A real-time detection system for a shield tail sealing grease cavity

Abstract

The application discloses a shield tail sealing grease cavity real-time detection system, and belongs to the technical field of shield construction monitoring, which comprises a multi-physical field monitoring module and an intelligent early warning analysis module, wherein the multi-physical field monitoring module comprises a pressure difference monitoring unit and a moisture monitoring unit, and the intelligent early warning analysis module comprises an edge data processing unit and a risk assessment unit; the pressure difference monitoring unit collects grease pressure data in the grease cavity and external environment water pressure data, establishes a pressure difference dynamic model and generates a pressure difference space-time distribution atlas, and marks a high-risk leakage area through abnormal identification and pressure calculation; the moisture monitoring unit generates a moisture intrusion path atlas in combination with a constructed moisture diffusion model and a three-dimensional coordinate system, and marks a potential leakage area; the edge data processing unit extracts dynamic characteristics, the risk assessment unit outputs a leakage risk grade through Bayesian network fusion analysis, and triggers a secondary early warning when triple abnormalities are detected, so that real-time and accurate monitoring and early warning of grease cavity leakage are realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of shield construction monitoring, in particular to a real-time detection system for a shield tail sealing grease cavity. BACKGROUND

[0002] In the process of shield construction, the sealing performance of the shield tail sealing grease cavity is crucial, which is directly related to whether the shield machine can operate safely and efficiently. In the prior art, there are many problems to be solved. For example, single-point pressure detection cannot accurately locate the leakage position, when the shield tail sealing appears problems, it is difficult to quickly determine the specific leakage point, thereby delaying the maintenance opportunity; the water detection also has the positioning difficulty, which is extremely unfavorable for timely discovery and processing of the sealing failure problem caused by water intrusion. At the same time, the multi-parameter independent early warning mode is easy to produce false alarm, since each parameter is judged by threshold value, in the complex shield construction environment, various factors interfere with each other, so that the threshold value judgment according to a single parameter is easy to produce an error alarm signal, which reduces the trust degree of the construction personnel to the early warning information, and affects the timely processing of the actual fault. In addition, when a large amount of monitoring data is transmitted to the cloud for processing, there is a problem of high delay, the shield construction is a dynamic and fast advancing process, and the real-time requirement is very high, and the delay of the cloud processing cannot meet the real-time monitoring requirement of the shield tail sealing state, it is difficult to provide accurate basis for the construction decision in time, which may cause serious engineering accidents. SUMMARY

[0003] In view of the deficiencies of the prior art, the application provides a real-time detection system for a shield tail sealing grease cavity, a multi-physical field monitoring module and an intelligent early warning analysis module, the multi-physical field monitoring module includes a pressure difference monitoring unit and a water monitoring unit, and the intelligent early warning analysis module includes an edge data processing unit and a risk assessment unit; the pressure difference monitoring unit collects grease pressure data in the grease cavity and external environment water pressure data, establishes a pressure difference dynamic model and generates a pressure difference space-time distribution map, and marks a high-risk leakage area through abnormal identification and pressure calculation; the water monitoring unit generates a water intrusion path map in combination with a constructed water diffusion model and a three-dimensional coordinate system, and marks a potential leakage area; the edge data processing unit extracts dynamic characteristics, and the risk assessment unit outputs a leakage risk grade through Bayesian network fusion analysis, triggers a secondary early warning when three abnormalities are detected, and realizes real-time and accurate monitoring and early warning of the grease cavity leakage.

[0004] To achieve the above purpose, the application provides the following technical scheme:

[0005] A real-time detection system for a shield tail sealing grease cavity, comprising: a multi-physical field monitoring module and an intelligent early warning analysis module; the multi-physical field monitoring module comprises a pressure difference monitoring unit and a water monitoring unit; the intelligent early warning analysis module comprises an edge data processing unit and a risk assessment unit;

[0006] The pressure difference monitoring unit collects oil pressure data in the oil chamber and external environment water pressure data, establishes a pressure difference dynamic model and generates a pressure difference spatiotemporal distribution map, then identifies a pressure imbalance point with the circumferential center of the oil chamber as the reference, calculates the pressure difference between the pressure imbalance point and a circumferentially symmetrical position point, generates a pressure difference abnormal distribution map, and locates abnormal deformation in combination with the principle of fluid mechanics, marks a high-risk leakage area based on a preset threshold, and outputs high-risk leakage area data and corresponding pressure monitoring signals;

[0007] The moisture monitoring unit collects moisture data at the bottom of the oil chamber, constructs a moisture diffusion model using a support vector machine algorithm and outputs a moisture three-dimensional distribution map, then establishes a three-dimensional coordinate system with the center of the bottom of the oil chamber as the origin, calculates the diffusion rate gradient of each monitoring point and generates a moisture intrusion path map, marks a potential leakage area based on a preset moisture rising rate threshold according to the moisture intrusion path map, and outputs potential leakage area data and corresponding moisture monitoring signals;

[0008] According to the high-risk leakage area data, pressure monitoring signals, potential leakage area data and moisture monitoring signals, the edge data processing unit extracts dynamic features using an isolation forest algorithm; the dynamic features include pressure difference change rate and moisture rising rate;

[0009] The risk assessment unit outputs a leakage risk level based on the extracted dynamic features and model parameters trained in combination with historical fault data through Bayesian network fusion analysis, and triggers a secondary warning when detecting three pieces of evidence of sudden viscosity reduction of the oil, sudden moisture rise and pressure difference overrun in the fusion analysis process.

[0010] Specifically, the pressure difference monitoring unit includes two types of sensors;

[0011] The first type of sensor is an array of at least two circles of pressure sensors uniformly arranged on the inner wall of the oil chamber, one circle near the tail brush being the first pressure array for collecting oil pressure data in the oil chamber, and the other circle being the second pressure array for assisting in verifying pressure distribution uniformity;

[0012] The second type of sensor is a group of external water pressure sensors buried on the outside of the tail shell corresponding to the position of the oil chamber for collecting external environment water pressure data;

[0013] The first type of sensor and the second type of sensor are both connected with a pressure data acquisition instrument; the pressure data acquisition instrument is used to convert the pressure signals collected by the sensors into digital pressure signals and complete the time synchronization upload of the internal chamber pressure and external water pressure data.

[0014] Specifically, the process of constructing a pressure difference dynamic model and generating a pressure difference spatiotemporal distribution map is as follows:

[0015] The pressure sensor array of the pressure difference monitoring unit and the external water pressure sensor group synchronously collect data, respectively generate the grease pressure data in the grease cavity and the external environment water pressure data, and transmit to the pressure data acquisition instrument;

[0016] The pressure data acquisition instrument calibrates the time stamp of the grease pressure data in the grease cavity and the external environment water pressure data, and converts the digital pressure signal after transmission to the pressure difference analysis processor;

[0017] The pressure difference analysis processor receives the synchronous digital pressure signal through the data interface, analyzes the measured grease pressure value of each cavity pressure sensor node and the external measured water pressure value of the corresponding position, calculates the measured pressure difference of each spatial position; the measured pressure difference is the difference between the external measured water pressure value and the measured grease pressure value;

[0018] Based on the fluid mechanics pressure transmission theory, the measured pressure difference data in time sequence is taken as the input parameter, the measured pressure difference data of continuous multiple time points and the corresponding spatial position information are associated, the time series analysis algorithm is adopted to fit the dynamic change trend of the measured pressure difference, the pressure difference dynamic model is established, and the theoretical pressure difference value of each point of the grease cavity at different time is output;

[0019] At the same time, the measured pressure difference of each node is compared with the theoretical pressure difference value of the corresponding position at the same real-time pressure data calculated by the pressure difference dynamic model one by one, and the measured and theoretical pressure difference deviation of each node is obtained;

[0020] Taking the spatial coordinate information of all sensor nodes and the corresponding measured and theoretical pressure difference deviation data as input, adopting three-dimensional interpolation algorithm for data fusion and spatial estimation, deducing the pressure difference deviation of all positions on the inner wall of the grease cavity, and combining the change trend of time dimension, the spatial distribution of pressure difference deviation and the dynamic change of time are combined, and the pressure difference space-time distribution map is converted; the pressure difference space-time distribution map reflects the pressure difference abnormal distribution and evolution trend of the whole grease cavity at different time.

[0021] Specifically, the pressure imbalance point is identified based on the circumferential center of the grease cavity, the pressure difference difference value between the pressure imbalance point and the circumferentially symmetrical position point is calculated, and the pressure difference abnormal distribution map is generated, including:

[0022] The pressure difference analysis processor scans the pressure difference space-time distribution map 360 degrees based on the circumferential center of the grease cavity, obtains the position information of all sensor nodes in the pressure difference space-time distribution map and the measured and theoretical pressure difference deviation;

[0023] If the absolute error of the measured and theoretical pressure difference of any sensor node deviates more than a first preset deviation threshold, the node is marked as a primary pressure imbalance point;

[0024] For the marked primary pressure imbalance point, the pressure difference analysis processor determines its symmetric position point relative to the circumferential center of the oil cavity, obtains the measured pressure difference of the symmetric position point, and then calculates the absolute error of the measured pressure difference of the primary pressure imbalance point and the symmetric position point to obtain a measured symmetric pressure difference difference value;

[0025] If the measured symmetric pressure difference difference value is greater than a second preset symmetric deviation threshold, the primary pressure imbalance point is confirmed as a final pressure imbalance point;

[0026] The pressure difference analysis processor collects all final pressure imbalance points and their corresponding measured symmetric pressure difference difference values, calculates the pressure difference anomaly coefficient of each final pressure imbalance point, and generates the pressure difference anomaly distribution map through the contour method.

[0027] Specifically, the positioning of the pressure weak area based on the principles of fluid mechanics includes:

[0028] The pressure difference analysis processor analyzes the correlation between the pressure difference anomaly coefficient and the fluid penetration risk based on the principles of fluid mechanics and the generated pressure difference anomaly distribution map, and locates the pressure weak area of the oil cavity; the pressure weak area is defined as a continuous area with a pressure difference anomaly coefficient greater than 1.2, i.e., an area where the external water pressure is significantly higher than the oil pressure in the cavity;

[0029] The located pressure weak area is analyzed, and the area in the pressure difference anomaly distribution map with a pressure difference anomaly coefficient value greater than a preset pressure difference anomaly threshold or the area with a measured symmetric pressure difference difference value greater than a preset pressure difference threshold is marked as the high-risk leakage area, and the high-risk leakage area data including the position, area, maximum pressure difference anomaly coefficient value, and maximum measured symmetric pressure difference difference value of the high-risk leakage area is output. At the same time, the real-time digital pressure signal corresponding to the triggering of the high-risk leakage area marking is extracted and output as a pressure monitoring signal; the preset pressure difference threshold is determined according to the oil resistance grade of the shield tail.

[0030] Specifically, the water monitoring unit includes a plurality of high-frequency capacitive moisture sensors arranged in the deposition area at the bottom of the oil cavity; the high-frequency capacitive moisture sensors are connected to a data collector, and the data collector is used to receive the moisture content digital signal collected by the sensors and upload it.

[0031] Specifically, the water monitoring unit collects water data at the bottom of the oil cavity, constructs a water diffusion model using a support vector machine algorithm, and outputs a water three-dimensional distribution map, including:

[0032] The sensor of the moisture monitoring unit collects moisture content data of multiple monitoring points at the bottom of the oil cavity, and transmits the moisture content digital signals converted from the moisture content data to the data collector;

[0033] The data collector transmits the moisture content digital signals to the moisture analysis processor;

[0034] The moisture analysis processor takes the values corresponding to the received moisture content digital signals as training samples, takes the coordinates of each monitoring point in the three-dimensional coordinate system established with the center of the bottom of the oil cavity as the input features, and takes the moisture content values of the corresponding monitoring points as the output labels, trains and learns by using a support vector machine regression algorithm, and constructs a moisture diffusion model for predicting the moisture content of any point in the oil cavity;

[0035] The moisture analysis processor performs three-dimensional grid decomposition on the oil cavity space based on the moisture diffusion model, predicts the moisture content of each point in the grid by using the moisture diffusion model, and generates the moisture three-dimensional distribution map.

[0036] Specifically, the three-dimensional coordinate system is established with the center of the bottom of the oil cavity as the origin, the diffusion rate gradient of each monitoring point is calculated, and the moisture intrusion path map is generated, which includes:

[0037] The moisture analysis processor calls the three-dimensional coordinate system established with the center of the bottom of the oil cavity as the origin;

[0038] Based on the moisture three-dimensional distribution map of the continuous multiple time series, the moisture content data of each monitoring point at different time points is extracted, and the moisture content change rate of each monitoring point in a unit of time, i.e., the diffusion rate, is calculated;

[0039] In combination with the spatial coordinates of each monitoring point and the corresponding diffusion rate, the diffusion rate gradient of each monitoring point on its spatial coordinates is calculated, and a gradient vector is obtained;

[0040] According to the gradient vectors of all monitoring points, the flow line tracking algorithm is used to simulate the moisture movement direction and path, and the moisture intrusion path map is generated; the moisture intrusion path map indicates the trend and channel of the moisture diffusion from the bottom of the cavity upward and around.

[0041] Specifically, the potential leakage area is marked based on the preset moisture rising rate threshold according to the moisture intrusion path map, which includes:

[0042] The moisture analysis processor analyzes the generated moisture intrusion path map, and identifies one or more core intrusion paths with the maximum diffusion rate gradient module length in the map;

[0043] Along the identified core intrusion path, the diffusion rate data of each segment on the core intrusion path is extracted;

[0044] The diffusion rate of each segment of the core invasion path is compared with a preset moisture rising rate threshold, and if the diffusion rate of any segment of the core invasion path is greater than the preset moisture rising rate threshold, the region corresponding to the core invasion path is marked as the potential leakage area;

[0045] The moisture analysis processor outputs potential leakage area data containing the position, range and maximum diffusion rate value of the potential leakage area, and extracts the real-time moisture content signal corresponding to the triggering of the potential leakage area marking as a moisture monitoring signal.

[0046] Specifically, the risk assessment unit outputs the leakage risk level through Bayesian network fusion analysis based on the extracted dynamic features and model parameters trained based on historical failure data, including:

[0047] The risk assessment unit obtains the real-time dynamic features extracted by the edge data processing unit, and the Bayesian network model parameters trained based on historical failure data; the Bayesian network model parameters include the conditional probability table of the Bayesian network;

[0048] The risk assessment unit inputs the real-time dynamic features as evidence into the corresponding input nodes of the Bayesian network model, and starts the probability reasoning process in combination with the built-in conditional probability table;

[0049] The Bayesian network performs probability reasoning based on the input real-time dynamic features and the conditional probability table, and updates the posterior probability of all nodes in the Bayesian network;

[0050] The probability of each risk level corresponding to the output node of the Bayesian network is calculated through the inference result, and the highest probability level is taken as the final output result; the levels include low risk, medium risk and high risk.

[0051] Compared with the prior art, the beneficial effects of the present application are:

[0052] 1. The present application proposes a real-time detection system for shield tail sealing grease cavity, and optimizes and improves the architecture, operation steps and process, the system has the advantages of simple process, low investment and operation cost, and low production cost.

[0053] 2. The present application proposes a real-time detection system for shield tail sealing grease cavity, the system realizes omnidirectional real-time sensing of the grease cavity through a multi-physical field monitoring module, the pressure difference monitoring unit accurately locates the high-risk leakage area based on a pressure difference dynamic model and high-risk area analysis, and the moisture monitoring unit efficiently identifies the potential leakage area with the help of a support vector machine model and diffusion path tracking, so that multi-dimensional fault early warning from structural deformation to moisture intrusion is realized, and the comprehensiveness and accuracy of abnormal detection of the grease cavity are improved.

[0054] 3.The application provides a shield tail sealing grease cavity real-time detection system, a key dynamic feature is extracted by an edge data processing unit of an intelligent early warning analysis module, a risk assessment unit is combined with a Bayesian network fusion analysis technology, a quantified leakage risk level can be output, a secondary early warning can be triggered when three abnormalities of a sudden drop in grease viscosity, a sudden rise in moisture and a pressure difference exceeding a limit are detected, a full-process intelligentization from data acquisition, feature extraction to risk assessment is realized, and an operation risk caused by a leakage fault is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 FIG. 1 is a schematic diagram of a shield tail sealing grease cavity real-time detection system according to the application;

[0056] Figure 2 FIG. 2 is a principle flowchart of a shield tail sealing grease cavity real-time detection system according to the application;

[0057] Figure 3 FIG. 3 is a process flowchart for generating a pressure difference space-time distribution atlas of a shield tail sealing grease cavity real-time detection system according to the application. DETAILED DESCRIPTION

[0058] Embodiment 1

[0059] Referring to Figures 1-2 The application provides a shield tail sealing grease cavity real-time detection system, which comprises:

[0060] a multi-physical field monitoring module and an intelligent early warning analysis module; the multi-physical field monitoring module comprises a pressure difference monitoring unit and a moisture monitoring unit; the intelligent early warning analysis module comprises an edge data processing unit and a risk assessment unit;

[0061] The pressure difference monitoring unit collects grease pressure data in the grease cavity and external environment water pressure data, establishes a pressure difference dynamic model and generates a pressure difference space-time distribution atlas, then identifies a pressure imbalance point with the circumferential center of the grease cavity as a reference, calculates a pressure difference difference value between the pressure imbalance point and a circumferentially symmetrical position point, generates a pressure difference abnormal distribution atlas, locates a pressure weak area based on a fluid mechanics principle, marks a high-risk leakage area based on a preset threshold value, and outputs high-risk leakage area data and a corresponding pressure monitoring signal;

[0062] The moisture monitoring unit collects moisture data at the bottom of the grease cavity, constructs a moisture diffusion model by using a support vector machine algorithm and outputs a moisture three-dimensional distribution atlas, then establishes a three-dimensional coordinate system with the center of the bottom of the grease cavity as an origin, calculates diffusion rate gradients of each monitoring point and generates a moisture intrusion path atlas, marks a potential leakage area based on a preset moisture rising rate threshold value according to the moisture intrusion path atlas, and outputs potential leakage area data and a corresponding moisture monitoring signal;

[0063] According to the high-risk leakage area data, the pressure monitoring signal, the potential leakage area data and the moisture monitoring signal, the edge data processing unit extracts dynamic features using an isolation forest algorithm; the dynamic features include a pressure difference change rate and a moisture rise rate;

[0064] The isolation forest algorithm is prior art in the field and is not part of the inventive concept of the present application, and thus will not be described in detail.

[0065] It should be noted that the edge data processing unit is deployed at the edge of data acquisition, which can realize real-time processing and analysis of data, reducing the delay and bandwidth consumption of data transmission to the cloud; the isolation forest algorithm has efficient anomaly detection and feature extraction capabilities, which can accurately extract dynamic features reflecting the operating state changes of the grease chamber from a large amount of monitoring data; the pressure difference change rate reflects the imbalance trend of external water pressure and internal grease pressure, which is a key driving index of water intrusion; the moisture rise rate reflects the speed of water intrusion, and these two dynamic features are key indicators for evaluating the leakage risk of the grease chamber; by extracting these dynamic features, simple and key input parameters are provided for subsequent risk assessment, improving the efficiency and accuracy of risk assessment.

[0066] The risk assessment unit outputs a leakage risk level based on the extracted dynamic features, combined with model parameters trained based on historical failure data, through Bayesian network fusion analysis; when detecting three evidences of sudden drop of grease viscosity, sudden rise of moisture and pressure difference overrun during the fusion analysis process, a secondary warning is triggered.

[0067] It should be noted that the system architecture design strictly conforms to the core failure mechanism of the shield tail sealing: the multi-physical field monitoring module captures the imbalance between external water pressure and internal grease pressure through the pressure difference monitoring unit, and tracks the degradation process of grease performance after water intrusion through the moisture monitoring unit, both of which cooperatively realize full-chain monitoring from the risk source to the failure result; the intelligent early warning analysis module converts multi-dimensional parameters into quantitative risk levels through edge data processing and Bayesian network fusion, avoiding the limitations of traditional single parameter monitoring, ensuring early identification and accurate early warning of sealing failure risk, and providing a scientific basis for timely measures such as supplementing grease and adjusting pressure on site.

[0068] Embodiment 2

[0069] The pressure difference monitoring unit in this embodiment includes two types of sensors;

[0070] The first type of sensor is at least two rings of pressure sensor arrays uniformly arranged on the inner wall of the grease chamber, one ring near the shield tail brush is the first pressure array for collecting grease pressure data in the grease chamber, and the other ring is the second pressure array for assisting verification of pressure distribution uniformity;

[0071] The second type of sensor is an external water pressure sensor group buried outside the shield tail shell corresponding to the position of the grease cavity, used to collect external environment water pressure data;

[0072] The first type of sensor and the second type of sensor are connected with a pressure data acquisition instrument; the pressure data acquisition instrument is used to convert the pressure signals collected by the sensors into digital pressure signals, and complete the time synchronization upload of the intracavity pressure and external water pressure data, ensuring the consistency of the data in the time dimension.

[0073] It should be noted that in the present application, this double-circle array arrangement can realize the key monitoring of the high-risk area near the shield tail brush: the first pressure array can capture the local pressure anomaly caused by the wear of the shield tail brush and uneven distribution of grease at close range, and the second pressure array provides a reference for the global pressure distribution, and the combination of the two forms a pressure distribution verification mechanism; the corresponding arrangement of the external water pressure sensor group and the intracavity pressure sensor can directly calculate the pressure difference at the same spatial position, avoid the calculation error of the pressure difference caused by the position deviation, and greatly improve the accuracy of the pressure difference monitoring.

[0074] Please refer to Figure 3 , the process of constructing a pressure difference dynamic model and generating a pressure difference spatiotemporal distribution map is as follows:

[0075] A1: The pressure sensor array of the pressure difference monitoring unit and the external water pressure sensor group synchronously collect data, respectively generate the grease pressure data in the grease cavity and the external environment water pressure data, and transmit them to the pressure data acquisition instrument;

[0076] Further, the process of converting the grease pressure data in the grease cavity and the external environment water pressure data into the wavelength variation of the optical signal by the pressure sensor array and the external water pressure sensor group includes:

[0077] (1) The sensitive element of the pressure sensor array, such as the elastic diaphragm, is directly in contact with the inner wall of the grease cavity, and the sensitive element of the external water pressure sensor group is in contact with the water outside the shield tail; when the grease pressure in the grease cavity or the external environment water pressure changes, the pressure is transmitted to the surface of the sensitive element through the medium such as grease or water, causing the elastic diaphragm to produce a mechanical deformation proportional to the pressure variation: when the pressure rises, the diaphragm protrudes outward, and when the pressure decreases, the diaphragm is concave inward, and the degree of deformation increases linearly with the increase of the pressure variation;

[0078] (2) The elastic diaphragm and the packaging structure of the fiber grating are rigidly connected, the mechanical deformation of the diaphragm is transmitted to the body of the fiber grating through the packaging structure, when the diaphragm protrudes, the packaging structure generates an axial tensile force on the fiber grating, so that the period of the grating increases, when the diaphragm is depressed, the packaging structure generates an axial compression force on the fiber grating, so that the period of the grating decreases, in this process, the change amount of the grating period and the deformation amount of the elastic diaphragm are strictly linearly related, and then the change amount of the pressure is linearly related, wherein the packaging structure of the fiber grating refers to a metal sleeve or an epoxy resin bonding layer;

[0079] (3) The core reflection characteristic of the fiber grating is determined by the Bragg equation, and the center wavelength of the reflected light changes with the grating period and the effective refractive index of the fiber; wherein the Bragg equation is prior art content in the art, not the inventive scheme of the present application, and will not be described here;

[0080] (4) The wide-spectrum light source built-in the sensor emits a continuous spectrum of light signals, which is injected into the fiber grating through the fiber coupler, the fiber grating only reflects a specific wavelength light signal matching the current center wavelength of the reflected light, and the rest of the wavelength light signal is transmitted or absorbed, then the reflected light signal containing the wavelength shift information is transmitted to the light signal output end through the same coupler, at this time, the wavelength change amount of the output light signal has formed a one-to-one correspondence with the original pressure change amount, and the conversion of the pressure data to the wavelength change amount of the light signal is completed.

[0081] A2: The pressure data acquisition instrument calibrates the oil cavity oil pressure data and the external environment water pressure data with time stamps, and converts the digital pressure signals into digital pressure signals after transmission to the pressure difference analysis processor;

[0082] In the present application, the pressure data acquisition instrument of the pressure difference monitoring unit is connected with the pressure sensor array and the external water pressure sensor group, and is specially responsible for collecting the wavelength change amount of the light signal output by the sensor array, the pressure data acquisition instrument can demodulate and convert the light signal with high precision, accurately convert the wavelength change amount of the light signal into a digital pressure signal, and stably transmit the digital pressure signal to the pressure difference analysis processor; in the demodulation process, the pressure data acquisition instrument filters out noise interference through a signal processing algorithm, ensuring the quality of the digital pressure signal.

[0083] Further, the specific steps of A2 include:

[0084] (1) The pressure data acquisition instrument receives the analog pressure signals output by the pressure sensor array and the external water pressure sensor group through the industrial Ethernet interface, first amplifies the weak voltage signal output by the sensor to a standard range through a signal conditioning circuit to reduce the error of subsequent analog-to-digital conversion, in the present application, the weak voltage signal refers to 0-5V, and the standard range refers to 0-10V;

[0085] (2) Start the built-in synchronous clock module, assign an independent timestamp to each sensor signal, the timestamp accuracy reaches 1 ms, by comparing the timestamps of each sensor signal, eliminate the data whose time deviation caused by transmission delay exceeds 5 ms, ensure that the cavity pressure and external water pressure data at the same time correspond one by one;

[0086] (3) Use a 16-bit analog-to-digital converter to convert the analog pressure signal after conditioning into a digital pressure signal, the conversion rate is 1 kHz, and the average value is taken by multiple sampling during the conversion process, further reducing random noise interference, wherein the analog-to-digital converter is a prior art in the field, and is not the inventive scheme of the present application, which will not be described here;

[0087] (4) Perform validity check on the digital pressure signal: if any signal exceeds the normal measurement range for 3 times in a row, or the fluctuation of adjacent two data exceeds 0.1 MPa, it is marked as abnormal data, and the standby sensor data is automatically enabled, if there is no standby sensor, the sliding average of the previous 5 data is used instead, to ensure data continuity, wherein the normal measurement range refers to the cavity pressure 0-2 MPa, and the external water pressure 0-3 MPa;

[0088] (5) Encode the digital pressure signal according to the Modbus-TCP communication protocol, add data check bits to avoid data tampering or loss during transmission, wherein the Modbus-TCP communication protocol is a prior art in the field, and is not the inventive scheme of the present application, which will not be described here;

[0089] (6) The encoded digital pressure signal is transmitted to the pressure difference analysis processor through optical fiber, with timestamp information, to ensure time synchronization during subsequent pressure difference calculation.

[0090] A3: The pressure difference analysis processor receives the synchronized digital pressure signal through the data interface, analyzes the measured grease pressure value of each cavity pressure sensor node and the corresponding external measured water pressure value, and calculates the measured pressure difference of each spatial position; the measured pressure difference is the difference between the external measured water pressure value and the measured grease pressure value;

[0091] Further, the pressure difference analysis processor is linked with the shield machine PLC system through the data interface, and can obtain the working condition parameters such as the advancing speed and cutterhead rotating speed of the shield machine in real time. When the working condition parameters suddenly change, such as the advancing speed suddenly increases from 30 mm per minute to 50 mm per minute, the pressure difference calculation frequency is automatically increased to ensure that the pressure fluctuation caused by the change of working condition is captured. At the same time, the sensor nodes are numbered and bound with spatial coordinates during the analysis process, to ensure that each pressure difference data corresponds to the actual position one by one, providing a position reference for subsequent spatial distribution map generation.

[0092] A4: Based on the pressure transfer theory of fluid mechanics, the measured pressure difference data in time series is taken as the input parameter, the measured pressure difference data at continuous time points are associated with the corresponding spatial position information, the time series analysis algorithm is adopted to fit the dynamic change trend of the measured pressure difference, the pressure difference dynamic model is established, and the theoretical pressure difference values of each point of the grease cavity at different time points are output;

[0093] Further, the specific steps of A4 include:

[0094] (1) The fluid mechanics characteristics of the grease cavity are determined: the inside of the grease cavity is high-viscosity shield tail grease, and the outside is a mixture of underground water and mud, and the pressure transfer follows Darcy's law, based on which the correlation between the pressure difference and the fluid infiltration risk is determined, wherein Darcy's law is a prior art content in the field and is not the inventive scheme of the present application, and will not be described here;

[0095] (2) The pressure difference analysis processor receives two key data: one is the real-time pressure difference data after synchronization , and the other is the real-time working condition parameters transmitted by the shield machine PLC system: first, the two key data are preprocessed, including time synchronization calibration to ensure that the time stamps of the working condition parameters and the pressure difference data are completely consistent, and abnormal data elimination;

[0096] (3) Based on Darcy's law and actual working condition parameters, an initial pressure difference dynamic model is constructed, and the initial pressure difference dynamic model is preliminarily calibrated through historical normal working condition data to determine the coefficient range; the expression of the initial pressure difference dynamic model is the weighted sum of real-time working condition parameters and time; the historical normal working condition data are obtained through the construction period without leakage record;

[0097] (4) The autoregressive moving average model is used to dynamically fit the real-time pressure difference data for 5 minutes, the initial pressure difference dynamic model parameters are optimized, and the final pressure difference dynamic model is obtained, wherein the autoregressive moving average model is a prior art content in the field and is not the inventive scheme of the present application, and will not be described here;

[0098] (5) The real-time working condition parameters are substituted into the final pressure difference dynamic model to calculate the theoretical pressure difference values at different time points and different spatial positions.

[0099] It needs to be explained that the pressure difference dynamic model not only considers the pressure change trend in the time dimension, but also integrates the shield machine working condition parameters, which can effectively distinguish between normal pressure fluctuations caused by working condition changes and abnormal pressure differences caused by sealing failure, and avoid false judgments caused by adjustment of the advancing speed and changes in the cutter load; the theoretical pressure difference values output by the model provide a dynamic benchmark for judging whether the measured pressure difference is abnormal, which is more consistent with the actual construction scene than a fixed threshold.

[0100] A5: At the same time, the measured pressure difference of each node is compared with the theoretical pressure difference value of the corresponding position calculated by the pressure difference dynamic model under the same real-time pressure data, and the deviation of the measured and theoretical pressure difference of each node is obtained;

[0101] Further, through such point-to-point accurate comparison, the local pressure abnormal area can be accurately located, and such deviation calculation method can effectively amplify the subtle pressure abnormality. For example, if the theoretical pressure difference value of any node continuously exceeds 0.05 MPa, it indicates that the pressure difference change at this position exceeds the working condition influence range, and there is a risk of oil leakage, local increase of external water pressure; the deviation calculation result can also be used for sensor fault diagnosis. If the theoretical pressure difference value of any node suddenly increases to more than 0.2 MPa, and there is no obvious change in the adjacent nodes, it is judged as a sensor fault, and a sensor calibration process is triggered.

[0102] A6: Taking the spatial coordinate information of all sensor nodes and the corresponding measured and theoretical pressure difference deviation data as input, a three-dimensional interpolation algorithm is used for data fusion and spatial estimation, and the pressure difference deviation of all positions on the inner wall of the oil cavity is deduced. Combined with the change trend in the time dimension, the spatial distribution and time dynamic change of the pressure difference deviation are combined to convert into a pressure difference space-time distribution map; the pressure difference space-time distribution map reflects the pressure difference abnormal distribution and evolution trend of the entire oil cavity at different times.

[0103] Further, the specific steps of A6 include:

[0104] (1) Organize the basic data of all sensor nodes; the basic data includes two parts, one is the three-dimensional spatial coordinates of each node on the inner wall of the oil cavity, and a cylindrical coordinate system is established with the circumferential center of the oil cavity as the origin, and then the cylindrical coordinates are converted into rectangular coordinates; the other is the corresponding measured and theoretical pressure difference deviation of each node;

[0105] (2) Discretize the surface of the inner wall of the oil cavity into regular three-dimensional grids as the spatial carrier for estimating the measured and theoretical pressure difference deviation, and the specific division method is: divide the grid lines at an interval of 1° along the circumference, and divide the grid lines at an interval of 5 mm along the axis, and the intersection of the two forms a grid element, and the vertex of each grid element is an unknown point that needs to be estimated for the measured and theoretical pressure difference deviation, and the grid density is determined according to the number of sensor nodes and the size of the oil cavity, and the coordinates of each grid vertex are determined by linear interpolation, and then all the grid vertex coordinates form an ordered spatial point set as the output target point of the three-dimensional interpolation algorithm, wherein the linear interpolation is a prior art content in the field and is not the inventive scheme of the present application, and will not be described here;

[0106] (3) Data fusion and spatial estimation are performed by using the Kriging interpolation algorithm, including:

[0107] By analyzing the spatial correlation of the measured and theoretical pressure difference deviation of the known nodes, a spherical variation function is used to fit the spatial variation structure, wherein the spherical variation function is a prior art content in the field and is not the creative scheme of the present application, and will not be described here;

[0108] For each grid vertex, based on the measured and theoretical pressure difference deviation data of the surrounding known nodes and the variation function model, a weight coefficient is calculated, and the estimated value of the measured and theoretical pressure difference deviation of the vertex is obtained by weighted summation;

[0109] (4) The pressure deviation estimation value of the three-dimensional grid vertex is converted into a pressure difference space-time distribution map, and the specific process is as follows:

[0110] a. The color gradient method is used to represent the pressure deviation size, for example, blue represents negative deviation, red represents positive deviation, the color depth increases with the increase of the absolute value of the deviation, and the intermediate value is represented by green, and a mapping table of deviation value and RGB color value is established, wherein the positive deviation represents that the measured pressure difference is greater than the theoretical pressure difference value, and the negative deviation represents that the measured pressure difference is less than the theoretical pressure difference value, and the color gradient method is a prior art content in the field and is not the creative scheme of the present application, and will not be described here.

[0111] b. For each grid element, according to the pressure deviation values of its four vertices and the corresponding colors, the color of any point inside the element is calculated by using bilinear interpolation, and then all grid elements are spliced into a complete oil cavity inner wall surface pattern by using the surface rendering technology in computer graphics, wherein the bilinear interpolation and the surface rendering technology are prior art contents in the field and are not the creative scheme of the present application, and will not be described here.

[0112] c. Coordinate scales, color scales, and sensor node position markers are added to the map to visually display the spatial distribution of pressure deviation and trace the location of the original data.

[0113] Further, the three-dimensional interpolation algorithm can fully utilize the limited sensor node data, and expand to all positions of the oil cavity inner wall through spatial interpolation, realizing the conversion from discrete point data to continuous surface data, and generating a pressure difference space-time distribution map that intuitively reflects the abnormal distribution of the entire oil cavity inner wall pressure, so that the staff can grasp the overall pressure state of the oil cavity at a glance, providing a visual tool for quickly identifying abnormal areas.

[0114] The oil cavity circumferential center is taken as the reference to identify the pressure imbalance point, calculate the pressure difference difference value between the pressure imbalance point and the circumferentially symmetrical position point, and generate a pressure difference abnormal distribution map, which includes:

[0115] B1: the pressure difference analysis processor performs 360-degree circumferential scanning on the pressure difference space-time distribution map based on the circumferential center of the oil cavity, and obtains the position information of all sensor nodes in the pressure difference space-time distribution map and the deviation of the measured and theoretical pressure difference;

[0116] B2: if the absolute error of the deviation of the measured and theoretical pressure difference of any sensor node is greater than a first preset deviation threshold, the node is marked as a primary pressure imbalance point;

[0117] B3: for the marked primary pressure imbalance point, the pressure difference analysis processor determines the symmetric position point thereof with respect to the circumferential center of the oil cavity, obtains the measured pressure difference of the symmetric position point, and then calculates the absolute error of the measured pressure difference of the primary pressure imbalance point and the symmetric position point, to obtain a measured symmetric pressure difference difference value;

[0118] B4: if the measured symmetric pressure difference difference value is greater than a second preset symmetric deviation threshold, the primary pressure imbalance point is confirmed as a final pressure imbalance point;

[0119] It should be noted that, under normal circumstances, the pressure values at the symmetric positions of the inner wall of the oil cavity should be approximately equal. Through comparison of the symmetric positions, false judgments caused by sensor errors or local interference can be further excluded, the accuracy of pressure imbalance point identification is improved, the setting of the second preset symmetric deviation threshold provides a strict standard for confirmation of the final pressure imbalance point, and it is ensured that only the pressure points with structural abnormalities will be confirmed.

[0120] B5: the pressure difference analysis processor collects all final pressure imbalance points and corresponding measured symmetric pressure difference difference values, calculates the pressure difference anomaly coefficients of each final pressure imbalance point, and generates a pressure difference anomaly distribution map through contour method.

[0121] The calculation of the pressure difference anomaly coefficient can quantify the degree of high-risk area at the pressure imbalance point; the application of the contour method presents the high-risk area in the form of an intuitive map; and the pressure difference anomaly distribution map clearly shows the position, range and degree of the high-risk area of the inner wall of the oil cavity.

[0122] Further, the specific steps of B5 include:

[0123] (1) obtaining all final pressure imbalance points and corresponding measured symmetric pressure difference difference values;

[0124] (2) Select the sensor nodes without abnormal areas on the inner wall of the grease cavity, that is, the nodes not marked as abnormal points, extract the measured pressure difference value of the non-abnormal area nodes under the current real-time pressure, and calculate the average pressure difference value by the arithmetic average method as the pressure difference reference under the normal stress state of the grease cavity under the current working condition. If the number of selected non-abnormal area nodes is less than the preset threshold, such as less than 30% of the total number of nodes, the pressure difference dynamic model is called to input the current real-time pressure data and calculate the theoretical normal pressure difference value of each position of the grease cavity under the current working condition. The average value of all theoretical normal pressure difference values is taken as the pressure difference reference;

[0125] (3) For each final pressure imbalance point, the determined pressure difference reference is combined with the measured symmetric pressure difference difference value corresponding to the point to calculate the local pressure difference value: Since the measured symmetric pressure difference difference value reflects the pressure difference deviation of the final pressure imbalance point from the symmetric position, and the deviation is caused by the pressure abnormality of the point itself, the pressure difference reference is added to the measured symmetric pressure difference difference value to obtain the local pressure difference value of each final pressure imbalance point. This value directly reflects the actual pressure difference level at the abnormal point;

[0126] (4) Based on the obtained pressure difference reference and local pressure difference value, the pressure difference abnormality coefficient of each final pressure imbalance point is calculated: the pressure difference abnormality coefficient is the ratio of the local pressure difference value to the pressure difference reference. The larger the ratio, the more serious the deviation of the final pressure imbalance point from the normal state. The pressure abnormality degree can be quantified through the pressure difference abnormality coefficient;

[0127] (5) Taking the spatial coordinates of the final pressure imbalance points and the corresponding pressure difference abnormality coefficients as known sample points, the pressure difference abnormality coefficients of all positions on the inner wall of the grease cavity are estimated by the inverse distance weighted interpolation method: the weight is determined according to the spatial distance between each unknown position and the known sample point, wherein the closer the distance, the greater the weight. Then the pressure difference abnormality coefficient of the unknown position is calculated by weighting. The discrete sample point data is expanded to continuous pressure difference abnormality coefficient distribution data covering the entire domain of the inner wall of the grease cavity, realizing the representation of pressure abnormality distribution from point to plane. The inverse distance weighted interpolation method is a prior art content in the field and is not the inventive scheme of the present application, and will not be described here;

[0128] (6) According to the numerical range of the pressure difference abnormality coefficient, the levels are divided, including low risk, medium risk and high risk, each level corresponds to different colors, such as blue, yellow and red, which is convenient for visual differentiation;

[0129] (7) superimpose the contour lines of different levels on the three-dimensional model of the inner wall of the oil cavity, mark the pressure difference anomaly coefficient values corresponding to each contour line, add a color scale and coordinate scale to the edge of the atlas, and mark the actual position of the final pressure imbalance point with a triangle in the atlas, so that the original data and the interpolation results form a corresponding relationship.

[0130] The abnormal deformation is located based on the principle of fluid mechanics, and a high-risk leakage area is marked based on a preset threshold, including:

[0131] C1: a pressure difference analysis processor, based on the principle of fluid mechanics, analyzes the correlation between the pressure difference anomaly coefficient and the fluid penetration risk in combination with the generated pressure difference anomaly distribution atlas, and locates a pressure weak area of the oil cavity; the pressure weak area is defined as a continuous area with a pressure difference anomaly coefficient greater than 1.2, i.e., an area where the external water pressure is significantly higher than the oil pressure in the cavity, wherein the principle of fluid mechanics is a prior art content in the art and is not the inventive scheme of the present application, and will not be described here.

[0132] Further, the specific steps of C1 include:

[0133] (1) The pressure difference analysis processor extracts features from the generated pressure difference anomaly distribution atlas, including: the pressure difference anomaly coefficient value range of each area, the spatial coordinate range of the high-risk area, the gradient change rate of the pressure difference anomaly coefficient, and the coordinate point where the maximum pressure difference anomaly coefficient is located;

[0134] (2) Preprocess the extracted feature data: eliminate isolated high-value points in the atlas caused by interpolation errors, which are usually represented as a single grid node with a pressure difference anomaly coefficient significantly higher than all surrounding nodes, such as a node with a coefficient greater than or equal to 1.5 while all surrounding nodes have coefficients less than or equal to 1.0, and no continuous high-coefficient node is associated with it, which does not reflect the true pressure anomaly state, but only the interpolation calculation deviation. Then, through a smoothing algorithm, the gradient change of the pressure difference anomaly coefficient is modified to make the gradient change comply with the continuity law of pressure transmission in fluid mechanics, wherein the continuity law of pressure transmission is that the spatial change of the pressure difference anomaly coefficient should be smooth and transition, without sudden rise and fall without reason, and the smoothing algorithm is a prior art content in the art and is not the inventive scheme of the present application, and will not be described here.

[0135] (3) Based on the principle of fluid mechanics, an association between the pressure difference anomaly coefficient and the fluid penetration risk is established, including:

[0136] aLinear relationship between pressure difference and permeation flow: according to Darcy's law, under the condition that the characteristics of the medium such as grease and water body and the fixed permeation area are constant, the permeation flow is positively correlated with the pressure difference, and the pressure difference anomaly coefficient is equal to the ratio of the local pressure difference to the normal pressure difference benchmark, so the greater the pressure difference anomaly coefficient, the greater the local pressure difference, and the greater the fluid permeation flow, the higher the permeation risk;

[0137] b Pressure transmission characteristics of the grease cavity: the grease cavity is filled with high-viscosity shield tail grease, and the internal pressure of the cavity is balanced with the external water pressure under normal working conditions, and the high viscosity of the grease can effectively prevent the penetration of the external water body; when the pressure difference anomaly coefficient is greater than 1.2, the external water pressure is significantly higher than the grease pressure in the cavity, the anti-permeation ability of the grease cannot offset the water pressure effect, and the fluid permeation risk rises sharply, and the greater the coefficient, the more obvious the increase in the permeation risk, which provides a core basis for subsequent positioning of the pressure weak area;

[0138] (4) Based on the above principles, combined with the predefined pressure weak area definition, the positioning basis of the pressure weak area is determined: in the pressure difference anomaly distribution map, the area with a pressure difference anomaly coefficient greater than 1.2 is an area with an increased fluid permeation risk, which can be preliminarily determined as a pressure weak area; wherein the coordinate point corresponding to the maximum pressure difference anomaly coefficient corresponds to the position with the most serious permeation risk, which needs to be the focus of attention, wherein the predefined pressure weak area definition refers to a continuous area with a pressure difference anomaly coefficient greater than 1.2, i.e. an area where the external water pressure is higher than the grease pressure in the cavity;

[0139] (5) Spatial boundary identification of the high-risk area in the pressure difference anomaly distribution map, including:

[0140] a The high pressure difference anomaly threshold is set to 1.2, all grid nodes in the pressure difference anomaly distribution map are traversed, and the nodes with a pressure difference anomaly coefficient greater than or equal to 1.2 are marked as suspected weak nodes, and the area formed by all the suspected weak nodes is a suspected pressure weak area;

[0141] b Taking the extracted core high-value point as the starting point, the region growing algorithm is used to gradually merge the suspected weak nodes adjacent to the core high-value point and having a pressure difference anomaly coefficient greater than or equal to 1.2, forming multiple independent connected regions, each connected region being an independent suspected pressure weak area, and ensuring that no isolated suspected weak node is missed; the core high-value point refers to the coordinate point corresponding to the maximum pressure difference anomaly coefficient, wherein the region growing algorithm is a prior art content in the field and is not the creative scheme of the present application, and will not be described here;

[0142] cFor each suspected pressure weak area, its spatial boundary is extracted by an edge detection algorithm: in the circumferential dimension, the minimum and maximum angles covered by the area are determined, such as from 30 degrees to 60 degrees; in the axial dimension, the minimum and maximum heights covered by the area are determined, such as from 50 mm to 150 mm at the bottom of the grease cavity, and the range of each suspected pressure weak area is framed in the form of a rectangular boundary box, wherein the edge detection algorithm is a prior art content in the art and is not the inventive scheme of the present application, and will not be described here;

[0143] (6) The rationality of the suspected pressure weak area is verified in combination with the principle of fluid mechanics to exclude false areas caused by non-real penetration risk factors.

[0144] C2: The located pressure weak area is analyzed, the area with a pressure difference anomaly coefficient value greater than a preset pressure difference anomaly threshold value in the pressure difference anomaly distribution map or the area with a measured symmetrical pressure difference difference value greater than a preset pressure difference threshold value is marked as the high-risk leakage area, and high-risk leakage area data containing the position, area, maximum pressure difference anomaly coefficient value and maximum measured symmetrical pressure difference difference value of the high-risk leakage area are output. At the same time, the real-time digital pressure signal corresponding to the triggering of the high-risk leakage area marking is extracted, which is taken as a pressure monitoring signal and output. The preset pressure difference threshold value is determined according to the penetration resistance grade of the shield tail grease, and is usually set to 0.2 MPa.

[0145] The water content monitoring unit includes a plurality of high-frequency capacitive water content sensors arranged in the deposition area at the bottom of the grease cavity; the high-frequency capacitive water content sensors are connected with a data collector, and the data collector is used for receiving the water content digital signal collected by the sensors and uploading.

[0146] The water content monitoring unit collects water content data at the bottom of the grease cavity, constructs a water content diffusion model using a support vector machine algorithm, and outputs a water content three-dimensional distribution map, including:

[0147] D1: The sensors of the water content monitoring unit collect water content data at a plurality of monitoring points at the bottom of the grease cavity, which is converted into a water content digital signal and transmitted to the data collector;

[0148] The water content monitoring unit includes a plurality of high-frequency capacitive water content sensors arranged in the deposition area at the bottom of the grease cavity, which is an area where water is easy to accumulate. Arranging the sensors here can capture the water intrusion signal as early and accurately as possible. The arrangement of multiple sensors forms a multi-point monitoring network, which can comprehensively reflect the water distribution at the bottom of the grease cavity, avoiding the missed detection problem that may occur in single-point monitoring, and improving the reliability of water content monitoring.

[0149] D2: The data collector transmits the water content digital signal to a water content analysis processor;

[0150] D3: The moisture analysis processor takes the value corresponding to the received moisture content digital signal as a training sample, takes the coordinates of each monitoring point in the three-dimensional coordinate system established with the center of the bottom of the oil cavity as the origin as the input features, and takes the moisture content value of the corresponding monitoring point as the output label, and uses a support vector machine regression algorithm to train and learn, to construct a moisture diffusion model for predicting the moisture content of any point in the oil cavity. The support vector machine regression algorithm is prior art in the field and is not the inventive solution of the present application, and will not be described here.

[0151] D4: The moisture analysis processor performs three-dimensional grid decomposition on the oil cavity space based on the moisture diffusion model, predicts the moisture content of each point in the grid through the moisture diffusion model, and generates the moisture three-dimensional distribution map.

[0152] The three-dimensional grid decomposition divides the oil cavity space into small grid units, ensuring the spatial accuracy of the prediction; the moisture diffusion model provides a reliable basis for predicting the moisture content of each grid unit; and the generated moisture three-dimensional distribution map intuitively shows the spatial distribution of moisture in the oil cavity, enabling the staff to clearly understand the distribution state and aggregation area of moisture in the oil cavity, providing a visual reference for judging the degree and range of moisture intrusion.

[0153] Further, the specific steps of D4 include:

[0154] (1) Call the constructed moisture diffusion model to determine the core parameters and applicable scope of the model; the core principle of the moisture diffusion model is to fit the nonlinear law of spatial moisture distribution through the moisture data of known monitoring points;

[0155] (2) Establish a three-dimensional coordinate system with the center of the bottom of the oil cavity as the origin, and determine the grid range according to the actual size of the oil cavity to ensure that it covers the entire space of the oil cavity, including the bottom deposition area, the inner wall area, and the top space;

[0156] (3) Determine the grid size according to the required accuracy of moisture monitoring to form a cubic grid unit, and the vertices of each grid unit are the spatial points whose moisture content needs to be predicted;

[0157] (4) Calculate the three-dimensional coordinates of each grid vertex according to the grid division rules, and sort all vertex coordinates according to the spatial position to form an ordered three-dimensional coordinate point set as the input object for moisture content prediction;

[0158] (5) Using the water diffusion model to predict the moisture content of each grid vertex, including: converting the three-dimensional coordinates of the grid vertex into the feature grid used during model training to obtain standardized grid vertex coordinates; batch inputting the standardized grid vertex coordinates into the water diffusion model, and the water diffusion model calculates the moisture content prediction value of each vertex through the built-in support vector machine regression function; locally checking the prediction results, focusing on checking the prediction values of the grid vertices near the known monitoring points;

[0159] (6) Converting the moisture content prediction value of the grid vertex into an intuitive moisture three-dimensional distribution map, including: dividing levels according to the numerical range of the moisture content, and establishing a mapping relationship between the moisture content and the RGB color value; for each cubic grid unit, according to the moisture content prediction value of its eight vertices and the corresponding color, the color value of any point inside the unit is calculated using trilinear interpolation, and then through the volume rendering technology in computer graphics, all grid units are combined into a complete oil cavity three-dimensional space model, for the boundary areas such as the inner wall and the bottom of the oil cavity, the spatial distribution of the high moisture area is highlighted through transparency adjustment; adding a three-dimensional coordinate system scale, a moisture content color scale and key position markers in the map; through a smoothing algorithm, the sawtooth visual error of the grid edge is eliminated, the moisture distribution in the map is made to transition naturally, while the boundary definition of the high moisture area is preserved;

[0160] (7) Comparing the newly collected moisture monitoring data with the prediction value of the corresponding position in the moisture three-dimensional distribution map, calculating the overall average deviation, and if the deviation exceeds the preset threshold, triggering the model to be retrained, and optimizing the moisture diffusion model parameters with the latest data.

[0161] The three-dimensional coordinate system is established with the center of the oil cavity bottom as the origin, the diffusion rate gradient of each monitoring point is calculated, and the moisture intrusion path map is generated, including:

[0162] E1: The moisture analysis processor calls the established three-dimensional coordinate system with the center of the oil cavity bottom as the origin;

[0163] E2: Based on the moisture three-dimensional distribution map of the continuous multiple time series, the moisture content data of each monitoring point at different time points is extracted, and the moisture content change rate of each monitoring point in unit time, i.e. the diffusion rate, is calculated;

[0164] Further, the moisture content of each monitoring point in unit time is obtained by calculating the moisture content change amount difference of the monitoring point in unit time, and the moisture content change rate of each monitoring point in unit time is obtained by calculating the ratio of the moisture content of each monitoring point in unit time to the unit time, wherein the time interval between adjacent two time points as the calculation unit of unit time;

[0165] E3: combining the spatial coordinates of each monitoring point and the corresponding diffusion rate, calculating the diffusion rate gradient of each monitoring point at its spatial coordinates to obtain a gradient vector;

[0166] wherein the diffusion rate gradient reflects the rate of change of water diffusion speed in space, and the gradient vector indicates the direction in which the water diffusion speed changes fastest; by calculating the diffusion rate gradient, the diffusion law and dynamic mechanism of water in the oil cavity can be understood.

[0167] E4: according to the gradient vectors of all monitoring points, using the stream tracing algorithm to simulate the direction and path of water movement, and generating a water intrusion path map; the water intrusion path map indicates the trend and channel of water diffusion from the bottom of the cavity upward and around, wherein the stream tracing algorithm is prior art content in the art and is not the inventive scheme of the present application, and will not be described here.

[0168] Further, the stream tracing algorithm can simulate the flow path of water in the oil cavity according to the direction of the gradient vector, and intuitively show the trend and channel of water diffusion from the bottom of the cavity upward and around; the generation of the water intrusion path map enables the staff to clearly master the specific path of water intrusion.

[0169] According to the water intrusion path map, a potential leakage area is marked based on a preset water rising rate threshold, comprising:

[0170] F1: the water analysis processor analyzes the generated water intrusion path map and identifies one or more core intrusion paths with the maximum diffusion rate gradient module length in the map;

[0171] Further, the diffusion rate gradient module length reflects the size of the diffusion rate gradient, and the larger the module length, the more violent the change in water diffusion speed on the core intrusion path, which is the main channel of water intrusion; identifying the main diffusion path can focus on the water intrusion channel that poses the greatest threat to the safety of the oil cavity.

[0172] Further, the specific steps of F1 include:

[0173] (1) reading the generated water intrusion path map; the water intrusion path map contains diffusion rate gradient vector data of each point in the three-dimensional space of the oil cavity, and the gradient vector of each point is composed of three components , x , y , z ,

[0174] (2) calculating the diffusion rate gradient module length; the diffusion rate gradient module length is obtained based on the components of the gradient vector using the Euclidean distance calculation formula;

[0175] (3) Statistics the numerical range of the gradient modulus length of the whole domain, and divide it into low modulus length area, medium modulus length area and high modulus length area from low to high, each interval corresponds to different path significance level, wherein the high modulus length area usually corresponds to the active path of water intrusion, and is the key area of path identification;

[0176] (4) The density clustering algorithm is used to cluster the grid nodes in the high modulus length area, and the nodes with continuous space and modulus length higher than the threshold value are aggregated into multiple independent high modulus length areas, each clustering area represents a core zone of a potential intrusion path, and the density and average modulus length of clustering reflect the activity degree of the path. The density clustering algorithm is prior art content in the art and is not the inventive scheme of the present application, and will not be described here;

[0177] (5) In each high modulus length clustering area, the node with the maximum gradient modulus length is selected as the seed point of the path, which is the position with the most drastic change in diffusion rate on the path, and usually corresponds to the vicinity of the source of water intrusion or the key channel;

[0178] (6) Starting from the seed point, the path tracking is performed in the opposite direction of the gradient vector, and the specific manner is that: the node with the minimum included angle and the second largest modulus length in the gradient vector direction of the current node is searched in the neighborhood of the current node, and is taken as the next node of the path; repeat the process until the boundary of the grease cavity or the modulus length falls below the threshold value of the low modulus length area, and a complete potential path is formed;

[0179] (7) For all the extracted potential paths, analyze their spatial intersection relationship, if two paths intersect at any node and the subsequent tracking direction is consistent, merge them into one path; if the path appears branching, i.e. extends in multiple directions from a node, take the branch with the largest modulus length as the main path, and the rest as secondary branches, and finally form a path topology network composed of main paths and secondary branches, wherein the main path is a potential intrusion path;

[0180] (8) For each potential path, calculate the average value of the gradient modulus length of all nodes, and the higher the average value, the higher the overall activity degree of the path;

[0181] (9) Sort the paths from high to low according to the average value of the gradient modulus length, and select the top N paths as the core intrusion paths;

[0182] (10) In the three-dimensional model of the inner wall of the grease cavity, draw the core intrusion paths with thick colored lines, wherein the core intrusion paths adopt bright red color, the secondary paths adopt orange color, and the line width is proportional to the peak value of the path modulus length, which can intuitively distinguish the importance of the paths.

[0183] F2: Extract the diffusion rate data of each segment of the core intrusion path along the identified core intrusion path;

[0184] F3: comparing the diffusion rate of each segment of the core invasion path with a preset moisture rising rate threshold value, if the diffusion rate of any segment of the core invasion path is greater than the preset moisture rising rate threshold value, marking the region corresponding to the core invasion path as the potential leakage zone;

[0185] F4: the moisture analysis processor outputs potential leakage zone data containing the position, range and maximum diffusion rate value of the potential leakage zone, and extracts the real-time moisture content signal corresponding to the triggering of the potential leakage zone marking as a moisture monitoring signal.

[0186] The risk assessment unit outputs the leakage risk level through Bayesian network fusion analysis based on the extracted dynamic features and model parameters trained based on historical failure data, including:

[0187] G1: the risk assessment unit acquires the real-time dynamic features extracted by the edge data processing unit and the Bayesian network model parameters trained based on historical failure data; the Bayesian network model parameters include the conditional probability table of the Bayesian network, wherein the Bayesian network model is a prior art in the field and is not the inventive scheme of the present application, and will not be described here;

[0188] G2: the risk assessment unit inputs the real-time dynamic features as evidence into the corresponding input nodes of the Bayesian network model, and starts the probability reasoning process in combination with the built-in conditional probability table;

[0189] G3: the Bayesian network performs probability reasoning based on the input real-time dynamic features and the conditional probability table, and updates the posterior probability of all nodes in the Bayesian network;

[0190] G4: calculate the probability of each risk level corresponding to the output nodes of the Bayesian network through the reasoning result, and take the highest probability level as the final output result; the levels include low risk, medium risk and high risk.

[0191] The embodiments of the present application are described above in combination with the drawings, but the present application is not limited to the above specific embodiments, and the above specific embodiments are only illustrative and not limiting, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments without departing from the purpose of the present application and the scope of protection, which are all within the protection of the present application.

Claims

1. A real-time detection system for a shield tail sealing grease cavity, characterized in that, The application relates to a multi-physical field monitoring module and an intelligent early warning analysis module. The multi-physical field monitoring module comprises a pressure difference monitoring unit and a moisture monitoring unit. The intelligent early warning analysis module comprises an edge data processing unit and a risk assessment unit. The pressure difference monitoring unit collects oil pressure data in an oil cavity and external environment water pressure data, constructs a pressure difference dynamic model, generates a pressure difference space-time distribution graph, identifies a pressure imbalance point based on a circumferential center of the oil cavity, calculates a pressure difference difference value between the pressure imbalance point and a circumferentially symmetrical position point, generates a pressure difference abnormal distribution graph, locates a pressure weak area based on a fluid mechanics principle, marks a high-risk leakage area based on a preset threshold value, and outputs high-risk leakage area data and a corresponding pressure monitoring signal. The moisture monitoring unit collects moisture data at the bottom of the oil cavity, constructs a moisture diffusion model by using a support vector machine algorithm, and outputs a moisture three-dimensional distribution graph, then establishes a three-dimensional coordinate system with the bottom center of the oil cavity as the origin, calculates diffusion rate gradients of all monitoring points, generates a moisture intrusion path graph, marks a potential leakage area based on a preset moisture rising rate threshold value according to the moisture intrusion path graph, and outputs potential leakage area data and a corresponding moisture monitoring signal. The edge data processing unit extracts dynamic characteristics by using an isolation forest algorithm according to the high-risk leakage area data, the pressure monitoring signal, the potential leakage area data and the moisture monitoring signal. The dynamic characteristics include a pressure difference change rate and a moisture rising rate. The risk assessment unit outputs a leakage risk grade based on the extracted dynamic characteristics and model parameters trained by combining historical fault data, and outputs a leakage risk grade through Bayesian network fusion analysis. When three pieces of evidence, namely, sudden viscosity reduction of oil, sudden moisture rise and pressure difference overrun, are detected in the fusion analysis process, a secondary early warning is triggered. The pressure difference monitoring unit comprises two types of sensors.

2. The real-time detection system for the grease cavity of the shield tail seal as claimed in claim 1, wherein, The first type of sensor is a pressure sensor array arranged uniformly on the inner wall of the oil cavity in at least two circles. One circle close to the tail brush is a first pressure array for collecting oil pressure data in the oil cavity, and the other circle is a second pressure array for assisting in verifying pressure distribution uniformity. The second type of sensor is an external water pressure sensor group embedded on the outer side of the tail shell corresponding to the position of the oil cavity, for collecting external environment water pressure data. The first type of sensor and the second type of sensor are connected with a pressure data acquisition instrument. The pressure data acquisition instrument is used for converting pressure signals collected by the sensors into digital pressure signals, and completing time synchronization uploading of the cavity pressure and the external water pressure data. The process of constructing a pressure difference dynamic model and generating a pressure difference space-time distribution graph is as follows:

3. The real-time detection system for the grease cavity of the shield tail seal as claimed in claim 2, characterized in that, The pressure sensor array of the pressure difference monitoring unit and the external water pressure sensor group synchronously collect data, generate oil pressure data in the oil cavity and external environment water pressure data respectively, and transmit the data to a pressure data acquisition instrument. The pressure data acquisition instrument calibrates the time stamp of the oil pressure data in the oil cavity and the external environment water pressure data, converts the data into digital pressure signals, and transmits the digital pressure signals to a pressure difference analysis processor. ​ The pressure difference analysis processor receives the synchronized digital pressure signals through a data interface, analyzes the measured grease pressure value of each cavity internal pressure sensor node and the external measured water pressure value of the corresponding position, and calculates the measured pressure difference of each spatial position; the measured pressure difference is the difference between the external measured water pressure value and the measured grease pressure value; Based on the pressure transmission theory of fluid mechanics, the measured pressure difference data in time series is taken as an input parameter, the measured pressure difference data at continuous multiple time points are associated with the corresponding spatial position information, a time series analysis algorithm is used to fit the dynamic change trend of the measured pressure difference, a pressure difference dynamic model is established, and the theoretical pressure difference values of each point of the grease cavity at different time points are output; Meanwhile, the measured pressure difference of each node is compared with the theoretical pressure difference value of the corresponding position at the same real-time pressure data calculated by the pressure difference dynamic model, and the measured and theoretical pressure difference deviations of each node are obtained; Taking the spatial coordinate information and the corresponding measured and theoretical pressure difference deviation data of all sensor nodes as inputs, a three-dimensional interpolation algorithm is used for data fusion and spatial estimation, and the pressure difference deviation conditions of all positions on the inner wall of the grease cavity are deduced, and the spatial distribution and time dynamic change of the pressure difference deviation are combined to convert the pressure difference spatiotemporal distribution map; the pressure difference spatiotemporal distribution map reflects the pressure difference abnormal distribution and evolution trend of the entire grease cavity at different time points.

4. The real-time detection system for the grease cavity of the shield tail seal as claimed in claim 3, wherein, The pressure imbalance point is identified based on the circumferential center of the grease cavity, the pressure difference difference value between the pressure imbalance point and the circumferentially symmetrical position point is calculated, and the pressure difference abnormal distribution map is generated, including: The pressure difference analysis processor performs 360-degree circumferential scanning on the pressure difference spatiotemporal distribution map based on the circumferential center of the grease cavity, and obtains the position information and the measured and theoretical pressure difference deviation of all sensor nodes in the pressure difference spatiotemporal distribution map; If the absolute error of the measured and theoretical pressure difference deviation of any sensor node is greater than a first preset deviation threshold, the node is marked as a primary pressure imbalance point; For the marked primary pressure imbalance point, the pressure difference analysis processor determines the symmetrical position point thereof with respect to the circumferential center of the grease cavity, obtains the measured pressure difference of the symmetrical position point, and then calculates the absolute error of the measured pressure difference between the primary pressure imbalance point and the symmetrical position point to obtain a measured symmetrical pressure difference difference value; If the measured symmetrical pressure difference difference value is greater than a second preset symmetrical deviation threshold, the primary pressure imbalance point is confirmed as a final pressure imbalance point; The pressure difference analysis processor collects all final pressure imbalance points and their corresponding measured symmetrical pressure difference difference values, calculates the pressure difference abnormality coefficients of the final pressure imbalance points, and generates the pressure difference abnormal distribution map by the contour method.

5. The real-time detection system for the grease cavity of the shield tail seal as claimed in claim 4, characterized in that, The pressure weak area is located based on the principle of fluid mechanics, and a high-risk leakage area is marked based on a preset threshold, including: The pressure difference analysis processor analyzes the correlation between the pressure difference anomaly coefficient and the fluid permeation risk based on the principle of fluid mechanics and in combination with the generated pressure difference anomaly distribution map, and locates the pressure weak area of the oil cavity; the pressure weak area is defined as a continuous area with a pressure difference anomaly coefficient greater than 1.2, i.e., an area where the external water pressure is significantly higher than the oil pressure in the cavity; The located pressure weak area is analyzed, and the area with a pressure difference anomaly coefficient greater than a preset pressure difference anomaly threshold value in the pressure difference anomaly distribution map or the area with a measured symmetrical pressure difference greater than a preset pressure difference threshold value is marked as the high-risk leakage area, and high-risk leakage area data including the position, area, maximum pressure difference anomaly coefficient value and maximum measured symmetrical pressure difference of the high-risk leakage area are output, and at the same time, the real-time digital pressure signal corresponding to the triggering of the high-risk leakage area marking is extracted as a pressure monitoring signal and output; the preset pressure difference threshold value is determined according to the oil resistance permeation grade of the shield tail.

6. The real-time detection system for the grease cavity of the shield tail seal as claimed in claim 5, wherein, The water monitoring unit includes a plurality of high-frequency capacitive moisture sensors arranged in the deposition area at the bottom of the oil cavity; the high-frequency capacitive moisture sensors are connected with a data collector, and the data collector is used to receive the moisture content digital signal collected by the sensors and upload the signal.

7. The real-time detection system for the grease cavity of the shield tail seal as claimed in claim 6, wherein, The water monitoring unit collects moisture data at the bottom of the oil cavity, constructs a moisture diffusion model using a support vector machine algorithm, and outputs a moisture three-dimensional distribution map, including: The sensors of the water monitoring unit collect moisture content data of a plurality of monitoring points at the bottom of the oil cavity, and transmit the data to the data collector after being converted into moisture content digital signals; The data collector transmits the moisture content digital signals to the water analysis processor; The water analysis processor takes the values corresponding to the received moisture content digital signals as training samples, takes the coordinates of each monitoring point in a three-dimensional coordinate system established with the center of the bottom of the oil cavity as the origin as input features, and takes the moisture content values of the corresponding monitoring points as output labels, trains and learns using a support vector machine regression algorithm, and constructs a moisture diffusion model for predicting the moisture content of any point in the oil cavity; The water analysis processor performs three-dimensional grid decomposition on the oil cavity space based on the moisture diffusion model, predicts the moisture content of each point in the grid through the moisture diffusion model, and generates the moisture three-dimensional distribution map.

8. The real-time detection system for a shield tail sealing grease cavity according to claim 7, wherein, The three-dimensional coordinate system is established with the center of the bottom of the oil cavity as the origin, the diffusion rate gradient of each monitoring point is calculated, and a moisture intrusion path map is generated, including: The water analysis processor calls the established three-dimensional coordinate system with the center of the bottom of the oil cavity as the origin; Based on the moisture three-dimensional distribution maps of a plurality of continuous time series, the moisture content data of each monitoring point at different time points is extracted, and the moisture content change rate of each monitoring point per unit time, i.e., the diffusion rate, is calculated; In combination with the spatial coordinates of each monitoring point and the corresponding diffusion rate, the diffusion rate gradient of each monitoring point on its spatial coordinates is calculated, and a gradient vector is obtained; According to the gradient vectors of all monitoring points, a streamline tracking algorithm is used to simulate the direction and path of water movement, and a water intrusion path atlas is generated; the water intrusion path atlas indicates the trend and channel of water spreading from the bottom of the cavity upward and around.

9. The real-time detection system for a shield tail sealing grease cavity according to claim 8, wherein, According to the water intrusion path atlas, a potential leakage area is marked based on a preset water rise rate threshold, including: The water analysis processor analyzes the generated water intrusion path atlas, and identifies one or more core intrusion paths with the maximum diffusion rate gradient module length in the atlas; Along the identified core intrusion path, diffusion rate data of each segment of the core intrusion path is extracted; The diffusion rate of each segment of the core intrusion path is compared with the preset water rise rate threshold, and if the diffusion rate of any segment of the core intrusion path is greater than the preset water rise rate threshold, the area corresponding to the core intrusion path is marked as the potential leakage area; The water analysis processor outputs potential leakage area data containing the location, range, and maximum diffusion rate value of the potential leakage area, and extracts the real-time water content signal corresponding to the triggering of the potential leakage area marking as the water monitoring signal.

10. The real-time detection system for a shield tail sealing grease cavity according to claim 9, wherein, The risk assessment unit outputs a leakage risk level through Bayesian network fusion analysis based on the model parameters trained based on the extracted dynamic features and historical failure data, including: The risk assessment unit obtains the real-time dynamic features extracted by the edge data processing unit, and the Bayesian network model parameters trained based on historical failure data; the Bayesian network model parameters include the conditional probability table of the Bayesian network; The risk assessment unit inputs the real-time dynamic features as evidence into the corresponding input nodes of the Bayesian network model, and starts the probability reasoning process in combination with the built-in conditional probability table; The Bayesian network performs probability reasoning according to the input real-time dynamic features and the conditional probability table, and updates the posterior probability of all nodes in the Bayesian network; Through the inference result, the probability of each risk level corresponding to the output node of the Bayesian network is calculated, and the level with the highest probability is taken as the final output result; the levels include low risk, medium risk, and high risk.

Images