Lubricating oil equipment state monitoring and early warning method and system

Abstract

The invention relates to the technical field of equipment testing, and discloses a lubricating oil equipment state monitoring and early warning method and system, and the method comprises the steps: synchronously collecting a time domain vibration signal and an operation temperature signal of lubricating oil equipment; performing fast Fourier transform on the time-domain vibration signal, converting the time-domain vibration signal into a frequency-domain vibration signal, and extracting a characteristic frequency component representing the operation state of the equipment; calculating a slope value of temperature change of the operation temperature signal in unit time to obtain a temperature dynamic change trend; inputting the characteristic frequency component and the temperature dynamic change trend into a preset weight fusion model to generate a comprehensive health degree state index; dynamically updating an early warning threshold value; comparing the comprehensive health degree state index with an early warning threshold value in real time, and judging a real-time state according to a comparison result; when the real-time state is in an abnormal state, multi-stage early warning is triggered; according to the invention, the accuracy of lubricating oil equipment state monitoring and early warning can be improved.

Description

Technical Field

[0001] This invention relates to the field of equipment testing technology, and in particular to a method and system for monitoring and early warning of lubricating oil equipment condition. Background Technology

[0002] Existing technologies have significant shortcomings in the signal processing and feature extraction stages of lubricating oil equipment condition monitoring. They fail to simultaneously collect and analyze time-domain vibration signals and operating temperature signals, instead only collecting a single type of signal or simply superimposing data, ignoring the synergistic influence of vibration and temperature parameters on equipment condition, and thus failing to comprehensively capture the characteristics of equipment operating status. Furthermore, they do not convert time-domain vibration signals to frequency-domain signals using fast Fourier transform and extract characteristic frequency components, relying solely on simple statistical analysis of time-domain signals, which makes it difficult to accurately identify frequency characteristics related to equipment faults. At the same time, they do not calculate the slope value of temperature changes through linear regression analysis, relying only on fixed temperature thresholds, which cannot accurately reflect the dynamic trend of temperature changes, resulting in incomplete and inaccurate characterization of equipment condition features.

[0003] Existing technologies have significant shortcomings in the health status assessment and early warning threshold setting stages. They fail to construct a weighted fusion model to nonlinearly integrate characteristic frequency components and dynamic temperature trends, relying instead on simple weighting or single-dimensional assessment to generate a health index, which cannot objectively reflect the overall health status of the equipment. Furthermore, they do not dynamically update early warning thresholds based on historical operating data, using only fixed thresholds for status determination, which cannot adapt to the changing state patterns during equipment operation, easily leading to delayed or false alarms. Finally, they lack a multi-level threshold system for status determination and do not combine state transition continuity with historical data for reliability assessment, simply comparing threshold values ​​to output status results, resulting in low accuracy and difficulty in precisely distinguishing between normal, warning, and abnormal states, thus failing to provide timely and reliable monitoring and early warning support for lubrication equipment. Summary of the Invention

[0004] This invention provides a method and system for monitoring and early warning of lubricating oil equipment status, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a method for monitoring and early warning of lubricating oil equipment condition, comprising:

[0006] S1. Synchronously acquire time-domain vibration signals and operating temperature signals of the lubrication equipment;

[0007] S2. Perform a fast Fourier transform on the time-domain vibration signal to convert the time-domain vibration signal into a frequency-domain vibration signal, and extract the characteristic frequency components that characterize the operating state of the equipment from the frequency-domain vibration signal.

[0008] S3. Calculate the slope of the temperature change of the operating temperature signal per unit time to obtain the dynamic temperature change trend of the operating temperature signal;

[0009] S4. Input the characteristic frequency components and the temperature dynamic change trend into a preset weighted fusion model to generate the comprehensive health status index of the lubricating oil equipment.

[0010] S5. Based on the status data of the lubricating oil equipment during the historical operating cycle, dynamically update the early warning threshold;

[0011] S6. Compare the comprehensive health status index with the warning threshold in real time, and determine the real-time status of the lubricating oil equipment based on the comparison result;

[0012] S7. When the real-time status is in an abnormal state, trigger a multi-level early warning for the lubricating oil equipment.

[0013] In a preferred embodiment, the step of performing a Fast Fourier Transform on the time-domain vibration signal to convert it into a frequency-domain vibration signal, and extracting characteristic frequency components representing the operating state of the equipment from the frequency-domain vibration signal, includes:

[0014] The power spectral density of the frequency domain vibration signal is estimated to obtain the power spectral density distribution of the frequency domain vibration signal;

[0015] Based on the power spectral density distribution, the characteristic frequency components of the device's operating state are calculated using the spectral centroid algorithm. The calculation formula for the characteristic frequency components is as follows:

[0016] ;

[0017] In the formula, The characteristic frequency component, For the first Frequency values ​​at each frequency point To indicate at frequency The power spectral density amplitude at that point, To analyze the total number of sampling points within the frequency band;

[0018] The characteristic frequency components are normalized to obtain the standardized characteristic frequency index of the lubricating oil equipment.

[0019] The standardized characteristic frequency index is used as the characteristic frequency component characterizing the operating state of the equipment.

[0020] In a preferred embodiment, calculating the slope of the temperature change of the operating temperature signal per unit time to obtain the dynamic temperature change trend of the operating temperature signal includes:

[0021] The operating temperature signal is discretized over time to obtain a sequence of temperature sample values ​​for the operating temperature signal.

[0022] Based on the temperature sampling value sequence, the slope value of the dynamic temperature change trend is calculated through linear regression analysis, wherein the formula for calculating the slope value is as follows:

[0023] ;

[0024] In the formula, The slope value is... For discrete time points, This is a temperature sample value. This is the average value of the temperature samples. It is the average value at different time points. This represents the total number of sampling points;

[0025] The slope value is smoothed to obtain the dynamic temperature change trend index of the operating temperature signal;

[0026] The temperature dynamic change trend index is mapped to the temperature dynamic change trend of the lubricating oil equipment.

[0027] In a preferred embodiment, the step of performing time-series discretization processing on the operating temperature signal to obtain a temperature sample value sequence of the operating temperature signal includes:

[0028] Based on the Nyquist sampling theorem, a fixed sampling window is set for the lubricating oil equipment;

[0029] Based on the fixed sampling window, the timing deviation of the operating temperature signal is eliminated to obtain a regular temperature sampling value sequence of the operating temperature signal;

[0030] The integrity of the regularized temperature sample value sequence is verified to obtain the temperature sample value sequence of the operating temperature signal.

[0031] In a preferred embodiment, the step of inputting the characteristic frequency components and the dynamic temperature change trend into a preset weighted fusion model to generate a comprehensive health status index for the lubricating oil equipment includes:

[0032] The dimensions of the characteristic frequency components are unified to obtain the standard frequency characteristic index of the characteristic frequency components.

[0033] The temperature dynamic change trend is normalized to obtain a standard temperature trend index for the temperature dynamic change trend.

[0034] Based on the operating stage characteristics of the lubricating oil equipment, assign corresponding dynamic weight coefficients to the standardized frequency characteristic index and the standard temperature trend index;

[0035] Based on the dynamic weighting coefficient, the standard frequency characteristic index and the standard temperature trend index are nonlinearly integrated to obtain the preliminary comprehensive health status index of the lubricating oil equipment.

[0036] Based on a preset threshold, the preliminary comprehensive health status index is evaluated using a confidence level assessment based on the preset threshold to obtain the comprehensive health status index of the lubricating oil equipment.

[0037] In a preferred embodiment, the step of nonlinearly integrating the standard frequency characteristic index and the standard temperature trend index based on the dynamic weighting coefficient to obtain the preliminary comprehensive health status index of the lubricating oil equipment includes:

[0038] Based on the dynamic weighting coefficient, the standard frequency characteristic index and the standard temperature trend index are assigned weights in a hierarchical manner to obtain the weighted frequency characteristics and the weighted temperature trend.

[0039] A nonlinear feature transformation is performed on the weighted frequency characteristics and the weighted temperature trend to obtain the enhanced feature representation of the lubricating oil equipment.

[0040] The enhanced feature representations are then subjected to feature interaction fusion to obtain a preliminary comprehensive health status index of the lubricating oil equipment.

[0041] In a preferred embodiment, dynamically updating the early warning threshold based on the status data of the lubricating oil equipment during historical operating cycles includes:

[0042] Extract the state data sequence within the historical operating cycle from the operating database of the lubricating oil equipment to construct the historical state dataset of the lubricating oil equipment;

[0043] Outliers in the historical state dataset are removed to obtain the regularized historical state data of the historical state dataset.

[0044] Trend features are extracted from the regularized historical state data to identify the long-term trend features of state changes in the lubricating oil equipment;

[0045] Based on the long-term trend characteristics, a preliminary early warning threshold for the lubricating oil equipment is generated;

[0046] The rationality of the preliminary warning threshold is verified to obtain the dynamic warning threshold of the lubricating oil equipment.

[0047] In a preferred embodiment, the step of comparing the comprehensive health status index with the warning threshold in real time and determining the real-time status of the lubricating oil equipment based on the comparison result includes:

[0048] A multi-level state determination threshold system for the lubricating oil equipment is constructed to accurately divide the operating state of the lubricating oil equipment into normal state, warning state, and abnormal state.

[0049] The comprehensive health status index is compared with the multi-level status judgment threshold system in real time to obtain the preliminary judgment result of the status level of the lubricating oil equipment.

[0050] A state transition continuity analysis is performed on the preliminary state level determination results, and the reliability of the determination results is evaluated in conjunction with the historical operating status data of the lubricating oil equipment.

[0051] Based on the credibility assessment results, the real-time status judgment conclusion of the lubricating oil equipment is output, and a corresponding confidence index is assigned to each status level.

[0052] In a preferred embodiment, the step of outputting the real-time status judgment conclusion of the lubricating oil equipment based on the confidence assessment result, and assigning a corresponding confidence index to each status level, includes:

[0053] Based on the statistical distribution characteristics of the regularized historical state data, a mapping relationship system between the state level and confidence index in the lubricating oil equipment is established.

[0054] The basic confidence level of the lubricating oil equipment is determined based on the degree of consistency between the current state determination result and the regularized historical state data.

[0055] The baseline confidence level is dynamically adjusted based on the operating time, maintenance records, and operating parameters of the lubricating oil equipment.

[0056] Based on the vibration characteristics, temperature trends, and regularized historical state data of the lubricating oil equipment, a confidence index for the lubricating oil equipment is generated.

[0057] To address the above problems, the present invention also provides a lubricating oil equipment condition monitoring and early warning system, the system comprising:

[0058] The signal synchronization acquisition module is used to synchronously acquire the time-domain vibration signal and operating temperature signal of the lubricating oil equipment;

[0059] The frequency domain feature extraction module is used to perform a fast Fourier transform on the time domain vibration signal, convert the time domain vibration signal into a frequency domain vibration signal, and extract the characteristic frequency components representing the operating state of the equipment from the frequency domain vibration signal.

[0060] The temperature trend analysis module is used to calculate the slope of the temperature change of the operating temperature signal per unit time, so as to obtain the dynamic temperature change trend of the operating temperature signal.

[0061] The health index generation module is used to input the characteristic frequency components and the temperature dynamic change trend into a preset weighted fusion model to generate the comprehensive health status index of the lubricating oil equipment.

[0062] The threshold update module is used to dynamically update the warning threshold based on the status data of the lubricating oil equipment in the historical operating cycle.

[0063] The status determination module is used to compare the comprehensive health status index with the warning threshold in real time, and determine the real-time status of the lubricating oil equipment based on the comparison result.

[0064] The early warning triggering module is used to trigger multi-level early warnings for the lubricating oil equipment when the real-time status is in an abnormal state.

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] 1. This invention provides comprehensive and accurate basic data support for lubricating oil equipment condition monitoring through multi-dimensional signal precision processing and in-depth feature extraction. It simultaneously collects time-domain vibration signals and operating temperature signals to ensure the synergy of these two key status signals; it converts the time-domain vibration signals into frequency-domain signals using a Fast Fourier Transform, and extracts standardized characteristic frequency components through power spectral density estimation and a spectral centroid algorithm to accurately capture the frequency characteristics of equipment operation; it discretizes the operating temperature signal over time, calculates the slope value through linear regression analysis and smooths it, accurately reflecting the dynamic trend of temperature changes and laying a solid foundation for equipment health status assessment.

[0067] 2. This invention significantly improves the accuracy and timeliness of lubricant equipment condition monitoring and early warning by leveraging a comprehensive health assessment and dynamic early warning mechanism. It nonlinearly integrates characteristic frequency components and dynamic temperature change trends using a weighted fusion model, assigning dynamic weights based on equipment operation stages to generate an objective and comprehensive health status index. Based on historical operating data, outliers are removed, long-term trend features are extracted, and early warning thresholds are dynamically updated to adapt to equipment condition change patterns. A multi-level condition determination system is constructed, combining the continuity of condition transitions with historical data for reliability assessment, accurately determining the real-time equipment status, and triggering multi-level early warnings when anomalies occur, providing timely and reliable decision support for equipment maintenance. Attached Figure Description

[0068] Figure 1This is a flowchart illustrating a lubricating oil equipment condition monitoring and early warning method according to an embodiment of the present invention;

[0069] Figure 2 A functional block diagram of a lubricating oil equipment condition monitoring and early warning system provided in an embodiment of the present invention;

[0070] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0071] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0072] This application provides a method for monitoring and early warning of lubricating oil equipment status. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for monitoring and early warning of lubricating oil equipment status can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0073] Reference Figure 1 The diagram shown is a flowchart illustrating a lubricating oil equipment condition monitoring and early warning method according to an embodiment of the present invention. In this embodiment, the lubricating oil equipment condition monitoring and early warning method includes:

[0074] S1. Synchronously acquire time-domain vibration signals and operating temperature signals of the lubrication equipment;

[0075] In this embodiment of the invention, the time-domain vibration signal and operating temperature signal of the lubricating oil equipment are collected simultaneously.

[0076] Select vibration and temperature sensors suitable for the operating conditions of the lubrication equipment. The vibration sensor must be able to accurately capture the vibration frequency and amplitude of the equipment, and the temperature sensor must be able to stably detect the temperature change range during equipment operation. Fix the vibration sensor near the core moving parts of the lubrication equipment, ensuring close contact between the sensor and the equipment surface to avoid signal attenuation during vibration transmission. Install the temperature sensor at a critical location in the lubrication oil circulation path to ensure direct sensing of the actual temperature of the lubrication oil and the contact points with the equipment, providing a stable and reliable hardware foundation for signal acquisition.

[0077] A synchronous signal acquisition system was built, which includes a signal conditioning module, a data transmission module, and a timing synchronization module. The signal conditioning module is used to filter and amplify the raw signals acquired by the sensors to eliminate noise caused by external electromagnetic interference. The data transmission module adopts a wired transmission method to ensure that there is no delay or distortion during signal transmission. The timing synchronization module sets a unified acquisition start time and acquisition interval for the vibration sensor and the temperature sensor, so that the signal acquisition actions of the two sensors are completely synchronized, avoiding signal misalignment caused by time difference.

[0078] Start the lubrication equipment and, once it reaches a stable operating state, activate the synchronous signal acquisition system. Vibration sensors continuously monitor the vibration of the equipment's core moving parts, converting the mechanical fluctuations into electrical signals. After processing by the signal conditioning module, these signals form a time-domain vibration signal reflecting the equipment's vibration state. Simultaneously, temperature sensors continuously monitor the equipment's operating temperature, converting temperature changes into corresponding electrical signals. After processing by the signal conditioning module, these signals form an operating temperature signal reflecting the equipment's temperature status.

[0079] During signal acquisition, the timing synchronization module continuously monitors the acquisition rhythm of the two sensors to ensure that the corresponding time-domain vibration signal and operating temperature signal can be acquired simultaneously at each acquisition point. The acquisition system continuously records the two signals according to the preset acquisition duration. All signals are stored sequentially in the order of acquisition time, forming a complete and synchronized set of time-domain vibration signal and operating temperature signal. This ensures that the two signals can accurately correspond during subsequent analysis and fully reflect the correlation of the equipment's operating status.

[0080] The beneficial effects are that simultaneously acquiring time-domain vibration and operating temperature signals from lubrication equipment enables the simultaneous capture of core data on mechanical vibration and thermal changes during equipment operation. This ensures complete coordination of these two key state signals in the time dimension, laying the foundation for subsequent multi-dimensional collaborative analysis. Vibration signals can reflect potential mechanical faults such as wear and loosening of core moving parts, while temperature signals can reflect thermal operating characteristics such as lubricating oil circulation efficiency and equipment load changes. The simultaneous acquisition of these two types of signals avoids the limitations of single-signal analysis and can comprehensively cover the key characterization dimensions of equipment operating status.

[0081] The synchronously acquired signals, after processing, can accurately present the correlation between vibration and temperature parameters, such as the synergistic changes in vibration abnormalities and temperature rise caused by mechanical failures. This provides complete and mutually verifiable data support for subsequent feature fusion and health assessment, effectively improving the comprehensiveness and accuracy of equipment status monitoring, and providing a reliable data prerequisite for accurate fault identification and early risk warning.

[0082] S2. Perform a fast Fourier transform on the time-domain vibration signal to convert the time-domain vibration signal into a frequency-domain vibration signal, and extract the characteristic frequency components that characterize the operating state of the equipment from the frequency-domain vibration signal.

[0083] In this embodiment of the invention, the step of performing a fast Fourier transform on the time-domain vibration signal to convert the time-domain vibration signal into a frequency-domain vibration signal, and extracting characteristic frequency components representing the operating state of the equipment from the frequency-domain vibration signal, includes:

[0084] The power spectral density of the frequency domain vibration signal is estimated to obtain the power spectral density distribution of the frequency domain vibration signal;

[0085] Based on the power spectral density distribution, the characteristic frequency components of the device's operating state are calculated using the spectral centroid algorithm. The calculation formula for the characteristic frequency components is as follows:

[0086] ;

[0087] In the formula, The characteristic frequency component, For the first Frequency values ​​at each frequency point To indicate at frequency The power spectral density amplitude at that point, To analyze the total number of sampling points within the frequency band;

[0088] The characteristic frequency components are normalized to obtain the standardized characteristic frequency index of the lubricating oil equipment.

[0089] The standardized characteristic frequency index is used as the characteristic frequency component characterizing the operating state of the equipment.

[0090] A comprehensive signal analysis is performed on the frequency domain vibration signal. A sliding window is used to traverse the entire frequency domain vibration signal, calculating the energy distribution within each window. The signal amplitude at each frequency point is squared, and combined with the signal sampling period, the power value at each frequency point is calculated, thus forming a power spectral density distribution covering the entire frequency range. This distribution clearly shows the energy proportion of different frequency components in the frequency domain vibration signal, providing an intuitive energy distribution basis for subsequent feature extraction.

[0091] Based on the obtained power spectral density distribution, the centroid algorithm is used to calculate the characteristic frequency components. First, all effective frequency points and their corresponding power values ​​in the power spectral density distribution are determined. Then, using the power value of each frequency point as a weight, a weighted calculation is performed on all frequency points to find the frequency corresponding to the centroid of the power distribution. This frequency can centrally reflect the frequency component with the most concentrated energy in the frequency domain vibration signal, which is the core characteristic frequency of the equipment's operating state, thus obtaining the characteristic frequency components of the equipment's operating state.

[0092] A unified normalization method is adopted. First, the value range of the characteristic frequency component is determined, and the maximum and minimum values ​​of the component are identified. Based on the difference between the maximum and minimum values, all values ​​of the characteristic frequency component are scaled proportionally to a fixed standard range, strictly maintaining the relative magnitude relationship between each characteristic frequency component during the scaling process. Through this process, the influence of frequency value differences under different equipment or different operating conditions is eliminated, resulting in standardized characteristic frequency indicators for lubricating oil equipment.

[0093] The standardized characteristic frequency index after normalization is directly determined as the characteristic frequency component representing the equipment's operating status. This component not only retains the core frequency information reflecting the equipment's operating status in the original frequency domain vibration signal, but also achieves unified numerical standardization through standardization, avoiding interference from irrelevant factors. This ensures that it can accurately and stably reflect the actual operating status of the lubricating oil equipment, providing a reliable characteristic basis for subsequent equipment status assessment and fault diagnosis.

[0094] The calculation of characteristic frequency components relies on the power spectral density distribution of the frequency domain vibration signal. The frequency value of the i-th frequency point comes from the frequency division of the frequency domain vibration signal. By dividing the entire analysis frequency band of the frequency domain vibration signal into uniform intervals, each independent frequency position after division is a frequency point, and its corresponding specific frequency value is the specific content of the parameter.

[0095] The power spectral density amplitude at a frequency point is derived from the power spectral density estimation of the frequency domain vibration signal. By traversing the frequency domain vibration signal through a sliding window, the energy distribution of the signal within each window is calculated. The signal amplitude corresponding to each frequency point is squared and then converted in combination with the signal sampling period to directly reflect the energy intensity of that frequency point in the frequency domain vibration signal.

[0096] The total number of sampling points within the analysis frequency band is determined before processing the frequency domain vibration signal. It is calculated based on the preset analysis frequency band range and frequency point splitting interval, thus clarifying the total number of frequency points involved in the calculation of characteristic frequency components and ensuring coverage of the entire frequency range to be analyzed.

[0097] This formula is used to accurately extract the core characteristic frequency components in the frequency domain vibration signal that can characterize the operating status of the equipment, providing a key basis for the assessment of the operating status and fault diagnosis of lubricating oil equipment.

[0098] The formula obtains the weighted frequency sum of all frequency points by multiplying the frequency value of each frequency point by the corresponding power spectral density amplitude and then summing the results. The power spectral density amplitude is used as the weight to highlight the influence of high energy intensity frequency points in the calculation, making the result more reflective of the core energy distribution of the signal.

[0099] By dividing the weighted sum of frequencies by the sum of the power spectral density amplitudes of all frequency points, the core frequency is located. The resulting characteristic frequency component is essentially the frequency corresponding to the centroid of the energy distribution in the frequency domain vibration signal, which can centrally reflect the frequency component where the vibration energy is most concentrated during equipment operation.

[0100] This characteristic frequency component can accurately reflect the operating status of lubricating oil equipment. When the equipment is operating normally, this frequency will remain stable. When the equipment malfunctions or becomes abnormal, the characteristic frequency component will shift or fluctuate significantly. Therefore, the calculation result of this formula provides objective and accurate quantitative support for judging whether the equipment is operating normally and identifying potential faults.

[0101] The beneficial effects are that by estimating the power spectral density of frequency domain vibration signals, the energy proportion of different frequency components can be clearly presented, the concentrated frequency band of equipment vibration energy can be accurately located, and an intuitive and accurate energy distribution basis can be provided for subsequent feature extraction, avoiding the omission of key frequency information.

[0102] The characteristic frequency components are calculated by using the spectral centroid algorithm. The power spectral density amplitude at each frequency point is used as the weight for weighted calculation, so that the result can reflect the most core frequency components of energy in the vibration signal. These components are directly related to the operating status of the equipment. They remain stable during normal operation and show obvious deviations during faults, which can accurately capture potential fault signs of the equipment.

[0103] Normalizing the characteristic frequency components can eliminate the dimensional differences in frequency values ​​under different equipment and operating conditions, unifying them to a fixed standard range. At the same time, it strictly preserves the relative magnitude relationship of each frequency component, ensuring the comparability and consistency of characteristic indicators and avoiding interference from irrelevant factors in the state judgment.

[0104] By using standardized characteristic frequency indicators as the core features to characterize the operating status of equipment, key information from frequency domain vibration signals is integrated, and the stability and reliability of the features are improved through standardized processing. This provides accurate and unified vibration dimension data support for subsequent calculation of comprehensive health status index and equipment status determination, effectively improving the accuracy and pertinence of equipment status monitoring.

[0105] S3. Calculate the slope of the temperature change of the operating temperature signal per unit time to obtain the dynamic temperature change trend of the operating temperature signal;

[0106] In this embodiment of the invention, calculating the slope of the temperature change of the operating temperature signal per unit time to obtain the dynamic temperature change trend of the operating temperature signal includes:

[0107] The operating temperature signal is discretized over time to obtain a sequence of temperature sample values ​​for the operating temperature signal.

[0108] Based on the temperature sampling value sequence, the slope value of the dynamic temperature change trend is calculated through linear regression analysis, wherein the formula for calculating the slope value is as follows:

[0109] ;

[0110] In the formula, The slope value is... For discrete time points, This is a temperature sample value. This is the average value of the temperature samples. It is the average value at different time points. This represents the total number of sampling points;

[0111] The slope value is smoothed to obtain the dynamic temperature change trend index of the operating temperature signal;

[0112] The temperature dynamic change trend index is mapped to the temperature dynamic change trend of the lubricating oil equipment.

[0113] The step of performing time-series discretization processing on the operating temperature signal to obtain a sequence of temperature sample values ​​for the operating temperature signal includes:

[0114] Based on the Nyquist sampling theorem, a fixed sampling window is set for the lubricating oil equipment;

[0115] Based on the fixed sampling window, the timing deviation of the operating temperature signal is eliminated to obtain a regular temperature sampling value sequence of the operating temperature signal;

[0116] The integrity of the regularized temperature sample value sequence is verified to obtain the temperature sample value sequence of the operating temperature signal.

[0117] The acquired operating temperature signal is divided into fixed time intervals, transforming the continuously changing temperature signal into discrete temperature data points. Each time interval corresponds to a unique temperature sample value. These sample values ​​are arranged sequentially in chronological order to form an ordered sequence of temperature sample values. This ensures that the sequence can completely reproduce the temperature state of the operating temperature signal at different times, providing a structured data foundation for subsequent slope calculation.

[0118] Using the temporal order of the temperature sampling value sequence as the horizontal reference dimension and the corresponding temperature sampling value as the vertical numerical dimension, a complete two-dimensional data distribution is constructed. Linear regression analysis is used to fit these data points, finding a straight line that best matches the distribution trend of all data points. The slope of this line directly reflects the temperature change over time. By analyzing the slope characteristics of the line, the corresponding slope value is calculated, which is the core quantitative data of the dynamic temperature change trend.

[0119] A smoothing method using moving averages is employed. A fixed-length sliding window is set, and the calculated slope values ​​are sequentially input into the window in chronological order. The average value of all slope values ​​within each window is then calculated, and this average value replaces the original slope value at the center of the window. This process is repeated window by window to process all slope values. This method eliminates random fluctuations and abnormal peaks in the slope values, resulting in a stable and continuous indicator of dynamic temperature change trends.

[0120] A correlation is established between dynamic temperature change trend indicators and the actual temperature change status of lubricating oil equipment. A positive indicator value represents an upward trend in equipment temperature over time, a negative indicator value represents a downward trend in temperature, and an indicator value approaching zero represents a relatively stable temperature. Based on the specific changes in the indicator, the real-time dynamic temperature change of the lubricating oil equipment during operation is accurately mapped, providing a key basis for assessing the equipment's operating status in terms of temperature.

[0121] The discrete time points are derived from the fixed time intervals set when the operating temperature signal is discretized into a time series. The specific time node corresponding to each temperature sample value is the discrete time point, which is arranged in chronological order to completely correspond to the acquisition time sequence of the temperature sample value sequence.

[0122] Temperature sampling values ​​are the core data obtained after the operating temperature signal has been discretized over time. The operating temperature of the equipment is collected in real time at each discrete time point by a temperature sensor, and the actual temperature status of the equipment at the corresponding moment is directly recorded. The temperature sampling value sequence is formed according to the collection time order.

[0123] The average temperature sample value is the result of summing all temperature sample values ​​in the temperature sample value sequence and dividing by the total number of sampling points. It reflects the average level of the equipment operating temperature throughout the entire sampling period and eliminates the influence of random fluctuations in individual sample values.

[0124] The average time point is the result of summing all discrete time points and dividing by the total number of sampling points. It represents the midpoint of the entire sampling period and serves as a benchmark for measuring the central tendency of the time dimension, providing a reference for analyzing the relationship between temperature and time.

[0125] The total number of sampling points is determined before the time series discretization processing of the operating temperature signal. It is calculated from the preset sampling duration and fixed time interval, which clarifies the total number of temperature sampling values ​​and discrete time points involved in the slope calculation, ensuring coverage of the complete analysis period.

[0126] This formula is used to accurately calculate the slope of the temperature change signal per unit time, quantify the rate of temperature change over time, and provide core quantitative basis for obtaining the dynamic trend of equipment temperature change.

[0127] The formula calculates the difference between each discrete time point and the average time point, as well as the difference between the corresponding temperature sample value and the average temperature sample value. Then, it multiplies the two sets of differences one by one and sums them up, accurately capturing the synergistic relationship between time and temperature changes, highlighting the contribution of temperature changes at different times to the overall trend.

[0128] The formula's denominator is constructed by summing the squares of the differences between each discrete time point and the average time point, thus establishing a benchmark reference in the time dimension. This benchmark is used to standardize the cooperative change relationship reflected by the numerator, ensuring that the final slope value accurately reflects the average temperature change per unit time.

[0129] The slope value obtained by calculating the ratio of the numerator to the denominator directly quantifies the dynamic rate of temperature change over time. The sign of the slope value corresponds to the trend of temperature increase or decrease, and the magnitude of the slope value corresponds to the speed of change. This calculation result provides objective quantitative support for determining whether equipment has abnormal temperature fluctuations and whether it is within the normal operating temperature range, and is a key temperature dimension indicator for analyzing the operating status of equipment.

[0130] The beneficial effects are that by performing time series discretization processing on the operating temperature signal, the continuous temperature changes are transformed into a structured temperature sampling value sequence, the temperature data at different times are completely preserved, the key information in the original signal is avoided, and an orderly and reliable data foundation is provided for subsequent slope calculation, ensuring that the temperature change pattern can be accurately captured.

[0131] Based on the temperature sampling value sequence, the slope value is calculated through linear regression analysis. By integrating the time and temperature correlation information of all sampling points, the linear trend that best fits the temperature change law is fitted, which quantitatively reflects the rate and direction of temperature change per unit time. Compared with the temperature comparison at a single time point, it can more objectively and comprehensively reflect the overall trend of temperature change and avoid misjudgment caused by local fluctuations.

[0132] Slope values ​​are smoothed to effectively eliminate abnormal slope values ​​caused by random interference and instantaneous fluctuations, making the dynamic temperature trend indicators more stable and continuous, accurately reflecting the true pattern of temperature changes, avoiding the influence of data noise on the judgment of equipment temperature status, and improving the reliability of trend indicators.

[0133] By mapping the dynamic temperature change trend index to the dynamic temperature change trend of lubricating oil equipment, a direct correlation is established between the index and the actual temperature state of the equipment. The specific trend of temperature rise, fall or stabilization is clearly presented, providing an accurate temperature dimension basis for subsequent fusion with characteristic frequency components to calculate the comprehensive health status index, and helping to comprehensively and accurately assess the operating status of the equipment.

[0134] By setting a fixed sampling window based on the Nyquist sampling theorem, the frequency characteristics of the operating temperature signal can be accurately matched, ensuring that the sampling interval is sufficient to fully capture the highest frequency components of temperature changes. This avoids the loss of key temperature change information due to sparse sampling, and also prevents data redundancy caused by excessive sampling, thus providing efficient and comprehensive raw temperature data for subsequent processing.

[0135] By eliminating timing deviations in the operating temperature signal through a fixed sampling window and calibrating the acquisition time point of each temperature data point, all data are arranged strictly according to a uniform time interval. This solves the timing disorder problems caused by acquisition delays and equipment jitter, forming a regular temperature sampling value sequence, ensuring data consistency in the time dimension, and providing a reliable foundation for subsequent analyses that depend on timing relationships, such as slope calculation.

[0136] The integrity of the regular temperature sampling value sequence is verified, and missing, duplicate or abnormally abrupt data points are comprehensively investigated. By supplementing missing data, eliminating redundant terms and correcting interference signals through reasonable interpolation, the final temperature sampling value sequence data is ensured to be complete and free of anomalies. This avoids the interference of invalid data on the calculation of temperature dynamic change trends, and further improves the accuracy and reliability of temperature correlation analysis results.

[0137] S4. Input the characteristic frequency components and the temperature dynamic change trend into a preset weighted fusion model to generate the comprehensive health status index of the lubricating oil equipment.

[0138] In this embodiment of the invention, the step of inputting the characteristic frequency components and the dynamic temperature change trend into a preset weighted fusion model to generate a comprehensive health status index of the lubricating oil equipment includes:

[0139] The dimensions of the characteristic frequency components are unified to obtain the standard frequency characteristic index of the characteristic frequency components.

[0140] The temperature dynamic change trend is normalized to obtain a standard temperature trend index for the temperature dynamic change trend.

[0141] Based on the operating stage characteristics of the lubricating oil equipment, assign corresponding dynamic weight coefficients to the standardized frequency characteristic index and the standard temperature trend index;

[0142] Based on the dynamic weighting coefficient, the standard frequency characteristic index and the standard temperature trend index are nonlinearly integrated to obtain the preliminary comprehensive health status index of the lubricating oil equipment.

[0143] Based on a preset threshold, the preliminary comprehensive health status index is evaluated using a confidence level assessment based on the preset threshold to obtain the comprehensive health status index of the lubricating oil equipment.

[0144] The preliminary comprehensive health status index of the lubricating oil equipment is obtained by nonlinearly integrating the standard frequency characteristic index and the standard temperature trend index based on the dynamic weighting coefficient, including:

[0145] Based on the dynamic weighting coefficient, the standard frequency characteristic index and the standard temperature trend index are assigned weights in a hierarchical manner to obtain the weighted frequency characteristics and the weighted temperature trend.

[0146] A nonlinear feature transformation is performed on the weighted frequency characteristics and the weighted temperature trend to obtain the enhanced feature representation of the lubricating oil equipment.

[0147] The enhanced feature representations are then subjected to feature interaction fusion to obtain a preliminary comprehensive health status index of the lubricating oil equipment.

[0148] A unified dimensional transformation method is adopted, referencing the typical value range of characteristic frequency components during normal operation of lubricating oil equipment, to scale the original characteristic frequency components proportionally to a fixed standard value range. During the scaling process, the distribution law of equipment vibration energy reflected by the characteristic frequency components is strictly maintained, eliminating only the numerical interference caused by dimensional differences under different operating conditions. The result is a standard frequency characteristic index of the characteristic frequency components that is numerically standardized and can be directly used for fusion calculations.

[0149] The temperature dynamic change trend data is standardized by first defining the reasonable fluctuation range of the temperature dynamic change trend during normal equipment operation, correcting abnormal trend values ​​that exceed this range according to boundary extreme values, and then mapping all trend values ​​proportionally to a numerical range consistent with the standard frequency characteristic index. This process unifies the numerical format of the temperature dynamic change trend with the standard frequency characteristic index while retaining the core trend information of temperature change, resulting in a standard temperature trend index.

[0150] A thorough analysis of the operational characteristics of lubricating oil equipment reveals the working characteristics of different stages, including startup, stable operation, and shutdown. During startup, temperature changes have a more significant impact on equipment health, thus assigning higher weight to standard temperature trend indicators. In stable operation, vibration characteristics have greater reference value, leading to increased weighting of standardized frequency characteristic indicators. During shutdown, the weights of both are balanced. Based on the differences in characteristics at each stage, corresponding dynamic weighting coefficients are dynamically adjusted and determined to ensure that the weighting allocation closely matches the actual operating state of the equipment.

[0151] Based on the assigned dynamic weighting coefficients, a nonlinear integration method is used to fuse standard frequency characteristic indicators and standard temperature trend indicators. First, each indicator is multiplied by its corresponding dynamic weighting coefficient to enhance the influence of high-weight indicators on the comprehensive result. Then, a nonlinear function is used to integrate the two weighted indicators. This function can highlight the sensitivity to changes when the indicator values ​​are abnormal, while smoothing out small fluctuations within the normal range. Finally, a preliminary comprehensive health status index that can initially reflect the overall health status of the equipment is obtained.

[0152] A preset threshold is set for the normal operation of the equipment. This threshold is determined based on historical health data statistics of a large number of similar lubricating oil equipment, clearly defining the boundaries between different health levels such as healthy, sub-healthy, and faulty. The preliminary comprehensive health status index is compared with the preset threshold to assess the health level range in which the index falls. At the same time, it is checked whether the index has any deviation due to data fluctuations. If the index is in the critical range, it is further verified by combining the historical operating data of the equipment. Finally, a comprehensive health status index that accurately and reliably reflects the actual health status of the lubricating oil equipment is obtained.

[0153] Based on the established dynamic weighting coefficients, a tiered weighting operation is performed on the standard frequency characteristic index and the standard temperature trend index. The standard frequency characteristic index is directly multiplied by its corresponding dynamic weighting coefficient to emphasize the weight of vibration characteristics on the equipment's health status during the current operating phase. Similarly, the standard temperature trend index is multiplied by its specific dynamic weighting coefficient to strengthen the weight of temperature change trends. Through this targeted weighting process, the values ​​of both indicators fully reflect their respective importance under the current operating conditions, ultimately yielding the weighted frequency characteristics and the weighted temperature trend.

[0154] Using pre-defined nonlinear transformation rules, the weighted frequency characteristics and weighted temperature trends are enhanced. For features with values ​​within the normal range, their original trends are preserved through a gentle transformation. For features with values ​​close to abnormal thresholds or exhibiting small fluctuations, the amplitude of their changes is amplified through nonlinear transformation to enhance their sensitivity to equipment health anomalies. For features that significantly deviate from the normal range, appropriate constraints are applied according to pre-defined rules to avoid the interference of extreme values ​​on the overall results. After this processing, an enhanced feature representation of the lubricating oil equipment is obtained that more accurately reflects differences in equipment health status.

[0155] A feature interaction and fusion mechanism is constructed to deeply combine the weighted frequency features and weighted temperature trends in the enhanced feature representation. First, a bidirectional correlation analysis is performed on the two enhanced features to uncover potential correlation patterns, such as the synergistic relationship between vibration anomalies and temperature abrupt changes. Then, through feature superposition and complementary operations, the core information of the two features is integrated, ensuring that the fused feature reflects both the equipment status at the vibration level and the impact of temperature changes. Finally, the fusion result is normalized to form a preliminary comprehensive health status index for lubricating oil equipment with unified numerical values ​​and complete information.

[0156] The beneficial effects are that unifying the dimensions of characteristic frequency components can eliminate the differences in the dimensions of frequency values ​​under different equipment and different operating conditions, unify them to a fixed standard range, ensure that the standard frequency characteristic indicators are comparable horizontally, avoid the deviation in fusion calculation caused by inconsistent dimensions, and lay a unified data foundation for the integration of multi-dimensional features.

[0157] The temperature dynamic change trend data is standardized to retain the core trend information of temperature change and map it to a numerical range consistent with the standard frequency characteristic index. This achieves the unification of the numerical format of the two types of key indicators, ensures the effectiveness of subsequent fusion calculations, and avoids the impact of single-dimensional data format differences on the comprehensive evaluation results.

[0158] Dynamic weighting coefficients are assigned based on the operational characteristics of lubricating oil equipment, which are aligned with the working characteristics of different stages such as equipment startup, stable operation, and shutdown. This allows high-impact factors to receive higher weights, enabling the comprehensive health status index to accurately match the status assessment needs of equipment at different stages and improve the index's adaptability to the actual operating status of the equipment.

[0159] By using dynamic weighting coefficients for nonlinear integration, the preliminary comprehensive health status index can accurately capture the core differences in equipment health status by strengthening the influence of high-weight indicators, highlighting the sensitivity of abnormal values, and smoothing small fluctuations within the normal range. Compared with simple weighting, it can more objectively reflect the overall health status of equipment.

[0160] The initial comprehensive health status index is assessed for confidence based on a preset threshold. The reliability of the index is verified by combining the preset health level classification boundaries, eliminating deviations caused by data fluctuations, and ensuring that the final comprehensive health status index is accurate and reliable, providing high-quality core quantitative basis for subsequent status determination and early warning.

[0161] By assigning weights in a hierarchical manner based on dynamic weight coefficients, the weight ratios of standard frequency characteristic indicators and standard temperature trend indicators can be accurately matched to the characteristics of different operating stages of lubricating oil equipment. This allows the weighted frequency characteristics and temperature trends to highlight the dimensions that have a greater impact on the health status of the equipment at the corresponding stage, avoiding the imbalance of feature importance caused by a single fixed weight, and providing basic data that fits the actual operating status of the equipment for subsequent integration.

[0162] By performing nonlinear feature transformation on the weighted features, key information can be enhanced and irrelevant interference can be suppressed. For features that are close to the abnormal threshold or have small fluctuations, the transformation amplifies their change range, thereby improving the sensitivity to equipment health anomalies. For features within the normal range, their original trend is maintained, while extreme values ​​that deviate significantly from the normal range are constrained. The resulting enhanced feature representation can more accurately highlight the core differences in equipment health status, thereby improving the recognizability and effectiveness of the features.

[0163] By fusing enhanced feature representations through feature interaction, we can deeply uncover the potential correlations between weighted frequency features and weighted temperature trends, such as the synergistic relationship between vibration anomalies and temperature abrupt changes. Through feature superposition and complementary operations, we can integrate the core information of both types of features, avoiding the limitations of single-dimensional features. The fused result encompasses both the mechanical state of the equipment at the vibration level and the thermal operating information at the temperature level, forming a preliminary comprehensive health status index that comprehensively reflects the overall health of the equipment. This provides high-quality intermediate data support for subsequent confidence assessments and the generation of the final health index.

[0164] S5. Based on the status data of the lubricating oil equipment during the historical operating cycle, dynamically update the early warning threshold;

[0165] In this embodiment of the invention, the step of dynamically updating the early warning threshold based on the status data of the lubricating oil equipment during its historical operating cycle includes:

[0166] Extract the state data sequence within the historical operating cycle from the operating database of the lubricating oil equipment to construct the historical state dataset of the lubricating oil equipment;

[0167] Outliers in the historical state dataset are removed to obtain the regularized historical state data of the historical state dataset.

[0168] Trend features are extracted from the regularized historical state data to identify the long-term trend features of state changes in the lubricating oil equipment;

[0169] Based on the long-term trend characteristics, a preliminary early warning threshold for the lubricating oil equipment is generated;

[0170] The rationality of the preliminary warning threshold is verified to obtain the dynamic warning threshold of the lubricating oil equipment.

[0171] Access the lubricating oil equipment's operational database, which stores all status data recorded during the equipment's past operation. This includes characteristic frequency data, dynamic temperature trend data, and corresponding comprehensive health status indices for different historical operating cycles. Filter the relevant data for complete historical operating cycles in chronological order, organize this data into periodic segments, and sequentially associate and arrange various status data within each cycle to construct a historical status dataset covering the equipment's long-term operational status. This ensures the dataset comprehensively reflects the equipment's operating condition at different times.

[0172] Statistical analysis methods were employed to detect outliers in the historical state dataset, examining each data point individually. By comparing the variation range of each data point with that of its neighboring data points within the same period, and with the overall distribution pattern of the dataset, isolated or abruptly changing data points exceeding the normal fluctuation range were identified. For detected outliers, verification was conducted using the equipment's operational records at the time of detection. Outliers confirmed to be caused by factors other than the equipment's inherent condition, such as sensor malfunctions or data transmission errors, were directly removed. Outliers caused by momentary equipment failures were retained and marked, ultimately resulting in well-distributed historical state data free from invalid interference.

[0173] The standardized historical status data is segmented according to the time dimension, dividing it into multiple continuous analysis periods. For the data within each period, the fluctuation patterns and overall trends of numerical changes are analyzed, extracting the status change trend for each period. These trends are then correlated and compared across periods to identify common patterns of change spanning multiple historical operating cycles. This method isolates the interference from short-term fluctuations, accurately identifying the long-term trend characteristics of lubricating oil equipment status over time. These characteristics reflect long-term changes such as equipment aging and performance degradation.

[0174] Based on the identified long-term trend characteristics and combined with the performance standards for normal equipment operation, the safe fluctuation range of the status data is determined. Referring to the rate of change and extreme values ​​of the status data in the long-term trend, and with the premise of ensuring safe equipment operation, a preliminary warning boundary value is set. This boundary value must cover the range of status changes during the normal aging process of the equipment, while also being able to promptly detect abnormal changes exceeding the normal decay rate, thus generating a preliminary warning threshold for the lubricating oil equipment.

[0175] Real-time status data of the lubrication equipment during its recent normal operating cycle is collected. The initial warning threshold is compared with this real-time data to verify whether the threshold can accurately distinguish between normal and abnormal states. If the real-time data is all within the safe range set by the initial warning threshold and effectively reflects the current normal operating status of the equipment, the threshold is considered basically reasonable. If there are cases where normal operating data exceeds the threshold or abnormal data is not captured by the threshold, the threshold is fine-tuned based on long-term trend characteristics and the current operating conditions of the equipment. Simultaneously, the scientific validity of the threshold is further verified by referring to the warning threshold setting standards of similar equipment, ultimately resulting in a dynamic warning threshold that accurately adapts to the current operating status and long-term change patterns of the equipment.

[0176] The beneficial effect is that by extracting the status data sequence within the historical operating cycle from the operating database of lubricating oil equipment to construct a historical status dataset, it can comprehensively collect the core status information of the equipment's past operation, covering key data such as characteristic frequencies, temperature trends, and health indices of different periods. This provides rich and complete historical data support for dynamically updating the early warning threshold, ensuring that the threshold update can conform to the long-term operating rules of the equipment.

[0177] Removing outliers from historical state datasets yields regularized historical state data. This effectively eliminates invalid interference data caused by factors other than the device's own state, such as sensor malfunctions and data transmission errors. This ensures that the data used for analysis is reasonably distributed, accurate, and reliable, and avoids outliers affecting the accuracy of long-term trend feature extraction. It also lays a high-quality data foundation for subsequent threshold calculations.

[0178] By extracting trend features from well-organized historical data and identifying long-term trend characteristics, interference from short-term fluctuations can be eliminated. This allows for the accurate capture of regular changes in equipment over time, such as aging and performance degradation. This enables the updating of warning thresholds to adapt to the natural evolution of equipment status, avoiding the drawbacks of using fixed thresholds that cannot cope with long-term changes in equipment.

[0179] The system generates preliminary warning thresholds based on long-term trend characteristics, ensuring that the thresholds cover the range of state changes during the normal aging process of the equipment. At the same time, it accurately locates the critical point of abnormal changes that exceed the normal decay rate. This avoids false warnings due to overly strict threshold settings, as well as warning delays due to overly broad threshold settings, thus achieving a precise match between the warning thresholds and the actual state changes of the equipment.

[0180] The rationality of the initial warning threshold is verified to obtain the dynamic warning threshold. By combining recent real-time status data and standards of similar equipment, the threshold is further verified and fine-tuned to improve its scientificity and adaptability. This ensures that the final dynamic warning threshold can fit the current operating conditions of the equipment in real time, significantly reducing the probability of false and missed warnings, and providing an accurate and reliable threshold reference for equipment status determination.

[0181] S6. Compare the comprehensive health status index with the warning threshold in real time, and determine the real-time status of the lubricating oil equipment based on the comparison result;

[0182] In this embodiment of the invention, the step of comparing the comprehensive health status index with the warning threshold in real time and determining the real-time status of the lubricating oil equipment based on the comparison result includes:

[0183] A multi-level state determination threshold system for the lubricating oil equipment is constructed to accurately divide the operating state of the lubricating oil equipment into normal state, warning state, and abnormal state.

[0184] The comprehensive health status index is compared with the multi-level status judgment threshold system in real time to obtain the preliminary judgment result of the status level of the lubricating oil equipment.

[0185] A state transition continuity analysis is performed on the preliminary state level determination results, and the reliability of the determination results is evaluated in conjunction with the historical operating status data of the lubricating oil equipment.

[0186] Based on the credibility assessment results, the real-time status judgment conclusion of the lubricating oil equipment is output, and a corresponding confidence index is assigned to each status level.

[0187] The real-time status determination conclusion of the lubricating oil equipment is output based on the confidence assessment result, and a corresponding confidence index is assigned to each status level, including:

[0188] Based on the statistical distribution characteristics of the regularized historical state data, a mapping relationship system between the state level and confidence index in the lubricating oil equipment is established.

[0189] The basic confidence level of the lubricating oil equipment is determined based on the degree of consistency between the current state determination result and the regularized historical state data.

[0190] The baseline confidence level is dynamically adjusted based on the operating time, maintenance records, and operating parameters of the lubricating oil equipment.

[0191] Based on the vibration characteristics, temperature trends, and regularized historical state data of the lubricating oil equipment, a confidence index for the lubricating oil equipment is generated.

[0192] Based on the dynamic early warning threshold of lubricating oil equipment, and combined with the performance standards for normal equipment operation and the rules for classifying fault risk levels, a multi-level state judgment threshold system is constructed. Using the dynamic early warning threshold as the core benchmark, three distinct threshold intervals are defined. The interval below the first threshold corresponds to a normal state, indicating that all equipment indicators meet operational requirements and there is no fault risk. The interval between the first and second thresholds corresponds to a warning state, indicating slight abnormal fluctuations in the equipment that require monitoring for subsequent changes. The interval above the second threshold corresponds to an abnormal state, indicating that the equipment has shown obvious signs of fault and requires timely handling. This system enables precise classification of the operating status of lubricating oil equipment.

[0193] A real-time dynamic comparison mechanism is established to continuously acquire the comprehensive health status index of lubricating oil equipment and match it in real time with three threshold intervals in a multi-level status judgment threshold system. Each time a new comprehensive health status index is acquired, its corresponding threshold interval is immediately determined. Based on the status level corresponding to the interval, a preliminary judgment result of the current operating status of the lubricating oil equipment is quickly obtained, ensuring the timeliness and continuity of the comparison process and enabling real-time capture of dynamic changes in equipment status.

[0194] A state transition continuity analysis is performed on the preliminary state level assessment results, tracking the preliminary assessment results across multiple consecutive time points to observe whether there are frequent switching or unreasonable jumps in state levels. If the state level suddenly jumps from a normal state to an abnormal state, it is necessary to verify whether there are data anomalies or sudden equipment failures. If the state level fluctuates repeatedly between the alert state and the normal state, the focus is on analyzing whether the fluctuation is due to changes in normal operating conditions. Simultaneously, historical operating status data of the lubrication equipment is retrieved, and the current preliminary assessment results are compared with the status data under the same historical period and similar operating conditions to evaluate the consistency between the current results and historical patterns, thus completing the reliability assessment of the assessment results.

[0195] Based on the credibility assessment results, if the preliminary judgment is highly consistent with the state transition pattern and historical operating data, and has high credibility, the corresponding real-time state judgment conclusion is directly output. If the credibility is low, further verification is required based on the equipment's real-time operating data, and the judgment conclusion is revised before output. Simultaneously, a corresponding confidence index is assigned to each state level. A high confidence index corresponds to a normal state, indicating that the judgment result is reliable. Note that a moderate state confidence index suggests the need for continuous monitoring. The confidence index for abnormal states is dynamically adjusted according to the degree of abnormality; the more severe the abnormality, the higher the confidence index, providing accurate state reference for subsequent equipment maintenance decisions.

[0196] A thorough analysis of the statistical distribution characteristics of historical status data was conducted, identifying the frequency of occurrence, duration, and correlation with actual equipment operating conditions for different status levels. For normal status, the proportion of historical data accurately reflecting fault-free equipment operation was statistically analyzed; for alert status, the probability distribution of its subsequent development into normal or abnormal status was calculated; for abnormal status, the accuracy rate of its correspondence with actual fault occurrences was statistically analyzed. Based on these statistical results, a one-to-one correspondence between status levels and confidence indices was established, forming a complete mapping system to ensure that confidence indices objectively reflect the reliability of each status level determination.

[0197] The current state assessment result is comprehensively compared with the standardized historical state data to analyze the matching situation of the current result in the historical data. If the current state level is frequently observed in the same period and under similar operating conditions in the past, and subsequent verifications all match the actual equipment state, then the consistency level is high, and a higher basic confidence level is assigned accordingly. If the current result occurs very infrequently in the historical data, or differs significantly from the historical state under similar operating conditions, then the consistency level is low, and a lower basic confidence level is assigned. The basic confidence level of the lubricating oil equipment is determined through consistency comparison.

[0198] The system retrieves the lubrication equipment's runtime records, maintenance files, and real-time operating parameters to dynamically adjust the baseline confidence level. If the equipment's runtime is short, there are recent complete maintenance records, and the operating parameters are stable within the normal range, it indicates that the equipment's operating status is stable, and the judgment result is more reliable, so the baseline confidence level is appropriately increased. If the equipment's runtime is long, the maintenance records are incomplete, or the operating parameters fluctuate greatly, it may lead to deviations in the status judgment, so the baseline confidence level is appropriately decreased to make the confidence level index more consistent with the equipment's current actual operating context.

[0199] This approach comprehensively considers real-time vibration characteristics, temperature trends, and well-organized historical state data of lubricating oil equipment, fusing multi-dimensional information. Vibration characteristics and temperature trends reflect the immediate operational status of the equipment. If both are stable and consistent with the current state assessment, the confidence level can be further strengthened. If slight fluctuations exist but do not exceed the normal range, the confidence level is adjusted based on experience in handling similar fluctuations in historical data. By integrating real-time state data with historical statistical patterns, the limitations of single-dimensional data are eliminated, generating a confidence index for lubricating oil equipment that comprehensively and accurately reflects the reliability of the current state assessment.

[0200] The beneficial effects are that by constructing a multi-level state determination threshold system and dividing it into three categories of operating states, the limitations of a single threshold determination are broken. By clarifying the boundary range of different states, the operating states of the equipment can be distinguished in a more refined manner, which not only avoids normal states from being misjudged as abnormal, but also prevents minor abnormalities from being ignored, making the state determination more targeted.

[0201] By comparing the comprehensive health status index with a multi-level threshold system in real time, the dynamic changes of the index can be quickly captured and matched with the corresponding status level, ensuring the timeliness of status determination and avoiding the lag caused by static comparison, allowing staff to keep abreast of the equipment's operating status.

[0202] By performing state transition continuity analysis on the preliminary judgment results, abnormal judgment situations such as frequent switching and unreasonable jumps can be identified. By combining historical operating status data to verify the consistency of the current results, misjudgments caused by data fluctuations and interference signals can be effectively filtered out, thereby improving the reliability of the judgment results.

[0203] Based on the reliability assessment, real-time status judgment conclusions are output, and reliability indicators are configured for each status level. This ensures that the judgment results not only clearly indicate the current status of the equipment, but also reflect the reliability of the results, providing staff with a more comprehensive decision-making reference. At the same time, the differentiated allocation of confidence indicators can help staff accurately grasp the handling priority of different statuses and improve equipment maintenance efficiency.

[0204] Establishing a mapping relationship system based on the statistical distribution characteristics of regularized historical state data can accurately correspond to state levels and confidence indices, enabling confidence indices to objectively reflect the historical reliability of each state level judgment result. This provides a unified reference standard that conforms to the operating rules of the equipment for subsequent confidence calculations, avoiding deviations caused by subjective settings.

[0205] The basic confidence level is determined by the degree of consistency between the current state judgment result and the regularized historical state data. By comparing the current result with the state patterns under the same period and similar working conditions in the past, it is ensured that the basic confidence level can truly reflect the rationality of the current judgment result. The higher the consistency, the more solid the confidence level base, providing a reliable benchmark for subsequent corrections.

[0206] By dynamically adjusting the base confidence level based on runtime, maintenance records, and operating parameters, the system can adapt to differences in the actual operating background of the equipment. The judgment results for equipment with stable operating status and standardized maintenance are more reliable, while the confidence level is appropriately lowered to make the confidence level index more in line with the actual operating conditions of the equipment and improve its accuracy.

[0207] By integrating vibration characteristics, temperature trends, and regular historical state data to generate a confidence index, a multi-dimensional fusion of real-time state data and historical statistical patterns is achieved. This approach considers both the current real-time operating status of the equipment and relies on historical data to ensure the stability of the results, effectively eliminating the limitations of single-dimensional data. The generated confidence index can comprehensively and reliably reflect the credibility of real-time state judgments, providing a more valuable basis for equipment maintenance decisions.

[0208] S7. When the real-time status is in an abnormal state, trigger a multi-level early warning for the lubricating oil equipment.

[0209] In this embodiment of the invention, when the real-time status is in an abnormal state, a multi-level early warning system for the lubricating oil equipment is triggered.

[0210] Based on the abnormal status judgment conclusions and confidence indices of lubricating oil equipment, a multi-level early warning system is constructed, clarifying the triggering conditions for different early warning levels. Abnormal statuses are divided into three levels according to severity: mild, moderate, and severe. Mild abnormalities correspond to confidence indices within the mild risk range, and the core operating parameters of the equipment do not exceed critical values. Moderate abnormalities correspond to confidence indices within the moderate risk range, with some key parameters showing significant fluctuations. Severe abnormalities correspond to confidence indices reaching the high-risk standard, with core parameters severely deviating from the normal range, ensuring a precise match between the early warning level and the degree of abnormality.

[0211] When the real-time status is determined to be abnormal, the system automatically extracts core data corresponding to the anomaly, such as the comprehensive health status index, characteristic frequency components, and dynamic temperature change trends, and combines this data with a confidence index to determine the specific warning level. If the core parameters slightly deviate from the normal range and the confidence index meets the criteria for mild anomaly, a mild warning is triggered; if the key parameters fluctuate significantly and the confidence index meets the requirements for moderate anomaly, a moderate warning is triggered; if the core parameters severely exceed the limits and the confidence index meets the conditions for severe anomaly, a severe warning is directly triggered, achieving automatic and accurate determination of the warning level.

[0212] Differentiated warning messages are generated for different warning levels. Mild warnings include the type of equipment anomaly, current key parameter values, and preliminary observation suggestions, clearly reminding staff to closely monitor equipment operation without immediate shutdown. Moderate warnings detail the anomaly characteristics, possible causes of the fault, and targeted inspection measures, requiring staff to complete equipment troubleshooting within a specified time. Severe warnings emphasize the urgency of the fault, potential risks and consequences, and emergency handling procedures, mandating immediate shutdown for repair. Warning messages are simultaneously pushed to equipment control terminals and staff mobile terminals in multiple formats, including text, audio-visual signals, etc., ensuring timely and effective information delivery.

[0213] Establish an early warning response tracking mechanism to monitor the progress of staff in handling early warning information in real time. For minor early warnings, regularly report changes in equipment operating parameters until the abnormal state is eliminated or upgraded. For moderate early warnings, set a processing time limit reminder; if not handled within the time limit, the early warning level will be automatically upgraded, and relevant personnel will be notified. For severe early warnings, activate the emergency linkage mechanism, simultaneously cut off the power supply to the equipment (if it is equipment that does not require continuous operation), and arrange professional maintenance personnel to arrive on site as soon as possible. Simultaneously, record the early warning trigger time, processing process, and results in the equipment operation database to provide data support for subsequent fault analysis and early warning optimization.

[0214] The beneficial effects are that a multi-level early warning system is constructed based on the abnormal state judgment conclusion and confidence index, different early warning levels are divided according to the severity and the triggering conditions are clearly defined, so as to achieve accurate matching between the early warning level and the degree of abnormality and avoid the problem that a single early warning mode cannot adapt to different fault urgency levels.

[0215] When the real-time status is determined to be abnormal, the system automatically extracts core data and combines it with confidence indicators to accurately determine the warning level. It can respond quickly without manual intervention, ensuring the timeliness and accuracy of warning triggering and buying time for fault handling.

[0216] Differentiated warning information is generated for different warning levels, clearly including key information such as the type of anomaly, parameter value, cause of failure, and handling suggestions. At the same time, multiple forms of push notifications are used to ensure that staff receive the information in a timely manner, avoiding delays in handling due to unclear information or untimely delivery.

[0217] Establish an early warning response tracking mechanism to implement targeted tracking management for different early warning levels. Mild early warnings will be continuously monitored, moderate early warnings will be subject to restrictions and reminders, and severe early warnings will be subject to emergency linkage. At the same time, all early warning-related information will be recorded to ensure that faults are effectively handled and to provide data support for subsequent fault analysis and early warning optimization.

[0218] The multi-level early warning system enables graded control of abnormal equipment conditions, allowing staff to accurately grasp the priority of handling based on the early warning level. It also reduces the risk of escalation of faults through differentiated handling measures, significantly improving the scientific and efficient nature of lubricant equipment maintenance and ensuring the safe and stable operation of the equipment.

[0219] like Figure 2 The diagram shown is a functional block diagram of a lubricating oil equipment condition monitoring and early warning system provided in an embodiment of the present invention.

[0220] The lubricating oil equipment condition monitoring and early warning system 100 of the present invention can be installed in an electronic device. Depending on the functions implemented, the lubricating oil equipment condition monitoring and early warning system 100 may include a signal synchronization acquisition module 101, a frequency domain feature extraction module 102, a temperature trend analysis module 103, a health index generation module 104, a threshold update module 105, a condition determination module 106, and an early warning triggering module 107. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0221] In this embodiment, the functions of each module / unit are as follows:

[0222] The signal synchronization acquisition module 101 is used to synchronously acquire the time-domain vibration signal and operating temperature signal of the lubricating oil equipment.

[0223] The frequency domain feature extraction module 102 is used to perform a fast Fourier transform on the time domain vibration signal, convert the time domain vibration signal into a frequency domain vibration signal, and extract the characteristic frequency components representing the operating state of the equipment from the frequency domain vibration signal.

[0224] The temperature trend analysis module 103 is used to calculate the slope value of the temperature change of the operating temperature signal per unit time, so as to obtain the dynamic temperature change trend of the operating temperature signal.

[0225] The health index generation module 104 is used to input the characteristic frequency components and the temperature dynamic change trend into a preset weighted fusion model to generate the comprehensive health status index of the lubricating oil equipment.

[0226] The threshold update module 105 is used to dynamically update the warning threshold based on the status data of the lubricating oil equipment in the historical operating cycle.

[0227] The status determination module 106 is used to compare the comprehensive health status index with the warning threshold in real time, and determine the real-time status of the lubricating oil equipment based on the comparison result.

[0228] The early warning triggering module 107 is used to trigger a multi-level early warning for the lubricating oil equipment when the real-time status is in an abnormal state.

[0229] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0230] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0231] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0232] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0233] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0234] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for monitoring and early warning of lubricating oil equipment condition, characterized in that, The method includes: S1. Synchronously acquire time-domain vibration signals and operating temperature signals of the lubrication equipment; S2. Perform a fast Fourier transform on the time-domain vibration signal to convert the time-domain vibration signal into a frequency-domain vibration signal, and extract the characteristic frequency components that characterize the operating state of the equipment from the frequency-domain vibration signal. S3. Calculate the slope of the temperature change of the operating temperature signal per unit time to obtain the dynamic temperature change trend of the operating temperature signal; S4. Input the characteristic frequency components and the temperature dynamic change trend into a preset weighted fusion model to generate the comprehensive health status index of the lubricating oil equipment. S5. Based on the status data of the lubricating oil equipment during the historical operating cycle, dynamically update the early warning threshold; S6. Compare the comprehensive health status index with the warning threshold in real time, and determine the real-time status of the lubricating oil equipment based on the comparison result; S7. When the real-time status is in an abnormal state, trigger a multi-level early warning for the lubricating oil equipment.

2. The lubricating oil equipment condition monitoring and early warning method as described in claim 1, characterized in that, The step of performing a Fast Fourier Transform on the time-domain vibration signal to convert it into a frequency-domain vibration signal, and extracting characteristic frequency components representing the operating state of the equipment from the frequency-domain vibration signal, includes: The power spectral density of the frequency domain vibration signal is estimated to obtain the power spectral density distribution of the frequency domain vibration signal; Based on the power spectral density distribution, the characteristic frequency components of the device's operating state are calculated using the spectral centroid algorithm. The calculation formula for the characteristic frequency components is as follows: ； In the formula, The characteristic frequency component, For the first Frequency values ​​at each frequency point To indicate at frequency The power spectral density amplitude at that point, To analyze the total number of sampling points within the frequency band; The characteristic frequency components are normalized to obtain the standardized characteristic frequency index of the lubricating oil equipment. The standardized characteristic frequency index is used as the characteristic frequency component characterizing the operating state of the equipment.

3. The lubricating oil equipment condition monitoring and early warning method as described in claim 1, characterized in that, The calculation of the slope value of the temperature change of the operating temperature signal per unit time, in order to obtain the dynamic temperature change trend of the operating temperature signal, includes: The operating temperature signal is discretized over time to obtain a sequence of temperature sample values ​​for the operating temperature signal. Based on the temperature sampling value sequence, the slope value of the dynamic temperature change trend is calculated through linear regression analysis, wherein the formula for calculating the slope value is as follows: ； In the formula, The slope value is... For discrete time points, This is a temperature sample value. This is the average value of the temperature samples. It is the average value at different time points. This represents the total number of sampling points; The slope value is smoothed to obtain the dynamic temperature change trend index of the operating temperature signal; The temperature dynamic change trend index is mapped to the temperature dynamic change trend of the lubricating oil equipment.

4. The lubricating oil equipment condition monitoring and early warning method as described in claim 3, characterized in that, The step of performing time-series discretization processing on the operating temperature signal to obtain a sequence of temperature sample values ​​for the operating temperature signal includes: Based on the Nyquist sampling theorem, a fixed sampling window is set for the lubricating oil equipment; Based on the fixed sampling window, the timing deviation of the operating temperature signal is eliminated to obtain a regular temperature sampling value sequence of the operating temperature signal; The integrity of the regularized temperature sample value sequence is verified to obtain the temperature sample value sequence of the operating temperature signal.

5. The lubricating oil equipment condition monitoring and early warning method as described in claim 1, characterized in that, The step of inputting the characteristic frequency components and the dynamic temperature change trend into a preset weighted fusion model to generate a comprehensive health status index for the lubricating oil equipment includes: The dimensions of the characteristic frequency components are unified to obtain the standard frequency characteristic index of the characteristic frequency components. The temperature dynamic change trend is normalized to obtain a standard temperature trend index for the temperature dynamic change trend. Based on the operating stage characteristics of the lubricating oil equipment, assign corresponding dynamic weight coefficients to the standardized frequency characteristic index and the standard temperature trend index; Based on the dynamic weighting coefficient, the standard frequency characteristic index and the standard temperature trend index are nonlinearly integrated to obtain the preliminary comprehensive health status index of the lubricating oil equipment. Based on a preset threshold, the preliminary comprehensive health status index is evaluated using a confidence level assessment based on the preset threshold to obtain the comprehensive health status index of the lubricating oil equipment.

6. The lubricating oil equipment condition monitoring and early warning method as described in claim 5, characterized in that, The preliminary comprehensive health status index of the lubricating oil equipment is obtained by nonlinearly integrating the standard frequency characteristic index and the standard temperature trend index based on the dynamic weighting coefficient, including: Based on the dynamic weighting coefficient, the standard frequency characteristic index and the standard temperature trend index are assigned weights in a hierarchical manner to obtain the weighted frequency characteristics and the weighted temperature trend. A nonlinear feature transformation is performed on the weighted frequency characteristics and the weighted temperature trend to obtain the enhanced feature representation of the lubricating oil equipment. The enhanced feature representations are then subjected to feature interaction fusion to obtain a preliminary comprehensive health status index of the lubricating oil equipment.

7. The lubricating oil equipment condition monitoring and early warning method as described in claim 1, characterized in that, The dynamic updating of the early warning threshold based on the status data of the lubricating oil equipment during its historical operating cycle includes: Extract the state data sequence within the historical operating cycle from the operating database of the lubricating oil equipment to construct the historical state dataset of the lubricating oil equipment; Outliers in the historical state dataset are removed to obtain the regularized historical state data of the historical state dataset. Trend features are extracted from the regularized historical state data to identify the long-term trend features of state changes in the lubricating oil equipment; Based on the long-term trend characteristics, a preliminary early warning threshold for the lubricating oil equipment is generated; The rationality of the preliminary warning threshold is verified to obtain the dynamic warning threshold of the lubricating oil equipment.

8. The lubricating oil equipment condition monitoring and early warning method as described in claim 1, characterized in that, The step of comparing the comprehensive health status index with the warning threshold in real time and determining the real-time status of the lubricating oil equipment based on the comparison result includes: A multi-level state determination threshold system for the lubricating oil equipment is constructed to accurately divide the operating state of the lubricating oil equipment into normal state, warning state, and abnormal state. The comprehensive health status index is compared with the multi-level status judgment threshold system in real time to obtain the preliminary judgment result of the status level of the lubricating oil equipment. A state transition continuity analysis is performed on the preliminary state level determination results, and the reliability of the determination results is evaluated in conjunction with the historical operating status data of the lubricating oil equipment. Based on the credibility assessment results, the real-time status judgment conclusion of the lubricating oil equipment is output, and a corresponding confidence index is assigned to each status level.

9. The lubricating oil equipment condition monitoring and early warning method as described in claim 8, characterized in that, The real-time status determination conclusion of the lubricating oil equipment is output based on the confidence assessment result, and a corresponding confidence index is assigned to each status level, including: Based on the statistical distribution characteristics of the regularized historical state data, a mapping relationship system between the state level and confidence index in the lubricating oil equipment is established. The basic confidence level of the lubricating oil equipment is determined based on the degree of consistency between the current state determination result and the regularized historical state data. The baseline confidence level is dynamically adjusted based on the operating time, maintenance records, and operating parameters of the lubricating oil equipment. Based on the vibration characteristics, temperature trends, and regularized historical state data of the lubricating oil equipment, a confidence index for the lubricating oil equipment is generated.

10. A lubricating oil equipment condition monitoring and early warning system, used to implement the lubricating oil equipment condition monitoring and early warning method according to claim 1, the system comprising: The signal synchronization acquisition module is used to synchronously acquire the time-domain vibration signal and operating temperature signal of the lubricating oil equipment; The frequency domain feature extraction module is used to perform a fast Fourier transform on the time domain vibration signal, convert the time domain vibration signal into a frequency domain vibration signal, and extract the characteristic frequency components representing the operating state of the equipment from the frequency domain vibration signal. The temperature trend analysis module is used to calculate the slope of the temperature change of the operating temperature signal per unit time, so as to obtain the dynamic temperature change trend of the operating temperature signal. The health index generation module is used to input the characteristic frequency components and the temperature dynamic change trend into a preset weighted fusion model to generate the comprehensive health status index of the lubricating oil equipment. The threshold update module is used to dynamically update the warning threshold based on the status data of the lubricating oil equipment in the historical operating cycle. The status determination module is used to compare the comprehensive health status index with the warning threshold in real time, and determine the real-time status of the lubricating oil equipment based on the comparison result. The early warning triggering module is used to trigger multi-level early warnings for the lubricating oil equipment when the real-time status is in an abnormal state.

Images