Equipment state monitoring method based on multi-source information fusion

Abstract

The invention discloses an equipment state monitoring method based on multi-source information fusion, and the method comprises the steps: enabling a real-time sensor to collect the operation parameter data, vibration parameter data and environment parameter data of equipment, collecting the historical state data of the equipment with a recording label, and carrying out the preprocessing of the data; time domain features, frequency domain features and working condition features of the data are extracted according to a layering mode, and dynamic weight coefficients of all the features are set; calculating a reference threshold value by adopting a specific model; evaluating the health condition of the current equipment by adopting a depth measurement method, calculating an equipment health factor, and calculating a trend compensation item; and obtaining an equipment state monitoring dynamic threshold based on the reference threshold, the health correction item and the trend compensation item. The invention further discloses an equipment state monitoring device based on multi-source information fusion, corresponding equipment and a storage medium. According to the equipment state monitoring method based on multi-source information fusion provided by the embodiment of the invention, the accuracy, real-time performance and reliability of equipment state monitoring can be effectively improved.

Description

Technical Field

[0001] The present application relates to the technical field of equipment health management, and more specifically, to an equipment status monitoring method based on multi-source information fusion. Background Art

[0002] In modern industrial production, the stable operation of various types of equipment is crucial for ensuring production efficiency, product quality, and safety. With the continuous improvement of equipment automation and the increasing complexity of equipment structure and functions, the requirements for equipment status monitoring technology are also increasing. During operation, equipment is affected by a variety of factors, such as varying operating conditions, complex environmental factors, and performance degradation, all of which can cause equipment failure. Failure of critical equipment not only results in significant economic losses but can also trigger safety incidents and pose serious risks to personnel and equipment. Therefore, a technology is needed that can accurately and in real time monitor equipment status and detect potential failures in advance.

[0003] Currently, equipment status monitoring primarily uses a fixed threshold method. This method pre-sets a fixed threshold and compares the monitored equipment operating parameters with this threshold to determine whether the equipment is in a normal state. However, the fixed threshold method has many limitations. Because it cannot adapt to the operational changes of equipment under multiple operating conditions and environments, when the operating conditions or environment change, the original fixed threshold may no longer be applicable, resulting in high rates of missed alarms and false alarms. In addition, while existing dynamic threshold technology can partially alleviate the above problems, it typically relies on only a single signal source for threshold adjustment and is still limited by its dependence on historical data. It has difficulty handling the coupled interaction problems between parameters, and its high computational complexity makes it difficult to meet the real-time requirements of complex equipment status monitoring and cannot provide a reliable basis for equipment fault diagnosis and prediction. Summary of the Invention

[0004] In response to at least one defect or improvement need in the prior art, the present invention provides an equipment status monitoring method based on multi-source information fusion, which can solve at least one of the problems existing in the above-mentioned background technology.

[0005] To achieve the above objectives, according to a first aspect of the present invention, a method for equipment status monitoring based on multi-source information fusion is provided, the method comprising: Real-time sensors collect equipment operating parameter data, vibration parameter data, and environmental parameter data, collect equipment historical status data with record tags, and pre-process the data; Extract the time domain features, frequency domain features and working condition features of the data in a hierarchical manner, and set the dynamic weight coefficient of each feature; Calculate the benchmark threshold using a specific model; Use deep measurement methods to evaluate the health of current equipment, calculate equipment health factors, and calculate trend compensation items; An equipment status monitoring dynamic threshold is obtained based on the baseline threshold, the health correction item, and the trend compensation item.

[0006] Furthermore, in the above-mentioned equipment status monitoring method based on multi-source information fusion, the data preprocessing includes filtering out redundant data, synchronizing data clocks, performing segmented coding according to working conditions, and normalizing features.

[0007] Furthermore, in the above-mentioned equipment status monitoring method based on multi-source information fusion, extracting data in a hierarchical manner specifically includes: The time domain features are extracted using RMS, peak, kurtosis and impulse factor methods; The frequency domain features are extracted using envelope spectrum and wavelet energy entropy methods; The operating condition characteristics are extracted using a rotational speed, load, and environment matrix coding method.

[0008] Furthermore, in the above-mentioned equipment status monitoring method based on multi-source information fusion, the reference threshold value calculation using a specific model specifically includes: The following model is used for baseline threshold Calculation

[0009] in, For the i eigenvalues, is the experience compensation item under the current working condition, is a weight coefficient, which is self-updated through online learning.

[0010] Furthermore, in the above-mentioned equipment status monitoring method based on multi-source information fusion, the calculation of the equipment health factor and the calculation of the trend compensation item specifically include: The equipment health factor is calculated using the following method

[0011]

[0012] in is the feature vector at the current moment, is the baseline state feature vector; The baseline state feature vector is the ideal state of the equipment when it is healthy, and is obtained through training with historical normal data; Will Normalized to [0,1], the closer to 1, the healthier the equipment is; the closer to 0, the more serious the abnormality of the equipment. Calculating trend compensation

[0013]

[0014] in, Time window The sequence of changes in the health factor of internal equipment, is the output of the long short-term memory network, extracting its Dependent features within time, output trend-related feature values, is a dynamic decay factor used to weight the importance of recent trends and is a time decay function or a learned parameter.

[0015] Furthermore, the equipment status monitoring method based on multi-source information fusion, wherein the dynamic threshold value for equipment status monitoring is obtained based on the reference threshold value, health correction item, and trend compensation item, specifically includes: The dynamic threshold Calculated by the following method

[0016] in, It is the sensitivity coefficient of health correction, which controls the adjustment range of real-time health status to the baseline value.

[0017] Furthermore, the above-mentioned equipment status monitoring method based on multi-source information fusion, after obtaining the equipment status monitoring dynamic threshold, further includes: When a false alarm occurs, the threshold relaxation coefficient is automatically increased; When an underreporting occurs, feature weight redistribution is triggered based on the sliding window incremental learning mechanism, and the window size is set to a preset number of samples. When the window is filled, part of the model is retrained using the samples in the window, while retaining the weights of time domain features, frequency domain features, and operating condition features to achieve continuous optimization and updating of the model.

[0018] According to a second aspect of the present invention, there is also provided an equipment status monitoring device based on multi-source information fusion, characterized in that it includes: The data acquisition module is used to collect the equipment's operating parameter data, vibration parameter data, and environmental parameter data through real-time sensors, collect equipment historical status data with record tags, and pre-process the data; The feature extraction module is used to extract the time domain features, frequency domain features and working condition features of the data in a hierarchical manner and set the dynamic weight coefficient of each feature; A first calculation module is used to calculate a benchmark threshold using a specific model; The second calculation module is used to evaluate the health status of the current equipment using a depth measurement method, calculate the equipment health factor, and calculate the trend compensation item; The dynamic threshold acquisition module is used to obtain the equipment status monitoring dynamic threshold based on the reference threshold, health correction item and trend compensation item.

[0019] According to the third aspect of the present invention, there is also provided an equipment status monitoring device based on multi-source information fusion, which includes at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit performs the steps of any one of the above methods.

[0020] According to the fourth aspect of the present invention, a storage medium is also provided, which stores a computer program that can be executed by an equipment status monitoring device based on multi-source information fusion. When the computer program runs on the equipment status monitoring device based on multi-source information fusion, the equipment status monitoring device based on multi-source information fusion executes the steps of any one of the above-mentioned methods.

[0021] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art: The present invention provides an equipment status monitoring method based on multi-source information fusion. Through a series of steps, including multi-source data collection and preprocessing, feature extraction and weight coefficient setting, baseline threshold calculation, health assessment and trend compensation, and dynamic threshold acquisition, this method can effectively improve the accuracy, real-time performance, and reliability of equipment status monitoring. This method can adaptively adjust monitoring thresholds under different operating conditions and environments, reducing both missed and false alarm rates, providing strong technical support for equipment fault diagnosis and prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 A flowchart of an equipment status monitoring method based on multi-source information fusion provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0025] The terms "first," "second," "third," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish between different objects, not to describe a particular order. Furthermore, the terms "including," "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or elements is not limited to the listed steps or elements, but may optionally include steps or elements not listed, or may optionally include other steps or elements inherent to the process, method, product, or apparatus.

[0026] Figure 1 A flow chart of an equipment status monitoring method based on multi-source information fusion provided in an embodiment of the present application is shown as follows: Figure 1 As shown, an embodiment of the present application provides an equipment status monitoring method based on multi-source information fusion, comprising the following steps: Real-time sensors collect equipment operating parameter data, vibration parameter data, and environmental parameter data, collect equipment historical status data with record tags, and pre-process the data; Extract the time domain features, frequency domain features and working condition features of the data in a hierarchical manner, and set the dynamic weight coefficient of each feature; Calculate the benchmark threshold using a specific model; Use deep measurement methods to evaluate the health of current equipment, calculate equipment health factors, and calculate trend compensation items; An equipment status monitoring dynamic threshold is obtained based on the baseline threshold, the health correction item, and the trend compensation item.

[0027] Specifically, in the industrial field, equipment status monitoring is crucial for ensuring production safety and improving production efficiency. The present application provides an equipment status monitoring method based on multi-source information fusion, which can improve the accuracy, real-timeness and reliability of monitoring.

[0028] First, multiple sensors are used to collect real-time data on the equipment's operating, vibration, and environmental parameters. Operating parameter data typically includes temperature, pressure, flow rate, current, and rotational speed. Vibration parameter data includes acceleration, velocity, and displacement. Environmental parameter data includes ambient temperature and humidity. Furthermore, historical equipment status data is collected, tagged with labels indicating normal or abnormal conditions. The collected data often contains noise and interference, and may be incomplete or overly large. Therefore, data preprocessing is necessary to ensure the accuracy of subsequent analysis. This preprocessing involves filtering out redundant data irrelevant to data mining, such as duplicate, meaningless, or irrelevant data to the monitoring objectives. The data is then clock-synchronized to eliminate errors caused by asynchronous sampling times. Finally, the data is segmented and encoded according to different operating conditions, and the features are normalized. This allows data of different dimensions and ranges to be compared and analyzed on the same scale.

[0029] After data preprocessing, the time domain features, frequency domain features, and operating condition features of the data are extracted in a hierarchical manner. Time domain features usually include root mean square, peak value, kurtosis, pulse factor, etc. These features can reflect the statistical characteristics of the signal in the time domain. Frequency domain features include envelope spectrum, wavelet energy entropy, etc., which can reveal the distribution and energy characteristics of the signal in the frequency domain. Operating condition features are mainly matrix codes such as speed, load, and environment. These features reflect the operating status of the equipment under different operating conditions. In order to highlight important features and suppress unimportant features, a dynamic weight coefficient is set for each feature. The dynamic weight coefficient can be adjusted according to the importance of the feature under different operating conditions and the changing trend of the feature, so that the status of the equipment can be more accurately reflected in subsequent analysis.

[0030] A specific model is used to calculate the baseline threshold. This threshold is determined based on the equipment's characteristic values ​​during normal operation and provides a reference standard for subsequent condition monitoring. The calculation model typically incorporates empirical compensation terms based on current operating conditions to adapt to varying operating conditions. Weight coefficients can be automatically updated through online learning to adapt to changes in equipment characteristics and operating environment.

[0031] The current health status of equipment is assessed using deep metrics. The equipment's health factor can be calculated to quantify its health. The equipment health factor is typically derived by comparing the current feature vector with the baseline feature vector. The health factor generally ranges from [0, 1], with values ​​closer to 1 indicating healthier equipment and values ​​closer to 0 indicating serious equipment anomalies. A trend compensation term is then calculated, taking into account the changing trend of the health factor over time. Using methods such as long short-term memory networks (LSTMs), the dependent features within the time window are extracted to predict possible future health status changes.

[0032] Based on the baseline threshold, health correction term, and trend compensation term, dynamic thresholds for equipment status monitoring are derived. Dynamic thresholds can be adjusted based on the equipment's real-time health and operating trends, enabling real-time, accurate monitoring of equipment status. By comparing actual characteristic values ​​with dynamic thresholds, abnormal equipment status can be promptly detected, providing a basis for fault diagnosis and prediction.

[0033] The present invention provides an equipment status monitoring method based on multi-source information fusion. Through a series of steps, including multi-source data collection and preprocessing, feature extraction and weight coefficient setting, baseline threshold calculation, health assessment and trend compensation, and dynamic threshold acquisition, this method can effectively improve the accuracy, real-time performance, and reliability of equipment status monitoring. This method can adaptively adjust monitoring thresholds under different operating conditions and environments, reducing both missed and false alarm rates, providing strong technical support for equipment fault diagnosis and prediction.

[0034] Optionally, in the equipment status monitoring method based on multi-source information fusion provided in an embodiment of the present application, the data preprocessing includes filtering out redundant data, synchronizing data clocks, performing segmented coding according to working conditions, and normalizing features.

[0035] Specifically, collected equipment data often contains a large amount of redundant information. This redundant data may originate from repeated sensor sampling, background noise unrelated to condition monitoring, or brief disturbances during equipment operation. The goal of filtering out redundant data is to retain valid information closely related to equipment status, reduce data volume, and improve the efficiency and accuracy of subsequent analysis. Specific methods include employing filtering algorithms, such as low-pass, high-pass, and band-pass filters. Based on the data characteristics and monitoring requirements, appropriate filters are selected and a suitable cutoff frequency is set to filter out high-frequency noise or low-frequency interference outside the cutoff frequency range, retaining the valid frequency components that primarily reflect changes in equipment status. Furthermore, data cleaning techniques can be used to identify and remove obviously abnormal or erroneous data points, such as values ​​outside the physical range or values ​​that differ significantly from adjacent data points. These data points may be caused by sensor failures, signal transmission errors, and other factors. Filtering out this redundant data helps improve data quality.

[0036] Because different sensors may have different sampling frequencies and sampling times, the collected data may not be completely synchronized in time. To accurately analyze equipment status changes at different times and the relationships between parameters, data clock synchronization is necessary. Clock synchronization can be based on interpolation algorithms such as linear, polynomial, and spline interpolation. This ensures that data from different parameters can be compared and correlated on the same time scale in subsequent analysis, avoiding misjudgments or information loss due to time asynchrony.

[0037] When equipment operates under different operating conditions, its parameters exhibit different characteristics. Therefore, segmented encoding of data according to operating conditions is an important step in achieving accurate status monitoring. First, it is necessary to define and classify the various operating conditions of the equipment. During the data preprocessing stage, the collected continuous data is divided into different operating condition segments based on pre-set operating condition judgment conditions, such as the combination range of parameters such as speed, load, and ambient temperature. By segmenting the data according to the operating conditions, subsequent feature extraction and analysis can be more targeted, and the status of the equipment can be accurately judged based on the characteristic change patterns under different operating conditions, thereby improving the accuracy of monitoring.

[0038] Feature normalization converts feature data of different dimensions and numerical ranges to the same or similar scales to facilitate subsequent data analysis and modeling. Common normalization methods include standardization and range normalization. Standardization transforms the data into a standard normal distribution with a mean of 0 and a standard deviation of 1. Range normalization linearly transforms the data to a specified interval, typically [0, 1]. Feature normalization eliminates the impact of large numerical differences between features on analysis results, ensuring that each feature has equal importance in subsequent model training and status assessment, thereby improving the performance and generalization capabilities of monitoring models.

[0039] Through the above data preprocessing steps, the quality and availability of equipment status monitoring data can be effectively improved, laying a solid foundation for subsequent feature extraction, threshold calculation, and status evaluation, thereby achieving accurate and reliable monitoring of equipment status.

[0040] Optionally, in the equipment status monitoring method based on multi-source information fusion provided in an embodiment of the present application, extracting data in a hierarchical manner specifically includes: The time domain features are extracted using RMS, peak, kurtosis and impulse factor methods; The frequency domain features are extracted using envelope spectrum and wavelet energy entropy methods; The operating condition characteristics are extracted using a rotational speed, load, and environment matrix coding method.

[0041] Specifically, in the embodiment of the present application, extracting data in a hierarchical manner is the core link to achieve accurate monitoring of equipment status.

[0042] Time-domain features reflect the statistical characteristics of equipment operating parameters in the time domain and are important basic information for condition monitoring. In this embodiment, the root mean square (RMS), peak value, kurtosis, and impulse factor methods are primarily used to extract time-domain features. RMS is a key metric for measuring the effective value of a signal and reflects the average energy level of the signal. To calculate this metric, the acquired time-domain signal is first squared, then averaged, and finally the square root is taken. The peak value represents the maximum amplitude of the signal within a specific time range and is crucial for detecting impact-induced faults in equipment. In signal monitoring, if the signal amplitude at a given moment is significantly higher than the peak value during normal operation, it may indicate that the equipment has experienced a sudden impact load or a localized, severe fault. Real-time peak value monitoring allows for rapid detection of such abnormal events. Kurtosis describes the shape of a signal's probability distribution and reflects the strength of the impact component in the signal. Calculating kurtosis requires raising the signal to the fourth power and taking the statistical average. In equipment condition monitoring, the kurtosis value increases when a signal contains a high number of impact pulses. Monitoring kurtosis changes can provide early warning of faults. The pulse factor is the ratio of the signal's peak value to its mean value, which highlights the pulse components in the signal. When intermittent faults occur in equipment, the pulse factor will change.

[0043] Frequency domain features can reveal the distribution and energy characteristics of equipment operation signals in the frequency domain, and help to discover periodic components in the signal and fault characteristics of specific frequencies. The embodiment of the present application uses two methods, envelope spectrum and wavelet energy entropy, to extract frequency domain features. Envelope spectrum analysis is an effective method for feature extraction of modulated signals. In equipment status monitoring, many fault signals usually appear in the form of modulated signals. By first performing a Hilbert transform on the original signal to obtain an analytical signal, then taking its envelope, and then performing a Fourier transform on the envelope, the envelope spectrum can be obtained. Wavelet energy entropy is a feature extraction method based on wavelet transform, which can reflect the complexity of the energy distribution of the signal at different scales. Wavelet transform decomposes the signal into different scales, calculates the proportion of energy at each scale to the total energy, and then calculates the entropy value based on these proportions. In equipment status monitoring, when the operating status of the equipment changes, the energy distribution of the signal at different scales will also change, and wavelet energy entropy can capture this change.

[0044] Operating condition characteristics reflect the operating state of equipment under different operating conditions and are of great significance for implementing adaptive condition monitoring. The present embodiment utilizes a matrix encoding method for speed, load, and environment to extract operating condition characteristics. Speed ​​is a key operating parameter for rotating equipment. Quantizing and encoding the speed signal can convert it into a feature vector. By monitoring changes in the speed encoding, changes in the equipment's operating speed state can be understood. This can be combined with other features for comprehensive analysis to meet the needs of condition monitoring at different speeds. Load magnitude directly impacts equipment characteristics such as operating stress and energy consumption. Matrix encoding of the load can transform it into an analyzable feature form. When the load suddenly increases or decreases, monitoring strategies can be adjusted promptly to improve the accuracy of condition monitoring. Equipment operating environment, such as temperature and humidity, can affect its performance and lifespan. Matrix encoding of environmental parameters can comprehensively reflect the impact of environmental factors on equipment status. For example, an environmental matrix can be constructed, where rows represent different environmental parameters (such as temperature and humidity), columns represent different value ranges or levels for the parameters, and matrix elements represent weights or influencing factors for corresponding environmental conditions. Environmental matrix coding can be used to quantitatively evaluate the comprehensive impact of environmental factors on equipment status, so that they can be taken into account in status monitoring and the reliability of monitoring results can be improved.

[0045] Through the above-mentioned hierarchical feature extraction method, the characteristic information of the equipment in the time domain, frequency domain and working conditions can be comprehensively and accurately obtained, providing rich data support for subsequent dynamic threshold calculation and status monitoring, and realizing accurate and real-time monitoring of the equipment status.

[0046] Optionally, in the equipment status monitoring method based on multi-source information fusion provided in an embodiment of the present application, the reference threshold value calculation using a specific model specifically includes: The following model is used for baseline threshold Calculation

[0047] in, For the i eigenvalues, is the experience compensation item under the current working condition, is a weight coefficient, which is self-updated through online learning.

[0048] Specifically, the present embodiment adopts the following model to perform the benchmark threshold Calculation

[0049] in, For the iThese eigenvalues ​​can be extracted from the time domain, frequency domain and working condition characteristics. It is the experience compensation item under the current working condition, which is used to adjust the impact of different working conditions on the threshold. are weight coefficients, which are continuously updated through online learning mechanisms to adapt to changes in the equipment's operating status and environmental factors.

[0050] As time goes by and more data accumulates, the weight coefficients can be dynamically adjusted through online learning. For example, if the speed feature is found to be more sensitive to faults in subsequent runs, the weight coefficients can be adjusted through the learning algorithm.

[0051] Optionally, in the equipment status monitoring method based on multi-source information fusion provided in an embodiment of the present application, the calculation of the equipment health factor and the calculation of the trend compensation item specifically include: The equipment health factor is calculated using the following method

[0052]

[0053] in is the feature vector at the current moment, is the baseline state feature vector; The baseline state feature vector is the ideal state of the equipment when it is healthy, and is obtained through training with historical normal data; Will Normalized to [0,1], the closer to 1, the healthier the equipment is; the closer to 0, the more serious the abnormality of the equipment. Calculating trend compensation

[0054]

[0055] in, Time window The sequence of changes in the health factor of internal equipment, is the output of the long short-term memory network, extracting its Dependent features within time, output trend-related feature values, is a dynamic decay factor used to weight the importance of recent trends and is a time decay function or a learned parameter.

[0056] Specifically, the embodiment of the present application uses the following method to calculate the equipment health factor:

[0057]

[0058] in It is the feature vector at the current moment, which can be obtained by collecting multi-source data of equipment in real time and extracting features. It can be expressed as a numerical value extracted from sensors, logs or other data. These data include time domain features, frequency domain features and working condition features, etc. It is the baseline state feature vector, which is obtained through training with a large amount of historical normal data and represents the ideal feature mean under the healthy state of the equipment.

[0059] Calculated using the cosine distance formula and The cosine distance measures the difference in direction between two vectors and reflects the degree of deviation between the current state of the equipment and the baseline health state.

[0060] Calculating health factors , and normalize it to the interval [0, 1]. The closer the value is to 1, the healthier the equipment is; the closer it is to 0, the more serious the abnormality is. The value is 0.8, which means that the equipment is in good condition, but its development trend still needs to be monitored.

[0061] The embodiment of the present application uses the following method to calculate the trend compensation term

[0062]

[0063] Set time window , collect the change sequence of equipment health factor within the window , this sequence reflects the recent trend of changes in the health status of the equipment. The above sequence is input into the long short-term memory network middle. The network can automatically learn the long-term dependencies in the sequence and extract the trend characteristics of the changes in health factors. Introducing dynamic attenuation factors , whose value range is between [0, 1]. This factor is used to weight the importance of recent trends and can be set based on a time decay function or obtained through a learning algorithm.

[0064] After completing the equipment health factor and trend compensation After the calculation of the , they are integrated into the overall framework of equipment condition monitoring. Value combined with the baseline threshold and The dynamic monitoring threshold is then compared to determine whether the equipment is in a normal state. If the value falls below the dynamic threshold, an early warning mechanism is triggered, reminding maintenance personnel to pay attention to and inspect the equipment in a timely manner. This dynamic monitoring method based on health factors and trend compensation can effectively improve the accuracy and timeliness of monitoring.

[0065] Optionally, in the equipment status monitoring method based on multi-source information fusion provided in an embodiment of the present application, obtaining a dynamic threshold for equipment status monitoring based on the baseline threshold, health correction item, and trend compensation item specifically includes: The dynamic threshold Calculated by the following method

[0066] in, It is the sensitivity coefficient of health correction, which controls the adjustment range of real-time health status to the baseline value.

[0067] Specifically, during equipment status monitoring, dynamic thresholds can be adaptively adjusted based on the equipment's real-time health status and operating trends, enabling precise monitoring and early warning. In the equipment status monitoring system, accurate judgment of the equipment's status can be achieved by comparing the monitored real-time characteristic values ​​of the equipment with the dynamic thresholds. When the real-time characteristic value exceeds the dynamic threshold, the system immediately issues a warning signal, alerting maintenance personnel to promptly pay attention to and inspect the equipment. The calculation of the dynamic threshold comprehensively considers the equipment's real-time health status and historical trends, providing maintenance personnel with more comprehensive status information. Based on the changing trends of the dynamic threshold, maintenance personnel can predict potential equipment failures in advance, rationally arrange maintenance plans, and reduce repair costs and downtime.

[0068] Optionally, the equipment status monitoring method based on multi-source information fusion provided in an embodiment of the present application, after obtaining the equipment status monitoring dynamic threshold, further includes: When a false alarm occurs, the threshold relaxation coefficient is automatically increased; When an underreporting occurs, feature weight redistribution is triggered based on the sliding window incremental learning mechanism, and the window size is set to a preset number of samples. When the window is filled, part of the model is retrained using the samples in the window, while retaining the weights of time domain features, frequency domain features, and operating condition features to achieve continuous optimization and updating of the model.

[0069] Specifically, when a false alarm occurs, the system automatically increases the threshold relaxation coefficient to reduce the false alarm rate.

[0070] Initially, the threshold relaxation coefficient is set to a small value. During monitoring, if the system issues a warning signal but the equipment is found to be in normal condition upon inspection, it is considered a false alarm. In this case, the system automatically increases the threshold relaxation coefficient, adjusting the step size based on actual conditions. The dynamic threshold is recalculated using the new threshold relaxation coefficient, and condition monitoring continues.

[0071] When a false negative occurs, the system triggers feature weight redistribution based on the sliding window incremental learning mechanism to improve the detection capability of the model.

[0072] The sliding window size is set to a preset number of samples. The window size can be adjusted based on actual conditions, for example, 200 samples. During monitoring, the actual equipment status is used as a label to collect new sample data. If the system fails to issue a timely warning signal, but the equipment is actually experiencing an abnormal state, it is considered a missed warning. The sample data within the window is used to retrain some model parameters, while retaining the weights of time domain features, frequency domain features, and operating condition features. Condition monitoring continues using the updated model parameters to improve the model's ability to detect equipment abnormalities.

[0073] To verify the effect of model optimization, the following methods can be used: Using equipment operation data from the past period, compare the optimized model with the original model and calculate changes in false alarm and missed alarm rates. If both false alarm and missed alarm rates are reduced after the optimization, the model optimization effect is significant. In the actual equipment operation environment, continuously monitor the equipment status and record the comparison between the warning signals issued by the system and the actual equipment status. If the optimized model can more accurately predict equipment anomalies and reduce false alarms and missed alarms, the model optimization has achieved the desired effect.

[0074] The equipment status monitoring method based on multi-source information fusion provided in this application can achieve continuous optimization and updating of the model after obtaining dynamic thresholds through false alarm and missed alarm processing mechanisms. This not only improves the accuracy and reliability of equipment status monitoring, but also reduces the false alarm and missed alarm rates, providing strong guarantees for the stable operation of industrial equipment.

[0075] The present application also provides an equipment status monitoring device based on multi-source information fusion, including: The data acquisition module is used to collect the equipment's operating parameter data, vibration parameter data, and environmental parameter data through real-time sensors, collect equipment historical status data with record tags, and pre-process the data; The feature extraction module is used to extract the time domain features, frequency domain features and working condition features of the data in a hierarchical manner and set the dynamic weight coefficient of each feature; A first calculation module is used to calculate a benchmark threshold using a specific model; The second calculation module is used to evaluate the health status of the current equipment using a depth measurement method, calculate the equipment health factor, and calculate the trend compensation item; The dynamic threshold acquisition module is used to obtain the equipment status monitoring dynamic threshold based on the reference threshold, health correction item and trend compensation item.

[0076] The present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above method. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a microdrive, a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.

[0077] It should be noted that for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

[0078] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0079] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interface, and the indirect coupling or communication connection of the device or unit can be electrical or other forms.

[0080] The units described as separate components may or may not be physically separate, and the components shown as units 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 these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0081] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0082] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a memory and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk, or optical disk, etc., various media that can store program code.

[0083] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0084] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure herein, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art that are not recorded in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

[0085] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0086] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for equipment status monitoring based on multi-source information fusion, characterized in that: The following steps are involved: Real-time sensors collect equipment operating parameter data, vibration parameter data, and environmental parameter data, collect equipment historical status data with record tags, and pre-process the data; Extract the time domain features, frequency domain features and working condition features of the data in a hierarchical manner, and set the dynamic weight coefficient of each feature; Calculate the benchmark threshold using a specific model; Use deep measurement methods to evaluate the health of current equipment, calculate equipment health factors, and calculate trend compensation items; An equipment status monitoring dynamic threshold is obtained based on the baseline threshold, the health correction item, and the trend compensation item.

2. The equipment status monitoring method based on multi-source information fusion according to claim 1, characterized in that: The data preprocessing includes filtering out redundant data, synchronizing data clocks, performing segmented coding according to working conditions, and normalizing features.

3. The equipment status monitoring method based on multi-source information fusion according to claim 1, characterized in that: The data extraction in a hierarchical manner specifically includes: The time domain features are extracted using RMS, peak, kurtosis and impulse factor methods; The frequency domain features are extracted using envelope spectrum and wavelet energy entropy methods; The operating condition characteristics are extracted using a rotational speed, load, and environment matrix coding method.

4. The equipment status monitoring method based on multi-source information fusion according to claim 1, characterized in that: The use of a specific model to calculate the benchmark threshold specifically includes: The following model is used for baseline threshold Calculation in, For the i eigenvalues, is the experience compensation item under the current working condition, is a weight coefficient, which is self-updated through online learning.

5. The equipment status monitoring method based on multi-source information fusion according to claim 4, characterized in that: The calculation of the equipment health factor and the calculation of the trend compensation item specifically include: The equipment health factor is calculated using the following method in is the feature vector at the current moment, is the baseline state feature vector; The baseline state feature vector is the ideal state of the equipment when it is healthy, and is obtained through training with historical normal data; Will Normalized to [0,1], the closer to 1, the healthier the equipment is; the closer to 0, the more serious the abnormality of the equipment. Calculating trend compensation in, Time window The sequence of changes in the health factor of internal equipment, is the output of the long short-term memory network, extracting its Dependent features within time, output trend-related feature values, is a dynamic decay factor used to weight the importance of recent trends and is a time decay function or a learned parameter.

6. The equipment status monitoring method based on multi-source information fusion according to claim 5, characterized in that: The obtaining of the equipment status monitoring dynamic threshold based on the reference threshold, the health correction item, and the trend compensation item specifically includes: The dynamic threshold Calculated by the following method in, It is the sensitivity coefficient of health correction, which controls the adjustment range of real-time health status to the baseline value.

7. The equipment status monitoring method based on multi-source information fusion according to claim 1, characterized in that: After obtaining the dynamic threshold value of equipment status monitoring, it also includes: When a false alarm occurs, the threshold relaxation coefficient is automatically increased; When an underreporting occurs, feature weight redistribution is triggered based on the sliding window incremental learning mechanism, and the window size is set to a preset number of samples. When the window is filled, part of the model is retrained using the samples in the window, while retaining the weights of time domain features, frequency domain features, and operating condition features to achieve continuous optimization and updating of the model.

8. An equipment status monitoring device based on multi-source information fusion, characterized in that: include: The data acquisition module is used to collect the equipment's operating parameter data, vibration parameter data, and environmental parameter data through real-time sensors, collect equipment historical status data with record tags, and pre-process the data; The feature extraction module is used to extract the time domain features, frequency domain features and working condition features of the data in a hierarchical manner and set the dynamic weight coefficient of each feature; A first calculation module is used to calculate a benchmark threshold using a specific model; The second calculation module is used to evaluate the health status of the current equipment using a depth measurement method, calculate the equipment health factor, and calculate the trend compensation item; The dynamic threshold acquisition module is used to obtain the equipment status monitoring dynamic threshold based on the reference threshold, health correction item and trend compensation item.

9. An equipment status monitoring device based on multi-source information fusion, characterized in that: The method comprises at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit executes the steps of the method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: It stores a computer program that can be executed by an equipment status monitoring device based on multi-source information fusion. When the computer program runs on the equipment status monitoring device based on multi-source information fusion, the equipment status monitoring device based on multi-source information fusion executes the steps of the method described in any one of claims 1 to 7.

Images