Mechanical seal fault diagnosis method and apparatus, electronic device, and program product

Abstract

The present disclosure relates to the technical field of fluid seals, and in particular to a mechanical seal fault diagnosis method and apparatus, an electronic device, and a program product. The method comprises: on the basis of a preset time interval, collecting high-pressure leakage pipe flow, a pressure upstream of a second-stage seal, and a pressure upstream of a third-stage seal of a reactor coolant pump shaft seal system, and preprocessing the high-pressure leakage pipe flow, the pressure upstream of the second-stage seal, and the pressure upstream of the third-stage seal to obtain a signal feature vector of the reactor coolant pump shaft seal system; and inputting the signal feature vector into a pre-trained classifier to obtain a diagnosis result of the reactor coolant pump shaft seal system. In this way, the problems that conventional mechanical seal inspection methods are difficult to realize real-time and comprehensive monitoring of a shaft seal state and cannot meet the requirements of modern industry are solved. By constructing a digital twin model of a shaft seal, real-time monitoring and intelligent diagnosis of the operation state of the shaft seal are realized, thereby improving the accuracy and real-time performance of shaft seal fault diagnosis, reducing the unexpected downtime, and lowering maintenance costs.

Description

Mechanical seal fault diagnosis methods, devices, electronic equipment and procedures

[0001] Cross-references to related applications

[0002] This disclosure is based on and claims priority to Chinese Patent Application No. 2024118667403, filed on December 18, 2024, the entire contents of which are incorporated herein by reference. Technical Field

[0003] This disclosure relates to the field of fluid sealing technology, and in particular to a mechanical seal fault diagnosis method, device, electronic equipment and program product. Background Technology

[0004] Nuclear energy, as a clean and efficient energy source, occupies an important position in the global energy structure. The main nuclear pump, as one of the key pieces of equipment in a nuclear power plant, directly affects the safe operation of the plant. The shaft seal, as a crucial component of the main nuclear pump, directly impacts the pump's sealing performance and efficiency. Traditional shaft seal fault diagnosis methods rely on periodic inspections and maintenance, which are costly, inefficient, and difficult to accurately monitor the operating status of the shaft seal. Furthermore, due to the unique working environment of the main nuclear pump, traditional detection methods have limitations, making it difficult to achieve real-time and comprehensive monitoring of the shaft seal's condition. With the development of Industry 4.0 and intelligent manufacturing, higher demands are placed on real-time monitoring and intelligent diagnosis of equipment status; traditional fault diagnosis methods can no longer meet the needs of modern industry. Summary of the Invention

[0005] This disclosure provides a mechanical seal fault diagnosis method, device, electronic equipment, and program product to solve the problems that traditional mechanical seal detection methods have certain limitations, making it difficult to achieve real-time and comprehensive monitoring of the shaft seal status and failing to meet the needs of modern industry.

[0006] The first aspect of this disclosure provides a method for diagnosing mechanical seal faults, comprising the following steps: collecting the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of a nuclear main pump shaft seal system based on a preset time interval; preprocessing the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal to obtain a signal feature vector of the nuclear main pump shaft seal system; inputting the signal feature vector into a pre-trained classifier to obtain a diagnostic result of the nuclear main pump shaft seal system, wherein the pre-trained classifier is obtained by generating a training set of the nuclear main pump shaft seal system under multiple states using a digital twin model, and training an initial classifier using the training set; the digital twin model is constructed from a preset surrogate model and a hydraulic model.

[0007] In some embodiments, before inputting the signal feature vector into a pre-trained classifier to obtain the diagnostic result of the nuclear main pump shaft seal system, the process includes: establishing a physical model and a hydraulic model of the nuclear main pump shaft seal system; determining the value range of the sealing health index and sealing pressure difference of the nuclear main pump shaft seal system; uniformly sampling a first input sample within the value range using a preset sampling method; obtaining a first output sample corresponding to the first input sample using the physical model; initializing an initial proxy model using the first input sample and the first output sample to obtain the preset proxy model; and inputting the preset proxy model into the hydraulic model to obtain the digital twin model.

[0008] In some embodiments, after obtaining the digital twin model, the process includes: determining the normal and abnormal ranges of the sealing health index and the flow resistance of the throttling coil in the main nuclear pump shaft seal system; uniformly sampling a second input sample using the Latin hypercube sampling method within the normal and abnormal ranges, inputting the second input sample into the digital twin model to obtain a second output sample; dividing the second input sample and the second output sample into a training set and a test set, training an initial classifier using the training set to obtain a trained classifier, and testing the recall rate of the trained classifier using the test set; if the recall rate of the trained classifier does not meet a preset condition, optimizing the hyperparameters of the trained classifier using a preset optimization method to obtain the pre-trained classifier.

[0009] In some embodiments, after obtaining the diagnostic results of the nuclear main pump shaft seal system, the method further includes: sending the diagnostic results and displaying them to the target terminal.

[0010] A second aspect of this disclosure provides a mechanical seal fault diagnosis device, comprising: a data acquisition module, configured to acquire the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of a nuclear main pump shaft seal system based on a preset time interval, and to preprocess the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal to obtain a signal feature vector of the nuclear main pump shaft seal system; and a diagnosis module, configured to input the signal feature vector into a pre-trained classifier to obtain a diagnosis result of the nuclear main pump shaft seal system, wherein the pre-trained classifier is obtained by generating a training set of the nuclear main pump shaft seal system under multiple states using a digital twin model, and training an initial classifier using the training set, wherein the digital twin model is constructed from a preset surrogate model and a hydraulic model.

[0011] In some embodiments, before inputting the signal feature vector into a pre-trained classifier to obtain the diagnostic result of the nuclear main pump shaft seal system, the diagnostic module is further configured to: establish a physical model and a hydraulic model of the nuclear main pump shaft seal system; determine the value range of the sealing health index and sealing pressure difference of the nuclear main pump shaft seal system; uniformly sample a first input sample within the value range using a preset sampling method; obtain a first output sample corresponding to the first input sample using the physical model; initialize an initial proxy model using the first input sample and the first output sample to obtain the preset proxy model; and input the preset proxy model into the hydraulic model to obtain the digital twin model.

[0012] In some embodiments, after obtaining the digital twin model, the diagnostic module is further configured to: determine the normal range and abnormal range of the sealing health index and the flow resistance of the throttling coil in the main nuclear pump shaft seal system; uniformly sample a second input sample using the Latin hypercube sampling method within the normal range and the abnormal range, input the second input sample into the digital twin model to obtain a second output sample; divide the second input sample and the second output sample into a training set and a test set, train an initial classifier using the training set to obtain a trained classifier, and test the recall rate of the trained classifier using the test set; if the recall rate of the trained classifier does not meet a preset condition, optimize the hyperparameters of the trained classifier using a preset optimization method to obtain the pre-trained classifier.

[0013] In some embodiments, after obtaining the diagnostic results of the nuclear main pump shaft seal system, the diagnostic module is further configured to: send the diagnostic results and display them to the target terminal.

[0014] A third aspect of this disclosure provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the mechanical seal fault diagnosis method as described in the above embodiments.

[0015] A fourth aspect of this disclosure provides a computer program product, including a computer program executed by a processor to implement the mechanical seal fault diagnosis method as described in the above embodiments.

[0016] In the above embodiments, the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of the nuclear main pump shaft seal system are collected based on a preset time interval. After preprocessing these parameters, a signal feature vector of the nuclear main pump shaft seal system is obtained. This signal feature vector is then input into a pre-trained classifier to obtain the diagnostic results of the nuclear main pump shaft seal system. This solves the limitations of traditional mechanical seal detection methods, such as the difficulty in achieving real-time and comprehensive monitoring of the shaft seal status and the inability to meet the needs of modern industry. By constructing a digital twin model of the shaft seal, real-time monitoring and intelligent diagnosis of its operating status can be achieved, improving the accuracy and real-time performance of shaft seal fault diagnosis, reducing unexpected downtime, and lowering maintenance costs. This is of great significance for ensuring the safe and efficient operation of nuclear power plants.

[0017] Additional aspects and advantages of this disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this disclosure. Attached Figure Description

[0018] The above and / or additional aspects and advantages of this disclosure will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, in which:

[0019] Figure 1 is a flowchart of a mechanical seal fault diagnosis method provided according to an embodiment of the present disclosure;

[0020] Figure 2 is a schematic diagram of the classification results of pump No. 1 according to an embodiment of the present disclosure;

[0021] Figure 3 is a schematic diagram of the classification results of pump No. 2 according to an embodiment of the present disclosure;

[0022] Figure 4 is a schematic diagram of a mechanical seal structure for a nuclear main pump according to an embodiment of the present disclosure;

[0023] Figure 5 is a flowchart of a mechanical seal fault diagnosis method according to an embodiment of the present disclosure;

[0024] Figure 6 is an example diagram of a mechanical seal fault diagnosis device according to an embodiment of the present disclosure;

[0025] Figure 7 is a schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure. Detailed Implementation

[0026] Embodiments of this disclosure are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting this disclosure.

[0027] The following describes a mechanical seal fault diagnosis method, apparatus, electronic device, and program product according to embodiments of the present disclosure with reference to the accompanying drawings. Addressing the limitations of traditional mechanical seal detection methods mentioned in the background art, which struggle to achieve real-time and comprehensive monitoring of the shaft seal status and fail to meet the needs of modern industry, this disclosure provides a mechanical seal fault diagnosis method. In this method, the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of the nuclear main pump shaft seal system are collected based on a preset time interval. After preprocessing the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal, a signal feature vector of the nuclear main pump shaft seal system is obtained. This signal feature vector is then input to a pre-trained classifier to obtain the diagnostic result of the nuclear main pump shaft seal system. This solves the limitations of traditional mechanical seal detection methods, such as the difficulty in achieving real-time and comprehensive monitoring of the shaft seal status and the inability to meet the needs of modern industry. By constructing a digital twin model of the shaft seal, real-time monitoring and intelligent diagnosis of its operating status can be achieved, improving the accuracy and real-time performance of shaft seal fault diagnosis, reducing unexpected downtime, and lowering maintenance costs. This is of great significance for ensuring the safe and efficient operation of nuclear power plants.

[0028] Digital twin technology enables real-time monitoring and prediction of equipment status by creating virtual models of physical equipment. This technology offers significant advantages in equipment health management, fault diagnosis, and maintenance strategy optimization. In the fault diagnosis of the shaft seal of a nuclear main pump, digital twin technology can simulate the operating environment of the shaft seal, analyze its working status in real time, and promptly identify potential fault risks.

[0029] The proposed fault diagnosis method for nuclear main pump shaft seals based on digital twin technology is applicable not only to nuclear power plants but also to multiple industries such as petroleum, chemical, and power, providing strong technical support for health management and fault prevention of various rotating equipment. With the continuous development and improvement of digital twin technology, this method is expected to become an important tool for intelligent diagnosis of industrial equipment, promoting the progress of Industry 4.0 and intelligent manufacturing.

[0030] Specifically, Figure 1 is a schematic flowchart of a mechanical seal fault diagnosis method provided in an embodiment of this disclosure.

[0031] Assuming the equipment's data acquisition system collects data online from certain key sensors at corresponding sampling rates, the following signal processing rules are set:

[0032] (1) Set a feature extraction interval duration T (determined according to analysis needs, typically 20 minutes. Too long a duration will cause the method to become sluggish, while too short a duration may lead to unstable feature extraction and increased computation due to too many intervals). Divide the intervals according to this duration, and preprocess the signal in each time period to extract feature vectors or determine invalid interval data. The content of the preprocessing should match the needs of the digital twin model. The following examples are all possible:

[0033] The average value of the signal from a single channel over a time period is used as one dimension of the feature vector.

[0034] The average value and standard deviation of the distribution of a single channel signal over a time period are used as the two dimensions of the feature vector;

[0035] The relationships between multi-channel signals are calculated (summation, difference, etc.) and used as one dimension of the feature vector;

[0036] A certain channel signal is processed by filtering and other methods according to its characteristics before steps such as calculating the average value are performed.

[0037] After performing empirical mode decomposition on a certain channel signal based on its characteristics, steps such as calculating the average value of each component are performed.

[0038] For known human-caused changes that are not due to faults, they can be marked as invalid data, and this range will not be included in subsequent processing;

[0039] ...

[0040] The i-th dimension of the signal feature vector of the j-th effective interval is denoted as . Let the total number of dimensions be M.

[0041] (2) The system monitors and processes the signals acquired in real time online. Whenever the duration of an interval ends, its feature vector x N It will be analyzed in an attempt to determine x N The specific category to which it belongs.

[0042] The following steps further describe the specific method for making the judgment:

[0043] (3) First, establish an efficient digital twin model of the mechanical seal, and use the model to generate data for each mode, with a sample size of N. m Digital twin models are based on N m The digital twin is used to generate samples for training, and then the new feature vector x can be generated. N The input model is used for identification (which can also be described as estimation, judgment, prediction, etc.) to obtain x. N The predicted label is usually expressed as a number, with each number representing a modality.

[0044] As shown in Figure 1, the mechanical seal fault diagnosis method includes the following steps:

[0045] In step S101, the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of the nuclear main pump shaft sealing system are collected based on a preset time interval. After preprocessing the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal, the signal feature vector of the nuclear main pump shaft sealing system is obtained.

[0046] In step S102, the signal feature vector is input to a pre-trained classifier to obtain the diagnostic results of the nuclear main pump shaft seal system. The pre-trained classifier is obtained by generating a training set of the nuclear main pump shaft seal system under various states using a digital twin model, and training an initial classifier using the training set. The digital twin model is constructed from a preset surrogate model and a hydraulic model.

[0047] There are a total of 8 diagnostic results, including Level 1 seal damage, Level 2 seal damage, Level 3 seal damage, Level 1-2 seal damage, Level 1-3 seal damage, Level 2-3 seal damage, Level 1-2-3 seal damage, and healthy.

[0048] In some embodiments, after obtaining the diagnostic results of the nuclear main pump shaft seal system, the method further includes: sending the diagnostic results and displaying them to the target terminal.

[0049] It should be noted that the target terminal includes, but is not limited to, smartphones, tablets, PDAs, and other terminal devices with data processing capabilities. Generally, a smart terminal refers to a terminal device with an independent operating system, which allows users to install software, games, and other programs provided by third-party service providers to continuously expand the functionality of the handheld device, and which can access wireless networks through mobile communication networks.

[0050] Specifically, in the online monitoring of the entire shaft seal, the sensor collects the channel signals of the main pump shaft seal system. The collected data is shown in Table 1, including the high-pressure leakage pipeline flow, the pressure before the second-stage seal, and the pressure before the third-stage seal.

[0051] Table 1

[0052] (1) Based on the meaning of the signal, the average value of the high pressure leakage pipeline flow, the pressure before the second-stage seal and the pressure before the third-stage seal of each channel of the nuclear main pump shaft sealing system is taken within a preset time interval to form a 3D feature. If the sensor failure is found, the data is marked as invalid data and no further calculation is performed, but the alarm is directly triggered, thereby obtaining the signal feature vector.

[0053] (2) In this embodiment of the present disclosure, the preset time interval length is 5 minutes, that is, the signal collected every 5 minutes will be processed as described above and converted into a signal feature vector or determined to be invalid. From the start of the system, once a valid signal feature vector is generated, it is judged using a pre-trained classifier to obtain the diagnostic result of the nuclear main pump shaft seal system, that is, to obtain the current fault type of the nuclear main pump shaft seal system.

[0054] To verify the effectiveness of the proposed method, this embodiment classifies and judges the annual operation monitoring data of pumps No. 1 and No. 3 of a certain unit. Pump No. 1, after analysis and preprocessing, yielded data samples in two states (including 20 samples of level 3 seal failure and 81 samples of level 3 seal failure). Pump No. 3 yielded data samples in three states (including 73 samples of level 3 seal failure, 2 samples of level 2 and 3 failure, and 21 samples of level 3 seal failure). These samples were classified using a classifier, and the results are shown in Figures 2 and 3. The values ​​in the figures represent the recall rate. It can be seen that the classification accuracy for pump No. 1 reached 80%, and the classification accuracy for pump No. 3 reached 96%. It can be observed that most of the samples were either healthy or had level 3 seal damage, and the classification of these two situations was very accurate.

[0055] After obtaining the diagnostic results of the nuclear main pump shaft seal system, the diagnostic results of the nuclear main pump shaft seal system are sent and displayed to the target terminal.

[0056] In some embodiments, before inputting the signal feature vector into a pre-trained classifier to obtain the diagnostic results of the nuclear main pump shaft seal system, the process includes: establishing a physical model and a hydraulic model of the nuclear main pump shaft seal system; determining the range of values ​​for the sealing health index and sealing pressure difference of the nuclear main pump shaft seal system; uniformly sampling a first input sample within the range using a preset sampling method; obtaining a first output sample corresponding to the first input sample using the physical model; initializing an initial surrogate model using the first input sample and the first output sample to obtain a preset surrogate model; and inputting the preset surrogate model into the hydraulic model to obtain a digital twin model.

[0057] Specifically, the schematic diagram of the fault monitoring method for the nuclear main pump shaft seal system is shown in Figures 4 and 5. First, a digital twin of each stage of the mechanical seal in the nuclear main pump shaft seal system is established. The process of establishing the digital twin includes numerical simulation and surrogate model acceleration. It should be noted that for other mechanical seals, the twin establishment method includes, but is not limited to, various simulation techniques or machine learning fitting methods; the acceleration method is used to quickly calculate the simulation results, and the method is not limited to surrogate models. In this embodiment, a physical model of a single-stage mechanical seal is established using pure numerical simulation with the finite element method. The input to the physical model is the pressure difference Δp between each stage of the seal. i The sealing health index T at all levels aiThe physical model outputs the leakage rate q for each stage of the seal, along with the geometric and physical properties of the seal. i To improve computational efficiency, the pressure-flow relationship of the throttling coil in the sealing system is linearized. The pressure-flow relationship of each stage of the mechanical seal should satisfy the following relationship: q i =f(Δp) i T ai )

[0058] In the formula, i = 1, 2, 3. f(·) is the surrogate model: the Kriging model, q i Δp represents the leakage rate for each seal stage. i q represents the pressure differential for each stage of the mechanical seal. ij r ij =k ij Δp ij +b ij

[0059] In the formula, (i,j)∈{(1,2),(1,3),(2,5),(3,5)}, q ij For the flow rate of each coil, r ij For the flow resistance of each coil, Δp ij k is the pressure difference for each coil. ij and b ij The linearization parameter can be calibrated using the pressure-flow relationship of the coil under standard conditions.

[0060] Next, a hydraulic model of the entire three-stage mechanical seal is constructed, whose pressure-flow relationship should be governed by the following equation: q1+q 12 -q2-q 25 =0 q2+q 13 -q3-q 35 =0

[0061] The main pump shaft seal only monitored the flow rate q in the leaking pipeline. H It is the sum of the flushing water flow rates of the second and third stage mechanical seals: q H =q 25 +q 35

[0062] Solving the above system of equations yields the complete hydraulic model of the shaft seal, as long as T is determined. ai and r ij Then q can be calculated. ij q i q H The pressure before the second-stage seal inlet is p2, and the pressure before the third-stage seal inlet is p3.

[0063] Due to the limitations of numerical simulation computation efficiency, this embodiment uses the Kriging surrogate model, which requires initialization training. Based on engineering experience, the sealing health index T for each level is determined. ai and the pressure difference Δp of each stage of the mechanical seal i The range of values ​​is determined, and 200 input parameters are uniformly sampled within this range using a preset sampling method, such as the Latin Hypercube (LHD) sampling method (Gaussian sampling can also be used). These 200 input parameters are used as the first input samples. The numerical simulation model is used to obtain 200 sealing leakage calculation results, which are the first output samples. These 200 input and output samples are used as the training set to initialize the Kriging proxy model. The Kriging proxy model is initialized using the training set to obtain the preset proxy model.

[0064] By inputting the preset proxy model into the hydraulic model, a digital twin model of the main pump shaft seal can be obtained.

[0065] In some embodiments, after obtaining the digital twin model, the process includes: determining the normal and abnormal ranges of the sealing health index and the flow resistance of the throttling coil in the main pump shaft seal system; uniformly sampling a second input sample using the Latin hypercube sampling method within the normal and abnormal ranges, inputting the second input sample into the digital twin model to obtain a second output sample; dividing the second input sample and the second output sample into a training set and a test set, training an initial classifier using the training set to obtain a trained classifier, and testing the recall rate of the trained classifier using the test set; if the recall rate of the trained classifier does not meet a preset condition, optimizing the hyperparameters of the trained classifier using a preset optimization method to obtain a pre-trained classifier.

[0066] Specifically, embodiments of this disclosure utilize a digital twin model to obtain a training set of the entire sealing system under various states. First, based on engineering experience, the sealing health index T for each level is confirmed. ai and the flow resistance r of each coil ij The normal and abnormal value ranges are defined. There are a total of 8 sealing states for shaft seals (including primary seal damage, secondary seal damage, tertiary seal damage, primary and secondary seal damage, primary and tertiary seal damage, secondary and tertiary seal damage, primary, secondary and tertiary seal damage, and healthy). The Latin Hypercube sampling method (LHD) is used to determine the sealing health index T for each sealing state at each level. ai and the flow resistance r of each coil ij 1000 input samples are uniformly sampled within both the normal and abnormal value ranges, forming the second input sample. A digital twin model is then used for rapid calculation to ultimately generate an output of 8000 samples (q). H(p2 and p3), that is, to obtain the second output sample, and finally use 8000 input and output samples as the training set and validation set of the initial classifier.

[0067] In this embodiment of the disclosure, a support vector machine (SVM) can be used as the initial classifier, or other machine learning models (neural networks, KNN, etc.) can be used as the initial classifier.

[0068] In a further embodiment of this disclosure, 8000 input and output samples are randomly divided into a training set and a test set (the division ratio is 8:2, or other ratios may be used). An initial classifier is trained using the training set to obtain a trained classifier. The recall rate of the trained classifier is tested using the test set. If the recall rate of the trained classifier does not meet the preset conditions, the hyperparameters of the trained classifier are optimized using a preset optimization method. For example, the optimal hyperparameters of the support vector machine can be obtained using methods such as cross-validation, thus obtaining a pre-trained classifier.

[0069] It should be noted that the preset conditions can be thresholds set by the user, thresholds obtained through a limited number of experiments, or thresholds obtained through a limited number of computer simulations; no specific limitations are made here.

[0070] Therefore, the beneficial effects of this disclosure include:

[0071] First, it is universally applicable to all types of mechanical seals, as long as the physical quantities measured by the sensors can be generated by the digital twin model.

[0072] Second, by making full use of digital twins to generate a large amount of data for training machine learning models, the problem of insufficient fault data in actual engineering is avoided.

[0073] Third, it can make full use of monitoring data from multiple channels to make a comprehensive judgment on the sealing status, rather than just making a local judgment based on a single channel sensor, thus improving the accuracy of fault diagnosis.

[0074] Fourth, the mechanical seal of the main nuclear pump is crucial to the safe operation and economic benefits of the entire nuclear power plant. Monitoring its condition and providing early warning of faults can prevent major accidents and is of great significance to national strategic security.

[0075] According to the mechanical seal fault diagnosis method proposed in this disclosure, the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of the nuclear main pump shaft seal system are collected based on a preset time interval. After preprocessing the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal, a signal feature vector of the nuclear main pump shaft seal system is obtained. This signal feature vector is then input into a pre-trained classifier to obtain the diagnostic result of the nuclear main pump shaft seal system. This solves the limitations of traditional mechanical seal detection methods, such as the difficulty in achieving real-time and comprehensive monitoring of the shaft seal status and the inability to meet the needs of modern industry. By constructing a digital twin model of the shaft seal, real-time monitoring and intelligent diagnosis of its operating status can be achieved, improving the accuracy and real-time performance of shaft seal fault diagnosis, reducing unexpected downtime, and lowering maintenance costs. This is of great significance for ensuring the safe and efficient operation of nuclear power plants.

[0076] Next, the mechanical seal fault diagnosis device proposed according to the embodiments of this disclosure is described with reference to the accompanying drawings.

[0077] Figure 6 is a block diagram of a mechanical seal fault diagnosis device according to an embodiment of the present disclosure.

[0078] As shown in Figure 6, the mechanical seal fault diagnosis device 10 includes: a data acquisition module 100 and a diagnosis module 200.

[0079] The acquisition module 100 is used to acquire the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of the nuclear main pump shaft sealing system based on a preset time interval, and to obtain the signal feature vector of the nuclear main pump shaft sealing system after preprocessing the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal; the diagnosis module 200 is used to input the signal feature vector into a pre-trained classifier to obtain the diagnosis result of the nuclear main pump shaft sealing system. The pre-trained classifier is obtained by generating a training set of the nuclear main pump shaft sealing system under various states using a digital twin model, and training an initial classifier using the training set. The digital twin model is constructed from a preset surrogate model and a hydraulic model.

[0080] In some embodiments, before inputting the signal feature vector into a pre-trained classifier to obtain the diagnostic results of the nuclear main pump shaft seal system, the diagnostic module 200 is further configured to: establish a physical model and a hydraulic model of the nuclear main pump shaft seal system; determine the value range of the sealing health index and sealing pressure difference of the nuclear main pump shaft seal system; uniformly sample the first input sample within the value range using a preset sampling method; obtain the first output sample corresponding to the first input sample using the physical model; initialize the initial proxy model using the first input sample and the first output sample to obtain a preset proxy model; and input the preset proxy model into the hydraulic model to obtain a digital twin model.

[0081] In some embodiments, after obtaining the digital twin model, the diagnostic module 200 is further configured to: determine the normal and abnormal ranges of the sealing health index and the flow resistance of the throttling coil in the main pump shaft seal system; within the normal and abnormal ranges, uniformly sample the second input sample using the Latin hypercube sampling method, input the second input sample into the digital twin model to obtain the second output sample; divide the second input sample and the second output sample into a training set and a test set, train an initial classifier using the training set to obtain a trained classifier, and test the recall rate of the trained classifier using the test set; if the recall rate of the trained classifier does not meet the preset conditions, optimize the hyperparameters of the trained classifier using a preset optimization method to obtain a pre-trained classifier.

[0082] In some embodiments, after obtaining the diagnostic results of the nuclear main pump shaft seal system, the diagnostic module 200 is further configured to: send the diagnostic results and display them to the target terminal.

[0083] It should be noted that the foregoing explanation of the embodiment of the mechanical seal fault diagnosis method also applies to the mechanical seal fault diagnosis device of this embodiment, and will not be repeated here.

[0084] According to the mechanical seal fault diagnosis device proposed in this disclosure, the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of the nuclear main pump shaft seal system are collected based on a preset time interval. After preprocessing the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal, a signal feature vector of the nuclear main pump shaft seal system is obtained. This signal feature vector is then input into a pre-trained classifier to obtain the diagnostic result of the nuclear main pump shaft seal system. This solves the limitations of traditional mechanical seal detection methods, such as the difficulty in achieving real-time and comprehensive monitoring of the shaft seal status and the inability to meet the needs of modern industry. By constructing a digital twin model of the shaft seal, real-time monitoring and intelligent diagnosis of its operating status can be achieved, improving the accuracy and real-time performance of shaft seal fault diagnosis, reducing unexpected downtime, and lowering maintenance costs. This is of great significance for ensuring the safe and efficient operation of nuclear power plants.

[0085] Figure 7 is a schematic diagram of the structure of an electronic device provided in an embodiment of this disclosure. The electronic device may include:

[0086] The memory 701, the processor 702, and the computer program stored on the memory 701 and executable on the processor 702.

[0087] When the processor 702 executes the program, it implements the mechanical seal fault diagnosis method provided in the above embodiments.

[0088] Furthermore, electronic devices also include:

[0089] Communication interface 703 is used for communication between memory 701 and processor 702.

[0090] The memory 701 is used to store computer programs that can run on the processor 702.

[0091] The memory 701 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0092] If the memory 701, processor 702, and communication interface 703 are implemented independently, they can be interconnected via a bus to communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, only one thick line is used in Figure 7, but this does not indicate that there is only one bus or one type of bus.

[0093] Optionally, in a specific implementation, if the memory 701, processor 702, and communication interface 703 are integrated on a single chip, then the memory 701, processor 702, and communication interface 703 can communicate with each other through an internal interface.

[0094] The processor 702 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present disclosure.

[0095] This disclosure also provides a computer program product, including a computer program executed by a processor to implement the mechanical seal fault diagnosis method as described above.

[0096] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0097] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0098] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more N executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of this disclosure includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of this disclosure pertain.

[0099] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequential list of executable instructions for implementing logical functions, and can be specifically implemented in any computer program product for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer program product" can be any means that can contain, store, communicate, propagate, or transmit a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer program products include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic device, and portable optical disc read-only memory (CDROM). Furthermore, the computer program product can even be paper or other suitable medium on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0100] It should be understood that the various parts of this disclosure can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0101] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer program product, and when executed, it includes one or a combination of the steps of the method embodiments.

[0102] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer program product.

[0103] The aforementioned computer program product may be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this disclosure have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting this disclosure. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this disclosure.

Claims

1. A mechanical seal failure diagnosis method, wherein, Includes the following steps: The high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of the nuclear main pump shaft sealing system are collected based on a preset time interval. The signal feature vector of the nuclear main pump shaft sealing system is obtained after preprocessing the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal. The signal feature vector is input into a pre-trained classifier to obtain the diagnostic result of the nuclear main pump shaft seal system. The pre-trained classifier is obtained by generating a training set of the nuclear main pump shaft seal system under various states using a digital twin model, and training an initial classifier using the training set. The digital twin model is constructed from a preset surrogate model and a hydraulic model.

2. The mechanical seal failure diagnostic method according to claim 1, wherein Before inputting the signal feature vector into a pre-trained classifier to obtain the diagnostic results of the main pump shaft seal system, the process includes: Establish the physical and hydraulic models of the nuclear main pump shaft sealing system; The sealing health index and sealing pressure difference of the main nuclear pump shaft seal system are determined to be within a certain range. A first input sample is uniformly sampled within the range using a preset sampling method. The first output sample corresponding to the first input sample is obtained using the physical model. The preset proxy model is obtained by initializing the initial proxy model using the first input sample and the first output sample. The preset proxy model is input into the hydraulic model to obtain the digital twin model.

3. The mechanical seal failure diagnostic method according to claim 2, wherein After obtaining the digital twin model, the following are included: Determine the normal and abnormal ranges of the sealing health index and the flow resistance of the throttling coil in the main nuclear pump shaft sealing system; Within the normal value range and the outlier value range, a second input sample is uniformly sampled using the Latin hypercube sampling method, and the second input sample is input into the digital twin model to obtain a second output sample; The second input sample and the second output sample are divided into a training set and a test set. An initial classifier is trained using the training set to obtain a trained classifier. The recall rate of the trained classifier is tested using the test set. If the recall rate of the trained classifier does not meet the preset conditions, the hyperparameters of the trained classifier are optimized using a preset optimization method to obtain the pre-trained classifier.

4. The mechanical seal failure diagnostic method according to claim 1, wherein After obtaining the diagnostic results of the nuclear main pump shaft sealing system, the following steps are also included: The diagnostic results are sent and displayed to the target terminal.

5. A mechanical seal failure diagnostic device, wherein, include: The acquisition module is used to acquire the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal of the nuclear main pump shaft sealing system based on a preset time interval, and to obtain the signal feature vector of the nuclear main pump shaft sealing system after preprocessing the high-pressure leakage pipeline flow rate, the pressure before the second-stage seal, and the pressure before the third-stage seal. The diagnostic module is used to input the signal feature vector into a pre-trained classifier to obtain the diagnostic results of the nuclear main pump shaft seal system. The pre-trained classifier is obtained by generating a training set of the nuclear main pump shaft seal system under multiple states using a digital twin model, and training an initial classifier using the training set. The digital twin model is constructed from a preset surrogate model and a hydraulic model.

6. The mechanical seal failure diagnostic device according to claim 5, wherein Before inputting the signal feature vector into a pre-trained classifier to obtain the diagnostic result of the main pump shaft seal system, the diagnostic module is further configured to: Establish the physical and hydraulic models of the nuclear main pump shaft sealing system; The sealing health index and sealing pressure difference of the main nuclear pump shaft seal system are determined to be within a certain range. A first input sample is uniformly sampled within the range using a preset sampling method. The first output sample corresponding to the first input sample is obtained using the physical model. The preset proxy model is obtained by initializing the initial proxy model using the first input sample and the first output sample. The preset proxy model is input into the hydraulic model to obtain the digital twin model.

7. The mechanical seal failure diagnostic device according to claim 6, wherein After obtaining the digital twin model, the diagnostic module is further configured to: Determine the normal and abnormal ranges of the sealing health index and the flow resistance of the throttling coil in the main nuclear pump shaft sealing system; Within the normal value range and the outlier value range, a second input sample is uniformly sampled using the Latin hypercube sampling method, and the second input sample is input into the digital twin model to obtain a second output sample; The second input sample and the second output sample are divided into a training set and a test set. An initial classifier is trained using the training set to obtain a trained classifier. The recall rate of the trained classifier is tested using the test set. If the recall rate of the trained classifier does not meet the preset conditions, the hyperparameters of the trained classifier are optimized using a preset optimization method to obtain the pre-trained classifier.

8. The mechanical seal failure diagnostic device according to claim 5, wherein After obtaining the diagnostic results of the nuclear main pump shaft sealing system, the diagnostic module is further used for: The diagnostic results are sent and displayed to the target terminal.

9. An electronic device, comprising: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the mechanical seal fault diagnosis method as described in any one of claims 1-4.

10. A computer program product comprising a computer program, wherein the computer program, when executed by a processor, implements the mechanical seal fault diagnosis method as described in any one of claims 1-4.

Images