Nuclear main pump mechanical seal life prediction method, device, equipment and program product

Abstract

The invention relates to the technical field of fluid sealing, in particular to a nuclear main pump mechanical seal life prediction method, device and equipment and a program product, and the method comprises the steps: inputting a pressure mean value and a pressure standard deviation of mechanical seal inlet pressure of a to-be-predicted nuclear main pump within a preset time into a life prediction model of the to-be-predicted nuclear main pump, the service life prediction model is obtained after a preset neural network is trained through historical degradation state data of multiple target nuclear main pumps to obtain an initial service life prediction model, and the initial service life prediction model is constrained through a preset loss function. Therefore, the degradation process of the mechanical seal in actual operation can be more accurately constructed, the accuracy of the result of the life prediction model is improved, and the sustainable and healthy development of the nuclear power industry is further promoted.

Description

Technical Field

[0001] The present application relates to the field of fluid sealing technology, and in particular to a method, device, equipment and program product for predicting the life of a nuclear main pump mechanical seal. Background Art

[0002] Nuclear power plays an important role in energy development, and ensuring the safe and reliable operation of nuclear power plants is crucial to the continued healthy development of the industry. Nuclear main pumps must be able to operate safely and reliably in the face of various complex working conditions to avoid leakage incidents. Mechanical seals are one of the key components of nuclear main pumps, which are used to control the leakage rate of the sealed fluid within the permitted range. At present, in the operation and maintenance of nuclear power plants, the average service life is generally used to regularly inspect and repair equipment, but unreasonable maintenance cycles are prone to problems such as insufficient and excessive equipment maintenance, which greatly increases operating costs.

[0003] Therefore, it is necessary to study the remaining service life of the mechanical seal device of the nuclear main pump in order to detect potential faults in advance and take corresponding maintenance measures to maximize the life of the mechanical seal. At present, in the field of mechanical seal life prediction, the research is based on data-driven prediction methods. This type of method does not need to consider the professional knowledge of the system, but relies on the analysis and deep mining of data information to establish a model. The accuracy of the model depends on the quality of the data. Many years ago, some experts used the temperature and leakage of the mechanical seal end face to train artificial neural networks to predict the remaining life, or based on acoustic emission technology, they collected multiple sets of full life data through experiments and performed feature noise reduction and extraction, and finally combined SVR training to obtain a degradation model, and used the QPSO algorithm to optimize the SVR life prediction model parameters to improve the accuracy of model prediction, or used the cumulative distribution function CDF of the acoustic emission signal characteristics to represent the seal degradation process, and used the CNN model to estimate the RUL. Or the acoustic emission method is used to monitor the parameters such as the opening time and film thickness of the mechanical seal end face, and the gray particle filter method is proposed to predict the remaining life. Alternatively, the wear amount of the sealing end face is taken as the research object, and its wear degradation process is described based on the Gamma process to predict its reliability and reliable life under different failure thresholds.

[0004] However, although the above purely data-driven methods can achieve good results under specific working conditions, the working conditions of mechanical seals are complex, and a large amount of complete data is required to achieve accurate prediction of the model. Although accelerated degradation tests can solve the problem of insufficient data to a certain extent, it cannot fundamentally mine the complete characteristic information contained in the seal degradation process, and it has limitations in physical interpretability. In addition, most data-driven models improve the accuracy of the prediction model by finding the optimal parameters through optimization algorithms, without considering the direct relationship between the model parameters and the physical characteristics of the system. Summary of the invention

[0005] The present application provides a method, device, equipment and program product for predicting the life of a nuclear main pump mechanical seal to solve the problems that the cost of nuclear main pump equipment is too high, resulting in few devices and failures, insufficient data samples to ensure the high accuracy of the life prediction model, and complex mechanical seal operating conditions, making it difficult for pure data-driven methods to capture the state degradation trend of the mechanical seal.

[0006] The first aspect of the present application provides a method for predicting the life of a nuclear main pump mechanical seal, comprising the following steps: monitoring the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time; inputting the pressure mean and pressure standard deviation of the mechanical seal inlet pressure into a life prediction model of the nuclear main pump to be predicted, and obtaining the degradation state and current life prediction value of the nuclear main pump to be predicted, wherein the life prediction model is obtained by training a preset neural network using historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and constraining the initial life prediction model using a preset loss function, wherein the preset loss function introduces the seal inlet pressure volatility.

[0007] Optionally, when obtaining the degradation state and current life prediction value of the nuclear main pump to be predicted, it includes: judging whether the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted; if the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted, then using the preset loss function to correct the life prediction model until the current life prediction value is less than or equal to the actual performance value of the nuclear main pump to be predicted, and obtaining a corrected life prediction model to predict the current life prediction value of the nuclear main pump to be predicted according to the corrected life prediction model.

[0008] Optionally, the preset loss function is: ;

[0009] in, is the root mean square error between the performance prediction value and the actual performance value during training, is the loss function, is the root mean square error, For the penalty item, is the pressure standard deviation, is the error when the performance prediction value is higher than the actual performance value, is the penalty factor.

[0010] Optionally, after obtaining the degradation state of the core main pump to be predicted and the current life prediction value, it includes: judging whether the current life prediction value is less than or equal to a preset threshold, or judging whether the health status of the mechanical seal is lower than a preset health standard based on the degradation state of the core main pump to be predicted; if the current life prediction value is less than or equal to the preset threshold, or the health status of the mechanical seal is lower than the preset health standard, triggering a life warning signal, and sending the life warning signal to a preset terminal.

[0011] A second aspect of the present application provides a nuclear main pump mechanical seal life prediction device, including: a monitoring module, used to monitor the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time; a prediction module, used to input the pressure mean and pressure standard deviation of the mechanical seal inlet pressure into the life prediction model of the nuclear main pump to be predicted, and obtain the degradation state and current life prediction value of the nuclear main pump to be predicted, wherein the life prediction model is obtained by training a preset neural network using historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and constraining the initial life prediction model using a preset loss function, wherein the preset loss function introduces the seal inlet pressure volatility.

[0012] Optionally, when the degradation state and current life prediction value of the nuclear main pump to be predicted are obtained, the prediction module is also used to: determine whether the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted; if the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted, then use the preset loss function to correct the life prediction model until the current life prediction value is less than or equal to the actual performance value of the nuclear main pump to be predicted, and obtain a corrected life prediction model to predict the current life prediction value of the nuclear main pump to be predicted according to the corrected life prediction model.

[0013] Optionally, the preset loss function is: ;

[0014] in, is the root mean square error between the performance prediction value and the actual performance value during training, is the loss function, is the root mean square error, For the penalty item, is the pressure standard deviation, is the error when the performance prediction value is higher than the actual performance value, is the penalty factor.

[0015] Optionally, after obtaining the degradation state and current life prediction value of the core main pump to be predicted, the prediction module is also used to: determine whether the current life prediction value is less than or equal to a preset threshold, or determine whether the health status of the mechanical seal is lower than a preset health standard based on the degradation state of the core main pump to be predicted; if the current life prediction value is less than or equal to the preset threshold, or the health status of the mechanical seal is lower than the preset health standard, then trigger a life warning signal, and send the life warning signal to a preset terminal.

[0016] The third aspect of the present application provides an electronic device, comprising: 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 method for predicting the life of a nuclear main pump mechanical seal as described in the above embodiment.

[0017] The fourth aspect of the present application provides a computer program product, including a computer program, which is executed by a processor to implement the method for predicting the life of a nuclear main pump mechanical seal as described in the above embodiment.

[0018] In the above implementation, the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time are monitored, and the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time are input into the life prediction model of the nuclear main pump to be predicted, so as to obtain the degradation state and current life prediction value of the nuclear main pump to be predicted, wherein the life prediction model is obtained by training a preset neural network using historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and constraining the initial life prediction model using a preset loss function. Thus, the problems that the equipment cost of the nuclear main pump is too high, resulting in few equipment and few failures, the data sample is insufficient to ensure the high accuracy of the life prediction model, and the mechanical seal working conditions are complex, resulting in the difficulty of the pure data-driven method to capture the state degradation trend of the mechanical seal, etc., can be solved, the degradation process of the mechanical seal in actual operation can be constructed more accurately, and the accuracy of the life prediction model results can be improved. By accurately predicting the current life prediction value of the mechanical seal, potential faults can be discovered in advance, and corresponding maintenance measures can be taken, which helps to avoid unexpected shutdowns and equipment damage, improve the safe and reliable operation of the nuclear main pump to be predicted, and promote the sustainable and healthy development of the nuclear power industry.

[0019] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which: Figure 1 A flowchart of a method for predicting the life of a nuclear main pump mechanical seal according to an embodiment of the present application; Figure 2 (a) is a schematic diagram of the fluctuation of the pressure before sealing of the third-stage mechanical seal according to an embodiment of the present application, Figure 2 (b) is a schematic diagram showing the fluctuation of the standard deviation of the pressure before sealing of the third-stage mechanical seal; Figure 3 is a schematic diagram of a life degradation model of different training sets of a No. 1 pump according to an embodiment of the present application; Figure 4 This is an example diagram of a nuclear main pump mechanical seal life prediction device according to an embodiment of the present application; Figure 5 Schematic diagram of the structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0022] The following describes the method, device, equipment and program product for predicting the life of the mechanical seal of the nuclear main pump of the embodiment of the present application with reference to the accompanying drawings. In view of the problem mentioned in the above background technology that the cost of the nuclear main pump equipment is too high, resulting in fewer equipment and fewer failures, the data samples are not enough to ensure the high accuracy of the life prediction model, and the mechanical seal working conditions are complex, which makes it difficult for the pure data-driven method to capture the state degradation trend of the mechanical seal, the present application provides a method for predicting the life of the mechanical seal of the nuclear main pump, in which the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time are monitored, and the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time are input into the life prediction model of the nuclear main pump to be predicted, and the degradation state and current life prediction value of the nuclear main pump to be predicted are obtained, wherein the life prediction model is obtained by training a preset neural network using historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and the initial life prediction model is constrained by a preset loss function. As a result, the problems of the high cost of nuclear main pump equipment, resulting in few equipment and few failures, insufficient data samples to ensure the high accuracy of the life prediction model, and complex working conditions of mechanical seals, which make it difficult for pure data-driven methods to capture the state degradation trend of mechanical seals, are solved. The degradation process of mechanical seals in actual operation can be constructed more accurately, and the accuracy of the life prediction model results can be improved. By accurately predicting the remaining service life of mechanical seals, potential faults can be discovered in advance and corresponding maintenance measures can be taken, which helps to avoid unexpected shutdowns and equipment damage, improve the safe and reliable operation of the predicted nuclear main pumps, and promote the sustainable and healthy development of the nuclear power industry.

[0023] Specifically, Figure 1 A schematic flow chart of a method for predicting the life of a nuclear main pump mechanical seal provided in an embodiment of the present application.

[0024] like Figure 1 As shown, the method for predicting the life of the mechanical seal of the nuclear main pump includes the following steps: In the nuclear main pump mechanical seal, since the pressure signal contains rich status information, its changes can be associated with its life and health status. The reduction in pressure indicates the degradation of the mechanical seal performance. Therefore, the embodiment of the present application infers the life and health status of the mechanical seal by analyzing the change pattern and trend of the pressure signal.

[0025] The model based on the combination of physical information and data-driven can obtain a more comprehensive and accurate prediction of the state degradation trend. By integrating physical knowledge into the data-driven model, the degradation state change law of the mechanical seal can be better captured and the accuracy of the prediction can be improved.

[0026] In the nuclear main pump mechanical seal, the change in pressure volatility can be used as a physical information to improve the prediction accuracy of the model. The inlet pressure of the nuclear main pump mechanical seal is generally controlled to fluctuate within a stable range. When the mechanical seal deteriorates such as end face damage, the seal inlet pressure will be significantly reduced, and the leakage will increase accordingly. When the secondary seal of the mechanical seal is worn, the following performance of the sealing ring will be reduced, resulting in increased fluctuations in the seal inlet pressure. Generally speaking, damage to the secondary seal will lead to severe wear of the end face, that is, an increase in the volatility of the seal inlet pressure often indicates that the inlet pressure will decrease. Therefore, the embodiment of the present application adds this physical information to the loss function of the model, which can correct the model to improve the accuracy of the life prediction model.

[0027] Since there are few sets of nuclear main pump equipment and few failures, the embodiment of the present application adopts a life prediction model based on degradation data to build a model. This model only needs to use the historical operation data of the equipment and directly characterize the degradation state of the equipment by monitoring variables. By using the historical operation information of the equipment, a performance degradation trajectory model of the equipment is established and its changing trend is predicted. The time from the degradation state of the equipment reaching the preset failure threshold from the current moment is the remaining life of the equipment.

[0028] The embodiment of the present application proposes the following scheme to achieve life prediction of the mechanical seal of the nuclear main pump: (1) The pressure signal is selected as the indicator of the model to establish the state degradation model of the mechanical seal. By analyzing the historical data of the pressure signal, the state degradation trajectory of the mechanical seal is tracked and its change trend is predicted.

[0029] (2) Considering that pressure fluctuation often indicates a decrease in pressure, the physical information of fluctuation is added to the loss function of the prediction model. Here, fluctuation is represented by standard deviation. When the standard deviation exceeds a certain threshold, if the predicted life value given by the model is higher than the actual value, the model is corrected to meet the physical law that the larger the pressure fluctuation value, the shorter the predicted life.

[0030] Specifically, in step S101, the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the core main pump to be predicted within a preset time are monitored.

[0031] It should be understood that the embodiment of the present application tracks the state degradation trajectory of the mechanical seal and predicts its changing trend by analyzing the changing trend of the pressure signal, that is, monitoring the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the core main pump to be predicted within a preset time through a pressure sensor.

[0032] In step S102, the pressure mean and pressure standard deviation of the mechanical seal inlet pressure are input into the life prediction model of the nuclear main pump to be predicted, and the degradation state and current life prediction value of the nuclear main pump to be predicted are obtained, wherein the life prediction model is obtained by training a preset neural network using historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and the initial life prediction model is constrained by a preset loss function, wherein the seal inlet pressure volatility is introduced into the preset loss function.

[0033] In some embodiments, the preset loss function is: ; (1)

[0034] in, is the root mean square error between the performance prediction value and the actual performance value during training, is the loss function, is the root mean square error, For the penalty item, is the pressure standard deviation, is the error when the performance prediction value is higher than the actual performance value, is the penalty factor.

[0035] Specifically, the embodiment of the present application establishes a polynomial-based neural network (i.e., a preset neural network) to achieve the preliminary construction of the life prediction model, and uses the historical degradation state data of multiple target nuclear main pumps to train the preset neural network to obtain an initial life prediction model. In the initial life prediction model, there are two loss functions to constrain the model, thereby obtaining a life prediction model, a loss function Used to calculate the mean square error of the model, another loss function A penalty term is added on the basis of the mean square error to reflect the constraints of physical information on the model.

[0036] During the specific implementation process, the neural network in the life prediction model expanded and transformed the pressure mean and pressure standard deviation of the mechanical seal inlet pressure by generating a polynomial characteristic matrix, and then obtained the weight vector and made a prediction by minimizing the loss function, thereby obtaining the degradation state and current life prediction value of the nuclear main pump to be predicted.

[0037] Optionally, in some embodiments, when the degradation state and current life prediction value of the nuclear main pump to be predicted are obtained, it includes: judging whether the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted; if the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted, then using a preset loss function to correct the life prediction model until the current life prediction value is less than or equal to the actual performance value of the nuclear main pump to be predicted, and obtaining a corrected life prediction model to predict the current life prediction value of the nuclear main pump to be predicted according to the corrected life prediction model.

[0038] Among them, the true performance value is the life value of the nuclear main pump to be predicted before the life prediction is carried out.

[0039] Optionally, in some embodiments, after obtaining the degradation state of the nuclear main pump to be predicted and the current life prediction value, it includes: judging whether the current life prediction value is less than or equal to a preset threshold, or judging whether the health status of the mechanical seal is lower than a preset health standard based on the degradation state of the nuclear main pump to be predicted; if the current life prediction value is less than or equal to the preset threshold, or the health status of the mechanical seal is lower than the preset health standard, then triggering a life warning signal, and sending the life warning signal to a preset terminal.

[0040] Among them, the preset threshold value and the preset health standard can be pre-set by the user, can be obtained through a limited number of experiments, or can be obtained through a limited number of computer simulations, and are not specifically limited here.

[0041] In an embodiment of the present application, if the current life prediction value given by the life prediction model is higher than the true performance value, the life prediction model needs to be penalized. If the current life prediction value is less than the true performance value, the current life prediction value is used as the current life prediction value.

[0042] Specifically, if the current life prediction value is greater than the actual performance value, the life prediction model is corrected using a preset loss function until the new life prediction value is less than or equal to the actual performance value to obtain a corrected life prediction model to satisfy the physical law that the larger the pressure fluctuation value, the shorter the corresponding predicted life. The pressure mean and pressure standard deviation of the mechanical seal inlet pressure are re-input into the corrected life prediction model to obtain the current life prediction value of the nuclear main pump to be predicted.

[0043] Finally, determine whether the current life prediction value is less than or equal to the preset threshold, or determine whether the health status of the mechanical seal is lower than the preset health standard based on the degradation state of the nuclear main pump to be predicted. If the current life prediction value is less than or equal to the preset threshold, or the health status of the mechanical seal is lower than the preset health standard, the life warning signal is triggered and sent to the preset terminal.

[0044] For example, assuming that the preset threshold of the life of the nuclear main pump is 5 years, and the health standard of the mechanical seal is set to 90% (the actual situation may vary depending on the specific equipment and technical requirements). If the health status of the mechanical seal is lower than 90%, it is considered that the health status of the nuclear main pump is lower than the preset health standard.

[0045] If the current life prediction value of a nuclear main pump is predicted to be 4 years, which is less than the preset threshold value of 5 years, or the health status of the mechanical seal of a nuclear main pump is detected to be 85%, which is lower than the preset health standard of 90%, in any case, a life warning signal will be triggered. Once the life warning signal is triggered, the system will automatically send this information to the preset terminal, such as the operation and maintenance center of the nuclear power plant, the mobile phone or email of the manager, etc. After receiving the warning signal, the operation and maintenance personnel can take immediate measures, such as arranging maintenance, replacing spare parts or conducting further inspections, to ensure the safe operation of the nuclear power plant.

[0046] A specific embodiment of this application is as follows: Taking a fluid dynamic pressure nuclear main pump mechanical seal as an example, the design value of the pressure before the third-stage mechanical seal is 5.4MPa. After the three nuclear main pumps of a nuclear power plant were started, the inlet pressure of the third-stage mechanical seal was about 5.2MPa, which was within the normal range. However, after the equipment had been running for a period of time, the inlet pressure of the third-stage mechanical seals of the three main pumps showed abnormal declines to varying degrees; at the same time, the abnormal decline in the inlet pressure of the third-stage mechanical seal was accompanied by an increase in the leakage of the third-stage mechanical seal, which showed that the inlet pressure of the mechanical seal was directly related to the failure of the mechanical seal. Therefore, the inlet pressure of the third-stage mechanical seal was determined as a key indicator for judging the health status of the mechanical seal.

[0047] The sampling interval of the pressure sensor is 5 seconds, and the mean and standard deviation every five minutes are extracted from the raw data, such as Figure 2 shown. Figure 2 (a) shows the fluctuation of the mean pressure before the seal of the third-stage mechanical seal of the three pumps. Figure 2 (b) shows the standard deviation of the pressure before sealing.

[0048] Through the analysis of actual machine operation data, it is found that there is a good correlation between the standard deviation of partial pressure over a period of time and future performance degradation. The size of the standard deviation can also reflect the volatility of the pressure of the equipment. The greater the volatility, the worse the sealing performance, the easier the end face is to wear, and the easier it is for the inlet pressure to drop. Figure 2 It can be observed that the time period with a larger standard deviation just corresponds to the pressure drop. This is because when the friction and wear of the sealing ring material is too great, the ability of the dynamic ring to follow the static ring decreases, resulting in increased leakage and seal failure. The inlet pressure will drop, resulting in amplified unstable characteristics and increased pressure fluctuations.

[0049] Based on two different loss functions, as shown in Equation (1), the degradation state of each pump is obtained. According to the proposed scheme, Figure 3 The life degradation model based on pressure data under different training sets of pump No. 1 is shown. The black dots correspond to the training set part, and the dotted lines correspond to the life degradation model based on pressure data. The degradation trajectory of the loss function, the dotted line corresponds to the The degradation trajectory of the loss function shows that: The prediction of the dotted line is more consistent with the degradation trajectory of the seal, which is due to the introduction of the physical information of pressure volatility in the prediction model. This method not only better captures the law of changes in the degradation state of the equipment, but also effectively improves the accuracy and reliability of the model by correcting the model loss function. With the mechanism of penalizing the model when the abnormal degree of pressure volatility exceeds the predetermined threshold, the model is more in line with the actual situation and accurately predicts the life of the equipment. As the number of training sets continues to increase, the accuracy of the model continues to improve. This shows that as more training samples are included in the model learning process, the model can more comprehensively understand the relationship between pressure volatility and equipment life. Therefore, gradually expanding the training set is of great significance for establishing a more powerful and reliable life prediction model. Based on long-term operation and maintenance experience, the mechanical seal inlet pressure of 4.7MPa is set as the lower limit threshold to characterize the health status of the third-stage mechanical seal of the nuclear main pump. Figure 3It is represented by a solid line. By calculating the time when the current time reaches this lower limit threshold, the remaining service life of the equipment can be obtained, which provides a practical and operational basis for model prediction. This information can not only serve as an important indicator of the health status of the equipment, but also an effective basis for the prediction of the remaining service life. According to the nuclear main pump mechanical seal life prediction method proposed in the embodiment of the present application, the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time are monitored, and the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within the preset time are input into the life prediction model of the nuclear main pump to be predicted, and the degradation state and current life prediction value of the nuclear main pump to be predicted are obtained, wherein the life prediction model is obtained by training a preset neural network using historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and the initial life prediction model is constrained by a preset loss function. In this way, the problems such as the high cost of nuclear main pump equipment, resulting in few equipment and few failures, insufficient data samples to ensure the high accuracy of the life prediction model, and complex working conditions of mechanical seals, which make it difficult for pure data-driven methods to capture the state degradation trend of mechanical seals, are solved. Integrating physical information into the data-driven model can more accurately construct the degradation process of mechanical seals in actual operation, improve the accuracy of the life prediction model results, and accurately predict the remaining service life of mechanical seals. Potential faults can be discovered in advance and corresponding maintenance measures can be taken, which helps to avoid unexpected shutdowns and equipment damage, improve the safe and reliable operation of nuclear main pumps, and promote the sustainable and healthy development of the nuclear power industry.

[0050] Next, a nuclear main pump mechanical seal life prediction device proposed according to an embodiment of the present application will be described with reference to the accompanying drawings.

[0051] Figure 4 It is a block diagram of a nuclear main pump mechanical seal life prediction device according to an embodiment of the present application.

[0052] like Figure 4 As shown, the nuclear main pump mechanical seal life prediction device 10 includes: a monitoring module 100 and a prediction module 200.

[0053] Among them, the monitoring module 100 is used to monitor the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time; the prediction module 200 is used to input the pressure mean and pressure standard deviation of the mechanical seal inlet pressure into the life prediction model of the nuclear main pump to be predicted, and obtain the degradation state and current life prediction value of the nuclear main pump to be predicted, wherein the life prediction model is obtained by training a preset neural network using historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and the initial life prediction model is constrained by a preset loss function, wherein the preset loss function introduces the volatility of the seal inlet pressure.

[0054] Optionally, in some embodiments, when the degradation state and current life prediction value of the nuclear main pump to be predicted are obtained, the prediction module is also used to: determine whether the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted; if the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted, then use a preset loss function to correct the life prediction model until the current life prediction value is less than or equal to the actual performance value of the nuclear main pump to be predicted, and obtain a corrected life prediction model to predict the current life prediction value of the nuclear main pump to be predicted according to the corrected life prediction model.

[0055] Optionally, in some embodiments, the preset loss function is: ;

[0056] in, is the root mean square error between the performance prediction value and the actual performance value during training, is the loss function, is the root mean square error, For the penalty item, is the pressure standard deviation, is the error when the performance prediction value is higher than the actual performance value, is the penalty factor.

[0057] Optionally, in some embodiments, after obtaining the degradation state of the core main pump to be predicted and the current life prediction value, the prediction module 200 is also used to: determine whether the current life prediction value is less than or equal to a preset threshold, or determine whether the health status of the mechanical seal is lower than a preset health standard based on the degradation state of the core main pump to be predicted; if the current life prediction value is less than or equal to the preset threshold, or the health status of the mechanical seal is lower than the preset health standard, then trigger a life warning signal, and send the life warning signal to a preset terminal.

[0058] It should be noted that the above explanation of the embodiment of the method for predicting the life of the mechanical seal of the nuclear main pump is also applicable to the device for predicting the life of the mechanical seal of the nuclear main pump of this embodiment, and will not be repeated here.

[0059] According to the nuclear main pump mechanical seal life prediction device proposed in the embodiment of the present application, the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time are monitored, and the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within the preset time are input into the life prediction model of the nuclear main pump to be predicted, so as to obtain the degradation state and current life prediction value of the nuclear main pump to be predicted, wherein the life prediction model is obtained by training a preset neural network with historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and the initial life prediction model is constrained by a preset loss function. As a result, the problems of the high cost of nuclear main pump equipment, resulting in few equipment and few failures, insufficient data samples to ensure the high accuracy of the life prediction model, and complex working conditions of mechanical seals, which make it difficult for pure data-driven methods to capture the state degradation trend of mechanical seals, are solved. The degradation process of mechanical seals in actual operation can be constructed more accurately, and the accuracy of the life prediction model results can be improved. By accurately predicting the current life prediction value of the mechanical seal, potential faults can be discovered in advance and corresponding maintenance measures can be taken, which helps to avoid unexpected shutdowns and equipment damage, improve the safe and reliable operation of the predicted nuclear main pumps, and promote the sustainable and healthy development of the nuclear power industry.

[0060] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include: A memory 501 , a processor 502 , and a computer program stored in the memory 501 and executable on the processor 502 .

[0061] When the processor 502 executes the program, the nuclear main pump mechanical seal life prediction method provided in the above embodiment is implemented.

[0062] Furthermore, the electronic device further comprises: The communication interface 503 is used for communication between the memory 501 and the processor 502 .

[0063] The memory 501 is used to store computer programs that can be executed on the processor 502 .

[0064] The memory 501 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0065] If the memory 501, the processor 502 and the communication interface 503 are implemented independently, the communication interface 503, the memory 501 and the processor 502 can be connected to each other through a bus and 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. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0066] Optionally, in a specific implementation, if the memory 501, the processor 502 and the communication interface 503 are integrated on a chip, the memory 501, the processor 502 and the communication interface 503 can communicate with each other through an internal interface.

[0067] The processor 502 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0068] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned method for predicting the life of a nuclear main pump mechanical seal.

[0069] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0070] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0071] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0072] 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 specifically implemented in any computer program product for use with 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 combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, a "computer program product" can be any device that can contain, store, communicate, propagate or transmit a program for use with an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer program products (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic device), a portable computer disk box (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). Furthermore, the computer program product may even be 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 and then editing, interpreting or, if necessary, processing it in another suitable manner, and then storing it in a computer memory.

[0073] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, N 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.

[0074] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer program product, which, when executed, includes one or a combination of the steps of the method embodiment.

[0075] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer program product.

[0076] The computer program product mentioned above may be a read-only memory, a disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for predicting the life of a nuclear main pump mechanical seal, characterized in that: The following steps are involved: Monitor the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time; The pressure mean and pressure standard deviation of the mechanical seal inlet pressure are input into the life prediction model of the nuclear main pump to be predicted to obtain the degradation state and current life prediction value of the nuclear main pump to be predicted, wherein the life prediction model is obtained by training a preset neural network using historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and the initial life prediction model is constrained by a preset loss function, wherein the seal inlet pressure volatility is introduced into the preset loss function.

2. The method according to claim 1, characterized in that When the degradation state and current life prediction value of the nuclear main pump to be predicted are obtained, it includes: Determine whether the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted; If the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted, the life prediction model is corrected using the preset loss function until the new life prediction value is less than or equal to the actual performance value of the nuclear main pump to be predicted, and a corrected life prediction model is obtained to predict the current life prediction value of the nuclear main pump to be predicted based on the corrected life prediction model.

3. The method according to claim 1, characterized in that The preset loss function is: ； in, is the root mean square error between the performance prediction value and the actual performance value during training, is the loss function, is the root mean square error, For the penalty item, is the pressure standard deviation, is the error when the performance prediction value is higher than the actual performance value, is the penalty factor.

4. The method according to claim 1, characterized in that: After obtaining the degradation state and current life prediction value of the nuclear main pump to be predicted, it includes: Determine whether the current life prediction value is less than or equal to a preset threshold, or determine whether the health status of the mechanical seal is lower than a preset health standard based on the degradation state of the core main pump to be predicted; If the current life prediction value is less than or equal to the preset threshold, or the health condition of the mechanical seal is lower than the preset health standard, a life warning signal is triggered and sent to a preset terminal.

5. A nuclear main pump mechanical seal life prediction device, characterized in that: include: A monitoring module, used to monitor the pressure mean and pressure standard deviation of the mechanical seal inlet pressure of the nuclear main pump to be predicted within a preset time; A prediction module is used to input the pressure mean and pressure standard deviation of the mechanical seal inlet pressure into the life prediction model of the nuclear main pump to be predicted, so as to obtain the degradation state and current life prediction value of the nuclear main pump to be predicted, wherein the life prediction model is obtained by training a preset neural network using historical degradation state data of multiple target nuclear main pumps to obtain an initial life prediction model, and constraining the initial life prediction model using a preset loss function, wherein the preset loss function introduces the volatility of the seal inlet pressure.

6. The device according to claim 5, characterized in that When the degradation state and current life prediction value of the nuclear main pump to be predicted are obtained, the prediction module is further used to: Determine whether the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted; If the current life prediction value is greater than the actual performance value of the nuclear main pump to be predicted, the life prediction model is corrected using the preset loss function until the current life prediction value is less than or equal to the actual performance value of the nuclear main pump to be predicted, and a corrected life prediction model is obtained to predict the current life prediction value of the nuclear main pump to be predicted based on the corrected life prediction model.

7. The device according to claim 5, characterized in that The preset loss function is: ； in, is the root mean square error between the performance prediction value and the actual performance value during training, is the loss function, is the root mean square error, For the penalty item, is the pressure standard deviation, is the error when the performance prediction value is higher than the actual performance value, is the penalty factor.

8. The device according to claim 5, characterized in that After obtaining the degradation state and current life prediction value of the nuclear main pump to be predicted, the prediction module is further used to: Determine whether the current life prediction value is less than or equal to a preset threshold, or determine whether the health status of the mechanical seal is lower than a preset health standard based on the degradation state of the core main pump to be predicted; If the current life prediction value is less than or equal to the preset threshold, or the health condition of the mechanical seal is lower than the preset health standard, a life warning signal is triggered and sent to a preset terminal.

9. An electronic device, characterized in that: It comprises 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 method for predicting the life of a nuclear main pump mechanical seal as described in any one of claims 1 to 4.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for predicting the life of a nuclear main pump mechanical seal is implemented as described in any one of claims 1 to 4.

Images