Energy equipment health state assessment method and device, electronic equipment and medium

Abstract

The embodiment of the invention relates to the technical field of fault assessment of energy equipment, and provides an energy equipment health state assessment method and device, electronic equipment and a medium, and the method comprises the steps: collecting actual operation data and actual health states of target energy equipment under different working conditions, and carrying out the analysis to obtain fault features under different working conditions; inputting the fault features under different working conditions into a matching network to obtain fault information under different working conditions output by the matching network; constructing a mechanism model based on the actual operation data and the actual health state, and obtaining mechanism characteristics based on the mechanism model; reconstructing fault information under different working conditions by taking the mechanism characteristics as guidance to obtain a fault characteristic system of the whole life cycle of the target energy equipment; and performing real-time health state evaluation on the target energy equipment based on the fault feature system. Therefore, high-quality analysis of early fault information in energy equipment, consistency alignment of fault data distribution under different working conditions and fault information joint modeling are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of fault assessment of energy equipment, and in particular to a method, device, electronic equipment and medium for assessing the health status of energy equipment. Background Art

[0002] During operation, energy equipment often faces complex operating conditions such as high temperature, high pressure, and severe corrosion, leading to increasingly prominent problems of aging and performance degradation of key equipment components. At the same time, the operation and maintenance of energy equipment are often restricted by multiple factors such as environmental conditions, operating load, and maintenance resources. Traditional regular maintenance and passive fault repair models are difficult to meet the requirements of high reliability and low downtime rates. In particular, failures in energy equipment are often hidden and sudden, and there may only be a short warning window from initial minor anomalies to functional failure. Through intelligent assessment technology, it is possible to utilize large amounts of real-time operating data to fully perceive the dynamic changes in the health status of equipment. In combination with advanced artificial intelligence algorithms, a health assessment model for the entire life cycle of the equipment can be constructed, thereby achieving accurate diagnosis of the status of key components and life prediction, providing a scientific basis for proactive maintenance and operational optimization of the equipment.

[0003] Traditional mechanistic model-based approaches can provide clear reliability and degradation analysis by describing the operating principles and physical processes of equipment. However, they often face challenges in implementation and parameter uncertainty in complex systems. Purely data-driven approaches, relying on large-scale, real-time data for rapid assessment and prediction, show great potential for intelligent evaluation of energy equipment. However, they also face challenges such as large data requirements and limited model generalization.

[0004] Furthermore, existing solutions lack high-quality analysis technology for early-stage fault information from energy equipment, making it difficult to accurately and efficiently extract effective fault information from early-stage fault data, impacting the accuracy of intelligent assessment models for the health status of energy equipment. The lack of joint modeling technology for multi-operating-condition fault information for energy equipment makes it difficult to effectively match accumulated fault knowledge under different operating conditions and use it for joint assessment modeling, impacting the generalizability of intelligent assessment models for the health status of energy equipment. Summary of the Invention

[0005] The present invention provides a method, device, electronic device and medium for evaluating the health status of energy equipment, which are used to solve the problems in the prior art, such as the lack of high-quality analysis technology for early fault information of energy equipment, difficulty in accurately and efficiently mining effective fault information from early fault data, and impact on the accuracy of the intelligent evaluation model for the health status of energy equipment; the lack of joint modeling technology for multi-working condition fault information of energy equipment, difficulty in effectively matching the fault knowledge accumulated under different working conditions and using it for joint evaluation modeling, and impact on the generalization of the intelligent evaluation model for the health status of energy equipment, thereby achieving high-quality analysis of early fault information in energy equipment, consistent alignment of fault data distribution under different working conditions, and joint modeling of fault information.

[0006] The present invention provides a method for evaluating the health status of energy equipment, comprising: Collect actual operating data and actual health status of target energy equipment under different working conditions; Obtaining fault characteristics under different working conditions based on the actual operating data and actual health status analysis, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health states; Inputting the fault characteristics under the different working conditions into the matching network to obtain the fault information under the different working conditions output by the matching network; Building a mechanism model based on the actual operating data and the actual health status, and obtaining mechanism characteristics based on the mechanism model, wherein the mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment; Reconstructing the fault information under the different working conditions based on the mechanism characteristics to obtain a fault feature system for the entire life cycle of the target energy equipment; A real-time health status assessment is performed on the target energy equipment based on the fault characteristic system.

[0007] In one possible implementation, the method further includes: The actual operation data is collected through multiple monitoring channels; Analyzing the correlation between the actual operating data and the actual health status; defining a first correlation loss function by maximizing the correlation between the actual operating data and the actual health status; Analyzing the correlation between the actual operation data collected by the multiple monitoring channels; By minimizing the correlation between the actual operation data, a second correlation loss function is defined; Constructing a third loss function based on the first correlation loss function and the second correlation loss function; The actual operating data and actual health status are analyzed based on the third loss function to obtain fault characteristics under different working conditions.

[0008] In one possible implementation, the method further includes: The fault characteristics under the different working conditions are input into the matching network, and the fault characteristics under the different working conditions are fused through the dynamic multi-supervisory sensor, dynamic weight matrix and bias vector in the matching network to obtain the fault information under the different working conditions output by the matching network.

[0009] In one possible implementation, the method further includes: The weights in the dynamic weight matrix and the bias vector are optimized based on the fault characteristics and adjustable network parameters under the different working conditions.

[0010] In one possible implementation, the method further includes: Extracting actual operating data under the different working conditions and fault characteristic parameters corresponding to the actual health status; Constructing a mechanism model based on the fault characteristic parameters; The physical principles and failure mechanisms of the target energy equipment are analyzed based on the mechanism model to obtain mechanism characteristics.

[0011] In one possible implementation, the method further includes: Guided by the mechanism characteristics, the fault information under the different working conditions and the mechanism characteristics are reconstructed by a weighted summation method; The reconstructed features are adjusted using an optimization algorithm, and a fault feature system for the entire life cycle of the target energy equipment is generated based on the optimized features.

[0012] In one possible implementation, the method further includes: Use the fault feature system as the input feature of the random forest algorithm to train the random forest model; Based on the random forest model, the real-time health status of the target energy equipment is classified and regressed to obtain a health status assessment result of the target energy equipment.

[0013] The present invention also provides an energy equipment health status assessment device, comprising the following modules: The acquisition module is used to collect the actual operating data and actual health status of the target energy equipment under different working conditions; An analysis module, configured to obtain fault characteristics under different operating conditions based on the actual operating data and actual health status analysis; An input module, configured to input the fault characteristics under the different working conditions into the matching network, and obtain the fault information under the different working conditions output by the matching network; A construction module, configured to construct a mechanism model based on the actual operation data and the actual health status, and obtain mechanism characteristics based on the mechanism model; The construction module is further configured to reconstruct the fault information under the different working conditions based on the mechanism characteristics to obtain a fault feature system for the entire life cycle of the target energy equipment; An evaluation module is used to perform real-time health status evaluation on the target energy equipment based on the fault characteristic system.

[0014] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for assessing the health status of energy equipment as described above is implemented.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described energy equipment health status assessment methods.

[0016] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for assessing the health status of energy equipment as described above is implemented.

[0017] The energy equipment health status assessment method, device, electronic device and medium provided by the present invention collect actual operating data and actual health status of the target energy equipment under different working conditions; obtain fault characteristics under different working conditions based on the actual operating data and actual health status analysis, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health states; input the fault characteristics under different working conditions into a matching network to obtain fault information under different working conditions output by the matching network; construct a mechanism model based on the actual operating data and actual health status, and obtain mechanism characteristics based on the mechanism model, wherein the mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment; reconstruct the fault information under the different working conditions guided by the mechanism characteristics to obtain a fault characteristic system for the entire life cycle of the target energy equipment; and perform real-time health status assessment on the target energy equipment based on the fault characteristic system. Compared with the existing technology, there is a lack of high-quality analysis technology for early fault information of energy equipment, which makes it difficult to accurately and efficiently mine effective fault information from early fault data, affecting the accuracy of the intelligent assessment model of the health status of energy equipment; there is a lack of joint modeling technology for multi-working condition fault information of energy equipment, which makes it difficult to effectively match the fault knowledge accumulated under different working conditions and use it for joint assessment modeling, affecting the generalization of the intelligent assessment model of the health status of energy equipment. This solution proposes a high-quality and low-rank early fault information analysis method to address the problem that early fault information in energy equipment is difficult to parse. To address the problem that fault information of multi-working condition energy equipment is difficult to match, a dynamic multi-supervisory sensor for multi-working condition fault information fusion is proposed to realize dynamic matching of fault information under multiple working conditions, achieve high-quality analysis of early fault information in energy equipment, and achieve consistent alignment of fault data distribution under different working conditions and joint modeling of fault information. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 This is one of the flow charts of the energy equipment health status assessment method provided by the present invention.

[0020] Figure 2 It is a schematic diagram of the process of S12 provided by the present invention.

[0021] Figure 3 This is the second flow chart of the energy equipment health status assessment method provided by the present invention.

[0022] Figure 4 This is a flow chart for evaluating the health status of energy equipment guided by the mechanism provided by the present invention.

[0023] Figure 5 It is a structural schematic diagram of the energy equipment health status assessment device provided by the present invention.

[0024] Figure 6 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0025] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0026] To facilitate understanding of the embodiments of the present invention, specific embodiments will be further explained below with reference to the accompanying drawings. The embodiments do not limit the embodiments of the present invention.

[0027] Figure 1 This is one of the flow charts of the method for evaluating the health status of energy equipment provided by the present invention, such as Figure 1 As shown, the method includes the following: S11. Collect actual operating data and actual health status of target energy equipment under different working conditions.

[0028] In an embodiment of the present invention, for the actual operating data collected from energy equipment under multiple operating conditions, intelligent analysis of weak monitoring signals coupled with early faults is implemented to obtain high-quality, low-rank fault features, that is, the feature data has high quality and the correlation between features is low.

[0029] Specifically, first, actual operating data collection: Energy equipment generates a large amount of data during operation. This data includes, but is not limited to, equipment operating parameters (such as temperature, pressure, flow, current, and voltage), sensor signals (such as vibration, noise, and oil composition), and equipment operation logs. This data can reflect the equipment's performance under different operating conditions. When collecting this data, ensuring its accuracy and completeness is crucial. This can be achieved through multiple monitoring channels. This is typically achieved using specialized data acquisition systems and sensor networks. For example, in a power generation device, temperature and vibration sensors installed in key locations collect real-time data such as temperature changes and vibration amplitude during operation.

[0030] Acquiring the actual health status of energy equipment: On the one hand, the actual health status of energy equipment can be determined through expert judgment based on the equipment's operating performance and historical maintenance records, such as healthy, early failure, or serious failure. On the other hand, key components of the equipment can be tested using testing equipment and technologies, such as non-destructive testing and oil analysis, to determine the actual health status of the equipment. For example, oil analysis can detect the metal particle content in the equipment's lubricating oil to determine the wear of the equipment's internal components and further determine the equipment's health status.

[0031] S12. Analyze and obtain fault characteristics under different working conditions based on the actual operating data and actual health status.

[0032] Before analysis, collected operational data can be preprocessed, including data cleaning, normalization, and noise reduction. Data cleaning primarily removes outliers and missing values; normalization converts data to a consistent dimension or range for ease of analysis; and noise reduction removes noise signals to improve data quality. For example, for vibration signal data, filtering algorithms can be used to remove high-frequency noise while retaining useful signal components.

[0033] Furthermore, fault-related features are extracted from the preprocessed actual operating data. These features can be based on physical models, such as thermodynamic and mechanical parameters of the equipment, or data-driven features, such as time-domain signal characteristics (mean, variance, kurtosis, etc.) or frequency-domain characteristics (spectral peak, bandwidth, etc.). When extracting features, it is necessary to consider the operating principles and fault mechanisms of the energy equipment and select features that effectively reflect the health status of the equipment. For example, when analyzing motor faults, time-domain features such as the mean and variance of the current can be extracted from the motor's current signal, as well as frequency-domain features such as the characteristic frequency components in the current spectrum.

[0034] Furthermore, based on the actual health status labels, the extracted features are analyzed to identify fault characteristics associated with different health statuses. Fault characteristics can be identified through statistical analysis methods, machine learning algorithms, and other methods. For example, cluster analysis can be used to classify feature data under different health statuses, identify typical characteristics within each category, and thus obtain fault characteristics under different operating conditions.

[0035] S13: Inputting the fault characteristics under the different working conditions into a matching network to obtain fault information under the different working conditions output by the matching network.

[0036] A matching network is a neural network structure used for feature matching and fusion. It can match fault signatures under different operating conditions, thereby integrating fault information. The design of a matching network requires consideration of characteristics such as the dimensionality and distribution of fault signatures, selecting an appropriate network structure and parameters. For example, a multilayer perceptron can be used as the basic structure of a matching network, and performance can be improved by adding batch normalization layers and activation functions.

[0037] Fault signatures under different operating conditions are input into the matching network. The network then matches and fuses the input fault signatures based on learned matching rules. The matching network outputs matched fault information, which better reflects the commonalities and differences between faults under different operating conditions. For example, for motor fault signatures under different load conditions, the matching network can match the fault signatures under different loads and output unified fault information, providing a more accurate basis for subsequent fault diagnosis.

[0038] S14: constructing a mechanism model based on the actual operation data and the actual health status, and obtaining mechanism characteristics based on the mechanism model.

[0039] A mechanism model is a mathematical model based on the physical principles, operating mechanisms, and failure mechanisms of energy equipment. It can describe the normal operation process and failure evolution of the equipment. When constructing a mechanism model, it is necessary to conduct in-depth research on the physical structure and operating principles of the energy equipment. The model parameters and form are then determined based on actual operating data and health status information. For example, for a heat exchanger, a heat exchange model can be established based on heat exchange principles. The model parameters are determined using parameters such as inlet and outlet temperatures and flow rates from actual operating data, thereby constructing a mechanism model that reflects the heat exchanger's operating status.

[0040] Furthermore, based on the constructed mechanism model, through model simulation and calculation, mechanism characteristics related to the equipment health status are obtained. These mechanism characteristics can be key parameters and state variables in the model, reflecting the health status of the equipment from a physical perspective. For example, in the mechanism model of a heat exchanger, parameters such as heat exchange efficiency and fouling thermal resistance can be used as mechanism characteristics. Through model calculation and simulation, the values ​​of these mechanism characteristics can be obtained, providing a physical basis for subsequent fault diagnosis.

[0041] S15. Reconstruct the fault information under the different working conditions based on the mechanism characteristics to obtain a full life cycle fault feature system for the target energy equipment.

[0042] Mechanism-guided intelligent assessment methods leverage in-depth research on the failure mechanisms of energy equipment. By constructing mechanistic models that reflect the equipment's status and degradation behavior, they combine theoretical analysis with data-driven approaches to address deficiencies caused by insufficient historical samples or differences in sample environments. In this approach, the characteristic parameters output by the mechanistic model can be used to supplement or constrain intelligent model training, significantly enhancing the model's robustness and generalization, thereby enabling accurate assessment of the health status of energy equipment within diverse and heterogeneous data samples.

[0043] By combining the physical interpretation capability of the mechanism model with the flexibility of the data-driven model, on the one hand, the mechanism model is used to provide solid theoretical support under conditions of insufficient samples or abnormal data, and on the other hand, data-driven supplementation and optimization are used to supplement and optimize the modeling accuracy of the complex operation behavior of equipment.

[0044] In this embodiment of the present invention, fault information output by the matching network under different operating conditions is reconstructed, guided by mechanism characteristics. The purpose of this reconstruction is to combine fault information with mechanism characteristics to obtain a more accurate and comprehensive fault signature system. This reconstruction process can be achieved through methods such as feature fusion and feature weighting. For example, a weighted summation of fault information and mechanism characteristics can be performed to obtain a reconstructed fault signature, where the weight can be determined based on the importance of the mechanism characteristics and the fault information.

[0045] Furthermore, through the reconstruction of fault information, the fault feature system of the target energy equipment can be obtained throughout the entire life cycle of the equipment. This fault feature system not only includes the fault features of the equipment under different operating conditions, but also integrates the mechanism features, which can more comprehensively reflect the changes in the health status of the equipment. This full life cycle fault feature system can serve as the basis for subsequent health status assessment and fault diagnosis, providing strong support for equipment maintenance and management. For example, for a wind power equipment, the full life cycle fault feature system obtained through reconstruction can include the fault features of the equipment at different wind speeds and different operating times, as well as the mechanism features such as blade fatigue and gearbox wear obtained based on the mechanism model, thereby providing comprehensive fault feature support for the full life cycle management of wind power equipment.

[0046] S16. Perform real-time health status assessment on the target energy equipment based on the fault characteristic system.

[0047] The random forest classification / regression algorithm is used to conduct intelligent assessment of the health status of energy equipment.

[0048] The fault feature system is used as the input feature of the random forest algorithm to train the random forest model. By training the random forest model, the real-time health status of the target energy equipment is classified and regressed to obtain the health status assessment result of the target energy equipment.

[0049] Optionally, the health status of the equipment can be interpreted and analyzed based on the health assessment results. For example, if the model predicts that the equipment is in an early stage of failure, feature importance analysis can be used to determine which features contribute most to this prediction, providing maintenance personnel with specific fault diagnosis information. Simultaneously, the assessment results are fed back to the equipment's operation management system, enabling timely maintenance measures or adjustments to operational strategies.

[0050] Application of Assessment Results: Real-time health status assessment results can provide a scientific basis for equipment maintenance decisions. For example, if equipment is in an early stage of failure, maintenance can be scheduled in advance based on the assessment results to prevent further deterioration, thereby reducing downtime and repair costs. Based on the health status assessment results, equipment operating parameters can be optimized and adjusted. For example, if the assessment results indicate potential risks under current operating conditions, operating parameters can be adjusted (such as reducing load or adjusting temperature) to extend the equipment's service life. Real-time health status assessments can detect potential equipment failures in advance, issue early warning signals, and alert maintenance personnel to take timely measures to prevent sudden failures from seriously impacting equipment and production systems.

[0051] The energy equipment health status assessment method provided by the present invention collects actual operating data and actual health status of the target energy equipment under different working conditions; obtains fault characteristics under different working conditions based on the actual operating data and actual health status analysis, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health status; inputs the fault characteristics under different working conditions into a matching network to obtain fault information under different working conditions output by the matching network; constructs a mechanism model based on the actual operating data and actual health status, and obtains mechanism characteristics based on the mechanism model, wherein the mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment; reconstructs the fault information under the different working conditions guided by the mechanism characteristics to obtain a fault characteristic system for the entire life cycle of the target energy equipment; and performs real-time health status assessment on the target energy equipment based on the fault characteristic system. Compared with the existing technology, there is a lack of high-quality analysis technology for early fault information of energy equipment, which makes it difficult to accurately and efficiently mine effective fault information from early fault data, affecting the accuracy of the intelligent assessment model of the health status of energy equipment; there is a lack of joint modeling technology for multi-working condition fault information of energy equipment, which makes it difficult to effectively match the fault knowledge accumulated under different working conditions and use it for joint assessment modeling, affecting the generalization of the intelligent assessment model of the health status of energy equipment. This method proposes a high-quality and low-rank early fault information analysis method to address the problem that early fault information in energy equipment is difficult to parse. To address the problem that fault information of multi-working condition energy equipment is difficult to match, a dynamic multi-supervisory sensor for multi-working condition fault information fusion is proposed to realize dynamic matching of fault information under multiple working conditions, achieve high-quality analysis of early fault information in energy equipment, and achieve consistent alignment of fault data distribution under different working conditions and joint modeling of fault information.

[0052] Figure 2 It is a schematic flow chart of S12 provided by the present invention, such as Figure 2 As shown, the method includes the following: S21. Analyze the correlation between the actual operating data and the actual health status.

[0053] S22. Define a first correlation loss function by maximizing the correlation between the actual operating data and the actual health status.

[0054] Analyze the correlation between each monitoring signal (actual operation data) and the current corresponding health status label of the energy equipment (such as healthy, early failure, serious failure, etc.), and define the first correlation loss function by maximizing the correlation between the two, which is expressed as follows: in, Indicates the s-way monitoring signal, yIndicates the current health status label of the energy equipment. express and y The mutual information is used to characterize the correlation between the two. express The sample size is M , represents the sample probability distribution.

[0055] S23: Analyze the correlation between the actual operation data collected by the multiple monitoring channels.

[0056] S24. Define a second correlation loss function by minimizing the correlation between the actual operation data.

[0057] Analyze the correlation between the monitoring signals and define the second correlation loss function by minimizing the correlation between the two, which is expressed as follows: S25. Construct a third loss function based on the first correlation loss function and the second correlation loss function.

[0058] Furthermore, based on the first and second correlation loss functions obtained above, a high-quality low-rank parsing strategy for early fault signals is constructed. The third loss function is defined as follows: S26. Analyze the actual operating data and actual health status based on the third loss function to obtain fault characteristics under different working conditions.

[0059] Furthermore, according to the actual health status label, the extracted features are analyzed based on the third loss function to find out the fault features associated with different health states under different working conditions.

[0060] To address the difficulty of parsing early-stage fault information in energy equipment, the present invention proposes a high-quality, low-rank method for parsing early-stage fault information. This method analyzes the correlation between monitoring signals and the corresponding health status labels of the energy equipment, as well as the correlation between individual monitoring signals. By maximizing the correlation between monitoring signals and labels and minimizing the correlation between individual monitoring signals, a high-quality, low-rank loss function is defined. This method effectively mitigates the redundancy between fault features while ensuring the accuracy of early-stage fault feature extraction, achieving high-quality parsing of early-stage fault information in energy equipment.

[0061] Figure 3 This is the second flow chart of the energy equipment health status assessment method provided by the present invention, such as Figure 3 As shown, the method includes: S31. Input the fault characteristics under the different working conditions into the matching network, and fuse the fault characteristics under the different working conditions through the dynamic multi-supervisory sensor, dynamic weight matrix and bias vector in the matching network to obtain the fault information under the different working conditions output by the matching network.

[0062] A matching network is a neural network structure used for feature matching and fusion. It can match fault signatures under different operating conditions, thereby integrating fault information. The design of a matching network requires consideration of characteristics such as the dimensionality and distribution of fault signatures, selecting an appropriate network structure and parameters. For example, a multilayer perceptron can be used as the basic structure of a matching network, and performance can be improved by adding batch normalization layers and activation functions.

[0063] Fault signatures under different operating conditions are input into the matching network. The network then matches and fuses the input fault signatures based on learned matching rules. The matching network outputs matched fault information, which better reflects the commonalities and differences between faults under different operating conditions. For example, for motor fault signatures under different load conditions, the matching network can match the fault signatures under different loads and output unified fault information, providing a more accurate basis for subsequent fault diagnosis.

[0064] Specifically, a dynamic multi-supervisory sensor is designed for multi-condition fault information fusion, which is defined as follows: in, Indicates historical multi-condition fault information and current operating fault information Dynamic multi-supervisory perceptrons. are the activation functions and batch normalization layers.

[0065] Design dynamic weight matrix and the bias vector , which can be expressed as: in, K Indicates the input channel.

[0066] Furthermore, the weights in the dynamic weight matrix and bias vector are optimized based on the fault characteristics and adjustable network parameters under different working conditions. It is not fixed, it is related to the input and network parameters. The optimization process is as follows: in, , is an adjustable constraint parameter. The model can be forced to learn feature representations from a diverse training set, thereby mining more rich information than linear static convolution.

[0067] In this way, dynamic matching of multi-operating condition fault information can be achieved.

[0068] S32: Extracting actual operating data under the different operating conditions and fault characteristic parameters corresponding to the actual health status.

[0069] In an embodiment of the present invention, time domain features such as mean, variance, kurtosis, skewness, etc. can be extracted from actual operating data. These features can reflect the statistical characteristics of the data. For example, the mean of the vibration signal can reflect the average vibration level of the equipment. The time domain signal is converted into a frequency domain signal through Fourier transform, and frequency domain features such as spectrum peak, bandwidth, harmonic frequency, etc. are extracted. These features can reflect the periodic vibration characteristics of the equipment. For example, harmonics of a specific frequency may appear when a motor fails. For non-stationary signals, time-frequency domain features such as wavelet transform coefficients can be extracted. These features can simultaneously reflect the time and frequency characteristics of the signal. Based on the actual health status label, features related to the health status are extracted. For example, for an early fault state, parameters that can reflect the characteristics of the initial stage of the fault are extracted, such as slight changes in the vibration signal.

[0070] Analyze the correlation between the extracted features and the actual health status, and select features that are highly correlated with the health status as fault characteristic parameters. For example, the Pearson correlation coefficient can be used to analyze the correlation between the kurtosis of the vibration signal and the health status of the equipment.

[0071] S33. Construct a mechanism model based on the fault characteristic parameters.

[0072] Thoroughly study the physical principles of the target energy equipment, including thermodynamics, mechanics, and electricity. For example, for generators, analyze their electromagnetic induction and mechanical rotation principles. Study the equipment's failure mechanisms under different operating conditions, such as fatigue, wear, and corrosion. For example, study the fatigue failure mechanisms of motor bearings under high load.

[0073] Build mathematical models based on physical principles and failure mechanisms. For example, a thermal balance model for a motor can be established to describe its temperature distribution under different loads. Use fault characteristic parameters to estimate the parameters of the mechanism model. For example, estimate the thermal resistance and thermal capacity parameters in the motor thermal balance model using temperature and current data from actual operating conditions.

[0074] Verify the mechanism model using simulation software to ensure that the model accurately reflects the operating status and fault characteristics of the equipment. For example, use finite element analysis software to verify the thermal balance model of the motor.

[0075] The mechanism model is verified through experimental data to ensure its accuracy and reliability. For example, the accuracy of the thermal balance model can be verified by experimentally measuring the temperature of the motor under different loads.

[0076] S34. Analyze the physical principles and failure mechanisms of the target energy equipment based on the mechanism model to obtain mechanism characteristics.

[0077] Using mechanism models, we can identify key parameters closely related to the health of the equipment. For example, in a motor thermal balance model, thermal resistance and thermal capacity are key parameters; their changes can reflect the motor's thermal state. We can also analyze the state variables in the mechanism model; these variables can reflect the equipment's operating status and fault evolution. For example, in a motor model, state variables such as temperature distribution and current density can reflect the motor's health.

[0078] Mechanistic feature extraction: Determine the values ​​of key parameters through simulation or calculation of the mechanism model. For example, calculate the thermal resistance and thermal capacity of a motor using a thermal balance model. Select mechanism features that effectively reflect the health of the device. For example, the thermal resistance of a motor is selected as a mechanism feature because changes in thermal resistance can reflect the motor's heat dissipation performance and thermal state.

[0079] S35. Guided by the mechanism characteristics, the fault information under the different working conditions and the mechanism characteristics are reconstructed by a weighted summation method.

[0080] like Figure 4 As shown, it is a flow chart of the mechanism-guided energy equipment health status assessment provided by the present invention. The embodiment of the present invention is guided by the mechanism characteristics to reconstruct the fault information under different working conditions output by the matching network. The purpose of reconstruction is to combine the fault information with the mechanism characteristics to obtain a more accurate and comprehensive fault feature system. The reconstruction process can be achieved through methods such as feature fusion and feature weighting. For example, the fault information and the mechanism characteristics can be weighted and summed to obtain the reconstructed fault characteristics, where the weight can be determined according to the importance of the mechanism characteristics and the fault information.

[0081] Furthermore, through the reconstruction of fault information, the fault feature system of the target energy equipment can be obtained throughout the entire life cycle of the equipment. This fault feature system not only includes the fault features of the equipment under different operating conditions, but also integrates the mechanism features, which can more comprehensively reflect the changes in the health status of the equipment. This full life cycle fault feature system can serve as the basis for subsequent health status assessment and fault diagnosis, providing strong support for equipment maintenance and management. For example, for a wind power equipment, the full life cycle fault feature system obtained through reconstruction can include the fault features of the equipment at different wind speeds and different operating times, as well as the mechanism features such as blade fatigue and gearbox wear obtained based on the mechanism model, thereby providing comprehensive fault feature support for the full life cycle management of wind power equipment.

[0082] S36. Use an optimization algorithm to adjust the reconstructed features, and generate a full life cycle fault feature system for the target energy equipment based on the optimized features.

[0083] S37. Use the fault feature system as the input feature of the random forest algorithm to train the random forest model.

[0084] This embodiment of the present invention uses a random forest classification / regression algorithm as a health status assessment model. Random forest is an ensemble learning algorithm that improves the classification or regression accuracy of a model by constructing multiple decision trees and integrating their results. It can handle high-dimensional data and is highly robust to noise and outliers. It also provides feature importance assessment, making it easier to understand which features contribute most to health status assessment.

[0085] During model training, the random forest model is trained using historical data (including fault signatures and corresponding health status labels under different operating conditions). During training, the model learns the mapping between fault signatures and health status, enabling it to accurately predict the health status of the device when new data is input.

[0086] S38. Based on the random forest model, classify and regress the real-time health status of the target energy equipment to obtain a health status assessment result of the target energy equipment.

[0087] The extracted real-time features are fed into a trained random forest model, which then outputs a health status assessment of the device. This assessment can be a categorical label (e.g., healthy, early failure, critical failure) or a continuous health status indicator (e.g., health index).

[0088] Furthermore, the model's output is used to interpret and analyze the equipment's health status. For example, if the model predicts an early stage of equipment failure, feature importance analysis can be used to determine which features contribute most to this prediction, providing maintenance personnel with specific fault diagnosis information. Simultaneously, the evaluation results are fed back to the equipment's operation management system, enabling timely maintenance measures or adjustments to operational strategies.

[0089] In response to the problem of difficulty in matching fault information of energy equipment under multiple working conditions, the embodiments of the present invention propose a dynamic multi-supervisory sensor for multi-condition fault information fusion to achieve dynamic matching of fault information under multiple working conditions. Compared with the existing technology, the knowledge-guided multi-condition fault information dynamic matching and reconstruction method proposed in the embodiments of the present invention achieves consistent alignment of fault data distribution under different working conditions and joint modeling of fault information.

[0090] The energy equipment health status assessment method provided by the present invention collects actual operating data and actual health status of the target energy equipment under different working conditions; obtains fault characteristics under different working conditions based on the actual operating data and actual health status analysis, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health status; inputs the fault characteristics under different working conditions into a matching network to obtain fault information under different working conditions output by the matching network; constructs a mechanism model based on the actual operating data and actual health status, and obtains mechanism characteristics based on the mechanism model, wherein the mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment; reconstructs the fault information under the different working conditions guided by the mechanism characteristics to obtain a fault characteristic system for the entire life cycle of the target energy equipment; and performs real-time health status assessment on the target energy equipment based on the fault characteristic system. Compared with the existing technology, there is a lack of high-quality analysis technology for early fault information of energy equipment, which makes it difficult to accurately and efficiently mine effective fault information from early fault data, affecting the accuracy of the intelligent assessment model of the health status of energy equipment; there is a lack of joint modeling technology for multi-working condition fault information of energy equipment, which makes it difficult to effectively match the fault knowledge accumulated under different working conditions and use it for joint assessment modeling, affecting the generalization of the intelligent assessment model of the health status of energy equipment. This method proposes a high-quality and low-rank early fault information analysis method to address the problem that early fault information in energy equipment is difficult to parse. To address the problem that fault information of multi-working condition energy equipment is difficult to match, a dynamic multi-supervisory sensor for multi-working condition fault information fusion is proposed to realize dynamic matching of fault information under multiple working conditions, achieve high-quality analysis of early fault information in energy equipment, and achieve consistent alignment of fault data distribution under different working conditions and joint modeling of fault information.

[0091] The energy equipment health status assessment device provided by the present invention is described below. The energy equipment health status assessment device described below and the energy equipment health status assessment method described above can be referenced to each other.

[0092] Figure 5 This is a schematic diagram of the structure of the energy equipment health status assessment device provided by the present invention, which specifically includes: The collection module 501 is used to collect the actual operating data and actual health status of the target energy equipment under different working conditions. For detailed description, please refer to the corresponding description of the above method embodiment, which will not be repeated here.

[0093] Analysis module 502 is configured to analyze the actual operating data and actual health status to obtain fault characteristics under different operating conditions, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health states. For detailed descriptions, please refer to the corresponding descriptions of the above method embodiments and will not be repeated here.

[0094] The input module 503 is used to input the fault characteristics under different working conditions into the matching network to obtain the fault information under different working conditions output by the matching network. Detailed descriptions refer to the corresponding descriptions of the above method embodiments, which will not be repeated here.

[0095] Construction module 504 is configured to construct a mechanism model based on the actual operating data and actual health status, and to obtain mechanism characteristics based on the mechanism model. The mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment. For detailed descriptions, please refer to the corresponding descriptions of the above method embodiments and will not be repeated here.

[0096] The construction module 504 is further configured to reconstruct the fault information under the different working conditions based on the mechanism characteristics to obtain a fault characteristic system for the entire life cycle of the target energy equipment. Detailed descriptions are provided in the corresponding descriptions of the above method embodiments, which will not be repeated here.

[0097] The evaluation module 505 is used to perform real-time health status evaluation on the target energy equipment based on the fault feature system. Detailed descriptions can be found in the corresponding descriptions of the above method embodiments, which will not be repeated here.

[0098] Figure 6 An example of a physical structure diagram of an electronic device is shown below. Figure 6As shown, the electronic device may include: a processor 810 , a communication interface 820 , a memory 830 and a communication bus 840 , wherein the processor 810 , the communication interface 820 and the memory 830 communicate with each other via the communication bus 840 . The processor 810 can call the logic instructions in the memory 830 to execute the energy equipment health status assessment method, which includes: collecting actual operating data and actual health status of the target energy equipment under different working conditions; analyzing the fault characteristics under different working conditions based on the actual operating data and actual health status, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health states; inputting the fault characteristics under different working conditions into the matching network to obtain fault information under different working conditions output by the matching network; constructing a mechanism model based on the actual operating data and actual health status, and obtaining mechanism characteristics based on the mechanism model, wherein the mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment; reconstructing the fault information under the different working conditions guided by the mechanism characteristics to obtain a fault characteristic system for the entire life cycle of the target energy equipment; and performing real-time health status assessment on the target energy equipment based on the fault characteristic system.

[0099] Furthermore, the logic instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0100] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the energy equipment health status assessment method provided by the above methods, the method including: collecting actual operating data and actual health status of the target energy equipment under different operating conditions; analyzing and obtaining fault characteristics under different operating conditions based on the actual operating data and actual health status, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health states; inputting the fault characteristics under different operating conditions into a matching network to obtain fault information under different operating conditions output by the matching network; constructing a mechanism model based on the actual operating data and actual health status, and obtaining mechanism characteristics based on the mechanism model, wherein the mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment; reconstructing the fault information under the different operating conditions guided by the mechanism characteristics to obtain a fault characteristic system for the entire life cycle of the target energy equipment; and performing real-time health status assessment on the target energy equipment based on the fault characteristic system.

[0101] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the energy equipment health status assessment method provided by the above-mentioned methods, the method comprising: collecting actual operating data and actual health status of the target energy equipment under different operating conditions; analyzing and obtaining fault characteristics under different operating conditions based on the actual operating data and actual health status, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health states; inputting the fault characteristics under different operating conditions into a matching network to obtain fault information under different operating conditions output by the matching network; constructing a mechanism model based on the actual operating data and actual health status, and obtaining mechanism characteristics based on the mechanism model, wherein the mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment; reconstructing the fault information under the different operating conditions guided by the mechanism characteristics to obtain a fault characteristic system for the entire life cycle of the target energy equipment; and performing real-time health status assessment on the target energy equipment based on the fault characteristic system.

[0102] The device embodiments described above are merely illustrative. 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, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0103] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0104] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for assessing the health status of energy equipment, characterized in that: include: Collect actual operating data and actual health status of target energy equipment under different working conditions; Obtaining fault characteristics under different working conditions based on the actual operating data and actual health status analysis, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health states; Inputting the fault characteristics under the different working conditions into the matching network to obtain the fault information under the different working conditions output by the matching network; Building a mechanism model based on the actual operating data and the actual health status, and obtaining mechanism characteristics based on the mechanism model, wherein the mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment; Reconstructing the fault information under the different working conditions based on the mechanism characteristics to obtain a fault feature system for the entire life cycle of the target energy equipment; A real-time health status assessment is performed on the target energy equipment based on the fault characteristic system.

2. The method according to claim 1, characterized in that The actual operation data is collected through multiple monitoring channels; The fault characteristics under different working conditions obtained based on the actual operating data and actual health status analysis include: Analyzing the correlation between the actual operating data and the actual health status; defining a first correlation loss function by maximizing the correlation between the actual operating data and the actual health status; Analyzing the correlation between the actual operation data collected by the multiple monitoring channels; By minimizing the correlation between the actual operation data, a second correlation loss function is defined; Constructing a third loss function based on the first correlation loss function and the second correlation loss function; The actual operating data and actual health status are analyzed based on the third loss function to obtain fault characteristics under different working conditions.

3. The method according to claim 2, characterized in that Inputting the fault characteristics under the different working conditions into the matching network to obtain the fault information under the different working conditions output by the matching network includes: The fault characteristics under the different working conditions are input into the matching network, and the fault characteristics under the different working conditions are fused through the dynamic multi-supervisory sensor, dynamic weight matrix and bias vector in the matching network to obtain the fault information under the different working conditions output by the matching network.

4. The method according to claim 3, characterized in that The method further comprises: The weights in the dynamic weight matrix and the bias vector are optimized based on the fault characteristics and adjustable network parameters under the different working conditions.

5. The method according to claim 3, characterized in that The constructing of a mechanism model based on the actual operation data and the actual health status, and obtaining mechanism characteristics based on the mechanism model, includes: Extracting actual operating data under the different working conditions and fault characteristic parameters corresponding to the actual health status; Constructing a mechanism model based on the fault characteristic parameters; The physical principles and failure mechanisms of the target energy equipment are analyzed based on the mechanism model to obtain mechanism characteristics.

6. The method according to claim 5, characterized in that The fault information under the different working conditions is reconstructed based on the mechanism characteristics to obtain a fault feature system for the entire life cycle of the target energy equipment, including: Guided by the mechanism characteristics, the fault information under the different working conditions and the mechanism characteristics are reconstructed by a weighted summation method; The reconstructed features are adjusted using an optimization algorithm, and a fault feature system for the entire life cycle of the target energy equipment is generated based on the optimized features.

7. The method according to claim 6, characterized in that The real-time health status assessment of the target energy equipment based on the fault feature system includes: Use the fault feature system as the input feature of the random forest algorithm to train the random forest model; Based on the random forest model, the real-time health status of the target energy equipment is classified and regressed to obtain a health status assessment result of the target energy equipment.

8. An energy equipment health status assessment device, characterized in that: include: The acquisition module is used to collect the actual operating data and actual health status of the target energy equipment under different working conditions; an analysis module, configured to obtain fault characteristics under different operating conditions based on the actual operating data and the actual health status, wherein the fault characteristics are characteristics of the operating data of the target energy equipment under different health states; An input module, configured to input the fault characteristics under the different working conditions into the matching network, and obtain the fault information under the different working conditions output by the matching network; a construction module, configured to construct a mechanism model based on the actual operation data and the actual health status, and obtain mechanism characteristics based on the mechanism model, wherein the mechanism model is used to analyze the physical principles and failure mechanisms of the target energy equipment; The construction module is further configured to reconstruct the fault information under the different working conditions based on the mechanism characteristics to obtain a fault feature system for the entire life cycle of the target energy equipment; An evaluation module is used to perform real-time health status evaluation on the target energy equipment based on the fault characteristic system.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the method for assessing the health status of energy equipment according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for assessing the health status of energy equipment according to any one of claims 1 to 7 is implemented.

Images