Bearing fault detection method and device of reactor coolant pump and related equipment

Abstract

The invention provides a bearing fault detection method and device for a reactor coolant pump and related equipment. The method comprises the steps that multiple pieces of equipment operation index data of a thrust bearing are acquired; performing dimensionless processing on the operation index data of each piece of equipment to obtain a corresponding dimensionless index matrix; performing fuzzification processing on each dimensionless index matrix to obtain the state probability of each piece of equipment operation index data in different equipment state levels; performing evidence fusion on the state probability of the operation index data of each device to obtain a target probability distribution function; and determining the bearing fault state of the reactor coolant pump according to the target probability distribution function. The bearing fault detection accuracy of the reactor coolant pump can be improved.

Description

Technical Field

[0001] This application relates to the field of nuclear power safety detection, and particularly to a method and device for detecting bearing faults of a reactor coolant pump and related equipment. Background Art

[0002] The reactor coolant pump, also known as the main pump, is one of the key equipment in the loop system of a pressurized water reactor nuclear power plant. Its function is to drive the coolant to circulate in the reactor coolant system and take away the heat of the reactor core. At the same time, the pressure-bearing components of the main pump are also an important part of the pressure boundary to prevent the leakage of radioactive substances. Therefore, the reliable operation of the main pump can ensure the continuous and safe operation of the nuclear power plant.

[0003] In the related art, the fault detection of the existing main pump in a nuclear power plant is to set temperature sensors on the thrust pads of the thrust bearing, and detect the fault state of the thrust bearing by collecting the temperature of the thrust pads, so as to carry out preventive maintenance and regular replacement. However, for the existing fault detection scheme, there have been several cases where the temperature of the main thrust bearing abnormally rises under the condition of main pump power failure, and it is found that the main thrust bearing is severely worn after disassembly. Therefore, the method of judging only by temperature cannot accurately monitor its operating state and cannot meet the fault diagnosis requirements of the main pump. Summary of the Invention

[0004] This application aims to solve at least one of the technical problems existing in the prior art. For this purpose, this application provides a method and device for detecting bearing faults of a reactor coolant pump and related equipment, which can improve the accuracy of detecting bearing faults of the reactor coolant pump.

[0005] In a first aspect, an embodiment of this application provides a method for detecting bearing faults of a reactor coolant pump, where the reactor coolant pump includes a thrust bearing, and the method includes:

[0006] Obtain a plurality of device operation index data of the thrust bearing;

[0007] Perform dimensionless processing on each piece of the device operation index data to obtain a corresponding dimensionless index matrix;

[0008] Perform fuzzy processing on each dimensionless index matrix to obtain the state probability of each piece of the device operation index data in different device state levels;

[0009] Perform evidence fusion on the state probabilities of each piece of the device operation index data to obtain a target probability assignment function;

[0010] Determine the bearing fault state of the reactor coolant pump according to the target probability assignment function.

[0011] In some embodiments, the thrust bearing is provided with a temperature sensor, an ultrasonic measurement sensor, a noise measurement sensor, and a vibration acceleration sensor. The multiple device operation index data includes bearing pad temperature data, bearing oil film thickness data, bearing noise data, and bearing vibration data. Obtaining the multiple device operation index data of the thrust bearing includes:

[0012] Collecting the temperature of the thrust pads of the thrust bearing through the temperature sensor to obtain the bearing pad temperature data;

[0013] Performing ultrasonic measurement on the thrust pads of the thrust bearing through the ultrasonic measurement sensor to obtain the bearing oil film thickness data;

[0014] Performing noise measurement on the outer ring of the bearing housing of the thrust bearing through the noise measurement sensor to obtain the bearing noise data;

[0015] Performing acceleration measurement on the outer ring of the bearing housing of the thrust bearing through the vibration acceleration sensor to obtain the bearing vibration data.

[0016] In some embodiments, the dimensionless index matrix includes a temperature index matrix, an oil film thickness index matrix, a bearing noise index matrix, and a bearing vibration index matrix. Performing dimensionless processing on each device operation index data to obtain the corresponding dimensionless index matrix includes:

[0017] Calculating the temperature index matrix according to a preset temperature initial value, a preset temperature threshold, and the bearing pad temperature data;

[0018] Calculating the oil film thickness index matrix according to a preset oil film thickness initial value, a preset oil film thickness threshold, and the bearing oil film thickness data;

[0019] Calculating the bearing noise index matrix according to a preset bearing noise initial value, a preset bearing noise threshold, and the bearing noise data;

[0020] Calculating the bearing vibration index matrix according to a preset bearing vibration initial value, a preset bearing vibration threshold, and the bearing vibration data.

[0021] In some embodiments, performing fuzzy processing on each dimensionless index matrix to obtain the state probabilities of each device operation index data in different device state levels includes:

[0022] Obtaining the membership function parameters of each device operation index data corresponding to the normal state, the warning state, and the fault state;

[0023] Calculate the normal state probability of each piece of equipment operation index data according to the membership function parameters in the normal state;

[0024] Calculate the warning state probability of each piece of equipment operation index data according to the membership function parameters in the warning state;

[0025] Calculate the fault state probability of each piece of equipment operation index data according to the membership function parameters in the fault state.

[0026] In some embodiments, the fusing the state probabilities of the various pieces of equipment operation index data to obtain a target probability assignment function includes:

[0027] Combine the normal state probability, the warning state probability, and the fault state probability of each piece of equipment operation index data to obtain an original probability assignment function for each piece of equipment operation index data;

[0028] Based on the preset importance weight of each piece of equipment operation index data, correct the original probability assignment function to obtain an intermediate probability assignment function;

[0029] Based on the Dempster-Shafer theory, fuse the intermediate probability assignment functions to obtain the target probability assignment function.

[0030] In some embodiments, the correcting the original probability assignment function based on the preset importance weight of each piece of equipment operation index data to obtain the intermediate probability assignment function includes:

[0031] Obtain the relative importance weight of each piece of equipment operation index data;

[0032] Calculate the confidence coefficient of each piece of equipment operation index data according to the relative importance weight;

[0033] Based on the confidence coefficient, correct the original probability assignment function of the corresponding piece of equipment operation index data to obtain the intermediate probability assignment function.

[0034] In some embodiments, the calculating the confidence coefficient of each piece of equipment operation index data according to the relative importance weight includes:

[0035] Obtain the maximum value in the relative importance weights;

[0036] Calculate the ratio of each relative importance weight to the maximum value to obtain the corresponding normalized weight;

[0037] Multiply the normalized weight by a preset conversion coefficient to obtain a confidence coefficient corresponding to each piece of the device operation index data.

[0038] In some embodiments, based on the confidence coefficient, modifying the original probability assignment function of the corresponding device operation index data to obtain the intermediate probability assignment function includes:

[0039] Multiply the state probabilities in different device state levels in the original probability assignment function by the corresponding confidence coefficients respectively to obtain the modified state probabilities;

[0040] Based on the confidence coefficient, calculate the state uncertainty of each piece of the device operation index data;

[0041] Combine the modified state probabilities and the state uncertainty to obtain the intermediate probability assignment function.

[0042] In some embodiments, determining the bearing fault state of the reactor coolant pump according to the target probability assignment function includes:

[0043] Obtain the maximum value of the state probability in the target probability assignment function and the device state level corresponding to the maximum value of the state probability;

[0044] Calculate the difference between the maximum value of the state probability and the maximum value of other state probabilities;

[0045] Judge whether the difference is greater than a first preset threshold, whether the state uncertainty is less than a second preset threshold, and whether the maximum value of the state probability is greater than the state uncertainty;

[0046] When the above judgment conditions are met, determine the device state level corresponding to the maximum value of the state probability as the bearing fault state of the reactor coolant pump.

[0047] In some embodiments, the method further includes:

[0048] Calculate the health degree corresponding to each piece of the device operation index data according to a preset health degree membership function;

[0049] Calculate the overall health degree of the thrust bearing according to the health degree of each piece of the device operation index data and a preset health degree weight coefficient;

[0050] When the health degree of any piece of the device operation index data is less than a preset health degree threshold, set the overall health degree to the minimum health degree less than the preset health degree threshold.

[0051] In some embodiments, the multiple device operation index data further includes bearing pad wear depth data, and the method further includes:

[0052] Performing interpolation and completion processing on the bearing pad wear depth data to obtain a wear depth sequence with equally spaced distribution;

[0053] Based on the grey prediction model, performing trend prediction on the wear depth sequence to obtain a predicted wear amount;

[0054] Calculating a wear penalty amount according to a preset operation cycle and a preset transient wear depth;

[0055] Adding the predicted wear amount and the wear penalty amount to obtain a corrected predicted wear amount;

[0056] Predicting the remaining service life based on the corrected predicted wear amount.

[0057] In some embodiments, the predicting the remaining service life based on the corrected predicted wear amount includes:

[0058] When the corrected predicted wear amount reaches the preset wear threshold, setting the remaining service life to zero;

[0059] When the corrected predicted wear amount does not reach the preset wear threshold, calculating the difference between the preset wear threshold and the corrected predicted wear amount, and predicting the remaining service life according to the difference and a preset wear rate.

[0060] In a second aspect, an embodiment of the present application provides a bearing fault detection device for a reactor coolant pump, including:

[0061] An acquisition module, configured to acquire multiple device operation index data of a thrust bearing;

[0062] A dimensionless processing module, configured to perform dimensionless processing on each of the device operation index data to obtain a corresponding dimensionless index matrix;

[0063] A fuzzification processing module, configured to perform fuzzification processing on each of the dimensionless index matrices to obtain the state probabilities of each of the device operation index data in different device state levels;

[0064] An evidence fusion module, configured to perform evidence fusion on the state probabilities of each of the device operation index data to obtain a target probability assignment function;

[0065] A determination module, configured to determine the bearing fault state of the reactor coolant pump according to the target probability assignment function.

[0066] In a third aspect, an embodiment of the present application provides an electronic device, including: a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the bearing fault detection method for a reactor coolant pump as described in any one of the embodiments of the first aspect of the present application.

[0067] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium. The storage medium stores a program, and when the program is executed by a processor, it implements the bearing fault detection method for a reactor coolant pump as described in any one of the embodiments of the first aspect of the present application.

[0068] The bearing fault detection method for a reactor coolant pump according to the embodiments of the present application has at least the following beneficial effects:

[0069] The bearing fault detection method for a reactor coolant pump according to the embodiments of the present application includes: obtaining a plurality of device operation index data of a thrust bearing; performing dimensionless processing on each device operation index data to obtain a corresponding dimensionless index matrix; performing fuzzy processing on each dimensionless index matrix to obtain the state probability of each device operation index data in different device state levels; performing evidence fusion on the state probabilities of each device operation index data to obtain a target probability assignment function; and determining the bearing fault state of the reactor coolant pump according to the target probability assignment function.

[0070] By obtaining a plurality of device operation index data of the thrust bearing, the present application can comprehensively monitor the operation state parameters of the thrust bearing. Then, by performing dimensionless processing on each device operation index data to obtain a corresponding dimensionless index matrix, the influence of the dimension and order of magnitude between different device operation index data can be eliminated, making each device operation index data comparable. Next, by performing fuzzy processing on each dimensionless index matrix to obtain the state probability of each device operation index data in different device state levels, the deterministic device operation index data can be converted into a fuzzy probability form, better characterizing the uncertainty of the device state. Subsequently, by performing evidence fusion on the state probabilities of each device operation index data to obtain a target probability assignment function, the relative importance of each device operation index data can be fully considered, realizing the effective fusion of multi-source information. Finally, by determining the bearing fault state of the reactor coolant pump according to the target probability assignment function, the fault mode of the thrust bearing can be accurately identified. Compared with the prior art that only relies on temperature sensors for state monitoring, the method provided by the embodiments of the present application realizes the comprehensive monitoring and accurate diagnosis of the thrust bearing state by fusing a plurality of device operation index data, improving the accuracy of bearing fault detection for the reactor coolant pump.

[0071] Additional aspects and advantages of the present application will be given in part in the following description, become apparent in part from the following description, or be learned by practice of the present application. Description of the Drawings

[0072] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description of embodiments in conjunction with the accompanying drawings, in which:

[0073] Figure 1 is a flowchart of an optional bearing fault detection method for a reactor coolant pump provided in an embodiment of the present application;

[0074] Figure 2 is a schematic structural diagram of an optional reactor coolant pump provided in an embodiment of the present application;

[0075] Figure 3 is a schematic structural diagram of an optional thrust bearing provided in an embodiment of the present application;

[0076] Figure 4 is a flowchart of another optional bearing fault detection method for a reactor coolant pump provided in an embodiment of the present application;

[0077] Figure 5 is a flowchart of another optional bearing fault detection method for a reactor coolant pump provided in an embodiment of the present application;

[0078] Figure 6 is a flowchart of another optional bearing fault detection method for a reactor coolant pump provided in an embodiment of the present application;

[0079] Figure 7 is a flowchart of another optional bearing fault detection method for a reactor coolant pump provided in an embodiment of the present application;

[0080] Figure 8 is a flowchart of another optional bearing fault detection method for a reactor coolant pump provided in an embodiment of the present application;

[0081] Figure 9 is a flowchart of another optional bearing fault detection method for a reactor coolant pump provided in an embodiment of the present application;

[0082] Figure 10 is a flowchart of another optional bearing fault detection method for a reactor coolant pump provided in an embodiment of the present application;

[0083] Figure 11 is a flowchart of another optional bearing fault detection method for a reactor coolant pump provided in an embodiment of the present application;

[0084] Figure 12Flowchart of another alternative bearing fault detection method for the reactor coolant pump provided by the embodiments of the present application;

[0085] Figure 13 Schematic diagram of an alternative health membership function provided by the embodiments of the present application;

[0086] Figure 14 Flowchart of another alternative bearing fault detection method for the reactor coolant pump provided by the embodiments of the present application;

[0087] Figure 15 Flowchart of another alternative bearing fault detection method for the reactor coolant pump provided by the embodiments of the present application;

[0088] Figure 16 Schematic diagram of the bearing fault detection device for the reactor coolant pump provided by the embodiments of the present application;

[0089] Figure 17 Schematic diagram of the hardware structure of the electronic device provided by the embodiments of the present application. Detailed implementation manners

[0090] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where 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 only for explaining the present application and should not be construed as limiting the present application.

[0091] In the description of the present application, the meaning of several is one or more, the meaning of multiple is two or more, greater than, less than, exceeding, etc. are understood not to include the present number, and above, below, within, etc. are understood to include the present number. If there is a description of first and second, it is only for the purpose of distinguishing technical features and should not be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features or implicitly indicating the sequence relationship of the indicated technical features.

[0092] In the description of the present application, it should be understood that the orientation or positional relationship involved, such as up, down, left, right, front, back, etc., is based on the orientation or positional relationship shown in the accompanying drawings. It is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the present application.

[0093] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "schematic embodiments", "examples", "specific examples", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0094] In the description of this application, it should be noted that unless otherwise clearly defined, words such as "arrangement", "installation", "connection", etc. should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above words in this application in combination with the specific content of the technical solution. In addition, the identification of specific steps below does not represent a limitation on the step sequence and execution logic. The execution sequence and execution logic between each step should be understood and inferred with reference to the content described in the embodiment.

[0095] The reactor coolant pump, also known as the main pump, is one of the key equipment in the loop system of a pressurized water reactor nuclear power plant. Its function is to drive the coolant to circulate in the reactor coolant system and take away the heat of the reactor core. At the same time, the pressure-bearing components of the main pump are also an important part of the pressure boundary to prevent the leakage of radioactive substances. Therefore, the reliable operation of the main pump can ensure the continuous and safe operation of the nuclear power plant. The thrust bearing is the core component of the main pump. Its type is a sliding bearing, which forms a lubricating oil film between the thrust disk and the thrust pad by using the hydrodynamic lubrication effect to achieve the separation and load-bearing of the relatively moving surfaces of the friction pair. The operating state and reliability of the thrust bearing directly affect the safe operation and economic benefits of the nuclear power plant. Achieving fault prediction and health management of the main pump thrust bearing is of great significance for reducing the fault consequences and risks of the main pump thrust bearing, extending its service life, and reducing the operation and maintenance costs.

[0096] In the related art, for the existing nuclear power plant main pump, a temperature sensor is set on the thrust pad of the thrust bearing to detect the fault state of the thrust bearing by collecting the temperature of the thrust pad, so as to carry out preventive maintenance and regular replacement. However, for the existing fault detection scheme, there have been several cases where the temperature of the main thrust bearing abnormally rises under the condition of main pump power failure, and it is found that the main thrust bearing is severely worn after disassembly. Therefore, the method of simply judging by temperature cannot accurately monitor its operating state and cannot meet the fault diagnosis of the main pump.

[0097] Based on this, by obtaining multiple device operation index data of the thrust bearing, the operation state parameters of the thrust bearing can be comprehensively monitored. Then, by performing dimensionless processing on each device operation index data, a corresponding dimensionless index matrix is obtained, which can eliminate the influence of the dimension and order of magnitude between different device operation index data, making the device operation index data comparable. Next, by performing fuzzification processing on each dimensionless index matrix, the state probability of each device operation index data in different device state levels is obtained, which can convert the deterministic device operation index data into a fuzzy probability form to better characterize the uncertainty of the device state. Subsequently, by performing evidence fusion on the state probabilities of each device operation index data, a target probability assignment function is obtained, which can fully consider the relative importance of each device operation index data and realize the effective fusion of multi-source information. Finally, by determining the bearing fault state of the reactor coolant pump according to the target probability assignment function, the fault mode of the thrust bearing can be accurately identified. Compared with the prior art that only relies on temperature sensors for state monitoring, the method provided by the embodiment of the present application realizes the comprehensive monitoring and accurate diagnosis of the thrust bearing state by fusing multiple device operation index data, improving the accuracy of bearing fault detection of the reactor coolant pump.

[0098] Please refer to Figure 1 , a method for detecting bearing faults of a reactor coolant pump provided by an embodiment of the present invention, the reactor coolant pump includes a thrust bearing, and the method may include, but is not limited to, the following steps 101 to step 105:

[0099] Step 101, obtain multiple device operation index data of the thrust bearing.

[0100] Step 102, perform dimensionless processing on each device operation index data to obtain a corresponding dimensionless index matrix.

[0101] Step 103, perform fuzzification processing on each dimensionless index matrix to obtain the state probability of each device operation index data in different device state levels.

[0102] Step 104, perform evidence fusion on the state probabilities of each device operation index data to obtain a target probability assignment function.

[0103] Step 105, determine the bearing fault state of the reactor coolant pump according to the target probability assignment function.

[0104] In step 101 of some embodiments, the reactor coolant pump, abbreviated as the main pump, is one of the key devices in the loop system of a pressurized water reactor nuclear power plant, and its function is to drive the coolant to circulate in the reactor coolant system and take away the core heat. Refer to Figure 2 , Figure 2FIG. 0 is a schematic structural diagram of a reactor coolant pump according to an embodiment of the present application. A flywheel is provided at the top to provide inertia; a radial bearing for the upper part of the motor is provided below to support the radial movement of the shaft; immediately following is a two-way thrust bearing for the motor to bear the axial load; the middle part is the stator and rotor assembly of the motor; below it is a radial bearing for the lower part of the motor, which also plays a role in supporting the radial movement; at the bottom are a coupling, a shaft seal system, a water-lubricated bearing for the lower part of the pump, and an impeller in sequence, where the impeller provides power for the coolant by rotating.

[0105] Exemplarily, the thrust bearing is a core component of the main pump. Refer to Figure 3 , Figure 3 FIG. 7 is a schematic structural diagram of the thrust bearing according to an embodiment of the present application. The thrust bearing mainly consists of a thrust disk, main thrust pads, secondary thrust pads, as well as components such as a bearing housing, a shaft, and a cover. The tribological failures of the thrust bearing mainly include two stages: early lubrication failure and later wear failure of the bearing pads. Various failure causes such as poor lubrication and uneven force lead to the thinning of the oil film, causing the migration of the lubrication state, the decrease of the oil film protection degree, and the evolution from hydrodynamic lubrication to mixed lubrication. This kind of failure is prone to friction and collision between the thrust disk and the bearing pads when the load changes suddenly or under large-load operation, and then induces bearing rubbing, bearing pad wear, burning of the bearing pads, etc. In the second stage, due to poor lubrication, local wear of the bearing pads occurs, and the surface topography of the pad blocks is damaged, and this irreversible wear will accumulate and lead to malignant failures such as the removal and shedding of the babbit alloy on the surface of the bearing pads.

[0106] Exemplarily, the equipment operation index data refers to the key parameters used to characterize the operation state of the thrust bearing. For example, it may include data such as temperature, oil film thickness, rubbing noise, and vibration. The change in temperature can reflect the working state of the thrust bearing. When failures such as poor lubrication and uneven force occur, the increase in friction will cause the temperature to rise abnormally. When the bearing rubs, abnormal noise will be generated, and the noise signal formed by the transmission of this impact energy through the structure can reflect the rubbing state of the bearing.

[0107] Please refer to Figure 4 , in some embodiments, the thrust bearing is provided with a temperature sensor, an ultrasonic measurement sensor, a noise measurement sensor, and a vibration acceleration sensor. The multiple equipment operation index data includes bearing pad temperature data, bearing oil film thickness data, bearing noise data, and bearing vibration data. Step 101 may include, but is not limited to, steps 201 to 204.

[0108] Step 201, collect the temperature of the thrust pads of the thrust bearing through the temperature sensor to obtain the bearing pad temperature data.

[0109] Step 202, perform ultrasonic measurement on the thrust pads of the thrust bearing through the ultrasonic measurement sensor to obtain the bearing oil film thickness data.

[0110] Step 203: Use a noise measurement sensor to measure the noise of the outer ring of the bearing housing of the thrust bearing to obtain bearing noise data.

[0111] Step 204: Use a vibration acceleration sensor to measure the acceleration of the outer ring of the bearing housing of the thrust bearing to obtain bearing vibration data.

[0112] In step 201 of some embodiments, temperature sensors are arranged on the upper and lower thrust pads of the thrust bearing to monitor the temperature change of the thrust pads in real time. When faults such as poor lubrication and uneven force occur, it will cause an increase in friction and lead to an abnormal rise in temperature. Therefore, the bearing pad temperature data is an important indicator reflecting the operating state of the thrust bearing.

[0113] In step 202 of some embodiments, since the thrust bearing uses the hydrodynamic lubrication effect to form a lubricating oil film between the thrust disk and the thrust pad to achieve the separation and load-bearing of the relative moving surfaces of the friction pair, an ultrasonic measurement sensor is arranged on the back of the thrust pad. The oil film thickness of the thrust bearing can be monitored by ultrasonic measurement to obtain bearing oil film thickness data. The change of the oil film thickness can directly reflect the lubrication state of the bearing and is a key parameter for predicting bearing faults.

[0114] In step 203 of some embodiments, a noise measurement sensor is arranged on the outer ring of the thrust bearing chamber to monitor the abnormal noise generated by the rubbing of the thrust bearing. When the bearing fails, abnormal collision noises often occur, and this kind of noise can be captured by the sensor through structural transmission. By analyzing the bearing noise data, it can be judged whether there are abnormal states such as rubbing in the bearing.

[0115] In step 204 of some embodiments, a vibration acceleration sensor is arranged on the outer ring of the thrust bearing chamber to monitor the vibration state of the thrust bearing. The vibration characteristics of the bearing during operation reflect its dynamic operating state. By measuring the acceleration of the outer ring of the bearing housing to obtain bearing vibration data, abnormal vibrations during the operation of the bearing can be detected in time, providing a basis for fault diagnosis.

[0116] Through the above steps 201 to 204, a variety of sensors are configured on the thrust bearing to collect data, establishing a complete data acquisition scheme. A temperature sensor is set on the thrust pad to monitor the bearing temperature in real time and promptly detect fault signs such as abnormal temperature rise. An ultrasonic measurement sensor is set on the back of the thrust pad to accurately obtain the change of the oil film thickness, directly reflecting the lubrication state of the bearing and providing a basis for preventing lubrication failure. A noise measurement sensor is set on the outer ring of the bearing housing to promptly capture the abnormal sound generated by bearing rubbing and effectively monitor the collision state of the bearing. At the same time, a vibration acceleration sensor is set on the outer ring of the bearing housing to obtain the vibration characteristics during the operation of the bearing and accurately grasp its dynamic operation state. Through the coordinated cooperation of multiple sensors, a comprehensive monitoring system covering temperature, oil film, noise, and vibration is constructed, significantly improving the monitoring ability of the operation state of the thrust bearing and providing reliable data support for fault diagnosis and early warning.

[0117] In step 102 of some embodiments, it is necessary to perform dimensionless processing on each piece of equipment operation index data obtained. The purpose of dimensionless processing is to unify equipment operation index data with different dimensions and different orders of magnitude to a comparable scale. To perform normalization processing on the input equipment operation index data, it is first necessary to initialize the thresholds, initial values, and present values of the input monitoring indicators, and establish a dimensionless cost-type matrix using the range transformation method. When performing normalization processing, the equipment operation index data can be divided into two categories: benefit-type indicators and cost-type indicators. For cost-type indicators such as bearing pad temperature data, bearing noise data, and bearing vibration data, that is, indicators whose values increase as the state deteriorates, if their present values are lower than the initial values, the indicator data is normalized to 0; for benefit-type indicators such as bearing oil film thickness data, that is, indicators whose values decrease as the state deteriorates, if their present values are higher than the initial values, they are initialized to 0. Through this dimensionless processing, the corresponding dimensionless index matrix can be obtained, establishing a unified evaluation basis for subsequent fuzzy processing.

[0118] Exemplarily, dimensionless processing is performed on each piece of equipment operation index data obtained, and the calculation is as follows:

[0119]

[0120] Among them, a i represents the equipment operation index data (including bearing pad temperature data, bearing noise data, bearing vibration data, bearing oil film thickness data), a iN represents the threshold value when the equipment operation index data fails (including temperature threshold, bearing noise threshold, bearing vibration threshold, oil film thickness threshold), a i0 represents the equipment operation index data a iInitial values (including the initial temperature value, the initial bearing noise value, the initial bearing vibration value, and the initial oil film thickness value), represent the normalized device operation index data. I1 is the set of benefit-type indicators, and I2 is the set of cost-type indicators.

[0121] Please refer to Figure 5 , in some embodiments, the dimensionless index matrix includes a temperature index matrix, an oil film thickness index matrix, a bearing noise index matrix, and a bearing vibration index matrix. Step 102 may include, but is not limited to, steps 301 to 304.

[0122] Step 301, calculate the temperature index matrix based on the preset initial temperature value, the preset temperature threshold, and the bearing pad temperature data.

[0123] Step 302, calculate the oil film thickness index matrix based on the preset initial oil film thickness value, the preset oil film thickness threshold, and the bearing oil film thickness data.

[0124] Step 303, calculate the bearing noise index matrix based on the preset initial bearing noise value, the preset bearing noise threshold, and the bearing noise data.

[0125] Step 304, calculate the bearing vibration index matrix based on the preset initial bearing vibration value, the preset bearing vibration threshold, and the bearing vibration data.

[0126] In step 301 of some embodiments, the bearing pad temperature data is an important parameter reflecting the working state of the thrust bearing. The initial temperature value represents the temperature reference when the thrust bearing is working normally, and the temperature threshold represents the maximum temperature limit that the thrust bearing can withstand. By comparing and calculating the collected bearing pad temperature data with the initial temperature value and the temperature threshold, a standardized temperature index matrix can be obtained, which is used to characterize the degree of influence of temperature change on the bearing state.

[0127] In step 302 of some embodiments, the initial oil film thickness value represents the reference oil film thickness when the thrust bearing is working normally, and the oil film thickness threshold represents the minimum oil film thickness required for the safe operation of the thrust bearing. By comparing and calculating the measured bearing oil film thickness data with the initial oil film thickness value and the oil film thickness threshold, a standardized oil film thickness index matrix can be obtained, which is used to characterize the degree of influence of oil film thickness change on the bearing lubrication state.

[0128] In step 303 of some embodiments, the initial bearing noise value represents the noise reference level when the thrust bearing is operating normally, and the bearing noise threshold represents the maximum acceptable noise limit for the thrust bearing. By comparing and calculating the collected bearing noise data with the initial bearing noise value and the bearing noise threshold, a standardized bearing noise index matrix can be obtained, which is used to characterize the degree of influence of noise changes on the bearing rubbing state.

[0129] In step 304 of some embodiments, the initial bearing vibration value represents the vibration reference level when the thrust bearing is operating normally, and the bearing vibration threshold represents the maximum acceptable vibration limit for the thrust bearing. By comparing and calculating the measured bearing vibration data with the initial bearing vibration value and the bearing vibration threshold, a standardized bearing vibration index matrix can be obtained, which is used to characterize the degree of influence of vibration changes on the dynamic characteristics of the bearing.

[0130] Through the above steps 301 to 304, by respectively performing standardized processing on the four types of equipment operation index data of the temperature, oil film thickness, noise, and vibration of the thrust bearing, a complete dimensionless index matrix is established. Compared with directly using the original data for state evaluation, by establishing a unified evaluation standard, the dimension differences between different types of data are eliminated, the accuracy and comparability of state evaluation are improved, and a standardized data foundation is laid for subsequent fault diagnosis.

[0131] In step 103 of some embodiments, the fuzzification process is a process of converting the deterministic dimensionless index matrix into a fuzzy probability form. The equipment state levels can include three levels: H1, H2, and H3, corresponding to three state levels of the bearing state from normal to faulty. The state probability represents the probability value of a certain equipment operation index data being in each state level.

[0132] Exemplarily, the state probabilities of the equipment operation index data in different equipment state levels are represented as M r (H N ); where r is the number of indexes, which can take 1, 2, 3, or 4, representing the bearing pad temperature data, bearing noise data, bearing vibration data, or bearing oil film thickness data respectively; H N represents the equipment state level, and the equipment state level can include three levels: H1, H2, and H3. For example: M r (H1) = 0.4, M r (H2) = 0.6, M r (H3) = 0 means that the probability of this equipment operation index data being in the H1 state level is 0.4, the probability of being in the H2 state level is 0.6, and the probability of being in the H3 state is 0.

[0133] Please refer to Figure 6, in some embodiments, step 103 may include, but is not limited to, steps 401 to 404.

[0134] Step 401, obtain the membership function parameters corresponding to each device operation index data in the normal state, warning state, and failure state.

[0135] Step 402, calculate the normal state probability of each device operation index data according to the membership function parameters in the normal state.

[0136] Step 403, calculate the warning state probability of each device operation index data according to the membership function parameters in the warning state.

[0137] Step 404, calculate the failure state probability of each device operation index data according to the membership function parameters in the failure state.

[0138] In step 401 of some embodiments, a Gaussian function can be used as the membership function, which has good mathematical properties and interpretability. For the four types of device operation index data, namely temperature, oil film thickness, noise, and vibration, it is necessary to determine their membership function parameters in the three states respectively. Each Gaussian function contains two key parameters: the mean μ and the variance σ. The mean μ reflects the central eigenvalue of each state, and the variance σ characterizes the degree of data dispersion. These parameters can be obtained in various ways: they can be obtained through statistical analysis based on a large amount of historical operation data; they can also be initially set according to expert experience and then continuously optimized and adjusted through actual operation data; they can also be determined through theoretical calculations in combination with device design specifications and operation requirements.

[0139] In step 402 of some embodiments, for the normalized temperature index matrix, oil film thickness index matrix, bearing noise index matrix, and bearing vibration index matrix, substitute them into the Gaussian function corresponding to the normal state for calculation respectively. Use each device operation index data as the independent variable for operation, and the calculation result represents the probability value that the index data belongs to the normal state. The larger the probability value, the closer the index is to the normal operation state. For example, when the oil film thickness is close to the design value, its normal state probability will be higher; when the temperature is within the normal working range, its normal state probability will also be higher.

[0140] In step 403 of some embodiments, the Gaussian function form is the same as in the normal state, but the membership function parameters corresponding to the warning state are used, that is, the mean and variance parameters. The warning state is a transitional state between the normal state and the fault state. The mean of the warning state should be set in the critical region between the normal state and the fault state, and the variance should consider the width of the warning interval. For example, for the temperature index, when the temperature begins to deviate from the normal range but has not reached the fault level, the probability of the warning state is relatively high; for the oil film thickness index, when the thickness begins to decrease but has not reached the dangerous level, the probability of the warning state will also increase accordingly. This setting can achieve early warning of the bearing state.

[0141] In step 404 of some embodiments, the determination of the fault state is particularly critical, and more stringent parameter criteria need to be set. The Gaussian function is also used for calculation, but the mean and variance parameters corresponding to the fault state are used. The determination of the parameters is based on the safety limits of the equipment and historical fault cases. For example, when the temperature exceeds the safety threshold, the probability of the fault state of the temperature index will increase significantly; when the oil film thickness is lower than the minimum allowable value, the probability of the fault state of the oil film thickness index will increase rapidly; when the noise or vibration exceeds the warning value, the probability of the fault state of the corresponding index will also increase. By calculating the probability of the fault state, the fault state of the bearing can be accurately identified, and a fault alarm can be issued in a timely manner.

[0142] Exemplarily, each dimensionless index matrix is fuzzified to obtain the state probability of each equipment operation index data in different equipment state levels, and the calculation is as follows:

[0143]

[0144] where a i represents the equipment operation index data, is the normalized equipment operation index data in the dimensionless index matrix, H N represents the equipment state level. The equipment state level can include three levels: H1, H2, and H3, corresponding to the normal state, warning state, and fault state respectively. P(H N |a i ) represents the state probability that the monitored index data a i (bearing pad temperature data, bearing noise data, bearing vibration data, or bearing oil film thickness data) belongs to the equipment state level H N . μ and σ represent the membership function parameters in different states. The membership function can use the Gaussian function, and μ and σ represent the mean and variance of the Gaussian function respectively.

[0145] Through the above steps 401 to 404, the probability values of each device operation index data in the normal state, warning state, and fault state are calculated respectively. The membership function parameters in each state are obtained, and the Gaussian function is used as the mathematical model to accurately describe the distribution characteristics of the device operation index data in different states. By calculating the normal state probability of the device operation index data, the healthy operation degree of the bearing can be quantitatively evaluated. By calculating the warning state probability of the device operation index data, the trend of bearing performance degradation can be detected early, providing a decision-making basis for preventive maintenance. By calculating the fault state probability of the device operation index data, it can be accurately judged whether the bearing fails, and corresponding measures can be taken in time. This method characterizes the state characteristics in the form of probability, not only considering the uncertainty of state judgment, but also realizing the continuous gradual change of state evaluation, improving the accuracy and reliability of state evaluation.

[0146] In step 104 of some embodiments, evidence fusion is performed on the state probabilities of each device operation index data. Based on the D-S evidence theory, the state probabilities of multiple device operation index data are comprehensively calculated, and the target probability assignment function is the final probability assignment result obtained after evidence fusion. Considering the differences in the relative importance among the device operation index data, a confidence coefficient can be introduced to correct the original probability assignment, and the relative weights of film thickness, temperature, noise, and vibration are used to reflect the relative importance of each device operation index data in the evidence synthesis process.

[0147] Please refer to Figure 7 , in some embodiments, step 104 may include, but is not limited to, steps 501 to 503.

[0148] Step 501, combine the normal state probability, warning state probability, and fault state probability of each device operation index data to obtain the original probability assignment function of each device operation index data.

[0149] Step 502, based on the preset importance weight of each device operation index data, correct the original probability assignment function to obtain the intermediate probability assignment function.

[0150] Step 503, perform evidence fusion on the intermediate probability assignment function based on the evidence theory to obtain the target probability assignment function.

[0151] In step 501 of some embodiments, systematic combination processing is required for the state probabilities of the four types of equipment operation index data calculated through membership function parameters. For each type of equipment operation index data, the probability values in three states (normal state, warning state, and failure state) are combined into a one-dimensional vector in a fixed order to form an original probability assignment function. For example, when the normal state probability of a certain equipment operation index data calculated through the membership function is 0.4, the warning state probability is 0.6, and the failure state probability is 0, its original probability assignment function can be expressed as [0.4, 0.6, 0]. The vectorized expression can completely retain the state characteristic information of each equipment operation index data.

[0152] In step 502 of some embodiments, in the bearing state assessment, the importance of each equipment operation index data is different. For example, as a core parameter directly reflecting the bearing lubrication state, the importance weight of the bearing oil film thickness data should be relatively high; as an important parameter reflecting the friction state, the weight of the bearing pad temperature data is the second; while the bearing noise data and the bearing vibration data are used as auxiliary judgment parameters. According to this importance difference, a confidence coefficient can be introduced to correct the original probability assignment.

[0153] Please refer to Figure 8 , in some embodiments, step 502 may include, but is not limited to, steps 601 to 603.

[0154] Step 601, obtain the relative importance weight of each equipment operation index data.

[0155] Step 602, calculate the confidence coefficient of each equipment operation index data according to the relative importance weight.

[0156] Step 603, based on the confidence coefficient, correct the original probability assignment function of the corresponding equipment operation index data to obtain an intermediate probability assignment function.

[0157] In step 601 of some embodiments, the characterization capabilities of different device operation index data for the bearing state vary. The bearing oil film thickness data directly reflects the lubrication state and wear degree of the bearing, and is the most direct index for judging the bearing health state. Therefore, its importance weight is the highest, and the relative importance weight can be set to the maximum, for example, it can be set to 0.5. The bearing pad temperature data can reflect the friction state and heat accumulation of the bearing, and is an important parameter for evaluating the bearing working state. Its importance weight can be set to 0.3. The bearing noise data can reflect the rubbing state of the bearing, but is greatly affected by the environment. Its importance weight can be set to 0.2. The bearing vibration data reflects the dynamic characteristics of the bearing and is also affected by external factors. Its importance weight can also be set to 0.2. The relative importance weights of the device operation index data can be optimized and adjusted based on long-term operation experience, expert opinions, or experimental data.

[0158] In step 602 of some embodiments, the relative importance weights of each device operation index data are converted into specific confidence coefficients to provide a quantitative basis for subsequent probability correction, while maintaining the relative importance relationship between the device operation index data and controlling the confidence coefficients within a reasonable range.

[0159] Please refer to Figure 9 , in some embodiments, step 602 may include, but is not limited to, steps 701 to 703.

[0160] Step 701, obtain the maximum value in the relative importance weights.

[0161] Step 702, calculate the ratio of each relative importance weight to the maximum value to obtain the corresponding normalized weight.

[0162] Step 703, multiply the normalized weight by a preset conversion coefficient to obtain the confidence coefficient corresponding to each device operation index data.

[0163] In step 701 of some embodiments, the four types of device operation index data, namely the bearing pad temperature data, the bearing noise data, the bearing vibration data, and the bearing oil film thickness data, are respectively assigned different importance weights. By comparing these four relative importance weights, the maximum value ω = max(W r ), r = 1, 2, 3, 4, where r is the index number, and W r is the relative importance weight of different device operation index data.

[0164] In step 702 of some embodiments, calculate the ratio of each relative importance weight to the maximum value to obtain the corresponding normalized weight. The calculation is as follows:

[0165]

[0166] Among them, is the normalized weight of the operation index data of different devices, W r is the relative importance weight of the operation index data of different devices, and ω is the maximum value in the relative importance weights.

[0167] In step 703 of some embodiments, according to a preset conversion coefficient, multiply the conversion coefficient by each normalized weight to obtain the final confidence coefficient. For example, 0.9 can be selected as the conversion coefficient, then the confidence coefficient not only ensures that the converted confidence coefficient has a suitable numerical range, but also reserves a certain uncertainty space.

[0168] Through the above steps 701 to 703, taking the maximum value in the relative importance weights as the benchmark, determining the standard scale for weight normalization, calculating the ratio of each relative importance weight to the maximum value to obtain the normalized weight, enabling comparison of each index on the same scale, maintaining the relative importance degree among the indexes, and multiplying the normalized weight by a preset conversion coefficient to obtain the confidence coefficient corresponding to each device operation index data, converting the qualitative importance judgment into a quantitative confidence coefficient, and improving the objectivity and reliability of the state assessment.

[0169] In step 603 of some embodiments, multiply the confidence coefficient of each device operation index data by its original probability assignment function to obtain the corrected state probability, and at the same time calculate the uncertainty. The original probability assignment functions of all device operation index data are converted into intermediate probability assignment functions containing uncertainty, laying a foundation for subsequent evidence fusion.

[0170] Please refer to Figure 10 , in some embodiments, step 603 may include, but is not limited to, steps 801 to 803.

[0171] Step 801, multiply the state probabilities in the original probability assignment function corresponding to different device state levels by the corresponding confidence coefficients respectively to obtain the corrected state probabilities.

[0172] Step 802, based on the confidence coefficient, calculate the state uncertainty of each device operation index data.

[0173] Step 803, combine the corrected state probability and the state uncertainty to obtain the intermediate probability assignment function.

[0174] In step 801 of some embodiments, for the original probability assignment function [M r (H1), M r (H2), M r(H3)], and multiply it with the corresponding confidence coefficient α r for correction, that is, the corrected state probability m r (H N ) = α r M r (H N ), r = 1, 2, 3, 4, and H N represents the equipment state level, and the equipment state level can include three levels: H1, H2, and H3.

[0175] In step 802 of some embodiments, the complement of the confidence coefficient is used to calculate the state uncertainty, that is, the state uncertainty is equal to 1 minus the confidence coefficient, that is, the state uncertainty m r (Θ) = 1 - α r . For example, for the bearing oil film thickness data, its confidence coefficient is 0.9, then the state uncertainty is 1 - 0.9 = 0.1; for the bearing pad temperature data, its confidence coefficient is 0.72, then the state uncertainty is 1 - 0.72 = 0.28. Based on the confidence coefficient, the state uncertainty of each equipment operation index data is calculated, ensuring that the equipment operation index data with higher importance has smaller uncertainty, while the equipment operation index data with lower importance retains a larger uncertainty space.

[0176] In step 803 of some embodiments, the corrected state probability and the state uncertainty are combined, that is, the state probability m r (H N ) and the state uncertainty m r (Θ) are combined to obtain an intermediate probability assignment function, denoted as [m r (H1), m r (H2), m r (H3), m r (Θ), 0].

[0177] Through the above steps 801 to 803, a systematic correction process is performed on the original probability assignment function, and a complete method for generating the intermediate probability assignment function is established. By multiplying the state probabilities at different device state levels in the original probability assignment function by the corresponding confidence coefficients respectively, the probability correction fully considers the importance differences of the device operation index data, ensuring that important indicators have a greater influence weight in the state evaluation. By calculating the state uncertainty of each device operation index data based on the confidence coefficient, a quantitative expression of uncertainty is achieved, enabling important indicators with lower importance to retain a larger uncertain space and improving the reliability of the evaluation. By combining the corrected state probabilities and state uncertainties, a standard intermediate probability assignment function is formed, which not only maintains the distribution characteristics of the state probabilities but also introduces the influence of uncertainties, providing a standardized data basis for subsequent evidence fusion.

[0178] Through the above steps 601 to 603, by obtaining the relative importance weights of each device operation index data, a quantitative standard reflecting the contribution differences of various indicators is established, enabling the importance of different types of indicators to be reasonably reflected. By calculating the confidence coefficients of each device operation index data based on the relative importance weights, the qualitative importance judgment is transformed into a quantitative correction parameter, providing an objective basis for probability correction. By correcting the original probability assignment function of the corresponding device operation index data based on the confidence coefficient, an intermediate probability assignment function is obtained, which not only maintains the characteristics of the original probability distribution but also introduces the influence of indicator importance, realizing the differential processing of data credibility and improving the accuracy and reliability of multi-source information fusion.

[0179] In step 503 of some embodiments, the classical fusion formula of D-S evidence theory is used to synthesize the intermediate probability assignment function. When fusing four device operation index data, a pairwise synthesis method can be adopted. For example, first fuse the bearing oil film thickness data and the bearing pad temperature data to obtain an intermediate result, then fuse this result with the bearing noise data, and finally fuse it with the bearing vibration data to finally obtain the target probability assignment function. Among them, the calculation of the synthesis of two pairwise device operation index data (i.e., m r (H N ) and m r+1 (H N )) is as follows:

[0180]

[0181]

[0182] Among them, m n (H N ) represents the state probability in the target probability assignment function, mn $(\Theta)$ represents the state uncertainty in the target probability assignment function, $m$ r (H N ) is the state probability of the intermediate probability assignment function, $m$ r (\Theta)$ is the state uncertainty of the intermediate probability assignment function, $r = 1, 2, 3, 4$, respectively representing the operating index data of four devices.

[0183] After fusing the operating index data of the four devices, the final target probability assignment function $[m$ n (H1), $m$ n (H2), $m$ n (H3), $m$ n (\Theta), 0]$ is obtained. This target probability assignment function comprehensively considers the state probabilities and importance differences of the operating index data of each device, and can comprehensively reflect the overall operating state of the bearing.

[0184] Through the above steps 501 to 503, by combining the normal state probability, early warning state probability, and fault state probability of the operating index data of each device, the original probability assignment function is obtained, and a complete state probability framework is established, so that the state information of each index is fully retained. By correcting the original probability assignment function based on the preset importance weights of the operating index data of each device, the intermediate probability assignment function is obtained, which realizes the reasonable consideration of the importance of different indexes and enables the core indexes to play a greater role in the evaluation. By performing evidence fusion on the intermediate probability assignment function based on the evidence theory, the target probability assignment function is obtained, which effectively integrates multi-source information, overcomes the limitations of single-index evaluation, and improves the comprehensiveness and accuracy of bearing state evaluation.

[0185] In step 105 of some embodiments, the target probability assignment function represents the final probability assignment result obtained through evidence fusion, and includes the probability values of the thrust bearing in the normal state, early warning state, and fault state. The bearing fault state refers to the current operating state of the bearing determined according to the magnitude of the probability value, and can be divided into the normal state, early warning state, and fault state.

[0186] Please refer to Figure 11 , in some embodiments, step 105 may include, but is not limited to, steps 901 to 904.

[0187] Step 901, obtain the maximum value of the state probability in the target probability assignment function and the device state level corresponding to the maximum value of the state probability.

[0188] Step 902, calculate the difference between the maximum value of the state probability and the maximum value of other state probabilities.

[0189] Step 903, determining whether the difference is greater than a first preset threshold, whether the state uncertainty is less than a second preset threshold, and whether the maximum state probability is greater than the state uncertainty.

[0190] Step 904: When the above judgment conditions are met, the equipment state level corresponding to the maximum state probability is determined as a bearing fault state of the reactor coolant pump.

[0191] In step 901 of some embodiments, the state probabilities in the target probability allocation function are compared, that is, [m n (H1),m n (H2),m n (H3),m n (Θ),0] n (H1),m n (H2),m n (H3) value, find the maximum value m n (H Nmax ). For example, when m in the target probability distribution function n (H1),m n (H2),m n When the value of (H3) is [0.7, 0.2, 0.1], the maximum state probability is 0.7, and the corresponding device state level is H1, that is, the normal state.

[0192] In step 902 of some embodiments, in the target probability distribution function, after removing the maximum state probability, the maximum value m is found from the remaining state probabilities. n (H Nmax2 ). For example, when the target probability distribution function is m n (H1),m n (H2),m n When the value of (H3) is [0.7, 0.2, 0.1], after removing the maximum state probability of 0.7, the maximum value of the remaining state probabilities is 0.2. Then calculate the difference between these two probability values m n (H Nmax )-m n (H Nmax2 ). This difference reflects the distinction between the most likely device status level and the second most likely device status level. The larger the difference, the higher the certainty of the device status level judgment.

[0193] In step 903 of some embodiments, a triple judgment is performed on the calculated difference, state uncertainty, and state probability maximum value. First, the difference m is judged. n (H Nmax )-m n (H Nmax2Is it greater than the first preset threshold ε0, indicating the significance requirement for distinguishing the device state level. Secondly, judge the state uncertainty m n (Θ) is less than the second preset threshold ε1, indicating the confidence requirement for the judgment result. Finally, judge whether the maximum state probability m n (H Nmax ) is greater than the state uncertainty m n (Θ), which characterizes the effectiveness of the state judgment. The triple judgment on the calculated difference, state uncertainty, and maximum state probability is as follows:

[0194]

[0195] Among them, m n (H Nmxx ) is the maximum state probability, and m n (H Nmax2 ) is the maximum state probability excluding m n (H Nmax ), and m n (Θ) is the state uncertainty value. In some embodiments, the value of ε0 is set to 0.001 and ε1 is 0.1.

[0196] In step 904 of some embodiments, when all the triple judgment conditions are satisfied, it can be considered that the evaluation result has sufficient reliability. At this time, the device state level H corresponding to the maximum state probability Nmax is determined as the bearing fault state of the reactor coolant pump, and at the same time, its confidence m n (H Nmax ) is output. For example, if the final judgment result is "the thrust bearing is in the H1 state with a confidence of 0.7", this indicates that the thrust bearing is currently in a normal operating state, and this judgment has a confidence of 0.7.

[0197] Through the above steps 901 to 904, by obtaining the maximum state probability in the target probability assignment function and the device state level corresponding to the maximum state probability, the most likely state of the bearing can be accurately found. By calculating the difference between the maximum state probability and the maximum values of other state probabilities, a quantitative index for state discrimination is established, avoiding the ambiguity of state judgment. By judging whether the difference is greater than the first preset threshold, whether the state uncertainty is less than the second preset threshold, and whether the maximum state probability is greater than the state uncertainty, a triple judgment mechanism is constructed to ensure the reliability of state evaluation. By determining the device state level corresponding to the maximum state probability as the bearing fault state when the judgment conditions are met, the final output of the state judgment is realized. Through the comprehensive judgment of multiple constraint conditions, the accuracy of the bearing fault detection of the reactor coolant pump is improved.

[0198] Through the above steps 101 to 105, multiple device operation index data of the thrust bearing can be obtained, and the operation state parameters of the thrust bearing can be comprehensively monitored. Then, by performing dimensionless processing on each device operation index data, a corresponding dimensionless index matrix can be obtained, which can eliminate the influence of dimensions and orders of magnitude between different device operation index data and make the device operation index data comparable. Next, by performing fuzzification processing on each dimensionless index matrix, the state probability of each device operation index data in different device state levels can be obtained, which can convert the deterministic device operation index data into a fuzzy probability form to better characterize the uncertainty of the device state. Subsequently, by performing evidence fusion on the state probabilities of each device operation index data, a target probability assignment function can be obtained, which can fully consider the relative importance of each device operation index data and realize the effective fusion of multi-source information. Finally, by determining the bearing fault state of the reactor coolant pump according to the target probability assignment function, the fault mode of the thrust bearing can be accurately identified. Compared with the prior art that only relies on temperature sensors for state monitoring, the method provided by the embodiment of the present application realizes the comprehensive monitoring and accurate diagnosis of the thrust bearing state by fusing multiple device operation index data, and improves the accuracy of detecting the bearing fault of the reactor coolant pump.

[0199] Please refer to Figure 12 , in some embodiments, the bearing fault detection method of the reactor coolant pump may further include, but is not limited to, steps 1001 to 1003.

[0200] Step 1001, calculate the health degree corresponding to each device operation index data according to a preset health degree membership function.

[0201] Step 1002, calculate the overall health degree of the thrust bearing according to the health degree of each device operation index data and a preset health degree weight coefficient.

[0202] Step 1003, when the health degree of any device operation index data is less than a preset health degree threshold, set the overall health degree to the minimum health degree less than the preset health degree threshold.

[0203] In step 1001 of some embodiments, the health degree membership function may adopt a linear function, and this function shows a continuous change from 0 to 100 with the change of the device operation index data. For the four types of device operation index data, namely bearing pad temperature data, bearing noise data, bearing vibration data, or bearing oil film thickness data, alarm values and early warning values are respectively set as the segmentation points of the health degree membership function, such as Figure 13As shown. When the device operation index data is within the normal operation range, the health level is maintained at a relatively high level; when the device operation index data exceeds the warning value, the health level begins to decline; when the device operation index data reaches the alarm value, the health level drops to the lowest. According to the preset health membership function, calculate the health level corresponding to each device operation index data, which can uniformly convert different types of device operation index data into standard health scores, realizing the quantitative evaluation of the bearing state.

[0204] In step 1002 of some embodiments, corresponding health weight coefficients are set for various device operation index data, and the sum of all weight coefficients is 100%. The setting of the weight coefficients is initially determined by the method of expert scoring and can be adjusted and optimized according to the actual usage later. When calculating the overall health level of the thrust bearing, multiply the health level of each device operation index data by its corresponding health weight coefficient, and then add up all the products to obtain the overall health level reflecting the overall state of the bearing, which not only considers the influence degree of each index but also realizes the effective integration of multi-source information.

[0205] In step 1003 of some embodiments, based on the principle of the barrel effect, when the health level of a certain device operation index data is lower than the preset health threshold (such as 60 points), regardless of the health levels of other index data, the overall health level of the thrust bearing is set to this lowest health value. This mechanism reflects the short-board effect, that is, the overall state of the bearing depends on the worst monitoring parameter. In this way, it can effectively prevent some abnormal indicators from being covered up by other normal index data, ensuring the conservativeness and reliability of the state evaluation and providing an effective guarantee for timely discovering and handling potential faults.

[0206] Through the above steps 1001 to 1003, by calculating the health level corresponding to each device operation index data according to the preset health membership function, different types of monitoring data are uniformly converted into standardized health scores, making various indicators comparable and consistent. By calculating the overall health level of the thrust bearing according to the health level of each device operation index data and the preset health weight coefficient, the comprehensive evaluation of multi-source information is realized, fully considering the difference in the influence degree of each index. By setting the overall health level to the minimum health level less than the preset health threshold when the health level of any device operation index data is less than the preset health threshold, the principle of the barrel effect is introduced, ensuring the conservativeness of the evaluation result and improving the accuracy of the thrust bearing state evaluation.

[0207] Please refer to Figure 14 , in some embodiments, the multiple device operation index data further includes the bearing pad wear depth data, and the bearing fault detection method of the reactor coolant pump may further include, but is not limited to, steps 1101 to 1105.

[0208] Step 1101: Interpolate and complete the bearing pad wear depth data to obtain a wear depth sequence with equally spaced distribution.

[0209] Step 1102: Based on the grey prediction model, predict the trend of the wear depth sequence to obtain the predicted wear amount.

[0210] Step 1103: Calculate the wear penalty amount according to the preset operation cycle and the preset transient wear depth.

[0211] Step 1104: Add the predicted wear amount and the wear penalty amount to obtain the corrected predicted wear amount.

[0212] Step 1105: Predict the remaining service life based on the corrected predicted wear amount.

[0213] In step 1101 of some embodiments, during the actual monitoring process, the bearing pad wear depth data obtained by the ultrasonic measurement sensor is often not equally spaced. To perform subsequent trend prediction, these unequally spaced data points need to be completed into an equally spaced sequence through interpolation. For example, the method of cubic polynomial fitting can be used. For each group of adjacent sample data points, a cubic curve is used for fitting. To ensure the uniqueness and continuity of the fitting result, the first-order and second-order derivatives at the sample data points need to be constrained to ensure the data continuity between the data points and at both ends of the interval.

[0214] In step 1102 of some embodiments, based on the grey prediction model (i.e., GM(1,1) model), the trend of the wear depth sequence is predicted. The prediction process is as follows: Generate a new data sequence with obvious trend by accumulating the wear depth sequence, establish a model for prediction according to the growth trend of the new data sequence, and then perform reverse calculation by subtraction to restore the original data sequence, thereby obtaining the predicted wear amount.

[0215] Exemplarily, the wear depth sequence is represented as x (0) =(x (0) (1),x (0) (2),…,x (0) (n)), where n is the number of sequence data. Accumulate x (0) to weaken the volatility and randomness of the random sequence, and obtain a new sequence x (1) =(x (1) (1),x (1) (2),…,x (1) (n)), where:

[0216]

[0217] Then, generate x (1)The adjacent mean equal-weight sequence z (1) ={z (1) (2), z (1) (3), …, z (1) (k)}, k = 2, 3, …, n, where: z (1) (k)=0.5x (1) (k - 1)+0.5x (1) (k), k = 2, 3, …, n; then, define the grey derivative of x (1) as: d(k)=x (0) (k)=x (1) (k)-x (1) (k - 1), then the differential equation of the grey prediction model GM(1, 1) is:

[0218] d(k)+az (1) (k)=b

[0219] It can also be expressed as:

[0220] x (0) (k)+az (1) (k)=b

[0221] where, x (0) (k) is the grey derivative, a is the development coefficient, z (1) (k) is called the whitenized background value, and b is called the grey action quantity.

[0222] It can be understood that substituting k = 2, 3, …, n into the differential equation of the grey prediction model GM(1, 1) respectively, we can obtain the system of equations:

[0223]

[0224] Let Y=(x (0) (2), x (0) (3), … x (0) (n)) T , u=(a, b) T .

[0225]

[0226] Y is called the data vector, B is called the data matrix, and μ is called the parameter vector. Then the GM(1, 1) model can be expressed as Y = Bμ. By the least square method, we can obtain: Making appropriate transformation to x (0) (k), according to the Newton-Leibniz formula, we can get:

[0227]

[0228] Continue to perform operations on z (1)(k) By performing a transformation, we can obtain:

[0229]

[0230] It can be understood that for the grey differential equation, if we consider x (0) at time instants k = 2, 3, …, n as a continuous variable t, the sequence x (1) can be regarded as a function of time t, denoted as x (1) = x (1) (t). And by making the grey derivative x (1) (k) correspond to the derivative and the background value z (1) (k) correspond to x (1) (t), we can obtain the whitened form of the GM(1, 1) grey differential equation:

[0231]

[0232] Denote the matrix formed by a and b as the grey parameter (a, b) T , the effective interval of a is (-2, 2). By solving for the parameters a and b, we can solve for x (1) (t), and then obtain the predicted value of x (0) . For the grey parameters that have been solved , substituting them into the whitened form equation and solving, we can obtain:

[0233]

[0234] By performing cumulative subtraction reduction on the obtained results, we can obtain the predicted value, that is, the predicted wear amount

[0235]

[0236] In step 1103 of some embodiments, during the actual operation of the thrust bearing, transient conditions such as power loss and coasting occur, which will lead to increased instantaneous wear. Assume that a power loss and coasting condition occurs once in each operating cycle T, and the depth of transient wear generated each time is h. The wear value brought by this transient event is evenly distributed during the prediction process, and the wear penalty W = h / T is calculated. This processing method takes into account the influence of special conditions on the bearing life.

[0237] In step 1104 of some embodiments, the predicted wear amount H(predict) obtained by predicting through the GM(1,1) model is added to the calculated wear penalty amount W to obtain the corrected predicted wear amount H(t + 1) = H(predict) + W. This correction takes into account the combined effects of normal wear and transient wear, making the prediction result closer to the actual situation. By introducing the correction mechanism of the penalty amount, the accuracy of wear prediction is improved.

[0238] In step 1105 of some embodiments, a life prediction model based on a neural network can be established. The input layer includes parameters such as the corrected predicted wear amount, the current wear rate, and the cumulative operating time; the hidden layer determines the network structure through training; the output layer is the remaining service life. The network is trained through a large amount of historical data to enable it to accurately predict the remaining life.

[0239] Please refer to Figure 15 , in some embodiments, step 1105 may include, but is not limited to, steps 1201 to 1202.

[0240] Step 1201, when the corrected predicted wear amount reaches the preset wear threshold, set the remaining service life to zero.

[0241] Step 1202, when the corrected predicted wear amount does not reach the preset wear threshold, calculate the difference between the preset wear threshold and the corrected predicted wear amount, and predict the remaining service life based on the difference and the preset wear rate.

[0242] In step 1201 of some embodiments, for the wear-resistant layer of the bearing pad, the preset wear threshold represents the maximum wear depth it can withstand. When the corrected predicted wear amount H(t + 1) reaches the preset wear threshold F, it indicates that the bearing can no longer operate safely. At this time, stop the machine for maintenance immediately, so the remaining service life is directly determined to be zero. This judgment mechanism embodies the principle of preventive maintenance, ensuring that the bearing is replaced in time before reaching the limit state and avoiding serious equipment accidents.

[0243] In step 1202 of some embodiments, when the corrected predicted wear amount H(t + 1) is less than the preset wear threshold F, it indicates that the bearing can still operate. At this time, first calculate the available wear allowance, that is, the difference between the wear threshold and the current wear amount [F - H(t + 1)]. Then consider the wear development speed V. By dividing the available wear allowance by the wear rate, the time required for the bearing to reach the wear limit can be obtained. Finally, subtract 1 from the calculation result to get the final remaining service life time = [F - H(t + 1)] / V - 1. This calculation method not only considers the current wear state but also the dynamic characteristics of wear development, and can provide an accurate time estimate for maintenance decisions.

[0244] Through the above steps 1201 to 1202, when the corrected predicted wear amount reaches the preset wear threshold, setting the remaining service life to zero realizes the accurate determination of the extreme state and avoids the risk of the bearing operating with defects. When the corrected predicted wear amount does not reach the preset wear threshold, calculating the difference between the preset wear threshold and the corrected predicted wear amount, and predicting the remaining service life based on the difference and the preset wear rate, a dynamic prediction mechanism based on the actual wear state is established, improving the accuracy of the remaining life prediction.

[0245] Through the above steps 1101 to 1105, by performing interpolation and completion processing on the bearing pad wear depth data to obtain an equally spaced wear depth sequence, the problem of unequal spacing distribution of the original data is solved, providing a standardized data basis for subsequent prediction. By performing trend prediction on the wear depth sequence based on the grey prediction model to obtain the predicted wear amount, reliable prediction under small sample data is realized. By calculating the wear penalty amount according to the preset operating cycle and the preset transient wear depth, the influence of special working conditions on wear is considered. By adding the predicted wear amount and the wear penalty amount to obtain the corrected predicted wear amount, the accuracy of the prediction result is improved. By predicting the remaining service life based on the corrected predicted wear amount, a complete life prediction chain is established. The method, through multi-step processing, considers both normal wear and the influence of transient working conditions, improving the accuracy and reliability of the remaining service life prediction.

[0246] Please refer to Figure 16 , the embodiment of the present application further provides a bearing fault detection device 1600 for a reactor coolant pump, which can implement the above-mentioned bearing fault detection method for the reactor coolant pump, including:

[0247] An acquisition module 1601, configured to acquire a plurality of device operation index data of the thrust bearing;

[0248] A dimensionless processing module 1602, configured to perform dimensionless processing on each device operation index data to obtain a corresponding dimensionless index matrix;

[0249] A fuzzification processing module 1603, configured to perform fuzzification processing on each dimensionless index matrix to obtain the state probability of each device operation index data in different device state levels;

[0250] An evidence fusion module 1604, configured to perform evidence fusion on the state probabilities of each device operation index data to obtain a target probability assignment function;

[0251] A determination module 1605, configured to determine the bearing fault state of the reactor coolant pump according to the target probability assignment function.

[0252] The bearing fault detection method of a reactor coolant pump according to an embodiment of the present application includes: obtaining a plurality of device operation index data of a thrust bearing; performing dimensionless processing on each device operation index data to obtain a corresponding dimensionless index matrix; performing fuzzification processing on each dimensionless index matrix to obtain the state probability of each device operation index data in different device state levels; performing evidence fusion on the state probabilities of each device operation index data to obtain a target probability assignment function; and determining the bearing fault state of the reactor coolant pump according to the target probability assignment function.

[0253] By obtaining a plurality of device operation index data of the thrust bearing in the present application, the operation state parameters of the thrust bearing can be comprehensively monitored. Then, by performing dimensionless processing on each device operation index data to obtain a corresponding dimensionless index matrix, the influence of the dimension and order of magnitude between different device operation index data can be eliminated, making each device operation index data comparable. Next, by performing fuzzification processing on each dimensionless index matrix to obtain the state probability of each device operation index data in different device state levels, the deterministic device operation index data can be transformed into a fuzzy probability form to better characterize the uncertainty of the device state. Subsequently, by performing evidence fusion on the state probabilities of each device operation index data to obtain a target probability assignment function, the relative importance of each device operation index data can be fully considered to realize the effective fusion of multi-source information. Finally, by determining the bearing fault state of the reactor coolant pump according to the target probability assignment function, the fault mode of the thrust bearing can be accurately identified. Compared with the prior art that only relies on temperature sensors for state monitoring, the method provided by the embodiment of the present application realizes the comprehensive monitoring and accurate diagnosis of the thrust bearing state by fusing a plurality of device operation index data, improving the accuracy of bearing fault detection of the reactor coolant pump.

[0254] Refer to Figure 17 , Figure 17 schematically shows the hardware structure of an electronic device in another embodiment. The electronic device includes:

[0255] A processor 1701, which can be implemented in ways such as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided by the embodiments of the present application;

[0256] The memory 1702 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 1702 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 1702, and the processor 1701 is used to call and execute the bearing fault detection method of the reactor coolant pump in the embodiments of this application.

[0257] The input / output interface 1703 is used to implement information input and output.

[0258] The communication interface 1704 is used to implement communication interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, Wi-Fi, Bluetooth, etc.).

[0259] The bus 1705 transmits information between the various components of the device (such as the processor 1701, the memory 1702, the input / output interface 1703, and the communication interface 1704).

[0260] Among them, the processor 1701, the memory 1702, the input / output interface 1703, and the communication interface 1704 achieve communication connections with each other inside the device through the bus 1705.

[0261] The embodiments of this application also provide a computer program product, which includes a computer program. The processor of the computer device reads and executes this computer program, so that the computer device executes to implement the above-mentioned bearing fault detection method of the reactor coolant pump.

[0262] Terms such as "first", "second", "third", "fourth", etc. (if any) in the specification of this disclosure and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances, so that the embodiments of this disclosure described here can be implemented in an order other than those illustrated or described here. In addition, the terms "include" and "contain" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0263] It should be understood that in this disclosure, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that three relationships can exist. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Here, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or its similar expressions refer to any combination of these items, including any combination of single item (one) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0264] It should be understood that in the description of the embodiments of this application, the meaning of a plurality (or multiple items) is more than two. Understandings such as greater than, less than, exceeding, etc. do not include the present number, and understandings such as above, below, within, etc. include the present number.

[0265] In several embodiments provided in this disclosure, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces, and the indirect coupling or communication connection of devices or units can be in electrical, mechanical, or other forms.

[0266] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0267] In addition, the functional units in each embodiment of this disclosure can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0268] When an integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present disclosure, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present disclosure. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0269] It should also be understood that the various embodiments provided in the embodiments of the present application can be combined arbitrarily to achieve different technical effects.

[0270] The above is a specific description of the embodiments of the present disclosure, but the present disclosure is not limited to the above embodiments. Those skilled in the art can also make various equivalent deformations or substitutions without departing from the spirit of the present disclosure, and these equivalent deformations or substitutions are all included within the scope defined by the claims of the present disclosure.

Claims

1. A method for detecting bearing faults of a reactor coolant pump, characterized in that, The reactor coolant pump includes a thrust bearing, and the method includes: Obtaining a plurality of device operation index data of the thrust bearing; Performing dimensionless processing on each of the device operation index data to obtain a corresponding dimensionless index matrix; Performing fuzzy processing on each of the dimensionless index matrices to obtain the state probabilities of each of the device operation index data in different device state levels; Performing evidence fusion on the state probabilities of each of the device operation index data to obtain a target probability assignment function; Determining the bearing fault state of the reactor coolant pump according to the target probability assignment function.

2. The bearing fault detection method for a reactor coolant pump according to claim 1, wherein The thrust bearing is provided with a temperature sensor, an ultrasonic measurement sensor, a noise measurement sensor, and a vibration acceleration sensor. The plurality of device operation index data includes bearing pad temperature data, bearing oil film thickness data, bearing noise data, and bearing vibration data. The obtaining of the plurality of device operation index data of the thrust bearing includes: Collecting the temperature of the thrust pads of the thrust bearing through the temperature sensor to obtain the bearing pad temperature data; Performing ultrasonic measurement on the thrust pads of the thrust bearing through the ultrasonic measurement sensor to obtain the bearing oil film thickness data; Performing noise measurement on the outer ring of the bearing housing of the thrust bearing through the noise measurement sensor to obtain the bearing noise data; Performing acceleration measurement on the outer ring of the bearing housing of the thrust bearing through the vibration acceleration sensor to obtain the bearing vibration data.

3. The bearing fault detection method of the reactor coolant pump according to claim 2, characterized in that The dimensionless index matrix includes a temperature index matrix, an oil film thickness index matrix, a bearing noise index matrix, and a bearing vibration index matrix. The performing of dimensionless processing on each of the device operation index data to obtain a corresponding dimensionless index matrix includes: Calculating the temperature index matrix according to a preset temperature initial value, a preset temperature threshold, and the bearing pad temperature data; Calculating the oil film thickness index matrix according to a preset oil film thickness initial value, a preset oil film thickness threshold, and the bearing oil film thickness data; Calculating the bearing noise index matrix according to a preset bearing noise initial value, a preset bearing noise threshold, and the bearing noise data; Calculating the bearing vibration index matrix according to a preset bearing vibration initial value, a preset bearing vibration threshold, and the bearing vibration data.

4. The bearing fault detection method for a reactor coolant pump according to claim 1, wherein The performing of fuzzy processing on each of the dimensionless index matrices to obtain the state probabilities of each of the device operation index data in different device state levels includes: Obtaining the membership function parameters of each of the device operation index data corresponding to the normal state, the warning state, and the fault state; Calculating the normal state probability of each of the device operation index data according to the membership function parameters in the normal state; Calculating the warning state probability of each of the device operation index data according to the membership function parameters in the warning state; Calculating the fault state probability of each of the device operation index data according to the membership function parameters in the fault state.

5. The bearing fault detection method for a reactor coolant pump according to claim 4, characterized in that, Performing evidence fusion on the state probabilities of the operation index data of each device to obtain a target probability assignment function, including: Combining the normal state probability, the early warning state probability, and the fault state probability of the operation index data of each device to obtain an original probability assignment function for the operation index data of each device; Based on the preset importance weight of the operation index data of each device, correcting the original probability assignment function to obtain an intermediate probability assignment function; Performing evidence fusion on the intermediate probability assignment function based on the evidence theory to obtain the target probability assignment function.

6. The bearing fault detection method for the reactor coolant pump according to claim 5, characterized in that, The step of correcting the original probability assignment function based on the preset importance weight of the operation index data of each device to obtain the intermediate probability assignment function includes: Obtaining the relative importance weight of the operation index data of each device; Calculating the confidence coefficient of the operation index data of each device according to the relative importance weight; Based on the confidence coefficient, correcting the original probability assignment function of the corresponding operation index data of the device to obtain the intermediate probability assignment function.

7. The bearing fault detection method for a reactor coolant pump according to claim 6, characterized in that, The step of calculating the confidence coefficient of the operation index data of each device according to the relative importance weight includes: Obtaining the maximum value in the relative importance weights; Calculating the ratio of each relative importance weight to the maximum value to obtain the corresponding normalized weight; Multiplying the normalized weight by a preset conversion coefficient to obtain the confidence coefficient corresponding to the operation index data of each device.

8. The bearing fault detection method for a reactor coolant pump according to claim 6, characterized in that, The step of correcting the original probability assignment function of the corresponding operation index data of the device based on the confidence coefficient to obtain the intermediate probability assignment function includes: Multiplying the state probabilities in different device state levels in the original probability assignment function by the corresponding confidence coefficients to obtain the corrected state probabilities; Calculating the state uncertainty of the operation index data of each device based on the confidence coefficient; Combining the corrected state probabilities and the state uncertainty to obtain the intermediate probability assignment function.

9. The bearing fault detection method for a reactor coolant pump according to claim 1, characterized in that, Determining the bearing fault state of the reactor coolant pump according to the target probability assignment function includes: Obtaining the maximum value of the state probability in the target probability assignment function and the device state level corresponding to the maximum value of the state probability; Calculating the difference between the maximum value of the state probability and the maximum value of other state probabilities; Judging whether the difference is greater than a first preset threshold, whether the state uncertainty is less than a second preset threshold, and whether the maximum value of the state probability is greater than the state uncertainty; When the above judgment conditions are met, determining the device state level corresponding to the maximum value of the state probability as the bearing fault state of the reactor coolant pump.

10. The bearing fault detection method for a reactor coolant pump according to claim 1, characterized in that, The method further includes: Calculating the health degree corresponding to the operation index data of each device according to a preset health degree membership function; Calculating the overall health degree of the thrust bearing according to the health degree of the operation index data of each device and a preset health degree weight coefficient. When the health degree of any one of the device operation index data is less than the preset health degree threshold, set the overall health degree to the minimum health degree less than the preset health degree threshold.

11. The bearing fault detection method for a reactor coolant pump according to claim 1, characterized in that, The multiple device operation index data further includes bearing pad wear depth data, and the method further includes: Performing interpolation and completion processing on the bearing pad wear depth data to obtain a wear depth sequence with equally spaced distribution; Based on the grey prediction model, performing trend prediction on the wear depth sequence to obtain a predicted wear amount; Calculating a wear penalty amount according to a preset operation cycle and a preset transient wear depth; Adding the predicted wear amount and the wear penalty amount to obtain a corrected predicted wear amount; Predicting the remaining service life based on the corrected predicted wear amount.

12. The bearing fault detection method for a reactor coolant pump according to claim 11, characterized in that, The predicting the remaining service life based on the corrected predicted wear amount includes: When the corrected predicted wear amount reaches the preset wear threshold, setting the remaining service life to zero; When the corrected predicted wear amount does not reach the preset wear threshold, calculating the difference between the preset wear threshold and the corrected predicted wear amount, and predicting the remaining service life according to the difference and a preset wear rate.

13. A bearing fault detection device for a reactor coolant pump, characterized in that, Including: An acquisition module, configured to acquire multiple device operation index data of a thrust bearing; A dimensionless processing module, configured to perform dimensionless processing on each of the device operation index data to obtain a corresponding dimensionless index matrix; A fuzzification processing module, configured to perform fuzzification processing on each of the dimensionless index matrices to obtain the state probabilities of each of the device operation index data in different device state levels; An evidence fusion module, configured to perform evidence fusion on the state probabilities of each of the device operation index data to obtain a target probability assignment function; A determination module, configured to determine the bearing fault state of a reactor coolant pump according to the target probability assignment function.

14. An electronic device, characterized in that, Including: A memory and a processor, where the memory stores a computer program, and the processor implements the bearing fault detection method of the reactor coolant pump according to any one of claims 1 to 12 when executing the computer program.

15. A computer-readable storage medium, characterized in that, The storage medium stores a program, and the program is executed by a processor to implement the bearing fault detection method of the reactor coolant pump according to any one of claims 1 to 12.

Images