Method, device, product and medium for detecting equipment based on wear elements of lubricating oil

Abstract

The application discloses a device detection method based on lubricating oil wear elements, a device, a product and a medium, and relates to the technical field of device state monitoring. The method comprises the following steps: acquiring a plasma spectrum signal and a plasma image signal; extracting a stability parameter of the plasma from the plasma image signal, and dividing the plasma spectrum signal into a high-confidence spectrum set and a low-confidence spectrum set; extracting a spectral line evolution mode from the high-confidence spectrum set; acquiring running condition parameters of a fan gear box at multiple time points, and establishing a modulation relationship between the running condition parameters and the spectral line evolution mode; performing condition correction on spectral line intensity in the low-confidence spectrum set based on the modulation relationship, so as to obtain corrected spectral line intensity; and matching the corrected spectral line intensity with the spectral line evolution mode, so as to predict the remaining life of the fan gear box. By implementing the technical scheme provided in the application, the reliability of the remaining life prediction of the device can be improved.

Description

Technical Field

[0001] This application relates to the technical field of equipment condition monitoring, specifically to an equipment detection method, equipment, product, and medium based on lubricating oil wear elements. Background Technology

[0002] The gearbox is the core transmission component of a wind turbine generator set, and its operating status directly affects the reliability and power generation efficiency of the unit. Monitoring the condition of the gearbox and predicting its remaining life can effectively prevent sudden failures and reduce maintenance costs.

[0003] Lubricating oil wear element analysis is one of the important methods for monitoring the condition of wind turbine gearboxes. By periodically collecting lubricating oil samples and detecting the content of metal elements released from the oil due to the wear of internal components, the degree of wear of the gearbox can be reflected. When the content of metal elements exceeds a set threshold, it is determined that the equipment has an abnormal wear risk. This method can obtain internal wear information without disassembling the equipment and has been widely used in engineering practice.

[0004] However, during actual operation, the operating conditions of wind turbine gearboxes, such as speed, load, and temperature, are in a state of dynamic change. The wear rate and wear characteristics under different operating conditions are significantly different. Furthermore, the oil detection data itself is also affected by various factors such as the measurement environment and sample condition. As a result, monitoring methods based on a single threshold or simple trend analysis are difficult to accurately assess the true wear condition of the equipment, thus affecting the reliability of remaining life prediction. Summary of the Invention

[0005] This application provides a method, equipment, product, and medium for equipment detection based on lubricating oil wear elements, which can improve the reliability of equipment remaining life prediction.

[0006] The first aspect of this application provides a device detection method based on lubricating oil wear elements, specifically including:

[0007] Plasma spectral signals and plasma image signals of lubricating oil samples at multiple time points of the wind turbine gearbox were acquired. The lubricating oil samples contained multiple different types of wear indicator elements generated by the wear of internal components of the wind turbine gearbox.

[0008] The stability parameters of the plasma are extracted from the plasma image signal, and the plasma spectral signal is divided into a high-confidence spectral set and a low-confidence spectral set based on the stability parameters.

[0009] The spectral evolution patterns of the wear indicator elements at multiple time points are extracted from the high-confidence spectral set, and the spectral evolution patterns reflect the changes in spectral intensity at different wear stages.

[0010] The operating condition parameters of the wind turbine gearbox at multiple time points are obtained, and the modulation relationship between the operating condition parameters and the spectral evolution mode is established.

[0011] Based on the modulation relationship, the intensity of spectral lines in the low-confidence spectrum set is corrected for operating conditions to obtain the corrected spectral line intensity.

[0012] The intensity of the corrected spectral line is matched with the spectral line evolution mode to identify the current wear stage of the wind turbine gearbox. Based on the spectral line evolution mode and modulation relationship corresponding to the wear stage, and combined with the current operating condition parameters, the remaining life of the wind turbine gearbox is predicted.

[0013] By employing the above technical solution, plasma spectral and image signals of lubricating oil samples are acquired. Stability parameters extracted from the plasma image signals are used to assess and classify the reliability of the spectral signals. Spectral line evolution patterns of wear indicator elements are extracted from high-reliability spectral data. The modulation relationship between operating conditions and these spectral line evolution patterns is established. Low-reliability spectral data undergoes operating condition correction. Wear stages are identified based on the matching results between the corrected spectral line intensities and evolution patterns. Finally, the remaining lifespan is predicted by combining the current operating conditions. This solution, by introducing plasma stability assessment and operating condition modulation correction mechanisms, effectively reduces the impact of plasma instability and changes in operating conditions on spectral data, improves the accuracy of wear state identification, and thus enhances the reliability of wind turbine gearbox remaining life prediction.

[0014] Optionally, dividing the plasma spectral signal into a high-confidence spectral set and a low-confidence spectral set based on the stability parameter includes:

[0015] Extract a first stability parameter reflecting the spatial distribution characteristics of the plasma and a second stability parameter reflecting the temporal evolution characteristics of the plasma from the plasma image signal;

[0016] The first deviation of the first stability parameter and the second deviation of the second stability parameter are calculated respectively. The deviation characterizes the degree of deviation of the plasma stability parameter from the ideal stable state.

[0017] Based on the first deviation and the second deviation, a spectral quality value is constructed. Plasma spectral signals with spectral quality values ​​lower than a preset quality threshold are classified into a high-confidence spectral set, and plasma spectral signals with spectral quality values ​​higher than or equal to the preset quality threshold are classified into a low-confidence spectral set.

[0018] By employing the aforementioned technical solution, a first stability parameter characterizing spatial distribution and a second stability parameter characterizing temporal evolution are extracted from plasma image signals. By calculating the deviation of these two stability parameters from the ideal stable state, a spectral quality value reflecting the overall stability of the plasma is constructed, and the reliability of the plasma spectral signal is classified based on this quality value. This scheme, by comprehensively considering the spatial and temporal stability characteristics of plasma, establishes a more comprehensive and accurate spectral signal quality assessment mechanism, providing a reliable data foundation for subsequent spectral line evolution mode extraction and operational condition correction, and further improving the accuracy of wear condition identification.

[0019] Optionally, predicting the remaining lifespan of the wind turbine gearbox based on the spectral evolution mode and modulation relationship corresponding to the wear stage, combined with the current operating condition parameters, includes:

[0020] Extract the wear evolution rate from the spectral evolution mode corresponding to the current wear stage;

[0021] The operating condition modulation coefficient of the operating condition parameters on the wear evolution rate is obtained according to the modulation relationship. The operating condition modulation coefficient is used to correct the wear evolution rate to obtain the corrected wear evolution rate.

[0022] Based on the wear evolution rate after the working condition correction, the evolution time required for the spectral intensity of the wear indicator element to evolve from the current value to reach the preset wear failure threshold is calculated, and the evolution time is taken as the remaining life of the wind turbine gearbox.

[0023] By employing the above technical solution, the wear evolution rate is extracted from the spectral evolution mode corresponding to the current wear stage. Based on the established modulation relationship, the operating condition modulation coefficient is obtained, and the wear evolution rate is corrected for operating conditions. Then, the remaining lifetime is predicted by calculating the evolution time required for the spectral intensity to reach the preset wear failure threshold from the current value. This solution establishes a quantitative relationship between operating condition parameters and the wear evolution rate, enabling dynamic adjustment of the wear rate prediction based on actual operating conditions. This achieves a more accurate and reliable remaining lifetime prediction while considering the impact of changes in operating conditions.

[0024] Optionally, the step of calculating the evolution time required for the spectral intensity of the wear indicator element to evolve from its current value to a preset wear failure threshold based on the wear evolution rate corrected for the operating conditions includes:

[0025] Identify the sequence of subsequent wear stages required for the wear indicator element to evolve from the current wear stage to the wear failure threshold based on the spectral line evolution pattern;

[0026] For each wear stage in the subsequent wear stage sequence, the stage termination spectral line intensity and stage wear evolution rate of each wear stage are extracted from the corresponding spectral line evolution mode;

[0027] For the first wear stage in the subsequent wear stage sequence, calculate the first stage spectral line intensity difference between the current value of the spectral line intensity of the wear indicator element and the stage termination spectral line intensity of the first wear stage. Based on the modulation relationship and the operating condition parameters at the current time, obtain the first stage operating condition modulation coefficient for calculating the evolution rate of the first wear stage. Use the first stage operating condition modulation coefficient to correct the stage wear evolution rate of the first wear stage to obtain the first stage corrected evolution rate. Divide the first stage spectral line intensity difference by the first stage corrected evolution rate to obtain the evolution duration of the first stage.

[0028] For each remaining wear stage in the subsequent wear stage sequence excluding the first wear stage, the intensity difference of the current stage spectral line between the stage termination spectral line intensity of the previous wear stage and the stage termination spectral line intensity of the current wear stage is calculated. Based on the stage evolution duration of the previous wear stage, the operating condition parameters at the start time of the current wear stage are predicted. Based on the operating condition parameters at the start time of the current wear stage, the corresponding current stage operating condition modulation coefficient is determined. The stage wear evolution rate of the current wear stage is corrected using the current stage operating condition modulation coefficient to obtain the current stage corrected evolution rate. The stage evolution duration of the current wear stage is obtained by dividing the current stage spectral line intensity difference by the current stage corrected evolution rate.

[0029] The evolution time is obtained by summing the evolution duration of the first stage with the stage evolution durations of all remaining wear stages.

[0030] By adopting the above technical solution, the complete sequence of subsequent stages required to reach the failure threshold from the current wear stage is first determined based on the spectral evolution model. The terminating spectral line intensity and wear evolution rate of each stage are then extracted from the spectral evolution model. For the first wear stage, the difference between the current spectral line intensity and the stage termination intensity is calculated, and the wear evolution rate is corrected based on the current operating parameters using the operating condition modulation coefficient, thereby calculating the evolution duration of the first stage. For each subsequent wear stage, the operating parameters at the start of the current stage are determined by considering the evolution duration of the previous stage, achieving dynamic updating of the operating parameters and accurately correcting the wear evolution rate of each stage accordingly. This scheme establishes temporal correlations between stages, achieving continuous evolution prediction of operating parameters. Furthermore, by employing segmented calculation and accumulation, it accurately characterizes the evolutionary characteristics of different wear stages, thus constructing an accurate remaining life prediction model that reflects the actual wear process.

[0031] Optionally, the method further includes:

[0032] Calculate the change in spectral intensity of the wear indicator element between adjacent time points, and compare the change in spectral intensity with the expected change in spectral intensity for the corresponding time interval in the spectral evolution mode;

[0033] When the absolute value of the change in spectral line intensity exceeds a preset sudden drop threshold, a lubricating oil replacement event is identified, and the time of occurrence of the lubricating oil replacement event is recorded.

[0034] The high-confidence spectral set and the low-confidence spectral set prior to the time of the lubricating oil replacement event are marked as the pre-oil change historical spectral dataset, and the spectral line evolution mode corresponding to the pre-oil change historical spectral dataset is used as the pre-oil change wear evolution benchmark.

[0035] Based on the plasma spectral signal and plasma image signal collected after the lubricating oil replacement event, quality classification is re-performed to obtain a high-confidence spectral set and a low-confidence spectral set after the oil change. The spectral evolution mode after the oil change is extracted from the high-confidence spectral set after the oil change.

[0036] By comparing the wear evolution benchmark before oil change with the spectral line evolution mode after oil change at the same operating time, the wear state change trend of the wind turbine gearbox before and after oil change is evaluated.

[0037] By employing the above technical solution, automatic identification of lubricating oil replacement events is achieved by monitoring the changes in spectral line intensity between adjacent time points and comparing them with expected changes. Based on this, the spectral data is segmented according to the oil change time, and the spectral line evolution patterns before and after the oil change are extracted. By comparing the spectral line intensity evolution rates under the same operating time, the trend of gearbox wear condition changes is assessed. This solution not only establishes a maintenance event identification mechanism based on abrupt changes in spectral line intensity characteristics, but also evaluates the effectiveness of maintenance measures by comparing the wear evolution characteristics before and after oil changes, providing a more comprehensive and continuous analytical perspective for equipment condition monitoring.

[0038] Optionally, after matching the corrected spectral line intensity with the spectral line evolution mode to identify the current wear stage, the method further includes:

[0039] The correction spectral intensity of each of the multiple different types of wear indicator elements is extracted, and the correction spectral intensity of each of the multiple different types of wear indicator elements is matched with the corresponding spectral evolution mode to obtain the wear stage determination result identified by each of the multiple different types of wear indicator elements.

[0040] The number of elements with consistent wear stage determination results among the multiple different types of wear indicator elements is counted, and the ratio of the number of elements with consistent wear stage determination results to the total number of multiple different types of wear indicator elements is used as the stage identification consistency index.

[0041] When the consistency index of the stage identification is lower than the preset consistency threshold, wear indicator elements that conflict with the wear stage determination results are identified, and the spectral evolution mode corresponding to the conflicting wear indicator elements is adjusted.

[0042] By employing the aforementioned technical solution, spectral intensity matching and wear stage determination are performed on multiple different types of wear indicator elements. The proportion of elements with consistent determination results is calculated as a consistency index, and the spectral evolution patterns of conflicting elements are adjusted when consistency is low. This solution, by establishing a consistency evaluation mechanism for multi-element wear stage determination, can not only identify wear indicator elements with biased determination results but also improve the accuracy of wear state identification based on multi-element comprehensive analysis by specifically adjusting their spectral evolution patterns.

[0043] Optionally, adjusting the spectral evolution mode corresponding to the conflicting wear indicator element includes:

[0044] Calculate the corrected spectral intensity of each of the multiple different types of wear indicator elements, and the matching degree between it and the spectral intensity range corresponding to each wear stage determined by the conflicting wear indicator elements. Select the wear stage determination result corresponding to the wear indicator element with the highest matching degree as the corrected current wear stage.

[0045] For the corrected current wear stage, the evolution trend of spectral intensity of the multiple different types of wear indicator elements over a historical preset time period is traced back. Abnormal elements that deviate from the normal evolution trend are identified, and the spectral evolution mode corresponding to the abnormal element is marked as an abnormal evolution mode and excluded from the subsequent remaining lifetime prediction.

[0046] By employing the aforementioned technical solution, the matching degree between the corrected spectral line intensity of each wear indicator element and the spectral line intensity range of the current wear stage is calculated. The element with the highest matching degree is selected as the corrected wear stage. Furthermore, by retrospectively analyzing the historical evolution trends of each element, anomalous elements deviating from the normal trend are identified and excluded from lifetime prediction. This solution establishes a wear stage correction mechanism based on matching degree and achieves automatic identification and removal of anomalous elements through historical evolution trend analysis, ensuring the reliability of the spectral line evolution models participating in lifetime prediction and improving the accuracy of the final prediction results.

[0047] In a second aspect, this application provides a device detection apparatus based on lubricating oil wear elements, the device detection apparatus based on lubricating oil wear elements comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory being used to store computer program code, the computer program code including computer instructions, the one or more processors calling the computer instructions to cause the device detection apparatus based on lubricating oil wear elements to perform the method described in the first aspect and any possible implementation thereof.

[0048] Thirdly, this application provides a computer program product containing instructions that, when run on a device detection device based on lubricating oil wear elements, cause the device detection device based on lubricating oil wear elements to perform the method described in the first aspect and any possible implementation thereof.

[0049] Fourthly, this application provides a computer-readable storage medium including instructions that, when executed on a device for detecting lubricating oil wear elements, cause the device for detecting lubricating oil wear elements to perform the method described in the first aspect and any possible implementation thereof. Attached Figure Description

[0050] Figure 1This is a schematic flowchart of a device detection method based on lubricating oil wear elements provided in an embodiment of this application;

[0051] Figure 2 This is a schematic diagram illustrating a spectral line evolution mode and wear stage identification provided in an embodiment of this application;

[0052] Figure 3 This is a schematic diagram illustrating a multi-stage remaining lifetime prediction provided in an embodiment of this application;

[0053] Figure 4 This is an exemplary hardware structure diagram of a device detection device based on lubricating oil wear elements provided in an embodiment of this application. Detailed Implementation

[0054] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0055] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0056] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0057] This application provides a device detection method based on lubricating oil wear elements, referencing... Figure 1 , Figure 1 This is a flowchart illustrating a device detection method based on lubricating oil wear elements according to an embodiment of this application, including steps S101 to S106, as follows:

[0058] S101: Acquire plasma spectral signals and plasma image signals of lubricating oil samples at multiple time points of the wind turbine gearbox. The lubricating oil samples contain multiple different types of wear indicator elements caused by the wear of internal components of the wind turbine gearbox.

[0059] Plasma spectral signals refer to the characteristic wavelength spectral signals emitted by transient plasma excited after pulsed laser ablation of lubricating oil samples using laser-induced breakdown spectroscopy. These signals include atomic and ion emission lines of wear-indicating elements such as iron, copper, aluminum, and chromium. Plasma image signals represent image data of the spatial morphology, brightness distribution, and temporal evolution of laser-induced plasma acquired by high-speed imaging equipment, reflecting the expansion characteristics and stability of the plasma plume. Wear-indicating elements refer to metallic elements that indicate the degree of wear on components such as gears, bearings, and drive shafts inside the gearbox. These elements are suspended or dissolved in the lubricating oil as micron or nanometer-sized particles due to friction and wear. Multiple time points refer to a discrete sequence of time points during equipment operation, where lubricating oil samples are collected according to a set monitoring cycle. The time interval is determined based on the intensity of equipment operation.

[0060] Specifically, at a preset sampling time, lubricating oil samples are extracted from the lubricating oil circulation system of the fan gearbox or from the oil tank sampling port. The samples are placed in a sample cell with a transparent window or dropped onto the substrate surface. A pulsed laser generates nanosecond or picosecond-level laser pulses, which are focused onto the surface or interior of the lubricating oil sample through a focusing lens. When the laser power density reaches the plasma breakdown threshold, the substances in the sample are instantly vaporized and ionized to form a high-temperature, high-density plasma plume. During the de-excitation process, excited-state atoms and ions in the plasma emit photons of specific wavelengths. This radiation is collected and transmitted to a spectrometer through a fiber optic probe or lens system. The spectrometer disperses the light signal and records the intensity distribution of different wavelengths through a photodetector array, forming a plasma spectral signal containing multiple elemental characteristic spectral lines. Simultaneously, an ICCD camera or a high-speed CMOS camera is positioned to the side or above the laser focal point to capture a sequence of emission images of the plasma plume with microsecond-level temporal and spatial resolution, obtaining the plasma image signal.

[0061] In some embodiments, signal acquisition based on laser-induced breakdown spectroscopy can be achieved in various ways. Optionally, an Nd:YAG pulsed laser is used to generate laser pulses with a wavelength of 1064 nm, a pulse width of 5 to 10 nanoseconds, and a single pulse energy of 50 to 100 millijoules. The laser is applied to the surface of lubricating oil placed in a sample cell through a focusing lens with a focal length of 50 mm. The resulting plasma radiation is collected by an optical fiber probe placed at a 45-degree angle to the laser path and transmitted to an echelle grating spectrometer. The spectrometer has a wavelength range covering 200 to 800 nm. After the plasma is generated, spectral acquisition is triggered with a delay of 1 to 2 microseconds and an integration time window of 2 to 5 microseconds is set. Simultaneously, an enhanced CCD camera is used to acquire a two-dimensional spatial distribution image of the plasma plume through an observation window perpendicular to the laser incident direction with an exposure time of 100 nanoseconds. Optionally, a lubricating oil sample is dropped onto the surface of a quartz or metal substrate to form a thin liquid film. This film is then ablated using a 532 nm frequency-doubled Nd:YAG laser or an ultraviolet laser. The laser is guided through a coaxial optical path while simultaneously collecting plasma radiation. A fiber-coupled high-resolution spectrometer records the spectral lines, and a gated enhanced CCD camera captures multiple frames of plasma evolution images at different delay times. By adjusting the gating timing, the complete time evolution process of the plasma from initial expansion to decay and disappearance can be obtained. It is understood that other laser parameter configurations and optical path designs can also be used to achieve plasma excitation and simultaneous dual-signal acquisition; this is not limited here.

[0062] S102: Extract the stability parameters of the plasma from the plasma image signal, and divide the plasma spectral signal into a high-confidence spectral set and a low-confidence spectral set based on the stability parameters.

[0063] Among them, stability parameters are used to represent the stability of laser-induced plasma during its formation and evolution, including parameters such as the repeatability of the spatial position of the plasma plume, morphological consistency, brightness fluctuation, and temporal evolution regularity. The high-confidence spectral set refers to the set of spectral data collected under stable plasma excitation conditions and with minimal fluctuations in the morphology and position of the plasma plume. These data are less affected by factors such as laser energy fluctuations, sample matrix effects, and environmental interference. The low-confidence spectral set represents spectral data obtained under unstable plasma conditions, which may have larger measurement deviations due to factors such as unstable laser pulse energy, sample surface inhomogeneity, plasma shielding effects, or environmental airflow disturbances. The plasma image signal contains information such as the spatial geometric features of the plasma plume, its spatial brightness distribution, and its temporal evolution morphology.

[0064] Specifically, the acquired plasma image signal sequence is processed frame by frame. The plasma luminescent region is extracted from the background using a grayscale thresholding method, and geometric parameters such as the centroid coordinates, equivalent circle diameter, major-to-minor axis ratio, and eccentricity of the plasma plume are calculated. The standard deviation of the centroid coordinates in plasma images generated by multiple consecutive laser pulses is statistically analyzed to obtain spatial position stability parameters. The spatial uniformity of pixel grayscale values ​​within the plasma region is calculated, using the coefficient of variation or entropy value as brightness distribution stability parameters. For temporal evolution stability, image sequences of plasma generated by a single laser pulse at different delay times are extracted, and temporal evolution characteristic parameters such as plasma plume expansion rate and decay rate are calculated. The consistency of evolution parameters across multiple pulses is compared to obtain temporal stability parameters. A comprehensive stability evaluation index is constructed using weighted combination or principal component analysis methods based on the extracted stability parameters. A stability threshold is set, and spectral signals with an index above the threshold are assigned to a high-confidence spectral set, while those below the threshold are assigned to a low-confidence spectral set.

[0065] In some embodiments, stability evaluation and spectral classification based on plasma images can be achieved in various ways. Optionally, each lubricating oil sample is excited by multiple laser pulses and corresponding plasma images are acquired. The standard deviations of the X and Y coordinates of the plasma centroid position in these images are calculated. When both standard deviations are less than 50 micrometers, the spatial position is considered stable. The relative standard deviation of the total plasma brightness is calculated, and when this value is less than 15%, the brightness is considered stable. Only when both spatial position and brightness are stable is the corresponding spectral signal assigned to the high-confidence set; otherwise, it is assigned to the low-confidence set. Optionally, an image similarity algorithm is used to calculate the structural similarity index between plasma images generated by multiple laser pulses of the same sample. The Hu moment or Zernike moment features of the plasma plume are extracted, and the Euclidean distance of the feature vectors between multiple pulses is calculated. When the similarity index is greater than 0.85 and the feature vector distance is less than a set threshold, the plasma morphology stability is considered good, and the corresponding spectrum is marked as high-confidence; otherwise, it is marked as low-confidence. It is understood that other image feature extraction and stability quantification methods can also be used to achieve quality classification of spectral data, which is not limited here.

[0066] S103: Extract the spectral line evolution patterns of wear indicator elements at multiple time points from a high-confidence spectral set. The spectral line evolution patterns reflect the changes in spectral line intensity at different wear stages.

[0067] The spectral line evolution model refers to the trend and stage-specific pattern of the radiation intensity of characteristic spectral lines of wear indicator elements detected by laser-induced breakdown spectroscopy as a function of gearbox operating time or cumulative operating hours. This model describes the complete evolution trajectory from initial commissioning of the equipment through the break-in period, stable wear period, and finally accelerated wear period. Spectral line intensity represents the numerical value of the radiation intensity of a specific element's characteristic wavelength spectral line in the plasma spectral signal, and this intensity has a quantitative correspondence with the mass concentration of the corresponding element in the lubricating oil. Wear stages refer to different periods in the gearbox wear process, divided according to differences in wear rate and wear mechanism, including the initial running-in wear stage, the normal steady-state wear stage, the abnormal wear stage, and the accelerated wear stage before failure. The element release rate and spectral line intensity growth rate differ significantly in each stage. The high-confidence spectral set contains spectral data obtained under stable laser plasma conditions, eliminating the interference of plasma fluctuations on elemental spectral line intensity measurements.

[0068] Specifically, key characteristic spectral lines of specific wear indicator elements are selected from a high-confidence spectral set. For example, for iron, the characteristic spectral lines are selected as Fe I atomic lines or Fe II ion lines with wavelengths of 371.99 nm or 438.35 nm. The peak intensity or integrated intensity of these spectral lines at each sampling time point is extracted to form a time-series spectral intensity data sequence for the element. This time-series data is preprocessed, including baseline subtraction, normalization, and outlier removal. Time series analysis methods are used to perform trend decomposition on the spectral intensity sequence to identify the overall upward trend and stage-specific changes in spectral intensity. The rate of change of spectral intensity within a local time period is calculated using a sliding window, and the curve of the rate of change over time is plotted. Based on the abrupt change points or inflection points of the rate of change, the entire time series is divided into several wear stages. Statistical modeling or function fitting is performed on the spectral intensity data of each wear stage to extract characteristic parameters such as the initial intensity, termination intensity, duration, average growth rate, and growth acceleration of each stage. The set of these stage-specific characteristic parameters constitutes the spectral evolution model of the element. The above analysis procedure was performed on various wear indicator elements in lubricating oil, such as iron, copper, chromium, and aluminum.

[0069] In some embodiments, the extraction of spectral line evolution patterns based on laser-induced breakdown spectroscopy can be achieved in various ways. Optionally, the 371.99 nm spectral line of iron is selected as the main indicator spectral line of gear wear. The original intensity sequence is smoothed using a five-point moving average. The first-order difference of the intensity at adjacent time points is calculated to obtain the incremental sequence. The incremental sequence is subjected to CUSUM accumulation and test to identify the mean abrupt change point. Based on the abrupt change point, the time series data is divided into three stages. The intensity-time data of each stage is fitted with least squares linearly or exponentially to obtain the slope parameter and intercept parameter of each stage. The fitting parameters of the three stages and the stage duration are combined to form the parameter vector of the spectral line evolution pattern. Optionally, unsupervised machine learning methods can be used to perform cluster analysis on the time-series data of spectral line intensity in a high-confidence spectral set. The K-means algorithm or hierarchical clustering algorithm can be used to divide the time-series data points into several categories according to intensity level and rate of change. Each category corresponds to a wear stage. The mean and standard deviation of the data within each category are calculated as the statistical characteristics of that stage. A state transition model is constructed based on the transition order and transition time between categories to describe the spectral line evolution pattern. It is understood that other signal processing and pattern recognition methods can also be used to extract the temporal evolution law of spectral lines; this is not limited here.

[0070] like Figure 2 As shown, Figure 2 This is a schematic diagram of spectral evolution mode and wear stage identification provided in this application embodiment. The diagram shows the complete wear evolution process of the wind turbine gearbox from initial commissioning to failure, with the running time as the horizontal axis and the spectral intensity of wear elements as the vertical axis. The spectral evolution mode curve (40) is divided into four typical segments: initial break-in stage (41), normal wear stage (42), accelerated wear stage (43), and severe wear stage (44). The spectral intensity growth rate of each stage is significantly different, reflecting the changing law of element release rate under different wear mechanisms. The wear failure threshold (45) marked by the horizontal dashed line in the figure represents the critical state in which the equipment needs to be repaired or the parts need to be replaced, which is the target endpoint for remaining life prediction. When the system collects a lubricating oil sample at the current moment and measures the corrected spectral intensity (46) by laser-induced breakdown spectroscopy, the intensity value can be matched and compared with the pre-established spectral evolution mode curve to accurately determine which wear stage the gearbox is currently in, providing a key state positioning basis for subsequent remaining life prediction.

[0071] S104: Obtain the operating condition parameters of the wind turbine gearbox at multiple time points and establish the modulation relationship between the operating condition parameters and the spectral evolution mode.

[0072] Among them, operating condition parameters refer to the working state parameters of the wind turbine gearbox during operation, including physical parameters such as spindle speed, input torque, output power, gearbox vibration acceleration, lubricating oil temperature, lubricating oil viscosity, ambient temperature, and ambient humidity, which affect the wear process and the concentration of wear particles in the lubricating oil. Modulation relationship is used to represent the influence mechanism of operating condition parameters on the evolution law of spectral line intensity measured by laser-induced breakdown spectroscopy, that is, the quantitative relationship between the differences in wear rate under different operating conditions and the resulting differences in the rate of change of spectral line intensity. Multiple time points refer to time nodes that correspond one-to-one with the lubricating oil sample collection time. At these times, the operating condition parameter values ​​from the equipment monitoring system need to be recorded or extracted synchronously. Spectral line evolution mode includes the variation characteristics of the spectral line intensity of wear indicator elements at different wear stages; this evolution process is significantly affected by operating condition parameters.

[0073] Specifically, through the SCADA monitoring system or distributed sensor network of the wind turbine gearbox, instantaneous measurements of operating parameters such as speed sensor data, torque sensor data, and temperature sensor data are simultaneously extracted at the moment of each lubricating oil sample collection, or the average and peak values ​​of the operating parameters within the 24 hours prior to the sampling time are calculated. The acquired operating parameters are time-aligned and correlated with the spectral line intensity data measured by laser-induced breakdown spectroscopy at the corresponding time, forming a paired sample set of operating parameters and spectral line intensities. A mapping function between operating parameters and the rate of change of spectral line intensity is established through statistical regression analysis or machine learning modeling methods. Using multiple linear regression or multiple nonlinear regression, the time derivative of spectral line intensity or the intensity difference between adjacent moments is used as the dependent variable, and operating parameters such as speed, load, and temperature are used as independent variables. A coefficient matrix is ​​obtained through regression fitting, which quantitatively characterizes the contribution weight of each operating parameter to the spectral line evolution rate. Alternatively, machine learning models such as neural networks and support vector machines can be used, with the operating condition parameter vector as input and the change in spectral line intensity as output for training. The trained model can predict the evolution rate of spectral line intensity based on given operating conditions, and the input-output mapping relationship of this model is the mathematical expression of the modulation relationship.

[0074] In some embodiments, the modulation relationship between operating conditions and the evolution of laser-induced breakdown spectral lines can be established in various ways. Optionally, historical operating data can be divided into three intervals according to rotational speed (low speed, medium speed, high speed), load (light load, medium load, heavy load), and lubricating oil temperature (low temperature, normal temperature, high temperature), forming 27 operating condition combinations. The unit-time increment of iron element spectral line intensity under each operating condition combination is statistically analyzed, and a mapping table between operating condition combinations and spectral line intensity growth rate is constructed. For operating conditions not in the mapping table, a three-dimensional interpolation method is used to estimate the corresponding growth rate. Optionally, a BP neural network is used to establish the modulation relationship model. The input layer has four neurons for rotational speed, torque, temperature, and vibration intensity, the hidden layer has 10 to 20 neurons, and the output layer has one neuron for the rate of change of spectral line intensity. The network is trained using operating condition parameters and corresponding spectral line intensity increment data from multiple historical time points, and the network weights are optimized using gradient descent. After training, the neural network can predict the evolution rate of spectral line intensity based on any combination of input operating condition parameters. It is understandable that other parametric or non-parametric modeling methods can also be used to quantify the modulation effect of operating parameters on spectral line evolution, which is not limited here.

[0075] S105: Based on the modulation relationship, the intensity of spectral lines in the low-confidence spectrum set is corrected according to the operating conditions to obtain the corrected spectral line intensity.

[0076] Operating condition correction refers to using established modulation relationships between operating condition parameters and the intensity of laser-induced breakdown spectra to correct low-confidence spectral data affected by plasma instability or laser energy fluctuations, eliminating or compensating for systematic biases in spectral intensity measurements caused by operating condition changes. The low-confidence spectrum set includes spectral data collected when the laser plasma is unstable, the plasma plume position is offset, or its shape is irregular; these data have low signal-to-noise ratios or poor repeatability. The corrected spectral intensity represents the spectral intensity value after operating condition compensation, corresponding to the spectral intensity that should be measured under standard or reference operating conditions, facilitating cross-sectional comparisons of data collected under different operating conditions. The modulation relationship includes a quantitative mapping model from the operating condition parameter vector to the spectral intensity deviation or rate of change.

[0077] Specifically, for each spectral sample in the low-confidence spectral set, the gearbox operating condition parameters at the corresponding acquisition time are extracted, including parameters such as rotational speed, load power, and lubricating oil temperature at that time. This operating condition parameter vector is input into the established modulation relationship model to calculate the spectral intensity deviation correction amount or proportional correction coefficient relative to the reference operating condition under this condition. The reference operating condition can be set as the rated operating condition, the average operating condition corresponding to the high-confidence spectral set, or the standard test operating condition. By adding or subtracting the deviation correction amount from the original measured spectral intensity, or by multiplying or dividing the original spectral intensity by the proportional correction coefficient, the normalized calibrated spectral intensity to the reference operating condition is obtained.

[0078] In some embodiments, condition-based correction of laser-induced breakdown spectral line intensity can be achieved through various methods. Optionally, a baseline operating condition is set as a rotational speed of 1500 rpm, a load power of 80% of the rated power, and a lubricating oil temperature of 60 degrees Celsius. For the actual operating parameters corresponding to the low-confidence spectrum, the theoretical difference in spectral line intensity between this operating condition and the baseline operating condition is calculated using the regression equation of the modulation relationship. The original spectral line intensity measured by the laser-induced breakdown spectrum is subtracted from this theoretical difference to obtain the corrected spectral line intensity equivalent to the baseline operating condition. For spectra containing multiple wear indicator elements, independent condition correction is performed on the spectral lines of each element. Optionally, a ratio correction method is used. The sample with the closest operating parameters in the high-confidence spectrum set is selected as the reference sample. The modulation ratio of the spectral line intensity between the actual operating condition of the low-confidence spectrum and the operating condition of the reference sample is calculated using a modulation relationship model. The original spectral line intensity of the low-confidence spectrum is divided by this modulation ratio to obtain the corrected spectral line intensity. This method can simultaneously compensate for measurement deviations caused by operating condition differences and plasma instability. It is understandable that other correction algorithms can also be used to normalize the working conditions and improve the quality of low-confidence laser-induced breakdown spectral data, which is not limited here.

[0079] S106: Match the intensity of the corrected spectral line with the spectral line evolution mode to identify the current wear stage of the wind turbine gearbox. Based on the spectral line evolution mode and modulation relationship corresponding to the wear stage, and combined with the current operating parameters, predict the remaining life of the wind turbine gearbox.

[0080] The corrected spectral intensity refers to the laser-induced breakdown spectral intensity value after operating condition correction. This value eliminates the dual effects of operating condition fluctuations and plasma state fluctuations, accurately reflecting the concentration level of wear elements in the lubricating oil. The spectral evolution model includes the trajectory of the laser-induced breakdown spectral intensity changes throughout the complete wear cycle from the initial operation of the gearbox to failure. This model is divided into multiple wear stages with different evolution rates. Each wear stage indicates the current wear level of the gearbox, with different stages corresponding to different wear mechanisms and remaining lifespans. Remaining lifespan represents the remaining operating time from the current moment until the gearbox wear level reaches a preset maintenance threshold or failure criterion, typically measured in hours or days. Operating condition parameters include variables such as rotational speed, load, and temperature that affect the subsequent wear rate.

[0081] Specifically, the spectral intensity of the latest lubricating oil sample is obtained using laser-induced breakdown spectroscopy and corrected for operating conditions. This intensity value is compared with the characteristic ranges of spectral intensity for each wear stage in the established spectral evolution model to determine which wear stage the current intensity value falls within, or the similarity between the current intensity and representative intensity values ​​for each stage is calculated to identify the current wear stage of the gearbox. After determining the current wear stage, the spectral intensity evolution rate parameters or evolution functions for this stage and subsequent stages are extracted from the spectral evolution model. Current and expected operating condition parameters of the gearbox are obtained, including planned operating speed gears, expected load distribution, and seasonal changes in ambient temperature. These parameters are input into a modulation relationship model to calculate the correction coefficients for the evolution rate of each wear stage. Using the corrected evolution rate, the increase in spectral intensity over time is gradually calculated from the current spectral intensity value until the spectral intensity reaches a preset failure threshold, which is determined based on historical failure cases or the equipment manufacturer's recommended value. The cumulative time required from the current moment until the spectral intensity reaches the failure threshold is the predicted remaining life of the gearbox.

[0082] In some embodiments, wear stage identification and remaining lifetime prediction based on laser-induced breakdown spectroscopy can be achieved in various ways. Optionally, the Euclidean distance is calculated between the intensity of the current condition-corrected 371.99 nm iron spectral line and the median intensity of the three wear stages in the spectral evolution mode. The stage with the smallest distance is selected as the current wear stage. The standard evolution rate is extracted from this stage. The condition correction coefficient is calculated based on the real-time speed and load of the gearbox through a regression equation of the modulation relationship. The actual evolution rate is obtained by multiplying the standard evolution rate by the correction coefficient. The failure threshold is set to three times the upper limit of the spectral line intensity of the normal steady-state wear stage. The difference between the current intensity and the failure threshold is calculated and divided by the actual evolution rate to obtain the remaining time of a single stage. If multiple wear stages need to be covered, the remaining times of each stage are accumulated to obtain the total remaining lifetime. Optionally, a dynamic Bayesian network prediction model based on laser-induced breakdown spectrum is established. The current wear stage, current corrected spectral line intensity, and current operating parameters are used as input observation variables. State inference is performed based on the prior probability distribution of the spectral line evolution mode and the conditional probability distribution of the modulation relationship. Monte Carlo simulation is used to generate samples of spectral line intensity evolution trajectories for future times. The distribution of the first time the failure threshold is exceeded in the samples is statistically analyzed, and the median or expected value of the distribution is taken as the predicted remaining lifetime, along with a confidence interval. It is understood that other prediction models and inference algorithms can also be used to achieve remaining lifetime estimation considering the characteristics of laser-induced breakdown spectral data and the influence of operating conditions; this is not limited here.

[0083] like Figure 3 As shown, Figure 3 This is a schematic diagram illustrating a multi-stage remaining lifetime prediction provided in an embodiment of this application. Figure 3 exist Figure 2 Based on the spectral evolution model, a refined prediction process is presented from the current point (50) until the wear failure threshold (45) is reached. The prediction path is divided into two consecutive wear evolution stages, Stage 1 and Stage 2. Each stage considers the influence of the actual operating conditions on the wear rate during that period. In Stage 1, the system calculates the operating condition modulation coefficient k1 based on the current and expected operating condition parameters such as rotational speed, load, and temperature through a pre-established operating condition modulation relationship. The standard wear evolution rate v1 of this stage is corrected to the actual evolution rate v1' (51), and then the time t1 required to complete this stage is calculated. Similarly, in Stage 2, the modulation coefficient k2 is calculated based on the new operating condition parameters determined by the evolution duration of the previous stage. The rate v2 is corrected to obtain v2' (53), and the time t2 is calculated. Finally, the total remaining lifetime RUL=t1+t2 (55) is obtained by accumulating the evolution times of each stage, realizing high-precision lifetime prediction under dynamic operating conditions.

[0084] Based on the above embodiments, as an optional embodiment, S102: the step of dividing the plasma spectral signal into a high-confidence spectral set and a low-confidence spectral set based on the stability parameter may specifically include the following steps:

[0085] S201: Extract the first stability parameter reflecting the spatial distribution characteristics of the plasma and the second stability parameter reflecting the temporal evolution characteristics of the plasma from the plasma image signal.

[0086] The acquired plasma image signals are preprocessed, and the plasma luminescent region is extracted and its centroid coordinates are calculated by grayscale thresholding. For plasma images generated by N consecutive laser pulses, the standard deviations σx and σy in the X and Y directions of the centroid coordinates are calculated. The relative standard deviation RSDd of the equivalent circle diameter of the plasma region is calculated, and the spatial variation coefficient CVg of the pixel grayscale values ​​within the plasma is calculated. σx, σy, RSDd, and CVg are used as the first stability parameters. Image sequences of plasma generated by a single laser pulse at different delay times t1, t2, ..., tm are extracted. The rate of change of plasma radius at adjacent time points is calculated to obtain the expansion velocity. The standard deviation σv of the expansion velocity for N pulses is calculated. The attenuation constant is obtained by fitting an exponential function to calculate the attenuation of the total plasma intensity over time. The coefficient of variation CVτ of the attenuation constant for N pulses is calculated. σv and CVτ are used as the second stability parameters.

[0087] S202: Calculate the first deviation of the first stability parameter and the second deviation of the second stability parameter respectively. The deviation characterizes the degree of deviation of the plasma stability parameter from the ideal stable state.

[0088] When calculating the first deviation, the ideal reference value of the first stability parameter is first determined. Reference values ​​σx0 and σy0 for the standard deviation of the centroid coordinates, RSDd0 for the relative standard deviation of the equivalent diameter, and CVg0 for the coefficient of variation of brightness are obtained statistically from historical high-quality laser-induced breakdown spectrum measurement data. The normalized difference between the actual measured value and the reference value is calculated. The first deviation D1 is calculated using the formula D1=w1×(σx-σx0) / σx0+w2×(σy-σy0) / σy0+w3×(RSDd-RSDd0) / RSDd0+w4×(CVg-CVg0) / CVg0, where w1, w2, w3, and w4 are weighting coefficients and satisfy w1+w2+w3+w4=1. When calculating the second deviation, the baseline value of the standard deviation of the expansion rate σv0 and the baseline value of the coefficient of variation of the attenuation constant CVτ0 are determined. The second deviation D2 is calculated by the formula D2=w5×(σv-σv0) / σv0+w6×(CVτ-CVτ0) / CVτ0, where w5 and w6 are weighting coefficients and satisfy w5+w6=1.

[0089] S203: Construct a spectral quality value based on the first deviation and the second deviation. Classify the plasma spectral signals with spectral quality values lower than the preset quality threshold into the high-confidence spectral set, and classify the plasma spectral signals with spectral quality values higher than or equal to the preset quality threshold into the low-confidence spectral set.

[0090] Use the weighted linear combination method to construct the spectral quality value Q. The calculation formula is Q = α×D1 + β×D2, where α and β are weight coefficients and satisfy α + β = 1. The weight coefficients are determined according to the importance of the influence of spatial stability and time stability on spectral quality and are calculated by the correlation analysis or principal component analysis method. Set the preset quality threshold Qth, and calculate the corresponding spectral quality value Q for each of the batch-collected plasma spectral signals one by one. When Q < Qth, it is determined that the plasma state stability of the spectral signal meets the requirements, and the spectral signal and its corresponding spectral line intensity data are classified into the high-confidence spectral set. When Q ≥ Qth, it is determined that the spectral signal is greatly interfered by plasma instability factors, and the spectral signal is classified into the low-confidence spectral set for subsequent working condition correction processing. The preset quality threshold Qth is determined by repeating the measurement experiment on the standard sample, and the critical Q value that makes the relative standard deviation of the spectral line intensity of the high-confidence spectral set less than 10% is selected as the threshold.

[0091] Based on the above embodiments, as an optional embodiment, step S106: Predict the remaining life of the fan gearbox based on the spectral evolution pattern and modulation relationship corresponding to the wear stage, combined with the current operating condition parameters, may specifically include the following steps:

[0092] S301: Extract the wear evolution rate from the spectral evolution pattern corresponding to the current wear stage; obtain the working condition modulation coefficient of the operating condition parameters on the wear evolution rate according to the modulation relationship, and use the working condition modulation coefficient to correct the wear evolution rate to obtain the wear evolution rate after working condition correction.

[0093] The slope of the spectral intensity-time curve for the current wear stage is read from the spectral evolution pattern corresponding to that stage, or the derivative of the spectral intensity with respect to time, dI / dt, is calculated as the wear evolution rate v0. The current operating parameters, including rotational speed n, load power P, and lubricating oil temperature T, are obtained. These operating parameters are substituted into a pre-established modulation relationship function f(n, P, T) to calculate the operating condition modulation coefficient k = f(n, P, T). The modulation relationship function is established by comparing the differences in wear evolution rates under different operating conditions, for example, using multiple linear regression k = a0 + a1×n + a2×P + a3×T, or polynomial regression k = b0 + b1×n + b2×n² + b3×P + b4×P×T. The wear evolution rate v0 is multiplied by the operating condition modulation coefficient k to calculate the corrected wear evolution rate v = k×v0, achieving an adaptive correction to the standard evolution rate.

[0094] S302: Calculate the evolution time required for the spectral intensity of the wear indicator element to evolve from the current value to the preset wear failure threshold based on the wear evolution rate after working condition correction, and use the evolution time as the remaining life of the wind turbine gearbox.

[0095] Wear indicator elements refer to metallic elements such as Fe, Cr, and Ni that can reflect the wear state of the gearbox. Their concentration in lubricating oil is positively correlated with the degree of wear, and their spectral intensity is detected by laser-induced breakdown spectroscopy.

[0096] Obtain the current measured value Ic of the spectral line intensity of the wear indicator element and the preset wear failure threshold If. Calculate the spectral line intensity difference ΔI = If - Ic, which represents the remaining evolution space from the current state to the failure state. Obtain the wear evolution rate v after operating condition correction from step S301. This rate has already considered the impact of the current operating conditions on the wear process. Divide the spectral line intensity difference by the operating condition-corrected wear evolution rate and calculate the evolution time t using the formula t = ΔI / v. Output the calculated evolution time t directly as the remaining life prediction result of the wind turbine gearbox.

[0097] Based on the above embodiments, as an optional embodiment, S302: the step of calculating the evolution time required for the spectral intensity of the wear indicator element to evolve from its current value to a preset wear failure threshold based on the wear evolution rate after working condition correction may specifically include the following steps:

[0098] S401: Identify the sequence of subsequent wear stages required for the wear indicator element to evolve from the current wear stage to the wear failure threshold based on the spectral evolution pattern; for each wear stage in the subsequent wear stage sequence, extract the stage termination spectral line intensity and stage wear evolution rate of each wear stage from the corresponding spectral evolution pattern.

[0099] Based on the preset wear stage division rules and the current spectral line intensity value Ic, the position interval of the current wear stage is located in the spectral line evolution mode curve, and the subsequent wear stages that need to be experienced from the current stage to the failure threshold If are identified. The identified wear stages are arranged in chronological order as a sequence of subsequent wear stages {stage 1, stage 2, ..., stage m}. For each wear stage i in the sequence, the spectral line intensity value corresponding to the end of the stage is read from the piecewise function or curve of the spectral line evolution mode, as the stage termination spectral line intensity Ii_end. The slope of the spectral line intensity-time curve for this stage is calculated, or the derivative of the evolution function for this stage is taken to obtain the stage wear evolution rate vi.

[0100] S402: For the first wear stage in the subsequent wear stage sequence, calculate the first-stage spectral line intensity difference between the current value of the spectral line intensity of the wear indicator element and the stage termination spectral line intensity of the first wear stage. Based on the modulation relationship and the operating condition parameters at the current time, obtain the first-stage operating condition modulation coefficient for calculating the evolution rate of the first wear stage. Use the first-stage operating condition modulation coefficient to correct the stage wear evolution rate of the first wear stage to obtain the first-stage corrected evolution rate. Divide the first-stage spectral line intensity difference by the first-stage corrected evolution rate to obtain the evolution duration of the first stage.

[0101] The first-stage spectral intensity difference represents the increment of the spectral intensity value from the current value to the end of the first wear stage, reflecting the evolution range required for the first stage. The first-stage operating condition modulation factor represents the correction factor of the operating condition parameters corresponding to the first wear stage to the standard wear evolution rate of that stage, used to reflect the impact of operating condition differences. The first-stage corrected evolution rate represents the actual evolution rate of the first wear stage after correction by the operating condition modulation factor, reflecting the true wear rate of that stage under specific operating conditions. The first-stage evolution duration represents the time required to complete the first wear stage evolution, and is the first time component in the remaining lifetime calculation.

[0102] Obtain the current value Ic of the wear indicator element spectral line intensity and the stage termination spectral line intensity I1_end of the first wear stage extracted from step S401. Calculate the spectral line intensity difference of the first stage using the formula ΔI1=I1_end-Ic. Obtain the operating condition parameters corresponding to the current moment, including the expected rotational speed n1, load power P1, and lubricating oil temperature T1 for this stage. These parameters are obtained from historical operating data statistics or operating plans. Substitute the operating condition parameters into the modulation relationship function to calculate the first stage operating condition modulation coefficient k1=f(n1, P1, T1). Multiply the stage wear evolution rate v1 of the first wear stage obtained from step S401 by the operating condition modulation coefficient k1, and calculate the first stage corrected evolution rate using the formula v1'=k1×v1. Divide the first stage spectral line intensity difference ΔI1 by the first stage corrected evolution rate v1', and calculate the first stage evolution duration t1 using the formula t1=ΔI1 / v1'.

[0103] S403: For each remaining wear stage in the subsequent wear stage sequence except the first wear stage, calculate the difference in spectral intensity between the stage termination spectral intensity of the previous wear stage and the stage termination spectral intensity of the current wear stage. Predict the operating parameters at the start of the current wear stage based on the stage evolution duration of the previous wear stage. Determine the corresponding current stage operating condition modulation coefficient based on the operating condition parameters at the start of the current wear stage. Use the current stage operating condition modulation coefficient to correct the stage wear evolution rate of the current wear stage to obtain the corrected evolution rate of the current stage. Divide the current stage spectral intensity difference by the corrected evolution rate of the current stage to obtain the stage evolution duration of the current wear stage. Summate the evolution duration of the first stage with the stage evolution durations of all remaining wear stages to obtain the evolution time.

[0104] The remaining wear stages refer to all subsequent stages in the wear stage sequence except the first wear stage, including stages two through m. The spectral intensity difference of the current stage represents the increment of spectral intensity from the start to the end of the currently processed wear stage, equal to the evolution amplitude of that stage. The start time of the current wear stage represents the time point at which the current processing stage begins, obtained by adding the evolution duration of all previously calculated stages to the current time. The current stage operating condition modulation coefficient represents the correction coefficient for the corresponding operating condition parameters of the current processing stage, used to adjust the evolution rate of that stage. The current stage corrected evolution rate represents the actual evolution rate of the current stage after correction by the operating condition modulation coefficient, reflecting the wear rate of that stage under the expected operating conditions.

[0105] Starting from the second wear stage, perform the following calculations sequentially for each stage i (i=2, 3, ..., m) in the subsequent wear stage sequence. Calculate the spectral intensity difference ΔIi=Ii_end-I(i-1)_end of the current stage, where Ii_end is the terminating spectral intensity of the current stage, and I(i-1)_end is the terminating spectral intensity of the previous stage. Calculate the start time of the current stage using the formula Tstart_i=t1+t2+...+t(i-1), where t1 to t(i-1) are the evolution durations of the preceding stages. Based on the start time Tstart_i of the current stage, obtain the corresponding operating condition parameters ni, Pi, and Ti from the operating condition prediction model or historical operating condition statistics. Substitute the operating condition parameters into the modulation relationship function to calculate the current stage operating condition modulation coefficient ki=f(ni, Pi, Ti). Obtain the stage wear evolution rate vi of the current stage from step S401, and calculate the corrected evolution rate of the current stage using the formula vi'=ki×vi. Divide the current stage spectral intensity difference ΔIi by the current stage corrected evolution rate vi', and calculate the current stage evolution duration ti using the formula ti=ΔIi / vi'. Sum the evolution durations of all wear stages and calculate the total evolution time T using the formula T=t1+t2+...+tm. This total evolution time serves as the remaining lifetime prediction result considering multi-stage evolution and operating condition effects.

[0106] Based on the above embodiments, as an optional embodiment, S103: Extracting the spectral evolution patterns of wear indicator elements at multiple time points from a high-confidence spectral set, whereby the spectral evolution patterns reflect the spectral intensity changes at different wear stages, this step also includes the case of an oil change event, specifically including the following steps:

[0107] S501: Calculate the change in spectral intensity of the wear indicator element between adjacent time points, and compare the change in spectral intensity with the expected change in spectral intensity for the corresponding time interval in the spectral evolution model.

[0108] Obtain the spectral line intensity measurement values \(I_n\) and \(I_{n + 1}\) at two adjacent time points \(t_n\) and \(t_{n+1}\), and calculate the change in spectral line intensity \(\Delta I=I_{n + 1}-I_n\) through subtraction. Extract the evolution rate \(v(t_n)\) for the corresponding time period from the established spectral line evolution pattern, which is obtained from the derivative or piecewise slope of the evolution curve. Multiply the evolution rate \(v(t_n)\) by the time interval \(\Delta t\) to calculate the expected change in spectral line intensity \(\Delta I_{exp}=v(t_n)\times\Delta t\). Calculate the difference \(\delta =|\Delta I-\Delta I_{exp}|\) or the ratio \(r=\Delta I / \Delta I_{exp}\) between the actual change and the expected change. When the difference \(\delta\) exceeds the preset deviation threshold or the ratio \(r\) deviates from 1 by more than the preset range, it is determined that the actual evolution deviates from the expected evolution pattern. For example, if the time interval is 10 hours and the evolution rate is 2 intensity units per hour, the expected change is 20 intensity units. If the actual change is -15 intensity units and the difference reaches 35 intensity units, it indicates a significant abnormal change.

[0109] S502: When the absolute value of the change in spectral line intensity exceeds the preset sudden drop threshold, identify the lubricant oil change event and record the occurrence time of the lubricant oil change event; mark the high-confidence spectral set and the low-confidence spectral set before the occurrence time of the lubricant oil change event as the historical spectral data set before oil change, and use the spectral line evolution pattern corresponding to the historical spectral data set before oil change as the wear evolution benchmark before oil change.

[0110] Judge whether the absolute value \(|\Delta I|\) of the change in spectral line intensity \(\Delta I\) calculated in step S501 exceeds the preset sudden drop threshold \(T_{hdrop}\), which is set as a negative change amount with a relatively large absolute value, for example, \(T_{hdrop} = 50\) intensity units. When \(|\Delta I|>T_{hdrop}\) and \(\Delta I<0\), it is determined that a lubricant oil change event occurs at time point \(t_{n + 1}\), and record this time as the oil change time \(T_{oil\_change}=t_{n + 1}\). Retrieve all the high-confidence spectral sets and low-confidence spectral sets stored in the system, screen all the spectral data with the acquisition time \(t < T_{oil\_change}\), and uniformly mark these data as the historical spectral data set before oil change \(D_{pre\_oil}\). Extract the spectral line intensity sequence \(\{I_1, I_2,\cdots, I_k\}\) and the corresponding time sequence \(\{t_1, t_2,\cdots, t_k\}\) of the wear-indicating elements from the high-confidence spectral set corresponding to the historical spectral data set before oil change \(D_{pre\_oil}\), and establish the spectral line intensity-time relationship function \(I_{pre}(t)\) through curve fitting or piecewise linear fitting. This function serves as the wear evolution benchmark before oil change.

[0111] S503: Based on the plasma spectral signals and plasma image signals collected after the occurrence time of the lubricant oil change event, re-perform quality grading to obtain the high-confidence spectral set after oil change and the low-confidence spectral set after oil change, and extract the spectral line evolution pattern after oil change from the high-confidence spectral set after oil change.

[0112] All plasma spectral and image signals acquired at times t ≥ Toil_change were selected and used as the original dataset after oil change. Plasma morphology parameters, including plasma area, aspect ratio, and eccentricity, were extracted from each plasma image signal after oil change to determine if each morphology parameter fell within a preset acceptable range. The spectral signal-to-noise ratio (SNR) was calculated for each plasma spectral signal after oil change using the formula SNR = Isignal / Inoise, where Isignal is the peak intensity of the spectral line and Inoise is the baseline noise standard deviation. Spectra with all plasma morphology parameters meeting the acceptable range and an SNR higher than the quality threshold SNRth were classified into the high-confidence spectrum set Dhigh_post after oil change, while spectra that did not meet the criteria were classified into the low-confidence spectrum set Dlow_post after oil change. The spectral intensity sequence {Ik+1, Ik+2, ..., Im} and the corresponding time sequence {tk+1, tk+2, ..., tm} of wear indicator elements were extracted from the high-confidence spectral set Dhigh_post after oil change. The time was calculated as the relative running time starting from the oil change moment Toil_change. The spectral intensity-time relationship function Ipost(t-Toil_change) was established using polynomial fitting, exponential fitting, or piecewise linear fitting methods. This function serves as the spectral evolution mode after oil change.

[0113] S504: Compare the wear evolution baseline before oil change with the spectral line evolution mode after oil change under the same operating time to evaluate the wear state change trend of the wind turbine gearbox before and after oil change.

[0114] Compare the wear evolution baseline Ipre(t) before oil change and the spectral evolution model Ipost(t) after oil change using the same runtime interval, for example, a runtime of 100 hours. Calculate the derivative of the wear evolution baseline function Ipre(t) before oil change at t=100 or calculate the average slope of the interval from t=0 to t=100 to obtain the evolution rate vpre=(Ipre(100)-Ipre(0)) / 100. Calculate the derivative of the spectral evolution model function Ipost(t-Toil_change) after oil change at a relative runtime of 100 hours or calculate the average slope to obtain the evolution rate vpost=(Ipost(100)-Ipost(0)) / 100. Calculate the rate of change of evolution rate ratio = vpost / vpre or the difference in evolution rate Δv = vpost-vpre. When ratio < 1 or Δv < 0, it is determined that the wear rate slows down after oil change, and the lubricating oil change has a positive effect. When ratio ≈ 1 or Δv ≈ 0, it is determined that the oil change has little impact on the wear rate. When ratio > 1 or Δv > 0, it is determined that the wear rate accelerates after oil change, indicating that there is abnormal wear in the equipment or a problem with the quality of the lubricating oil.

[0115] Based on the above embodiments, as an optional embodiment, S106: after the step of matching the correction spectral line intensity with the spectral line evolution mode to identify the current wear stage of the wind turbine gearbox, the method further includes a step of adjusting the spectral line evolution mode corresponding to the conflicting wear indicator elements, which may specifically include the following steps:

[0116] S601: Extract the correction spectral intensity of each of the multiple different types of wear indicator elements, and match the correction spectral intensity of each of the multiple different types of wear indicator elements with the corresponding spectral evolution mode to obtain the wear stage determination result of each of the multiple different types of wear indicator elements.

[0117] Extract the characteristic spectral lines of each element such as iron, copper, and aluminum from the currently collected plasma spectral signal. Perform internal standard element intensity ratio correction and baseline drift correction on the spectral line intensities of each element to obtain the corrected spectral line intensity values of each element, such as IFe_corr, ICu_corr, IAl_corr, etc. Read the spectral evolution pattern corresponding to each element. This pattern includes segmented intervals such as the intensity range [Inormal_min, Inormal_max] in the normal wear stage, the intensity range [Iaccel_min, Iaccel_max] in the accelerated wear stage, and the intensity range [Isevere_min, Isevere_max] in the severe wear stage, etc. For the iron element, determine which intensity range IFe_corr falls into. For example, when Iaccel_min ≤ IFe_corr < Iaccel_max, it is determined that the wear stage identified for the iron element is the accelerated wear stage. Perform the same matching judgment process for other elements such as copper and aluminum to obtain the wear stage determination results identified for each element. For example, the iron element is determined to be in the accelerated wear stage, the copper element is determined to be in the normal wear stage, and the aluminum element is determined to be in the accelerated wear stage. Store the wear stage determination results of each element as a result set {StageFe, StageCu, StageAl,...} for subsequent consistency analysis and comprehensive determination.

[0118] S602: Count the number of elements with consistent wear stage determination results among multiple different types of wear indicator elements, and use the ratio of the number of elements with consistent wear stage determination results to the total number of multiple different types of wear indicator elements as the stage identification consistency index.

[0119] Count the frequencies of each wear stage that appear in the wear stage determination result set {StageFe, StageCu, StageAl, StageCr,...} obtained in step S601. Calculate the number of elements corresponding to each wear stage. For example, the number of elements determined to be in the normal wear stage is Nnormal, the number of elements determined to be in the accelerated wear stage is Naccel, and the number of elements determined to be in the severe wear stage is Nsevere. Select the wear stage with the highest frequency as the dominant determination result, and the number of elements corresponding to this stage is the number of elements with consistent wear stage determination results Nmax = max(Nnormal, Naccel, Nsevere). Count the total number of multiple different types of wear indicator elements Ntotal. For example, if four elements, iron, copper, aluminum, and chromium, are detected, then Ntotal = 4. Calculate the stage identification consistency index Consistency = Nmax / Ntotal. For example, if three of the four elements are determined to be in the accelerated wear stage and one is determined to be in the normal wear stage, then Nmax = 3, Ntotal = 4, and Consistency = 3 / 4 = 0.75.

[0120] S603: When the stage recognition consistency index is lower than the preset consistency threshold, identify the wear indication elements with conflicting wear stage determination results, and adjust the spectral evolution pattern corresponding to the conflicting wear indication elements.

[0121] Judge whether the stage recognition consistency index Consistency calculated in step S602 is lower than the preset consistency threshold Thconsistency, for example, Thconsistency = 0.65. When Consistency < Thconsistency, start the conflict recognition process. Determine the wear stage with the highest frequency of occurrence in step S602 as the dominant determination result Stagemajor, traverse the set of wear stage determination results of all elements {StageFe, StageCu, StageAl, StageCr}, and mark the elements whose determination results are not equal to Stagemajor as the wear indication elements with conflicts, forming a conflict element set. For example, if the dominant determination result is accelerated wear and the copper element is determined to be normal wear, then the copper element is marked as a conflict element. For each element in the conflict element set, extract the spectral evolution pattern parameters corresponding to this element, including the boundary values of the intensity range of each stage and the evolution rate threshold. The methods for adjusting the spectral evolution pattern include: refitting the evolution curve based on the historical highly reliable spectral data of this element to calculate the new boundary values of the stage intensity range; correcting the stage division standard by referring to the element evolution statistical data of the same type of equipment; adjusting the evolution rate threshold according to the current operating conditions of the equipment. After the adjustment is completed, use the adjusted spectral evolution pattern to re - execute the matching determination process in step S601 to verify whether the adjustment effect improves the consistency index.

[0122] Based on the above - mentioned embodiments, as an optional embodiment, step S603: adjusting the spectral evolution pattern corresponding to the wear indication elements with conflicts may specifically include the following steps:

[0123] S701: Calculate the matching degrees between the corrected spectral intensities of multiple different types of wear indication elements and the spectral intensity ranges corresponding to the respective wear stages determined by the wear indication elements with conflicts, and select the wear stage determination result corresponding to the wear indication element with the highest matching degree as the corrected current wear stage.

[0124] For each type of wear indicator element, calculate the match degree between its corrected spectral line intensity and the intensity range of the current wear stage determined in step S601. The match degree calculation formula is: MatchDegree=1-|Icorr-Imid| / (Istage_max-Istage_min), where Icorr is the element's corrected spectral line intensity, Imid=(Istage_max+Istage_min) / 2 is the midpoint value of the stage intensity range, and Istage_max and Istage_min are the upper and lower limits of the stage intensity range. The match degree is 1 when the corrected spectral line intensity is exactly at the midpoint of the range, and the closer to the boundary, the lower the match degree. For example, the corrected spectral line intensity of iron element IFe_corr=150, the intensity range of the accelerated wear stage is [100, 200], the midpoint value is 150, and the match degree is 1-|150-150| / (200-100)=1. For all elements, including copper and aluminum, calculate the matching degree for their corresponding wear stage, resulting in a matching degree set {MatchDegreeFe, MatchDegreeCu, MatchDegreeAl, ...}. Compare the matching degree values ​​of each element and select the element with the highest matching degree. For example, if MatchDegreeFe=1.0, MatchDegreeCu=0.4, and MatchDegreeAl=0.9, then iron is selected. The wear stage determination result corresponding to the iron element with the highest matching degree is used as the corrected current wear stage.

[0125] S702: For the corrected current wear stage, backtrack the spectral intensity evolution trends of multiple different types of wear indicator elements over a historical preset time period, identify abnormal elements that deviate from the normal evolution trend, mark the spectral evolution modes corresponding to the abnormal elements as abnormal evolution modes, and exclude them from subsequent remaining lifetime predictions.

[0126] Read the corrected wear stage Stagefinal determined in step S701, and extract the normal evolution trend parameters from the standard spectral evolution pattern corresponding to this stage, including the evolution direction (rising or falling) and the normal evolution rate range [vnormal_min, vnormal_max]. Set a preset historical time period length Thyback, for example, Thyback=30 days, and backtrack from the current time tcurrent to the time tstart=tcurrent-Thyback. For each wear indicator element, extract the corrected spectral intensity sequence {I1, I2, ..., Ik} and the corresponding time series {t1, t2, ..., tk} of the element from the high-confidence spectrum set within the time period [tstart, tcurrent]. Perform linear fitting on the intensity-time series I(t)=a×t+b, and calculate the slope a of the fitted line as the actual evolution rate void of the element. Determine whether the actual evolution rate void satisfies: the evolution direction is consistent with the direction and vnormal_min≤vactual≤vnormal_max. When an element does not meet the above conditions, it is marked as an anomalous element. For example, copper should show an increasing trend and a rate greater than 5 intensity units / day during the accelerated wear phase, but the measured slope is -2, indicating a decrease, thus copper is determined to be an anomalous element. The spectral evolution mode corresponding to the anomalous element is marked as an anomalous evolution mode and the anomalous type is recorded. In subsequent remaining lifetime prediction calculations, the evolution data of this element is excluded, and only the evolution data of normal elements are used for lifetime prediction.

[0127] The following describes an exemplary device for detecting lubricating oil wear elements provided in an embodiment of this application. Figure 4 This is an exemplary hardware structure diagram of a device detection device based on lubricating oil wear elements provided in an embodiment of this application.

[0128] In some embodiments, the device for detecting lubricating oil wear elements is a computer device or includes a computer device in the device for detecting lubricating oil wear elements. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface; in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it implements the methods in the embodiments of this application.

[0129] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0130] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0131] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0132] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0133] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for detecting a device based on a wear element of lubricating oil, characterized by, The method includes: Plasma spectral signals and plasma image signals of lubricating oil samples at multiple time points of the wind turbine gearbox were acquired. The lubricating oil samples contained multiple different types of wear indicator elements generated by the wear of internal components of the wind turbine gearbox. The stability parameters of the plasma are extracted from the plasma image signal, and the plasma spectral signal is divided into a high-confidence spectral set and a low-confidence spectral set based on the stability parameters. The spectral evolution patterns of the wear indicator elements at multiple time points are extracted from the high-confidence spectral set, and the spectral evolution patterns reflect the changes in spectral intensity at different wear stages. The operating condition parameters of the wind turbine gearbox at multiple time points are obtained, and the modulation relationship between the operating condition parameters and the spectral evolution mode is established. Based on the modulation relationship, the intensity of spectral lines in the low-confidence spectrum set is corrected for operating conditions to obtain the corrected spectral line intensity. The intensity of the corrected spectral line is matched with the spectral line evolution mode to identify the current wear stage of the wind turbine gearbox. Based on the spectral line evolution mode and modulation relationship corresponding to the wear stage, and combined with the current operating condition parameters, the remaining life of the wind turbine gearbox is predicted.

2. The apparatus detection method based on lubricating oil wear elements according to claim 1, characterized in that, The process of dividing the plasma spectral signal into a high-confidence spectral set and a low-confidence spectral set based on the stability parameter includes: Extract a first stability parameter reflecting the spatial distribution characteristics of the plasma and a second stability parameter reflecting the temporal evolution characteristics of the plasma from the plasma image signal; The first deviation of the first stability parameter and the second deviation of the second stability parameter are calculated respectively. The deviation characterizes the degree of deviation of the plasma stability parameter from the ideal stable state. Based on the first deviation and the second deviation, a spectral quality value is constructed. Plasma spectral signals with spectral quality values ​​lower than a preset quality threshold are classified into a high-confidence spectral set, and plasma spectral signals with spectral quality values ​​higher than or equal to the preset quality threshold are classified into a low-confidence spectral set.

3. The apparatus detection method based on lubricating oil wear elements according to claim 1, characterized by, The prediction of the remaining lifespan of the wind turbine gearbox based on the spectral evolution mode and modulation relationship corresponding to the wear stage, combined with the current operating condition parameters, includes: Extract the wear evolution rate from the spectral evolution mode corresponding to the current wear stage; The operating condition modulation coefficient of the operating condition parameters on the wear evolution rate is obtained according to the modulation relationship. The operating condition modulation coefficient is used to correct the wear evolution rate to obtain the corrected wear evolution rate. Based on the wear evolution rate after the working condition correction, the evolution time required for the spectral intensity of the wear indicator element to evolve from the current value to reach the preset wear failure threshold is calculated, and the evolution time is taken as the remaining life of the wind turbine gearbox.

4. The equipment detection method based on lubricating oil wear elements according to claim 3, characterized in that, The calculation of the evolution time required for the spectral intensity of the wear indicator element to evolve from its current value to a preset wear failure threshold based on the wear evolution rate after the working condition correction includes: Identify the sequence of subsequent wear stages required for the wear indicator element to evolve from the current wear stage to the wear failure threshold based on the spectral line evolution pattern; For each wear stage in the subsequent wear stage sequence, the stage termination spectral line intensity and stage wear evolution rate of each wear stage are extracted from the corresponding spectral line evolution mode; For the first wear stage in the subsequent wear stage sequence, calculate the first stage spectral line intensity difference between the current value of the spectral line intensity of the wear indicator element and the stage termination spectral line intensity of the first wear stage. Based on the modulation relationship and the operating condition parameters at the current time, obtain the first stage operating condition modulation coefficient for calculating the evolution rate of the first wear stage. Use the first stage operating condition modulation coefficient to correct the stage wear evolution rate of the first wear stage to obtain the first stage corrected evolution rate. Divide the first stage spectral line intensity difference by the first stage corrected evolution rate to obtain the evolution duration of the first stage. For each remaining wear stage in the subsequent wear stage sequence excluding the first wear stage, the intensity difference of the current stage spectral line between the stage termination spectral line intensity of the previous wear stage and the stage termination spectral line intensity of the current wear stage is calculated. Based on the stage evolution duration of the previous wear stage, the operating condition parameters at the start time of the current wear stage are predicted. Based on the operating condition parameters at the start time of the current wear stage, the corresponding current stage operating condition modulation coefficient is determined. The stage wear evolution rate of the current wear stage is corrected using the current stage operating condition modulation coefficient to obtain the current stage corrected evolution rate. The stage evolution duration of the current wear stage is obtained by dividing the current stage spectral line intensity difference by the current stage corrected evolution rate. The evolution time is obtained by summing the evolution duration of the first stage with the stage evolution durations of all remaining wear stages.

5. The apparatus detection method based on lubricating oil wear elements according to claim 1, characterized by, The method further includes: Calculate the change in spectral intensity of the wear indicator element between adjacent time points, and compare the change in spectral intensity with the expected change in spectral intensity for the corresponding time interval in the spectral evolution mode; When the absolute value of the change in spectral line intensity exceeds a preset sudden drop threshold, a lubricating oil replacement event is identified, and the time of occurrence of the lubricating oil replacement event is recorded. The high-confidence spectral set and the low-confidence spectral set prior to the time of the lubricating oil replacement event are marked as the pre-oil change historical spectral dataset, and the spectral line evolution mode corresponding to the pre-oil change historical spectral dataset is used as the pre-oil change wear evolution benchmark. Based on the plasma spectral signal and plasma image signal collected after the lubricating oil replacement event, quality classification is re-performed to obtain a high-confidence spectral set and a low-confidence spectral set after the oil change. The spectral evolution mode after the oil change is extracted from the high-confidence spectral set after the oil change. By comparing the wear evolution benchmark before oil change with the spectral line evolution mode after oil change at the same operating time, the wear state change trend of the wind turbine gearbox before and after oil change is evaluated.

6. The apparatus detection method based on lubricating oil wear elements according to claim 1, characterized by, After matching the corrected spectral line intensity with the spectral line evolution mode to identify the current wear stage, the method further includes: The correction spectral intensity of each of the multiple different types of wear indicator elements is extracted, and the correction spectral intensity of each of the multiple different types of wear indicator elements is matched with the corresponding spectral evolution mode to obtain the wear stage determination result identified by each of the multiple different types of wear indicator elements. The number of elements with consistent wear stage determination results among the multiple different types of wear indicator elements is counted, and the ratio of the number of elements with consistent wear stage determination results to the total number of multiple different types of wear indicator elements is used as the stage identification consistency index. When the consistency index of the stage identification is lower than the preset consistency threshold, wear indicator elements that conflict with the wear stage determination results are identified, and the spectral evolution mode corresponding to the conflicting wear indicator elements is adjusted.

7. The apparatus detection method based on lubricating oil wear elements according to claim 6, characterized in that, The adjustment of the spectral line evolution mode corresponding to the conflicting wear indicator elements includes: Calculate the corrected spectral intensity of each of the multiple different types of wear indicator elements, and the matching degree between it and the spectral intensity range corresponding to each wear stage determined by the conflicting wear indicator elements. Select the wear stage determination result corresponding to the wear indicator element with the highest matching degree as the corrected current wear stage. For the corrected current wear stage, the evolution trend of spectral intensity of the multiple different types of wear indicator elements over a historical preset time period is traced back. Abnormal elements that deviate from the normal evolution trend are identified, and the spectral evolution mode corresponding to the abnormal element is marked as an abnormal evolution mode and excluded from the subsequent remaining lifetime prediction.

8. A device for detecting a device based on a wear element of a lubricating oil, characterized by The device detection device based on lubricating oil wear elements includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the device detection device based on lubricating oil wear elements to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on a device testing equipment based on lubricating oil wear elements, the device testing equipment based on lubricating oil wear elements performs the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on a device for detecting lubricating oil wear elements, the device for detecting lubricating oil wear elements performs the method as described in any one of claims 1-7.

Images