Intelligent monitoring and fault diagnosis method of power plant based on multi-source data fusion

Abstract

The present invention discloses a method for intelligent monitoring and fault diagnosis of power plants based on multi-source data fusion, which relates to the technical field of equipment fault analysis, including multi-source data collection, preprocessing and data reduction; data fusion of the reduced data to construct a fault analysis model; fault diagnosis output, and feedback adjustment of the model. The method for intelligent monitoring and fault diagnosis of power plants based on multi-source data fusion provided by the present invention organizes the pre-processed multi-source data into a high-dimensional data matrix, and uses the SVD-POD algorithm for order reduction processing, thereby reducing the data dimension and redundant information and increasing the processing speed. The feedback optimization diagnosis model realizes effective response to unknown working conditions and fault modes, ensures the accuracy and timeliness of fault diagnosis, has adaptive capabilities, can be dynamically adjusted according to new data, and improves the intelligence level and adaptability of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment failure analysis, and in particular to a method for intelligent monitoring and fault diagnosis of a power plant based on multi-source data fusion. Background Art

[0002] Traditional power plant monitoring systems mainly rely on data collection and analysis from a single data source, such as temperature sensors, vibration sensors, current and voltage sensors, etc. These systems detect the operating status and fault conditions of equipment through predetermined rules and thresholds. However, as the scale and complexity of power plants increase, monitoring systems with a single data source have gradually exposed their limitations and are difficult to fully and accurately reflect the true operating conditions of equipment. In recent years, multi-source data fusion technology has been introduced into the field of power plant monitoring. By integrating data from different sensors, it can achieve comprehensive perception of the operating status of equipment and fault diagnosis, greatly improving the accuracy and reliability of the monitoring system.

[0003] However, the existing multi-source data fusion technology still has some shortcomings in application. First, the presence of noise and outliers in the data acquisition and preprocessing stages may lead to a decline in data quality, thereby affecting the accuracy of fault diagnosis. Although existing studies have used methods such as wavelet transform for denoising, the computational complexity is high and the real-time performance is poor when processing large-scale, multi-dimensional data. Secondly, in the process of building a data fusion model, the weight distribution problem of each data source is often ignored, resulting in the credibility and importance of the fused data not being effectively reflected. In addition, most of the existing fault diagnosis models rely on preset fault modes and lack the ability to identify and process unknown fault modes. In practical applications, the equipment operating environment is complex and changeable, and the emergence of unknown fault modes is inevitable. Therefore, how to improve the adaptability and dynamic update capability of the fault diagnosis model is an urgent problem to be solved. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the existing power plant equipment detection technology based on the monitoring method of a single data source has the problems of low data quality, poor real-time data processing, high calculation complexity, and how to improve the adaptability of the fault diagnosis model.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for intelligent monitoring and fault diagnosis of power plants based on multi-source data fusion, including multi-source data collection, preprocessing and data reduction. Data fusion is performed on the reduced data to build a fault analysis model. Fault diagnosis output is performed, and feedback is provided to adjust the model.

[0007] As a preferred solution of the power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion described in the present invention, the multi-source data acquisition includes configuring sensors to acquire real-time data of power plant equipment operation, and obtaining temperature sensor data T(t), vibration sensor data V(t), current data I(t), and voltage data U(t).

[0008] As a preferred solution of the power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion described in the present invention, the preprocessing and data reduction include denoising the collected data, eliminating noise and outliers, using wavelet transform to denoise, normalizing the denoised data, storing the preprocessed multi-source data in a centralized database, constructing a high-dimensional data matrix, and organizing the preprocessed multi-source data into a high-dimensional data matrix X, where each row represents a sampling time, and each column represents a type of data, and the constructed high-dimensional data matrix X is reduced using the SVD-POD algorithm.

[0009] As a preferred solution of the power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion described in the present invention, the data fusion of the reduced-order data includes comprehensive processing of data from different sources to obtain unified fused data, which is expressed as:

[0010]

[0011] Among them, D fusion is the fused data, D i is the i-th source data, w i is the weight of the i-th source data. According to the credibility and importance of each data source, the weight of each data source is calculated and expressed as:

[0012]

[0013] Among them, C i is the credibility of the i-th source data, C j is the credibility of the j-th source data.

[0014] As a preferred solution of the power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion described in the present invention, the construction of the fault analysis model includes initializing neural network parameters, defining a loss function and optimizing the neural network, which is expressed as:

[0015]

[0016] in, is the loss function, m is the number of samples, y i is the actual output label of the i-th sample, h θ( x i ) is the model’s predicted value for the i-th sample, x (i) is the input feature of the i-th sample, λ is the regularization parameter, θ j is the jth parameter of the model.

[0017] Update the parameters:

[0018]

[0019] Among them, η is the learning rate.

[0020] As a preferred solution of the power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion described in the present invention, the fault diagnosis output includes inputting the fused data into the fault diagnosis model. If the input data matches the existing data in the database and is identified as a known operating condition and fault mode, the model directly performs diagnosis and outputs the results.

[0021] If the input data is an unknown working condition but matches an existing failure mode, it is identified as an unknown working condition but a known failure mode, and supplementary data is generated through dynamic simulation to update the model parameters.

[0022] If the input data is an unknown operating condition and an unknown failure mode, an alarm is sent to the operation and maintenance personnel to identify it as an unexpected risk.

[0023] As a preferred solution of the power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion described in the present invention, the feedback adjustment model includes transmitting the update signal back to the fault diagnosis model when the input data is unknown data, performing parameter optimization and model updating, which is expressed as:

[0024]

[0025] in, is the gradient of the loss function with respect to the neural network parameters, x new,i is the new working condition data, h θ＝1 ( x new ) For the model to respond to new input data x new The predicted output, y new,i For the new input data x new,i The actual target value of , K is the number of new input data.

[0026] Another object of the present invention is to provide an intelligent monitoring and fault diagnosis system for power plants based on multi-source data fusion, which can organize the pre-processed multi-source data into a high-dimensional data matrix, use the SVD-POD algorithm for order reduction processing, reduce data dimensions and redundant information, increase processing speed, and solve the problem of poor timeliness of current equipment diagnosis.

[0027] As a preferred solution of the power plant intelligent monitoring and fault diagnosis system based on multi-source data fusion described in the present invention, it includes: a data processing module, a fault analysis module, and a feedback module. The data processing module is used to collect multi-source data, and perform preprocessing and data reduction. The fault analysis module is used to perform data fusion on the reduced data and construct a fault analysis model. The smoothing processing module is used to perform fault diagnosis output and feedback adjustment model.

[0028] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a step of a power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion.

[0029] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion.

[0030] Beneficial effects of the present invention: The power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion provided by the present invention organizes the pre-processed multi-source data into a high-dimensional data matrix, and uses the SVD-POD algorithm for order reduction processing, thereby reducing the data dimension and redundant information and increasing the processing speed. The feedback optimization diagnosis model realizes the effective response to unknown working conditions and fault modes, ensures the accuracy and timeliness of fault diagnosis, has adaptive capabilities, can be dynamically adjusted according to new data, and improves the intelligence level and adaptability of the system. The present invention achieves better results in terms of timeliness, adaptability and automation level. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0032] Figure 1 An overall flow chart of a power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion provided for the first embodiment of the present invention.

[0033] Figure 2 A module diagram of a power plant intelligent monitoring and fault diagnosis system based on multi-source data fusion provided in the third embodiment of the present invention. DETAILED DESCRIPTION

[0034] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0035] Example 1, reference Figure 1 , which is an embodiment of the present invention, provides a power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion, comprising:

[0036] S1: Collect multi-source data and perform preprocessing and data reduction.

[0037] Furthermore, multi-source data collection includes configuring sensors to collect real-time data of power plant equipment operation, obtaining temperature sensor data T(t), vibration sensor data V(t), current data I(t), and voltage data U(t).

[0038] It should be noted that the preprocessing and data reduction include denoising the collected data, eliminating noise and outliers, using wavelet transform to denoise, normalizing the denoised data, storing the preprocessed multi-source data in a centralized database, constructing a high-dimensional data matrix, and organizing the preprocessed multi-source data into a high-dimensional data matrix X, where each row represents a sampling moment and each column represents a type of data, and the constructed high-dimensional data matrix X is reduced using the SVD-POD algorithm.

[0039] S2: Perform data fusion on the reduced-order data and build a fault analysis model.

[0040] Furthermore, data fusion of the reduced-order data includes comprehensive processing of data from different sources to obtain unified fused data, which is expressed as:

[0041]

[0042] Among them, D fusion is the fused data, D i is the i-th source data, w i is the weight of the i-th source data. According to the credibility and importance of each data source, the weight of each data source is calculated and expressed as:

[0043]

[0044] Among them, C i is the credibility of the i-th source data, C j is the credibility of the j-th source data.

[0045] It should be noted that building a fault analysis model includes initializing neural network parameters, defining a loss function and optimizing the neural network, which can be expressed as:

[0046]

[0047] in, is the loss function, m is the number of samples, y i is the actual output label of the i-th sample, h θ ( x i ) is the model’s predicted value for the i-th sample, x (i) is the input feature of the i-th sample, λ is the regularization parameter, θ j is the jth parameter of the model.

[0048] Update the parameters:

[0049]

[0050] Where η is the learning rate. By defining the loss function, the error of model prediction can be accurately evaluated, the model training process can be guided, and the model can be continuously optimized to reach the optimal state. The model parameters are optimized by the gradient descent method, so that the model can quickly converge to the optimal solution. The setting of the learning rate ensures that the step size of each parameter update is reasonable, which not only ensures the convergence speed, but also avoids over-adjustment.

[0051] S3: Perform fault diagnosis output and provide feedback to adjust the model.

[0052] Furthermore, the fault diagnosis output includes inputting the fused data into the fault diagnosis model. If the input data matches the existing data in the database and is identified as a known operating condition and failure mode, the model directly diagnoses and outputs the result. When the input data matches the existing data in the database, it means that the current operating status and failure mode of the equipment have been learned and identified by the model. At this point, the diagnostic model can directly use the trained parameters for fault identification and diagnosis, avoiding complex calculation processes and achieving rapid diagnosis. Since these operating conditions and failure modes have been verified and optimized in the model, the accuracy and reliability of the recognition results are high, which can effectively reduce the probability of misdiagnosis and missed diagnosis.

[0053] If the input data is an unknown working condition, but matches an existing fault mode, the unknown working condition but known fault mode is identified and supplementary data is generated through dynamic simulation to update the model parameters. In actual operation, the equipment may encounter new working conditions that were not included in the model training. However, if the failure modes corresponding to these new working conditions have been recorded in the database, it means that the nature of the fault has not changed. At this time, by generating supplementary data through dynamic simulation, the operating status of the equipment under the new working condition can be simulated, which helps to update the model parameters and enable the model to adapt to the new environment. This mechanism ensures that the model can continue to learn and optimize, and continuously improve its adaptability and diagnostic accuracy. This adaptive learning capability enables the diagnostic system to maintain efficient diagnostic performance in the face of changing working conditions.

[0054] If the input data is an unknown operating condition and an unknown failure mode, an alarm is issued to the operation and maintenance personnel to identify it as an unexpected risk. When the system cannot match any known operating conditions and failure modes, it means that the equipment may have encountered an unexpected operating state or a new failure mode. At this time, by issuing an alarm to the operation and maintenance personnel, they can be reminded in time to pay attention and take preventive measures to avoid possible further damage or shutdown of the equipment. The security of the system is greatly improved, ensuring that the operation and maintenance personnel can be informed of the abnormal status of the equipment and handle it at the first time, thereby reducing risks, ensuring the safety and stability of the equipment and production process, and avoiding excessive erroneous intelligent adjustments that cause the equipment status to be out of control.

[0055] It should be noted that the feedback adjustment model includes transmitting the update signal back to the fault diagnosis model when the input data is unknown data, performing parameter optimization and model updating, which can be expressed as:

[0056] The feedback adjustment model includes transmitting the update signal back to the fault diagnosis model for parameter optimization and model update when the input data is unknown data, which can be expressed as:

[0057]

[0058] in, is the gradient of the loss function with respect to the neural network parameters, x new,i is the new working condition data, h θ＝1 ( x new ) For the model to respond to new input data x new The predicted output, y new,i For the new input data x new,iThe actual target value of K is K, and K is the number of new input data. Gradient descent is a classic optimization algorithm that finds the optimal solution of the function by continuously adjusting parameters to gradually reduce the loss function. The introduction of new data error terms allows the model to make corresponding adjustments when facing new data, improves the adaptability and generalization ability of the model, and ensures that the model is constantly exposed to new data and working conditions during training, avoiding over-reliance on existing data, thereby improving the diversity and robustness of the model. In the process of model optimization, the adjustment of model parameters is guided by calculating the gradients of the loss function and the feedback function. The calculation of the gradient reflects the error of the current model under different data inputs, so that parameter optimization can be performed accurately. By introducing new working condition data into the feedback function, the system can generate supplementary data for new failure modes and dynamically adjust model parameters. This mechanism ensures that the system can cope with unknown working conditions and failure modes.

[0059] Example 2, an embodiment of the present invention, provides a method for intelligent monitoring and fault diagnosis of power plants based on multi-source data fusion. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0060] First, the experimental objects are the key equipment of a power plant, including the main engine, pumps, fans and generators. During the experiment, these equipment were tested in the whole process of multi-source data acquisition, preprocessing and data reduction, data fusion, fault analysis model construction, fault diagnosis output and feedback adjustment model.

[0061]

[0062]

[0063] The temperature of the host machine fluctuated greatly during the experiment, but through denoising and order reduction processing, the fused data obtained can better reflect the true state of the equipment, and the diagnostic results are more reliable. In traditional methods, noise and outliers often lead to misdiagnosis or missed diagnosis. In the experimental data, the pump was diagnosed as a fault state many times. The model was adjusted through feedback and the parameters were continuously optimized to ensure the accuracy of the diagnosis. Traditional methods are often difficult to adapt quickly when faced with dynamically changing working conditions. The method of the present invention achieves rapid adaptation to unknown working conditions through a dynamic feedback mechanism, significantly improving the real-time and accuracy of fault diagnosis. Overall, through multi-source data fusion and neural network optimization, the present invention shows significant advantages in the accuracy, real-time and adaptability of fault diagnosis.

[0064] Example 3, reference Figure 2 , which is an embodiment of the present invention, provides a power plant intelligent monitoring and fault diagnosis system based on multi-source data fusion, including a data processing module, a fault analysis module, and a feedback module.

[0065] The data processing module is used to collect multi-source data, perform preprocessing and data reduction. The fault analysis module is used to fuse the reduced data and build a fault analysis model. The smoothing module is used to output fault diagnosis and provide feedback to adjust the model.

[0066] If the function 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 storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for 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 of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0067] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0068] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0069] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0070] 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 preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for intelligent monitoring and fault diagnosis of power plants based on multi-source data fusion, characterized in that: include: Collect multi-source data, and perform preprocessing and data reduction; Perform data fusion on the reduced-order data and build a fault analysis model; Perform fault diagnosis output and provide feedback to adjust the model.

2. The method for intelligent monitoring and fault diagnosis of a power plant based on multi-source data fusion according to claim 1, characterized in that: The multi-source data collection includes configuring sensors to collect real-time data of power plant equipment operation, and obtaining temperature sensor data T(t), vibration sensor data V(t), current data I(t), and voltage data U(t).

3. The method for intelligent monitoring and fault diagnosis of a power plant based on multi-source data fusion according to claim 2, characterized in that: The preprocessing and data reduction include denoising the collected data, eliminating noise and outliers, denoising by wavelet transform, normalizing the denoised data, storing the preprocessed multi-source data in a centralized database, constructing a high-dimensional data matrix, organizing the preprocessed multi-source data into a high-dimensional data matrix X, where each row represents a sampling moment and each column represents a type of data, and reducing the constructed high-dimensional data matrix X by using an SVD-POD algorithm.

4. The method for intelligent monitoring and fault diagnosis of a power plant based on multi-source data fusion according to claim 3, characterized in that: The data fusion of the reduced-order data includes comprehensive processing of data from different sources to obtain unified fused data, which is expressed as: Among them, D fusion is the fused data, D i is the i-th source data, w i is the weight of the i-th source data. According to the credibility and importance of each data source, the weight of each data source is calculated and expressed as: Among them, C i is the credibility of the i-th source data, C j is the credibility of the j-th source data.

5. The method for intelligent monitoring and fault diagnosis of a power plant based on multi-source data fusion according to claim 4, characterized in that: The construction of the fault analysis model includes initializing the neural network parameters, defining the loss function and optimizing the neural network, which is expressed as: in, is the loss function, m is the number of samples, y i is the actual output label of the i-th sample, h θ (x i ) is the model’s predicted value for the i-th sample, x (i) is the input feature of the i-th sample, λ is the regularization parameter, θ j is the jth parameter of the model; Update the parameters: Among them, η is the learning rate.

6. The method for intelligent monitoring and fault diagnosis of a power plant based on multi-source data fusion according to claim 5, characterized in that: The fault diagnosis output includes inputting the fused data into the fault diagnosis model. If the input data matches the existing data in the database and is identified as a known operating condition and fault mode, the model directly performs diagnosis and outputs the result. If the input data is an unknown condition but matches an existing failure mode, the unknown condition but known failure mode is identified and supplementary data is generated through dynamic simulation to update the model parameters; If the input data is an unknown operating condition and an unknown failure mode, an alarm is sent to the operation and maintenance personnel to identify it as an unexpected risk.

7. The method for intelligent monitoring and fault diagnosis of a power plant based on multi-source data fusion according to claim 6, characterized in that: The feedback adjustment model includes transmitting the update signal back to the fault diagnosis model to perform parameter optimization and model update when the input data is unknown data, which is expressed as: in, is the gradient of the loss function with respect to the neural network parameters, x new,i is the new working condition data, h θ＝1 (x new ) is the model's response to the new input data x new The predicted output, y new,i For the new input data x new,i The actual target value of , K is the number of new input data.

8. A system using the power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion as claimed in any one of claims 1 to 7, characterized in that: Including data processing module, fault analysis module and feedback module; The data processing module is used to collect multi-source data and perform preprocessing and data reduction; The fault analysis module is used to perform data fusion on the reduced-order data and construct a fault analysis model; The smoothing processing module is used to perform fault diagnosis output and feedback adjustment model.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the power plant intelligent monitoring and fault diagnosis method based on multi-source data fusion described in any one of claims 1 to 7 are implemented.

Images