high-precision LCR table is reliably connected with the capacitance connection sensing stylus through special wires to ensure the stability and accuracy of signal transmission. As a professional electrical parameter measuring instrument, the LCR table can generate accurate test signals and measure the capacitance characteristics of the capacitance connection sensing stylus in real time. During the operation of the laser cutting equipment, as the cutting work progresses and the surrounding environment changes, the capacitance value of the capacitance connection sensing stylus will change accordingly. The LCR table continuously collects these capacitance value data at a very high sampling frequency, thus forming a capacitance value time sequence that changes continuously over time. This capacitance value time sequence can accurately reflect the capacitance state of the capacitance connection sensing stylus at different times, providing an important data basis for subsequent analysis of its working state and judgment of whether there are hidden faults, such as whether the capacitance connection sensing stylus is disturbed by external interference or its performance is attenuated.

[0025] Step S400: calling the cutting head state evaluation model to analyze the focusing mirror monitoring information and the capacitance value time sequence to obtain real-time cutting head index.

[0026] Specifically, a pre-constructed and well-trained cutting head state evaluation model is started, which integrates advanced algorithms and a large amount of experimental data, and has strong data analysis capability. First, the focusing mirror monitoring information (including mirror barrel reflectivity time sequence and lens light signal time sequence) obtained by the light analyzer and the capacitance value time sequence obtained by the LCR table are accurately input into the model. The focusing mirror evaluation unit in the model will carry out in-depth analysis on the focusing mirror monitoring information according to the built-in predetermined reflectivity reference and predetermined light signal reference, compare the differences between the data in the reflectivity time sequence and the light signal time sequence and the reference values, and through complex calculation and logical judgment, obtain the real-time focusing mirror state feedback coefficient, which can quantitatively reflect the current state of the focusing mirror. At the same time, the stylus evaluation unit will comprehensively analyze the capacitance value time sequence, extract key features from the time sequence data, and then calculate the real-time stylus state feedback coefficient to represent the real-time working state of the capacitance connection sensing stylus. Finally, the model will perform weighted and normalized processing on the two feedback coefficients, comprehensively consider the influence weight of the focusing mirror and the stylus on the overall state of the cutting head, and finally output the real-time cutting head index through a series of precise operations. The index can intuitively present the current comprehensive performance state of the cutting head, providing a key basis for subsequent judgment.

[0027] Step S500: judging whether the real-time cutting head index meets a predetermined index threshold.

[0028] Specifically, the accurate judgment of the real-time cutting head index and the predetermined index threshold is the key decision point of the entire air compressor fault diagnosis and early warning process. The predetermined index threshold is a standard threshold range or a specific value determined by in-depth analysis of a large amount of historical data and comprehensive statistics of the parameters of the cutting head corresponding to the laser cutting equipment in the normal running state. It comprehensively reflects the normal running indicators of the cutting head in the ideal working environment. After obtaining the cutting head index calculated in real time, it is quickly compared with the predetermined index threshold. If the real-time cutting head index is within the reasonable range specified by the predetermined index threshold, it indicates that the current working state of the cutting head is good, and the cooperative operation between the laser cutting equipment and the target air compressor is basically normal, and no fault warning intervention is required for the time being. However, if the real-time cutting head index exceeds the range of the predetermined index threshold, whether it is higher than the upper limit or lower than the lower limit, it means that the working state of the cutting head has abnormally fluctuated, which is caused by the quality problems of the compressed air provided by the target air compressor, such as impurity or dryness. At this time, the subsequent fault warning process will be further triggered according to this judgment result, so as to timely investigate the root cause of the problem and take corresponding corrective measures, so as to ensure the stable and efficient operation of the entire laser cutting system, and avoid serious consequences such as equipment damage or production interruption caused by the continuous existence of potential faults.

[0029] Step S600: If not satisfied, issue a warning instruction, and based on the warning instruction, the target air compressor is warned of failure.

[0030] Specifically, when it is determined that the real-time cutting head index does not satisfy the predetermined index threshold, the warning mechanism is triggered immediately. First, a warning instruction is issued quickly, which contains rich information such as the abnormal real-time cutting head index value, the current monitoring time, and the preliminary judgment of the possible fault type. Based on the warning instruction, the fault warning system will warn the target air compressor of failure in multiple intuitive ways. On the operation interface, a prominent warning pop-up window will pop up, prompting the operator with clear text and bright colors that the air compressor may have a fault risk, and detailed monitoring data will be displayed for reference. At the same time, an audible and visual alarm will be issued through a high-decibel sound signal and flashing lights. In addition, the warning information will also be recorded in the system log for subsequent tracing and analysis of the fault cause. This series of warning actions can timely remind the target air compressor to be checked, maintained or take appropriate emergency measures, effectively prevent the air compressor failure from further deterioration, and ensure the continuity and safety of the production process.

[0031] In one possible implementation manner, step S200 further includes:

[0032] Step S210: The optical analyzer includes a first analyzer and a second analyzer, wherein the first analyzer is arranged on the lens barrel of the laser cutting focusing mirror, and the second analyzer is arranged on the lens of the laser cutting focusing mirror.

[0033] Step S220: The reflectivity time sequence of the lens barrel is acquired by the first analyzer.

[0034] Step S230: The light signal time sequence of the lens is acquired by the second analyzer.

[0035] Step S240: The reflectivity time sequence and the light signal time sequence constitute the focusing mirror monitoring information.

[0036] Specifically, in the laser cutting equipment operating environment, the optical analyzer for monitoring the laser cutting focusing mirror is composed of two key parts, namely the first analyzer and the second analyzer. The first analyzer is precisely installed on the lens barrel of the laser cutting focusing mirror, and this arrangement enables it to be in close contact with the lens barrel, directly acquiring information related to the optical properties of the lens barrel surface. The second analyzer is placed in the lens part of the laser cutting focusing mirror according to its detection requirements, and is in the same optical path environment as the lens, which can effectively capture various optical phenomena exhibited by the lens during laser transmission, laying a foundation for subsequent accurate analysis of the lens state.

[0037] When the first analyzer is in operation, it actively emits a precisely controlled beam of detection light towards the lens barrel surface. When this beam of light contacts the lens barrel, a portion of the light is reflected back. The first analyzer is equipped with a highly sensitive light receiving element inside, which can accurately perceive the intensity of the reflected light. As time goes on, the first analyzer continuously performs light emission and reflected light intensity detection operations at certain time intervals during the continuous operation of the laser cutting equipment. Each detected reflected light intensity data is recorded by the system and arranged in chronological order, which gradually forms a complete lens barrel reflectivity time sequence. This time sequence data is like a dynamic record of the optical state of the lens barrel surface, which can reflect the reflectivity changes of the lens barrel at different times, for example, when the lens barrel surface is slightly contaminated or worn, the reflectivity time sequence may fluctuate accordingly.

[0038] The second analyzer is in a highly active state during the laser cutting process. When the lens passes through the laser beam, it will have multiple optical effects on the laser, such as transmission, reflection, and scattering, and its own optical properties will also affect the light signal, such as changes in phase, polarization state, and other characteristics. The second analyzer, with its advanced optical sensing technology, can capture these light signals containing rich lens state information in real time. After capturing the light signal, the second analyzer quickly performs a series of complex signal processing operations on it, including signal amplification, filtering, and analog-to-digital conversion, etc., to convert the light signal into a digital signal form. Similarly, as time goes on, these processed light signal data are recorded in chronological order, forming a lens light signal time sequence that accurately reflects the optical state changes of the lens at different time points. For example, when the lens has a slight scratch or coating damage, certain characteristic parameters in the light signal time sequence may change significantly.

[0039] When the lens barrel reflectivity time sequence obtained by the first analyzer and the lens light signal time sequence obtained by the second analyzer are both complete, the two time sequence data are integrated. The reflectivity time sequence provides information from the perspective of changes in the optical properties of the lens barrel surface, while the light signal time sequence provides supplementary information from the perspective of changes in the optical performance of the lens. Together, they form the focusing mirror monitoring information. This comprehensive information provides a detailed description of the state changes of the laser cutting focusing mirror during the entire laser cutting process, providing essential data for subsequent in-depth analysis of the health of the focusing mirror, prediction of potential failures, and optimization of the laser cutting process. It helps technicians to timely grasp the working state of the focusing mirror and ensures the efficient and stable operation of the laser cutting equipment.

[0040] In one possible implementation, step S400 further includes:

[0041] Step S410: Obtain the focusing mirror evaluation unit in the cutting head state evaluation model, wherein the focusing mirror evaluation unit stores a predetermined reflectivity reference and a predetermined optical signal reference.

[0042] Step S420: The focusing mirror evaluation unit combines the predetermined reflectivity reference and the predetermined optical signal reference to analyze the reflectivity time sequence and the optical signal time sequence to obtain a real-time focusing mirror state feedback coefficient.

[0043] Step S430: Analyze the capacitance value time sequence through the stylus evaluation unit in the cutting head state evaluation model to obtain a real-time stylus state feedback coefficient.

[0044] Step S440: Weight and normalize the real-time focusing mirror state feedback coefficient and the real-time stylus state feedback coefficient to obtain the real-time cutting head index.

[0045] Specifically, in the key stage of cutting head state evaluation, first, the focusing mirror evaluation unit is accurately positioned and obtained from the cutting head state evaluation model that has been constructed. This focusing mirror evaluation unit is an important module in the model for analyzing the state of the laser cutting focusing mirror, and it stores extremely critical predetermined reflectivity reference and predetermined optical signal reference inside. These reference data are determined through a large number of experimental tests, actual operation data statistics, and accurate research on the ideal working state of the laser cutting focusing mirror. The predetermined reflectivity reference specifies the reflectivity range and variation law that the laser cutting focusing mirror barrel should have under normal working conditions, and the predetermined optical signal reference specifies the standard characteristics of the optical signal generated or transmitted by the lens, such as light intensity, phase, polarization state, etc. The expected value range under normal state. These reference data provide an important reference standard for subsequent accurate evaluation of the focusing mirror.

[0046] When the reflectivity time series and the light signal time series are analyzed by the focusing mirror evaluation unit to obtain the real-time focusing mirror state feedback coefficient, the process is as follows: first, any reflectivity corresponding to any time point in the reflectivity time series is randomly selected, and when the reflectivity is within the predetermined reflectivity reference range, the evaluation unit immediately issues a first trend prediction instruction. Subsequently, according to the instruction, the reflectivity time series is analyzed for trends, such as using a least squares method or other fitting method to calculate the first trend slope of the reflectivity with respect to time. This slope can intuitively reflect the change trend of the barrel reflectivity, for example, a positive and gradually increasing slope may indicate that the barrel surface condition is gradually deteriorating. Next, any light signal corresponding to any time is extracted from the light signal time series, and when the light signal meets the predetermined light signal reference, a second trend prediction instruction is triggered. Then, the light signal time series is analyzed for trends in a similar manner, and the second trend slope of the light signal is obtained, which can reflect the change trend of the lens optical performance. Finally, the focusing mirror evaluation unit performs a variation weighting calculation on the first trend slope and the second trend slope, that is, different weights are assigned according to the importance of the two slopes on the overall state of the focusing mirror, and the variation factors such as the change amplitude of the slope are also considered. After comprehensive calculation, the real-time focusing mirror state feedback coefficient that can accurately represent the current state of the focusing mirror is obtained, which provides key data support for subsequent comprehensive evaluation of the state of the cutting head.

[0047] When the capacitance value time sequence is obtained, it is first pre-processed to check the integrity and accuracy of the data, and to remove possible outliers or noise interference. Then, various features of the capacitance value time sequence are extracted. The average value of the capacitance value is calculated, which can reflect the average level of the capacitance connection sensing probe in a period of time; then the variance and standard deviation are calculated to measure the dispersion degree of the capacitance value, that is, the fluctuation size, and larger fluctuation may indicate that the probe state is unstable or affected by external factors; the trend feature of the capacitance value is calculated, such as the slope of the capacitance value with respect to time obtained by linear regression, and if the slope is abnormal, it may indicate that the probe performance changes. At the same time, the periodicity feature of the capacitance value can be analyzed to see if there is a certain period of fluctuation, which helps to determine whether the probe is subject to periodic interference. These average values, variances, trend features, periodicity features, etc. are combined to form a feature set of the capacitance value time sequence. The probe state prediction model is trained in advance by a large amount of historical capacitance connection sensing probe data, which contains capacitance value time sequences in different states and corresponding probe actual state feedback coefficients. In the training model, a neural network algorithm is used, the input layer receives the feature data of the capacitance value time sequence, and after complex calculation and weight adjustment of the hidden layer, the output layer predicts the probe state feedback coefficient. When the current extracted capacitance value time sequence feature set is input into the trained probe state prediction model, the model will evaluate the current state of the capacitance connection sensing probe according to the patterns and rules learned inside, and output a predicted probe state feedback coefficient. This coefficient can quantitatively reflect the current working state of the probe, and compared with the normal or abnormal state patterns learned during the model training process, it provides a key basis for judging the overall performance of the cutting head, so as to calculate the real-time cutting head index together with the focus lens state feedback coefficient, and realize the comprehensive evaluation of the cutting head state.

[0048] After obtaining the real-time focus lens state feedback coefficient and the real-time probe state feedback coefficient respectively, in order to comprehensively evaluate the overall state of the cutting head, the two coefficients need to be weighted and normalized. The weighting process is to allocate the calculation proportion according to the importance weight of the focus lens and the probe in the overall function of the cutting head. For example, if the performance of the focus lens has a greater impact on the cutting quality and the stability of the equipment, then the proportion of its state feedback coefficient in the weighted calculation will be relatively high. The normalization process is to standardize the weighted coefficients to a unified numerical range, which is convenient for intuitive comparison and comprehensive evaluation, and finally obtains the real-time cutting head index. This real-time cutting head index is a quantitative index that can comprehensively and comprehensively reflect the current performance state of the cutting head, which provides the most critical basis for subsequent judgment of whether the cutting head is working normally and whether the target air compressor needs to be warned of failure. Technical personnel can make accurate decisions quickly according to this index, and ensure the stable operation of the entire laser cutting system.

[0049] In one possible implementation, step S420 further comprises:

[0050] Step S421: extracting any reflectivity corresponding to any time in the reflectivity time sequence.

[0051] Step S422: issuing a first trend prediction instruction when the any reflectivity meets the predetermined reflectivity criterion.

[0052] Step S423: performing trend analysis on the reflectivity time sequence according to the first trend prediction instruction to obtain a first trend slope.

[0053] Step S424: extracting any optical signal corresponding to any time in the optical signal time sequence.

[0054] Step S425: issuing a second trend prediction instruction when the any optical signal meets the predetermined optical signal criterion.

[0055] Step S426: performing trend analysis on the optical signal time sequence according to the second trend prediction instruction to obtain a second trend slope.

[0056] Step S427: the focusing mirror evaluation unit performs a variation weighting calculation on the first trend slope and the second trend slope to obtain the real-time focusing mirror state feedback coefficient.

[0057] Specifically, the focusing mirror evaluation unit first extracts any reflectivity value corresponding to any time point in the reflectivity time sequence data. This reflectivity value is a sample point in the entire reflectivity time sequence, representing the reflection ability of the lens barrel at that particular time.

[0058] Then, the extracted reflectivity value is compared with the predetermined reflectivity criterion pre-stored in the focusing mirror evaluation unit. The predetermined reflectivity criterion is a numerical range or standard value determined through a large number of experiments and statistical analysis of the reflectivity of the focusing mirror lens barrel in the normal working state. When the extracted any reflectivity is within this predetermined reflectivity criterion range, it indicates that the reflectivity of the lens barrel at this time meets the expected normal state, and the evaluation unit then issues a first trend prediction instruction.

[0059] Upon receiving the first trend prediction instruction, the evaluation unit begins to perform trend analysis on the entire reflectivity time series. Taking time as the independent variable and reflectivity value as the dependent variable, a straight line that best fits these data points is calculated by algorithms such as least squares. The slope of this straight line is the first trend slope. The first trend slope can reflect the trend of reflectivity over time. If the slope is positive and the value is large, it means that the optical properties of the lens barrel surface are gradually changing, for example, there are pollutants gradually accumulating to cause the reflectivity to rise; if the slope is negative, it means that the lens barrel surface is worn or the like, causing the reflectivity to decrease.

[0060] Similar to the reflectivity analysis, the focusing lens evaluation unit extracts any light signal corresponding to any time from the light signal time series. This light signal contains comprehensive information of the transmission, reflection and scattering of the lens at that moment. The extracted light signal is compared with the predetermined light signal reference, which is also a standard range or value determined based on in-depth research on the light signal characteristics of the lens in the normal working state. When the light signal meets the predetermined light signal reference, the second trend prediction instruction is triggered.

[0061] According to the second trend prediction instruction, the evaluation unit processes the light signal time series using similar trend analysis methods. For example, for the change of light intensity in the light signal over time, a trend curve of light intensity change is obtained by constructing a function relationship between light intensity and time, and using a suitable fitting algorithm, and then the slope of the curve, i.e. the second trend slope, is calculated. The second trend slope can reflect the trend of the optical performance of the lens over time, such as the abnormality of the light intensity change trend slope may indicate damage to the lens coating or changes in the internal structure of the lens.

[0062] After obtaining the first trend slope and the second trend slope, the focusing lens evaluation unit performs variation weighting calculation. The variation weighting takes into account factors such as the stability and change amplitude of the two slopes. If the first trend slope fluctuates greatly within a period of time, i.e. the variation degree is high, its weight in the weighting calculation may be adjusted accordingly to highlight its importance to the focusing lens state evaluation. At the same time, according to the relative importance of the focusing lens barrel reflectivity and the lens light signal in the overall performance of the focusing lens, different initial weights are assigned to the first trend slope and the second trend slope. For example, if the optical performance of the lens has a greater impact on the cutting accuracy, the initial weight of the second trend slope may be set higher. By multiplying the variation-adjusted weight with the two slopes and summing them up, the real-time focusing lens state feedback coefficient is finally obtained. This coefficient integrates the state information of the lens barrel and the lens, and can more accurately reflect the current overall working state of the focusing lens, providing a key basis for subsequent cutting head state evaluation.

[0063] In one possible implementation, step S423 further comprises:

[0064] Step 4231: plot a reflectivity scatter plot according to the reflectivity time sequence.

[0065] Step 4232: perform spline fitting on the reflectivity scatter plot to obtain a reflectivity fitted spline curve.

[0066] Step 4233: obtain a real-time slope based on the reflectivity fitted spline curve, and record the real-time slope as the first trend slope.

[0067] Step 4234: wherein performing spline fitting on the reflectivity scatter plot to obtain a reflectivity fitted spline curve comprises:

[0068] extracting a first scatter point set in the reflectivity scatter plot; performing polynomial fitting on the first scatter point set to obtain a first polynomial, and the first polynomial corresponds to a first spline curve; obtaining a first check scatter point set by combining the reflectivity scatter plot and the first scatter point set; performing check analysis on the first spline curve according to the first check scatter point set to obtain a first check result; if the first check result satisfies a predetermined check constraint, taking the first spline curve as the reflectivity fitted spline curve.

[0069] Specifically, when performing trend analysis on the reflectivity time sequence, first, a reflectivity scatter plot is plotted according to the obtained reflectivity time sequence. Taking time as the horizontal coordinate and reflectivity as the vertical coordinate, each data point in the reflectivity time sequence is marked in the plane rectangular coordinate system. For example, a point is marked at the coordinate (t1, r1) if the reflectivity at the time point t1 is r1, and so on. All data points in the reflectivity time sequence are marked in this way, and thus the reflectivity scatter plot is formed. These scatter points intuitively show the distribution of reflectivity changes over time, providing a visual basis for subsequent fitting analysis, and enabling preliminary observation of the general trend of reflectivity data, such as whether there is a linear trend, the size of fluctuation range, etc.

[0070] The first set of scatter points is extracted from the reflectivity scatter plot. The extraction method can be according to certain rules, such as equal interval selection, random selection of a certain number of points, etc., to ensure that the selected first set of scatter points can represent the characteristics of the entire scatter plot to some extent. Then the first set of scatter points is fitted with a polynomial, and the coefficients of the polynomial are determined through mathematical calculation to obtain the first polynomial. The curve corresponding to the first polynomial is the first spline curve. The purpose of polynomial fitting is to find a relatively smooth curve that can better approximate the trend of scatter point distribution. The degree of the polynomial will affect the fitting effect and the complexity of the curve, and needs to be balanced according to the characteristics of the data and actual needs. Generally, a polynomial of lower degree is selected to avoid overfitting the data and ensure that the curve can reflect the data trend and have a certain universality.

[0071] After obtaining the first spline curve, it needs to be verified to determine whether it can accurately fit the reflectivity time series data. First, the first verification set of scatter points is obtained by combining the reflectivity scatter plot and the first set of scatter points, that is, the points in the first set of scatter points are removed from the entire reflectivity scatter plot, and the remaining points constitute the first verification set of scatter points. Then, the first spline curve is verified and analyzed according to the first verification set of scatter points, and the distance or deviation of each point in the first verification set of scatter points to the first spline curve is calculated. The comprehensive situation of these distances or deviation values constitutes the first verification result. For example, the average value, variance, etc. of the distances of all verification scatter points to the curve are calculated to measure the fitting effect. If the first verification result meets the predetermined verification constraints, such as the average distance being less than the pre-set threshold, the variance being within an acceptable range, etc., it indicates that the first spline curve can better fit the entire reflectivity time series data. At this time, the first spline curve is taken as the reflectivity fitting spline curve, which is used for subsequent slope calculation and trend analysis. If the verification constraints are not met, the selection of the first set of scatter points or the parameters of the polynomial fitting may need to be adjusted, and the fitting and verification operations are performed again until the reflectivity fitting spline curve that meets the requirements is obtained.

[0072] After the reflectivity fitting spline curve is determined, a real-time slope is calculated based on the curve. The slope of each point on the reflectivity fitting spline curve is calculated by a mathematical method, such as a derivative operation. At the time point of the current analysis, the obtained slope value is the real-time slope, which is recorded as the first trend slope. The first trend slope can accurately reflect the change rate and trend direction of the reflectivity at the current time, and is of great significance for evaluating the state change of the focusing mirror barrel. If the first trend slope is positive and gradually increases, it means that the reflectivity of the barrel surface is continuously rising due to the deposition of substances on the barrel surface or other reasons. If the slope is negative, it means that the reflectivity of the barrel surface decreases due to wear or other factors, thereby helping technicians to understand the optical performance change of the focusing mirror barrel in a timely manner and providing key data basis for further judging the overall state of the focusing mirror.

[0073] In one possible implementation manner, step S600 further includes:

[0074] Step S610: determining whether the real-time focusing mirror state feedback coefficient meets a first threshold.

[0075] Step S620: if not, performing a pre-warning of a compressed air impurity fault of the target air compressor.

[0076] Specifically, in the air compressor fault diagnosis and pre-warning process, after the real-time focusing mirror state feedback coefficient is obtained, a judgment is immediately performed. At this time, the real-time focusing mirror state feedback coefficient is compared with a first threshold. The first threshold is a numerical range or a specific value determined based on a large amount of experimental data and an accurate study on the state of the focusing mirror during normal operation of the laser cutting equipment, and is an important standard for measuring whether the working state of the focusing mirror is normal, reflecting a reasonable interval in which the state feedback coefficient of the focusing mirror should be in under ideal working conditions. By comparing the real-time focusing mirror state feedback coefficient calculated with the first threshold, it can be preliminarily determined whether the focusing mirror is in a normal working state, thereby providing a basis for subsequent pre-warning of possible faults of the air compressor.

[0077] If it is judged that the real-time focusing mirror state feedback coefficient does not meet the first threshold, which indicates that the working state of the focusing mirror is abnormal, it is caused by impure compressed air. At this time, the target air compressor is warned of the impure compressed air fault, and the warning method is various and targeted. For example, a prominent warning pop-up window is popped up on the operation interface, and the words "impure compressed air fault warning" are clearly displayed in the pop-up window, and are marked with bright colors such as red to attract the attention of the operator. At the same time, a sound and light alarm of a specific frequency and volume, such as a high-pitched buzzing sound and a flashing red light, is emitted to ensure that the operator can detect it in time even if he is not near the operation interface. In addition, the warning information is also recorded in the system log in detail, including the warning time, the specific value of the focusing mirror state feedback coefficient, and the like, so as to provide a reference for subsequent technical personnel to troubleshoot and analyze, help to quickly locate the problem source, and take appropriate measures, such as checking the air filter system of the air compressor and replacing the filter element, so as to ensure the stable operation of the laser cutting equipment and avoid further damage to the equipment or affect the cutting quality due to the impure compressed air problem.

[0078] In a possible implementation manner, step S600 further includes:

[0079] Step S630: If yes, judging whether the real-time stylus state feedback coefficient meets a second threshold.

[0080] Step S640: If no, warning the target air compressor of a compressed air not dry fault.

[0081] Specifically, when it is determined that the real-time focusing mirror state feedback coefficient meets the first threshold, the system will focus on the evaluation of the real-time stylus state feedback coefficient, and compare it with the second threshold set in advance. The second threshold is also a key numerical limit determined through a large number of experiments and in-depth research on the state of the capacitive connection sensing stylus of the laser cutting equipment under normal working conditions. It can represent the reasonable range of the state feedback coefficient of the stylus under normal working conditions, and is an important basis for judging whether the stylus is working normally. Through this comparison operation, it can be further determined whether the overall state of the cutting head is good, and a more detailed basis is provided for the comprehensive evaluation of the equipment running condition, because the stylus state also plays a key role in the normal operation of the laser cutting equipment, and the change of the state feedback coefficient of the stylus can reflect whether the stylus is affected by adverse factors such as moisture.

[0082] If it is judged that the real-time stylus state feedback coefficient does not meet the second threshold, it means that the stylus state is abnormal, which is caused by the non-drying of compressed air. Therefore, the target air compressor is warned of the non-drying fault of compressed air. The warning mechanism is comprehensive and effective, and a clear and eye-catching warning prompt will be displayed on the operation interface, such as displaying "compressed air non-drying fault warning" in yellow background with black bold font, to ensure that the operator can quickly notice. At the same time, a continuous and rhythmic alarm sound, such as a short beep, is emitted, and a flashing yellow warning light is turned on, reminding the operator from multiple aspects of hearing and vision. In addition, detailed information of this warning is recorded in the system log, including the exact time of the warning, the real-time stylus state feedback coefficient value, and related equipment parameters. These records help technicians analyze the fault cause in depth later, take targeted measures, such as checking whether the drying equipment of the air compressor is running normally and whether the drying agent needs to be replaced, so as to timely eliminate the fault hidden danger, ensure the stable and efficient operation of the laser cutting equipment and the matching air compressor, and ensure that the production process is not affected.

[0083] In one possible implementation manner, step S430 further includes:

[0084] Step S431: collecting multi-domain features of the capacitance value time sequence to obtain a capacitance value multi-domain feature set.

[0085] Step S432: calling a stylus state prediction model to perform prediction analysis on the capacitance value multi-domain feature set to obtain a predicted stylus state feedback coefficient.

[0086] Step S433: recording the predicted stylus state feedback coefficient as the real-time stylus state feedback coefficient.

[0087] Step S434, wherein calling the stylus state prediction model to perform prediction analysis on the capacitance value multi-domain feature set to obtain the predicted stylus state feedback coefficient includes:

[0088] obtaining a historical capacitance-connected sensing stylus log; assembling training data sets according to a first log, wherein the first log refers to any one log record in the historical capacitance-connected sensing stylus log, and the first log includes a first capacitance value multi-domain feature set of a first capacitance value time sequence and a first stylus state feedback coefficient; and performing machine learning on the first capacitance value multi-domain feature set and the first stylus state feedback coefficient in the training data set to obtain the stylus state prediction model.

[0089] Specifically, in the process of analyzing the capacitance connection sensing stylus state to obtain real-time stylus state feedback coefficient, first, the mean value of the capacitance value is calculated, and all capacitance values in the capacitance value time sequence are added and divided by the number of data points to reflect the average capacitance level of the capacitance connection sensing stylus in a period of time. Then, the standard deviation of the capacitance value is determined, which measures the dispersion degree of the capacitance value by calculating the square root of the average of the square sum of the deviation of each capacitance value from the mean value. The larger the standard deviation, the more violent the fluctuation of the capacitance value, and the more unstable the stylus working state may be. Then, the maximum and minimum values of the capacitance value are determined, and the difference between them constitutes the range, which directly presents the change range of the capacitance value in the time sequence. In addition, the autocorrelation coefficient of the capacitance value is also calculated to observe the correlation between the capacitance values at different time lags and to determine whether the capacitance value change has a periodic rule or a specific trend. By integrating the mean value, standard deviation, range, autocorrelation coefficient and other characteristic data obtained by these calculations, a comprehensive capacitance value multi-domain feature set that can reflect the time sequence characteristics of the capacitance value from multiple dimensions is finally formed, providing a rich and valuable data basis for subsequent accurate stylus state evaluation.

[0090] The stylus state prediction model is retrieved, which is pre-constructed and fully trained based on a large amount of historical capacitance connection sensing stylus data, and contains rich association rules between capacitance value features and stylus state. The capacitance value multi-domain feature set collected in the previous step is accurately input into the stylus state prediction model, which will perform in-depth analysis and comprehensive consideration on each feature parameter in the feature set, such as mean value, standard deviation, range, autocorrelation coefficient, etc. According to the historical data patterns and internal logical relationships learned by it, through a series of complex mathematical operations and data processing processes, the current state of the capacitance connection sensing stylus is simulated, evaluated and predicted. Finally, the model outputs a predicted stylus state feedback coefficient, which quantitatively represents the current possible working state of the stylus and reflects its similarity to historical normal or abnormal states, providing key basis and reference for subsequent comprehensive judgment of the overall performance of the cutting head and the potential fault risk of the air compressor.

[0091] After the prediction analysis of the multi-domain feature set of the capacitance value by the stylus state prediction model is completed and the predicted stylus state feedback coefficient is successfully obtained, the predicted stylus state feedback coefficient is directly assigned to the name of the real-time stylus state feedback coefficient. This assignment operation is of great significance, enabling the prediction result obtained from the model to participate in the operation as the real-time stylus state feedback coefficient in the subsequent cutting head state comprehensive evaluation process. Together with the real-time focusing mirror state feedback coefficient, they will be included in the category of weighted normalization processing, and through the information about the stylus and focusing mirror states carried by each of them, they will jointly build a real-time cutting head index that can comprehensively and accurately reflect the current performance state of the cutting head, thereby providing indispensable key data support for judging whether the operation of the entire laser cutting equipment and the target air compressor is normal, and is a key link in the entire fault diagnosis and early warning chain.

[0092] In the process of calling the prediction analysis of the multi-domain feature set of the capacitance value by the stylus state prediction model to obtain the predicted stylus state feedback coefficient, first, the historical capacitance connection sensing stylus logs are obtained, and the first capacitance value multi-domain feature set of the first capacitance value time sequence and the first stylus state feedback coefficient are extracted from each first log. A large number of such first log data are integrated to construct a training data set. The decision tree algorithm constructs a decision tree model based on indicators such as information gain or Gini index. Starting from the root node, for each feature in the first capacitance value multi-domain feature set (such as capacitance mean, standard deviation, etc.), the information gain or Gini index is calculated, and the feature with the maximum information gain or minimum Gini index is selected as the splitting feature of the current node. For example, if the capacitance mean feature can most effectively distinguish samples corresponding to different stylus state feedback coefficients, then the capacitance mean is used as the splitting feature, and the training data set is divided into different subsets, each corresponding to a different value range of the capacitance mean. Then, for each subset, the above process is repeated to continue selecting the best splitting feature for division until the stopping condition is met, such as when the samples in the subset belong to the same class of stylus state feedback coefficient, or when the preset tree depth is reached. In this way, a decision tree model is constructed. When predicting a new multi-domain feature set of the capacitance value, start from the root node of the decision tree, and make judgments and move on the branches of the decision tree according to the feature values, finally reaching the leaf node. The stylus state feedback coefficient class corresponding to the leaf node is the prediction result, i.e., the predicted stylus state feedback coefficient. Different machine learning algorithms have their own characteristics and application scenarios, and in actual application, appropriate algorithms can be selected to construct the stylus state prediction model to achieve effective analysis and prediction of the multi-domain feature set of the capacitance value according to the nature of the data and the needs.

[0093] In the second embodiment, based on the same inventive concept as the fault diagnosis and early warning method for air compressors in the preceding embodiments, like Figure 2As shown, the application provides a fault diagnosis and early warning device for an air compressor, and the device and method embodiments in the application are based on the same inventive concept. The device comprises:

[0094] A cutting head acquisition module 10 is configured to acquire a cutting head of the laser cutting device, and the cutting head comprises a laser cutting focusing mirror and a capacitance connection sensing stylus.

[0095] A focusing mirror monitoring information acquisition module 20 is configured to dynamically monitor the laser cutting focusing mirror by an optical analyzer to obtain focusing mirror monitoring information.

[0096] A capacitance value time sequence acquisition module 30 is configured to dynamically monitor the capacitance connection sensing stylus by an LCR meter to obtain a capacitance value time sequence.

[0097] A real-time cutting head index acquisition module 40 is configured to call a cutting head state evaluation model to analyze the focusing mirror monitoring information and the capacitance value time sequence to obtain a real-time cutting head index.

[0098] A predetermined index threshold judgment module 50 is configured to judge whether the real-time cutting head index meets a predetermined index threshold.

[0099] A fault early warning module 60 is configured to issue a warning instruction if the predetermined index threshold is not met, and perform fault early warning on the target air compressor based on the warning instruction.

[0100] Further, the focusing mirror monitoring information module 20 further comprises:

[0101] An analyzer arrangement module is configured to arrange the optical analyzer to comprise a first analyzer and a second analyzer, wherein the first analyzer is arranged on a lens barrel of the laser cutting focusing mirror, and the second analyzer is arranged on a lens of the laser cutting focusing mirror.

[0102] A reflectivity time sequence acquisition module is configured to acquire a reflectivity time sequence of the lens barrel by the first analyzer.

[0103] An optical signal time sequence acquisition module is configured to acquire an optical signal time sequence of the lens by the second analyzer.

[0104] A focusing mirror monitoring information composition module is configured to compose the reflectivity time sequence and the optical signal time sequence to obtain the focusing mirror monitoring information.

[0105] Further, the real-time cutting head index acquisition module 40 further comprises:

[0106] A reference acquisition module is configured to acquire a focusing mirror evaluation unit in the cutting head state evaluation model, wherein the focusing mirror evaluation unit stores a predetermined reflectivity reference and a predetermined optical signal reference.

[0107] A time sequence analysis module is configured to analyze the reflectivity time sequence and the optical signal time sequence by the focusing mirror evaluation unit in combination with the predetermined reflectivity reference and the predetermined optical signal reference, to obtain a real-time focusing mirror state feedback coefficient.

[0108] A stylus state feedback coefficient acquisition module is configured to analyze the capacitance value time sequence by a stylus evaluation unit in the cutting head state evaluation model, to obtain a real-time stylus state feedback coefficient.

[0109] A weighted normalization module is configured to weight and normalize the real-time focusing mirror state feedback coefficient and the real-time stylus state feedback coefficient, to obtain the real-time cutting head index.

[0110] Further, the time sequence analysis module further comprises:

[0111] An arbitrary reflectivity extraction module is configured to extract an arbitrary reflectivity corresponding to an arbitrary time in the reflectivity time sequence.

[0112] A first trend prediction instruction issuing module is configured to issue a first trend prediction instruction when the arbitrary reflectivity meets the predetermined reflectivity reference.

[0113] A first trend slope acquisition module is configured to perform trend analysis on the reflectivity time sequence according to the first trend prediction instruction, to obtain a first trend slope.

[0114] An arbitrary optical signal extraction module is configured to extract an arbitrary optical signal corresponding to an arbitrary time in the optical signal time sequence.

[0115] A second trend prediction instruction issuing module is configured to issue a second trend prediction instruction when the arbitrary optical signal meets the predetermined optical signal reference.

[0116] A second trend slope acquisition module is configured to perform trend analysis on the optical signal time sequence according to the second trend prediction instruction, to obtain a second trend slope.

[0117] a variation weighting calculation module, configured to perform variation weighting calculation on the first trend slope and the second trend slope by the focusing mirror evaluation unit to obtain the real-time focusing mirror state feedback coefficient.

[0118] Further, the time sequence analysis module further comprises:

[0119] a reflectivity scatter plot drawing module, configured to draw a reflectivity scatter plot according to the reflectivity time sequence.

[0120] a reflectivity fitting spline curve acquisition module, configured to perform spline fitting on the reflectivity scatter plot to obtain a reflectivity fitting spline curve.

[0121] a real-time slope acquisition module, configured to obtain a real-time slope based on the reflectivity fitting spline curve, and record the real-time slope as the first trend slope.

[0122] wherein, the spline fitting on the reflectivity scatter plot to obtain the reflectivity fitting spline curve comprises:

[0123] a first scatter point set extraction module, configured to extract a first scatter point set in the reflectivity scatter plot.

[0124] a first polynomial acquisition module, configured to perform polynomial fitting on the first scatter point set to obtain a first polynomial, and the first polynomial corresponds to a first spline curve.

[0125] a first check scatter point set acquisition module, configured to obtain a first check scatter point set in combination of the reflectivity scatter plot and the first scatter point set.

[0126] a first check result acquisition module, configured to perform check analysis on the first spline curve according to the first check scatter point set to obtain a first check result.

[0127] a first check result judgment module, configured to take the first spline curve as the reflectivity fitting spline curve if the first check result meets a predetermined check constraint.

[0128] Further, the fault early warning module 60 further comprises:

[0129] a first threshold judgment module, configured to judge whether the real-time focusing mirror state feedback coefficient meets a first threshold.

[0130] The impurity early warning module is configured to perform early warning of a compressed air impurity fault of the target air compressor if the target air compressor does not meet the second threshold.

[0131] Further, the fault early warning module 60 further includes:

[0132] The second threshold judgment module is configured to judge whether the real-time stylus state feedback coefficient meets a second threshold if the target air compressor meets the first threshold.

[0133] The non-drying fault early warning module is configured to perform early warning of a compressed air non-drying fault of the target air compressor if the target air compressor does not meet the second threshold.

[0134] Further, the stylus state feedback coefficient acquisition module further includes:

[0135] The capacitance value multi-domain feature set acquisition module is configured to collect multi-domain features of the capacitance value time sequence to obtain a capacitance value multi-domain feature set.

[0136] The prediction analysis module is configured to invoke a stylus state prediction model to perform prediction analysis on the capacitance value multi-domain feature set to obtain a predicted stylus state feedback coefficient.

[0137] The feedback coefficient recording module is configured to record the predicted stylus state feedback coefficient as the real-time stylus state feedback coefficient.

[0138] The prediction analysis module is configured to invoke a stylus state prediction model to perform prediction analysis on the capacitance value multi-domain feature set to obtain a predicted stylus state feedback coefficient, including:

[0139] The historical capacitance-connected sensing stylus log acquisition module is configured to acquire a historical capacitance-connected sensing stylus log.

[0140] The training data group assembly module is configured to assemble a training data group according to a first log, wherein the first log refers to any one of the historical capacitance-connected sensing stylus logs, and the first log includes a first capacitance value multi-domain feature set of a first capacitance value time sequence and a first stylus state feedback coefficient.

[0141] The prediction model acquisition module is configured to perform machine learning on the first capacitance value multi-domain feature set and the first stylus state feedback coefficient in the training data group to obtain the stylus state prediction model.

[0142] It should be noted that the above-mentioned embodiment sequences of the present application are merely for description only, but not for representing the advantages and disadvantages of the embodiments. And the above-mentioned embodiments of the present specification have been described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0143] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0144] The specification and drawings are merely exemplary of the present application, and any and all modifications, variations, combinations or equivalents that are within the scope of the present application should be included. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A method for fault diagnosis and early warning for an air compressor, characterized in that, The fault diagnosis and early warning method for the air compressor is applied to a fault diagnosis and early warning device for the air compressor, which is in communication connection with a laser cutting equipment and is used with a target air compressor, and comprises the following steps: A cutting head of the laser cutting equipment is acquired, and the cutting head comprises a laser cutting focusing mirror and a capacitive connection sensing stylus; The laser cutting focusing mirror is dynamically monitored by an optical analyzer to obtain focusing mirror monitoring information; The capacitive connection sensing stylus is dynamically monitored by an LCR meter to obtain a capacitive value time sequence; The focusing mirror monitoring information and the capacitive value time sequence are analyzed by calling a cutting head state evaluation model to obtain a real-time cutting head index; It is judged whether the real-time cutting head index meets a predetermined index threshold; If not, an early warning instruction is issued, and the target air compressor is given a fault warning based on the early warning instruction; The focusing mirror monitoring information is obtained by dynamically monitoring the laser cutting focusing mirror by the optical analyzer, and comprises the following steps: The optical analyzer comprises a first analyzer and a second analyzer, wherein the first analyzer is arranged on a lens barrel of the laser cutting focusing mirror, and the second analyzer is arranged on a lens of the laser cutting focusing mirror; The reflectivity time sequence of the lens barrel is acquired by the first analyzer; The optical signal time sequence of the lens is acquired by the second analyzer; The reflectivity time sequence and the optical signal time sequence constitute the focusing mirror monitoring information; The focusing mirror monitoring information and the capacitive value time sequence are analyzed by calling the cutting head state evaluation model to obtain a real-time cutting head index, which comprises the following steps: A focusing mirror evaluation unit in the cutting head state evaluation model is acquired, wherein the focusing mirror evaluation unit has a predetermined reflectivity reference and a predetermined optical signal reference stored therein; The focusing mirror evaluation unit analyzes the reflectivity time sequence and the optical signal time sequence in combination with the predetermined reflectivity reference and the predetermined optical signal reference to obtain a real-time focusing mirror state feedback coefficient; The capacitive value time sequence is analyzed by a stylus evaluation unit in the cutting head state evaluation model to obtain a real-time stylus state feedback coefficient; The real-time focusing mirror state feedback coefficient and the real-time stylus state feedback coefficient after weighted and normalized processing are obtained to obtain the real-time cutting head index.

2. The method for diagnosing and warning of a fault of an air compressor according to claim 1, wherein The focusing mirror evaluation unit analyzes the reflectivity time sequence and the optical signal time sequence in combination with the predetermined reflectivity reference and the predetermined optical signal reference to obtain a real-time focusing mirror state feedback coefficient, which comprises the following steps: Any reflectivity corresponding to any time in the reflectivity time sequence is extracted; When the any reflectivity meets the predetermined reflectivity reference, a first trend prediction instruction is issued; The trend of the reflectivity time sequence is analyzed according to the first trend prediction instruction to obtain a first trend slope; Any optical signal corresponding to any time in the optical signal time sequence is extracted; When the any optical signal meets the predetermined optical signal reference, a second trend prediction instruction is issued; performing trend analysis on the light signal timing according to the second trend prediction instruction to obtain a second trend slope; the focusing mirror evaluation unit performs a variation weighting calculation on the first trend slope and the second trend slope to obtain the real-time focusing mirror state feedback coefficient.

3. The method for diagnosing and warning of a fault of an air compressor according to claim 2, wherein performing trend analysis on the reflectivity timing according to the first trend prediction instruction to obtain a first trend slope, including: plotting a reflectivity scatter plot according to the reflectivity timing; performing spline fitting on the reflectivity scatter plot to obtain a reflectivity fitted spline curve; obtaining a real-time slope based on the reflectivity fitted spline curve, and recording the real-time slope as the first trend slope; wherein performing spline fitting on the reflectivity scatter plot to obtain a reflectivity fitted spline curve, including: extracting a first scatter point set in the reflectivity scatter plot; performing polynomial fitting on the first scatter point set to obtain a first polynomial, and the first polynomial corresponds to a first spline curve; combining the reflectivity scatter plot and the first scatter point set to obtain a first check scatter point set; performing check analysis on the first spline curve according to the first check scatter point set to obtain a first check result; if the first check result meets a predetermined check constraint, taking the first spline curve as the reflectivity fitted spline curve.

4. The method for diagnosing and warning of a fault of an air compressor according to claim 3, wherein if not, issuing a warning instruction, and performing fault warning on the target air compressor based on the warning instruction, including: determining whether the real-time focusing mirror state feedback coefficient meets a first threshold; if not, performing a compressed air impurity fault warning on the target air compressor.

5. The method for diagnosing and warning of a fault of an air compressor according to claim 4, wherein including: if yes, determining whether the real-time stylus state feedback coefficient meets a second threshold; if not, performing a compressed air dryness fault warning on the target air compressor.

6. The method for diagnosing and warning of a fault of an air compressor according to claim 1, wherein performing analysis on the capacitance value timing through a stylus evaluation unit in the cutting head state evaluation model to obtain a real-time stylus state feedback coefficient, including: performing multi-domain feature collection on the capacitance value timing to obtain a capacitance value multi-domain feature set; calling a stylus state prediction model to perform prediction analysis on the capacitance value multi-domain feature set to obtain a predicted stylus state feedback coefficient; recording the predicted stylus state feedback coefficient as the real-time stylus state feedback coefficient; wherein calling a stylus state prediction model to perform prediction analysis on the capacitance value multi-domain feature set to obtain a predicted stylus state feedback coefficient, including: obtaining historical capacitance connection sensing stylus logs; assembling training data sets according to first logs, wherein the first logs refer to any one log record in the historical capacitance connection sensing stylus logs, and the first logs include a first capacitance value multi-domain feature set of a first capacitance value timing and a first stylus state feedback coefficient; performing machine learning on the first capacitance value multi-domain feature set and the first stylus state feedback coefficient in the training data set to obtain the stylus state prediction model.

7. A fault diagnosis and early warning device for an air compressor, characterized by, the device is used to implement the fault diagnosis and warning method for air compressors according to any one of claims 1-6, and the device includes: The cutting head acquisition module is configured to acquire a cutting head of the laser cutting device, and the cutting head comprises a laser cutting focusing mirror and a capacitance connection sensing stylus. The focusing mirror monitoring information acquisition module is configured to dynamically monitor the laser cutting focusing mirror by using an optical analyzer to obtain focusing mirror monitoring information. The capacitance value time sequence acquisition module is configured to dynamically monitor the capacitance connection sensing stylus by using an LCR meter to obtain a capacitance value time sequence. The real-time cutting head index acquisition module is configured to call a cutting head state evaluation model to analyze the focusing mirror monitoring information and the capacitance value time sequence to obtain a real-time cutting head index. The predetermined index threshold judgment module is configured to judge whether the real-time cutting head index meets a predetermined index threshold. The fault early warning module is configured to issue a warning instruction if the real-time cutting head index does not meet the predetermined index threshold, and perform fault early warning on the target air compressor based on the warning instruction.

Images