A method, system and terminal for predicting blockage based on a shield mud cake determination graph

Abstract

The application discloses a kind of based on shield mud cake determination graph jam prediction method, system and terminal, the method includes: obtaining target shield mud cake determination graph, mathematical modeling is carried out, generates two-dimensional coordinate space;The area proportion of each risk area in target shield mud cake determination graph is calculated, according to all area proportions, generate multiple sample points in two-dimensional coordinate space;According to the sample point label assignment requirement generated by target shield mud cake determination graph, according to sample point label assignment requirement, each sample point is divided into classification label, and generates sample data, according to sample data training target model;When carrying out real-time jam prediction, the liquid limit of slag soil in target environment, plastic limit and moisture content are acquired, pretreated, input into trained target model, and output jam prediction classification label.The application can obtain the data for training the model for identifying the mud cake risk of shield machine, so as to train the model to predict the mud cake risk of shield machine.

Description

TECHNICAL FIELD

[0001] The present application relates to a data generation and processing method, in particular to a blockage prediction method, system, terminal and computer readable storage medium based on a shield mud cake determination graph. BACKGROUND

[0002] In engineering practice, many key experience knowledge is usually stored in the form of analytical expressions and graphs. The experience formula, experience graph and classification standard commonly used in engineering design are convenient for intuitive presentation and quick reference, but due to the limitation of its form, the knowledge contained has not been fully utilized, which limits its application in intelligent analysis and data-driven model.

[0003] And in the current engineering, the shield method is a main construction method widely used in current tunnel construction. In the construction process, the "mud cake" problem is one of the common and important problems in shield tunneling, which may lead to the decrease of tunneling efficiency and the increase of cutter wear, and even cause serious consequences such as construction interruption. Therefore, accurate prediction of mud cake risk is crucial to ensure the efficiency and safety of shield construction.

[0004] However, at present, when predicting the risk of mud cake, due to the lack of samples, the prediction is inaccurate, which affects the safety of shield construction.

[0005] Therefore, the prior art still needs to be improved and developed. SUMMARY

[0006] The main purpose of the present application is to provide a blockage prediction method, system, terminal and computer readable storage medium based on a shield mud cake determination graph, which aims to solve the problem that in the prior art, when predicting the risk of mud cake, due to the lack of samples, the prediction is inaccurate, which affects the safety of shield construction.

[0007] To achieve the above purpose, the present application provides a blockage prediction method based on a shield mud cake determination graph, which comprises the following steps:

[0008] Obtain a target shield mud cake determination graph, mathematically model the target shield mud cake determination graph, and generate a two-dimensional coordinate space;

[0009] Calculate the area proportion of each risk area in the target shield mud cake determination graph, and generate a plurality of sample points in the two-dimensional coordinate space according to all the area proportions;

[0010] According to the sample point label assignment requirement generated according to the target shield mud cake judgment graph, a corresponding classification label is divided for each sample point, sample data is generated, a target model is trained according to the sample data, and a trained target model is obtained;

[0011] When real-time jam prediction is performed, the liquid limit, plastic limit and water content of the muck in the target environment are obtained, and after preprocessing the liquid limit, plastic limit and water content of the muck in the target environment, the trained target model is input for processing, and a jam prediction classification label is output.

[0012] Optionally, the target shield mud cake judgment graph is obtained, the target shield mud cake judgment graph is mathematically modeled, and a two-dimensional coordinate space is generated, specifically including:

[0013] A target shield mud cake judgment graph is obtained.

[0014] The numerical range of the difference between the plastic limit and the water content and the numerical range of the difference between the liquid limit and the water content in the target shield mud cake judgment graph are extracted.

[0015] The numerical boundary is determined according to the corresponding numerical range, and the two-dimensional coordinate space is generated according to the numerical boundary.

[0016] Optionally, the area proportion of each risk area in the target shield mud cake judgment graph is calculated, and a plurality of sample points are generated in the two-dimensional coordinate space according to all the area proportions, specifically including:

[0017] The area proportion of each risk area is calculated according to the range of each risk area in the target shield mud cake judgment graph.

[0018] The total number of preset sample points is obtained, and a preset number of sample points are generated in the two-dimensional coordinate space by using the scatter point method according to all the area proportions.

[0019] Optionally, the sample point label assignment requirement generated according to the target shield mud cake judgment graph specifically includes:

[0020] The target shield mud cake judgment graph is assigned according to the two-dimensional coordinate space, and a mathematical expression of the boundary of each risk area is constructed.

[0021] According to each of the mathematical expressions, the coordinate relationship of the points in each risk area is calculated, and according to each of the coordinate relationships, a sample point label assignment requirement is generated.

[0022] Optionally, the sample point label assignment requirement generated according to the target shield mud cake judgment graph specifically includes:

[0023] obtaining the coordinates of each sample point in the two-dimensional coordinate space, and dividing a corresponding classification label for each sample point according to the coordinates of each sample point and the sample point label assignment requirement;

[0024] generating sample data according to the classification label and the coordinates of each sample point, training a target model according to the sample data, and obtaining a trained target model.

[0025] Optionally, the generating sample data according to the classification label and the coordinates of each sample point, training a target model according to the sample data, and obtaining a trained target model, specifically includes:

[0026] generating multiple groups of data according to the classification label and the coordinates of each sample point, and obtaining the sample data by aggregating all the data;

[0027] inputting the coordinates of each group of data in the sample data into the target model, optimizing the target model according to the output of the target model and the classification label of each group of data, ending the training when a training condition is reached, and obtaining a trained target model.

[0028] Optionally, when performing real-time jam prediction, the liquid limit, plastic limit and water content of the slag soil in the target environment are obtained, and after preprocessing the liquid limit, plastic limit and water content of the slag soil in the target environment, the trained target model is inputted for processing, and a jam prediction classification label is outputted, specifically including:

[0029] When performing real-time jam prediction, the liquid limit, plastic limit and water content of the slag soil in the target environment are obtained, and the difference between the current liquid limit and water content is calculated according to the liquid limit and water content of the slag soil in the target environment, and the difference between the current plastic limit and water content is calculated according to the plastic limit and water content of the slag soil in the target environment;

[0030] generating a current coordinate according to the difference between the current liquid limit and water content and the difference between the current plastic limit and water content;

[0031] inputting the current coordinate into the trained target model, and outputting a jam prediction classification label.

[0032] In addition, to achieve the above-mentioned purpose, the application also provides a jam prediction system based on a shield mud cake determination graph, wherein the jam prediction system based on the shield mud cake determination graph comprises:

[0033] a two-dimensional coordinate space generation module, configured to obtain a target shield mud cake determination graph, perform mathematical modeling on the target shield mud cake determination graph, and generate a two-dimensional coordinate space;

[0034] The sample point generation module is configured to calculate area proportions of each risk region in the target shield mud cake judgment graph, and generate a plurality of sample points in the two-dimensional coordinate space according to all the area proportions.

[0035] The training module is configured to generate a sample point label assignment requirement according to the target shield mud cake judgment graph, divide a corresponding classification label for each sample point according to the sample point label assignment requirement, generate sample data, train a target model according to the sample data, and obtain a trained target model.

[0036] The application module is configured to, when performing real-time jam prediction, acquire a liquid limit, a plastic limit and a water content of the muck in a target environment, input the liquid limit, the plastic limit and the water content of the muck in the target environment after preprocessing to the trained target model for processing, and output a jam prediction classification label.

[0037] In addition, to achieve the above object, the application further provides a terminal, wherein the terminal comprises a memory, a processor and a shield mud cake judgment graph-based jam prediction program stored in the memory and executable on the processor, and the shield mud cake judgment graph-based jam prediction program implements the steps of the shield mud cake judgment graph-based jam prediction method when executed by the processor.

[0038] In addition, to achieve the above object, the application further provides a computer readable storage medium, wherein the computer readable storage medium stores a shield mud cake judgment graph-based jam prediction program, and the shield mud cake judgment graph-based jam prediction program implements the steps of the shield mud cake judgment graph-based jam prediction method when executed by a processor.

[0039] In the present application, the target shield mud cake judgment graph is obtained, the target shield mud cake judgment graph is mathematically modeled to generate a two-dimensional coordinate space; the area proportion of each risk area in the target shield mud cake judgment graph is calculated, a plurality of sample points are generated in the two-dimensional coordinate space according to all the area proportions; a sample point label assignment requirement is generated according to the target shield mud cake judgment graph, a corresponding classification label is divided for each sample point according to the sample point label assignment requirement, and sample data is generated, a target model is trained according to the sample data, and a trained target model is obtained; when real-time jamming prediction is performed, the liquid limit, plastic limit and water content of the muck in the target environment are obtained, and after preprocessing the liquid limit, plastic limit and water content of the muck in the target environment, the trained target model is input for processing, and a jamming prediction classification label is output. The knowledge in the shield mud cake judgment graph is converted into data form, then a large number of mud cake sample data with low fidelity characteristics can be constructed based on the graph, these data are used as quantitative expression of experience knowledge, the problem of insufficient mud cake samples in actual engineering is solved, thereby the quantity and quality of the mud cake sample database can be significantly improved, a mud cake risk prediction model with high prediction accuracy, i.e. the trained target model, is trained to predict the shield mud cake risk and ensure the safety of shield construction. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 is a flowchart of a preferred embodiment of the shield mud cake judgment graph based jamming prediction method of the present application;

[0041] Figure 2 is a schematic diagram of a shield mud cake judgment graph in the shield mud cake judgment graph based jamming prediction method of the present application;

[0042] Figure 3 is a schematic diagram of sample point distribution in a two-dimensional coordinate space in the shield mud cake judgment graph based jamming prediction method of the present application;

[0043] Figure 4 is a flowchart of a preferred embodiment of the shield mud cake judgment graph based jamming prediction method of the present application;

[0044] Figure 5 is a structural diagram of a preferred embodiment of the shield mud cake judgment graph based jamming prediction system of the present application;

[0045] Figure 6 is a structural diagram of a preferred embodiment of the terminal of the present application. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical scheme and advantages of the present application clearer and more explicit, the present application is further described in detail below with reference to the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application and not to limit the present application.

[0047] In engineering practice, many key empirical knowledge is usually stored in the form of analytical expressions, charts, etc. The empirical formulas, empirical charts and classification criteria commonly used in engineering design are convenient for intuitive presentation and quick reference, but due to the limitation of their form, the knowledge contained has not been fully utilized, limiting their application in intelligent analysis and data-driven models. In the current engineering, the shield method is a main construction method widely used in current tunnel construction. In the construction process, the "mud cake" problem is one of the common and important problems in shield tunneling, which may lead to decreased tunneling efficiency, increased cutter wear, and even serious consequences such as construction interruption. Therefore, accurate prediction of mud cake risk is crucial to ensure the efficiency and safety of shield construction. However, when predicting mud cake risk based on transfer learning or multi-fidelity modeling methods, the knowledge in the empirical chart cannot be combined, resulting in a lack of samples, limiting the performance of the model, and therefore the precise prediction of shield mud cake risk cannot be achieved.

[0048] In view of one or more of the above problems, the present application obtains a target shield mud cake determination chart, mathematically models the target shield mud cake determination chart, and generates a two-dimensional coordinate space; calculates the area proportion of each risk area in the target shield mud cake determination chart, generates a plurality of sample points in the two-dimensional coordinate space according to all the area proportions; generates sample point label assignment requirements according to the target shield mud cake determination chart, divides the corresponding classification label for each sample point according to the sample point label assignment requirements, and generates sample data, trains a target model according to the sample data, and obtains a trained target model; when performing real-time jam prediction, the liquid limit, plastic limit and water content of the target environment are obtained, and the liquid limit, plastic limit and water content of the target environment are preprocessed and input into the trained target model for processing, and the jam prediction classification label is output.

[0049] The jam prediction method based on the shield mud cake determination chart according to the preferred embodiment of the present application, as shown in Figure 1 The jam prediction method based on the shield mud cake determination chart includes the following steps:

[0050] Step S10, obtaining a target shield mud cake determination chart, mathematically modeling the target shield mud cake determination chart, and generating a two-dimensional coordinate space.

[0051] It should be noted that in the present application, the shield mud cake determination chart is an empirical chart for mud cake risk discrimination in shield construction, as shown in Figure 2As shown, it is a shield mud cake determination graph, which takes the difference between plastic limit and water content (WP-Wn) and the difference between liquid limit and water content (WL-Wn) as the coordinate axis, and classifies different slag conditions through the consistency index, wherein the consistency index adopted is the viscosity index (Ic) of the soil body, wherein Ip in the graph represents the plasticity index of the soil body. Through the target shield mud cake determination graph, a two-dimensional coordinate space is generated.

[0052] Further, the target shield mud cake determination graph is obtained, the target shield mud cake determination graph is mathematically modeled, and a two-dimensional coordinate space is generated, specifically comprising:

[0053] Obtaining a target shield mud cake determination graph;

[0054] Extracting the numerical range of the difference between the plastic limit and the water content and the numerical range of the difference between the liquid limit and the water content in the target shield mud cake determination graph;

[0055] Determine the numerical boundary according to the corresponding numerical range, and generate the two-dimensional coordinate space according to the numerical boundary.

[0056] Specifically, in the present application, after obtaining the target shield mud cake determination graph, the axis information presented in the target shield mud cake determination graph is converted into explicit numerical range and scale, the numerical range of the horizontal axis (the difference between the plastic limit and the water content) and the vertical axis (the difference between the liquid limit and the water content) in the target shield mud cake determination graph is quantified, and the numerical range is obtained; as Figure 2 As shown, the corresponding horizontal axis range is [-160%, 50%], and the vertical axis range is [-20%, 200%]. After obtaining the numerical range, the maximum and minimum values of the difference between the plastic limit and the water content are determined, and this range is taken as the numerical boundary of the horizontal axis; the maximum and minimum values of the difference between the liquid limit and the water content are determined, and this range is taken as the numerical boundary of the vertical axis. The two-dimensional coordinate space is generated by the corresponding numerical boundary, which is a plane space composed of horizontal coordinates, vertical coordinates and numerical boundaries, wherein the horizontal axis and the vertical axis are the difference between the plastic limit and the water content and the difference between the liquid limit and the water content, respectively.

[0057] It should be noted that for the soil, the liquid limit and the plastic limit are different water content stages of the soil, and the liquid limit water content is greater than the plastic limit water content, so the vertical coordinate of all the sample points generated thereafter is greater than the horizontal coordinate.

[0058] Step S20, calculating the area proportion of each risk area in the target shield mud cake determination graph, and generating a plurality of sample points in the two-dimensional coordinate space according to all the area proportions.

[0059] Specifically, in the present application, the area of each risk region in the target shield mud cake judgment graph is obtained, so that the corresponding scatter method is used to generate sample points, and the corresponding sample data is constructed by the sample points.

[0060] Further, the area proportion of each risk region in the target shield mud cake judgment graph is calculated, and a plurality of sample points are generated in the two-dimensional coordinate space according to all the area proportions, specifically including:

[0061] According to the range of each risk region in the target shield mud cake judgment graph, the area proportion of each risk region is calculated;

[0062] The total number of preset sample points is obtained, and a preset number of sample points is generated in the two-dimensional coordinate space by using the scatter method according to all the area proportions.

[0063] Specifically, in the target shield mud cake judgment graph, five risk regions are included, which are: well-dispersed region, slight blockage region, serious blockage region, moderate blockage region and cluster region, and the corresponding area proportion is calculated according to the range of each risk region. As shown in Figure 2 The shield mud cake judgment graph in Figure 2 The area of each risk region is calculated, and the area of the well-dispersed region is:

[0064] ;

[0065] The area of the slight blockage region is:

[0066] ;

[0067] The area of the moderate blockage region is:

[0068] ;

[0069] The area of the cluster region is:

[0070] ;

[0071] The area of the serious blockage region is:

[0072] ;

[0073] The area proportion of each risk area can be obtained from the area correspondence, and then the total number of preset sample points is set, and the corresponding five risk area proportions are used to generate a preset number of sample points in a two-dimensional coordinate space, wherein the sample points are randomly generated by the scattering method in the present application. That is, in the present application, a preset number of sample points are randomly generated in a two-dimensional coordinate space according to the area proportion by the scattering method, and the coordinate values of each sample point in the two-dimensional coordinate space are generated by a random function, and the random function is a standard uniform distribution pseudo-random number generator. The random number generator is a general term and can include two types of true random number generator and pseudo-random number generator. In an embodiment of the present application, the pseudo-random number generator is preferably used, which generates a numerical sequence with random statistical characteristics through an algorithm, but the generation process is deterministic. Compared with the pseudo-random number generator, the true random number generator depends on physical phenomena to generate numbers, which has true randomness.

[0074] In Figure 2 , the area ratio of the corresponding five risk areas is 30:128:144.8:66.67:87.5, so the corresponding sample points are generated in a two-dimensional coordinate space, as shown in Figure 3 , when the total number of preset sample points is set to 10000, the number of sample points generated in each area is 656, 2801, 3169, 1459 and 1915, respectively.

[0075] Step S30, generating sample point label assignment requirements according to the target shield mud cake judgment graph, dividing the corresponding classification label for each sample point according to the sample point label assignment requirements, and generating sample data, training the target model according to the sample data, and obtaining the trained target model.

[0076] Specifically, in the present application, the corresponding label is generated for each sample point after obtaining the sample point, wherein the label of each sample point is one of well dispersed, slightly blocked, seriously blocked, moderately blocked and lumped.

[0077] Further, the sample point label assignment requirements generated according to the target shield mud cake judgment graph specifically include:

[0078] The two-dimensional coordinate space is used to assign the target shield mud cake judgment graph, and the mathematical expression of the boundary of each risk area is constructed;

[0079] According to each of the mathematical expressions, the coordinate relationship of the points in each risk area is calculated, and the sample point label assignment requirements are generated according to each of the coordinate relationships.

[0080] Specifically, in the present application, each sample point can be expressed in the form of specific numerical values (x, y) in a two-dimensional coordinate space, so that the regions and boundary conditions in the two-dimensional coordinate space can be described and sampled by numerical methods, wherein x represents the horizontal axis of the two-dimensional coordinate space, and y represents the vertical axis of the two-dimensional coordinate space. The boundary of each risk region is the demarcation line of the viscous index Ic, which can be expressed as y=0%, y=-x, y=-3x and x=0% in mathematics. Figure 3

[0081] The corresponding mathematical expression can obtain the conditions that the sample points in each risk region need to satisfy, that is, the coordinate relationship.

[0082] As Figure 3 indicated in Figure 3 , the corresponding coordinate relationship is the sample point position in Table 1, and the corresponding label of each coordinate range can be divided according to the coordinate relationship.

[0083] Table 1: Sample point label assignment requirements

[0084]

[0085] The sample points in Figure 3 can be classified according to the corresponding classification labels by Table 1.

[0086] Further, according to the sample point label assignment requirements, the corresponding classification labels of each sample point are divided, and sample data is generated, and a target model is trained according to the sample data to obtain a trained target model, which specifically includes:

[0087] Obtain the coordinates of each sample point in the two-dimensional coordinate space, and according to the coordinates of each sample point, divide the corresponding classification labels of each sample point according to the sample point label assignment requirements;

[0088] According to the classification labels and the coordinates of each sample point, generate sample data, and train a target model according to the sample data to obtain a trained target model.

[0089] Specifically, after obtaining the sample point label assignment requirements, the classification labels of each sample point can be divided according to the coordinates of each sample point in the two-dimensional coordinate space, and the sample data can be generated through the classification labels and the corresponding coordinates.

[0090] Further, according to the classification labels and the coordinates of each sample point, generate sample data, and train a target model according to the sample data to obtain a trained target model, which specifically includes:

[0091] ​According to the classification label and the coordinates of each sample point, a plurality of groups of data are generated, and the sample data is obtained by aggregating all the data;

[0092] The coordinates of each group of data in the sample data are input into the target model, and the target model is optimized according to the output of the target model and the classification label of each group of data. When the training condition is reached, the training is ended, and the trained target model is obtained.

[0093] Specifically, a group of data can be generated according to the classification label and the coordinates of each sample point, and the sample data is obtained by aggregating all the data. The target model is trained by using the sample data. The target model is a model that can perform transfer learning or multi-fidelity modeling method.

[0094] In the training process, the coordinates are input into the model, and the classification label of the data in the sample data is compared with the output of the model to optimize the model parameters. When the training reaches the training condition, the training is ended, and the trained target model is obtained. The training condition is that the model accuracy meets the requirement or the model training times reaches the preset training times.

[0095] In an embodiment of the present application, the neural network can be trained according to the sample data, the model hyperparameters are found through Bayesian optimization, the structure of the model is set to a framework of 2 (input features, liquid limit-water content and plastic limit-water content) x 16 x 32 x 1 (blockage prediction classification label), the learning rate is 0.001, and the trained target model is obtained.

[0096] In step S40, when real-time blockage prediction is performed, the liquid limit, plastic limit and water content of the slag soil in the target environment are obtained, and after preprocessing the liquid limit, plastic limit and water content of the slag soil in the target environment, the preprocessed liquid limit, plastic limit and water content of the slag soil in the target environment are input into the trained target model for processing, and a blockage prediction classification label is output.

[0097] Specifically, in the present application, when the trained target model is obtained, in addition to being applied in actual working conditions, for engineering design optimization, the change trend of the mud cake blockage risk can be predicted in the design stage by simulating different construction parameter combinations (such as shield machine operation parameters, construction process, etc.). Based on the risk assessment results, the scheme with the lowest risk and the highest efficiency can be quickly screened out, realizing the organic combination of design and risk management, and improving the engineering safety and economic benefits.

[0098] Further, when real-time blockage prediction is performed, the liquid limit, plastic limit and water content of the slag soil in the target environment are obtained, and after preprocessing the liquid limit, plastic limit and water content of the slag soil in the target environment, the preprocessed liquid limit, plastic limit and water content of the slag soil in the target environment are input into the trained target model for processing, and a blockage prediction classification label is output. Specifically, it comprises:

[0099] In real-time clogging prediction, the liquid limit, plastic limit and water content of the slag in the target environment are obtained, the current liquid limit and water content difference is calculated according to the liquid limit and water content of the slag in the target environment, and the current plastic limit and water content difference is calculated according to the plastic limit and water content of the slag in the target environment;

[0100] According to the current liquid limit and water content difference and the current plastic limit and water content difference, the current coordinate is generated;

[0101] The current coordinate is input into the trained target model, and a clogging prediction classification label is output.

[0102] Specifically, in the present application, in actual working conditions, the liquid limit, plastic limit and water content of the slag in the target environment are obtained, the current liquid limit and water content difference is calculated according to the liquid limit and water content, and the current plastic limit and water content difference is calculated according to the plastic limit and water content, so as to be input into the target model as coordinates, and the model can directly output the classification label by using the information of the target shield mud cake determination graph absorbed during training, to realize real-time risk assessment.

[0103] Further, in the present application, by Figure 4 Further describe the process of the present application, in the present application, for the target shield mud cake determination graph, first mathematical modeling is carried out, then the coordinate axis range of the graph is extracted to determine the boundary condition, and then the sample points are generated, and the classification label of each sample point is determined, and then the data is structured, that is, the sample data is obtained, and the corresponding target model is trained to realize the mud cake risk prediction in actual working conditions.

[0104] The application obtains a target shield mud cake judgment graph, performs mathematical modeling on the target shield mud cake judgment graph, and generates a two-dimensional coordinate space; the area proportion of each risk area in the target shield mud cake judgment graph is calculated, a plurality of sample points are generated in the two-dimensional coordinate space according to all the area proportions; a sample point label assignment requirement is generated according to the target shield mud cake judgment graph, a corresponding classification label is divided for each sample point according to the sample point label assignment requirement, sample data is generated, a target model is trained according to the sample data, and a trained target model is obtained; when real-time jam prediction is performed, the liquid limit, plastic limit and water content of the muck in the target environment are obtained, and after preprocessing the liquid limit, plastic limit and water content of the muck in the target environment, the trained target model is input for processing, and a jam prediction classification label is output. The knowledge in the shield mud cake judgment graph is converted into data form, so a large number of mud cake sample data with low fidelity characteristics can be constructed based on the graph, these data are used as quantitative expression of experience knowledge, the problem of insufficient mud cake samples in actual engineering is solved, so the quantity and quality of the mud cake sample database can be significantly improved, and a mud cake risk prediction model with high prediction accuracy, that is, the trained target model, is trained to predict the shield mud cake risk.

[0105] Further, as shown in Figure 5 based on the jam prediction method based on the shield mud cake judgment graph, the application also correspondingly provides a jam prediction system based on the shield mud cake judgment graph, wherein the jam prediction system based on the shield mud cake judgment graph comprises:

[0106] The two-dimensional coordinate space generation module 51 is used to obtain a target shield mud cake judgment graph, perform mathematical modeling on the target shield mud cake judgment graph, and generate a two-dimensional coordinate space;

[0107] The sample point generation module 52 is used to calculate the area proportion of each risk area in the target shield mud cake judgment graph, and generate a plurality of sample points in the two-dimensional coordinate space according to all the area proportions;

[0108] The training module 53 is used to generate a sample point label assignment requirement according to the target shield mud cake judgment graph, divide a corresponding classification label for each sample point according to the sample point label assignment requirement, generate sample data, train a target model according to the sample data, and obtain a trained target model;

[0109] The application module 54 is used to obtain the liquid limit, plastic limit and water content of the muck in a target environment when real-time jam prediction is performed, preprocess the liquid limit, plastic limit and water content of the muck in the target environment, input the trained target model for processing after preprocessing, and output a jam prediction classification label.

[0110] Further, as shown in Figure 6 Further, based on the above shield mud cake determination graph based blockage prediction method and system, the present application also provides a terminal accordingly, which comprises a processor 10, a memory 20 and a display 30. Figure 6 Only some components of the terminal are shown, but it should be understood that all the shown components are not required, and more or less components can be implemented instead.

[0111] The memory 20 can be an internal storage unit of the terminal in some embodiments, such as a hard disk or a memory of the terminal. The memory 20 can also be an external storage device of the terminal in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc. equipped on the terminal. Further, the memory 20 can include both the internal storage unit and the external storage device of the terminal. The memory 20 is used to store application software and various data installed on the terminal, such as program codes of the terminal, etc. The memory 20 can also be used to temporarily store data that has been output or will be output. In an embodiment, the memory 20 stores a shield mud cake determination graph based blockage prediction program 40, which can be executed by the processor 10 to implement the shield mud cake determination graph based blockage prediction method of the present application.

[0112] The processor 10 can be a Central Processing Unit (CPU), a microprocessor or other data processing chip in some embodiments, which is used to run program codes or process data stored in the memory 20, such as to execute the shield mud cake determination graph based blockage prediction method, etc.

[0113] The display 30 can be an LED display, a liquid crystal display, a touch liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, etc. in some embodiments. The display 30 is used to display information of the terminal and to display visualized user interfaces.

[0114] In an embodiment, the above shield mud cake determination graph based blockage prediction method is implemented when the processor 10 executes the shield mud cake determination graph based blockage prediction program 40 in the memory 20.

[0115] The application also provides a computer readable storage medium, wherein the computer readable storage medium stores a blockage prediction program based on a shield mud cake determination graph, and the blockage prediction program based on the shield mud cake determination graph realizes the following steps when executed by a processor:

[0116] obtaining a target shield mud cake determination graph, mathematically modeling the target shield mud cake determination graph, and generating a two-dimensional coordinate space;

[0117] calculating the area proportion of each risk area in the target shield mud cake determination graph, generating a plurality of sample points in the two-dimensional coordinate space according to all the area proportions;

[0118] generating a sample point label assignment requirement according to the target shield mud cake determination graph, dividing a corresponding classification label for each sample point according to the sample point label assignment requirement, generating sample data, training a target model according to the sample data, and obtaining a trained target model;

[0119] when performing real-time blockage prediction, obtaining the liquid limit, plastic limit and water content of the muck in a target environment, preprocessing the liquid limit, plastic limit and water content of the muck in the target environment, inputting the preprocessed liquid limit, plastic limit and water content of the muck in the target environment into the trained target model for processing, and outputting a blockage prediction classification label.

[0120] The obtaining of the target shield mud cake determination graph, the mathematical modeling of the target shield mud cake determination graph, and the generation of the two-dimensional coordinate space specifically include:

[0121] obtaining a target shield mud cake determination graph;

[0122] extracting the numerical range of the difference between the plastic limit and the water content and the numerical range of the difference between the liquid limit and the water content in the target shield mud cake determination graph;

[0123] determining a numerical boundary according to the corresponding numerical range, and generating the two-dimensional coordinate space according to the numerical boundary.

[0124] The calculation of the area proportion of each risk area in the target shield mud cake determination graph and the generation of a plurality of sample points in the two-dimensional coordinate space according to all the area proportions specifically include:

[0125] calculating the area proportion of each risk area in the target shield mud cake determination graph according to the range of each risk area;

[0126] obtaining a preset total number of sample points, and generating a preset number of sample points in the two-dimensional coordinate space according to all the area proportions by using a scatter point method.

[0127] The sample point label assignment requirement according to the target shield mud cake judgment graph, specifically includes:

[0128] The target shield mud cake judgment graph is valued according to the two-dimensional coordinate space, and the mathematical expression of the boundary of each risk area is constructed;

[0129] According to each mathematical expression, the coordinate relationship of the points in each risk area is calculated, and according to each coordinate relationship, a sample point label assignment requirement is generated.

[0130] According to the sample point label assignment requirement, a corresponding classification label is assigned to each sample point, and sample data is generated, and a target model is trained according to the sample data to obtain a trained target model, specifically including:

[0131] The coordinates of each sample point in the two-dimensional coordinate space are obtained, and according to the coordinates of each sample point, a corresponding classification label is assigned to each sample point through the sample point label assignment requirement;

[0132] According to the classification label and the coordinates of each sample point, sample data is generated, and a target model is trained according to the sample data to obtain a trained target model.

[0133] According to the classification label and the coordinates of each sample point, sample data is generated, and a target model is trained according to the sample data to obtain a trained target model, specifically including:

[0134] According to the classification label and the coordinates of each sample point, a plurality of groups of data are generated, and all data are summarized to obtain the sample data;

[0135] The coordinates of each group of data in the sample data are input into the target model, and the target model is optimized according to the output of the target model and the classification label of each group of data, and when the training condition is reached, the training is ended, and a trained target model is obtained.

[0136] In real-time jam prediction, the liquid limit, plastic limit and water content of the muck in the target environment are obtained, and after preprocessing the liquid limit, plastic limit and water content of the muck in the target environment, they are input into the trained target model for processing, and the jam prediction classification label is output, specifically including:

[0137] In real-time jam prediction, the liquid limit, plastic limit and water content of the muck in the target environment are obtained, and according to the liquid limit and water content of the muck in the target environment, the difference between the current liquid limit and water content is calculated, and according to the plastic limit and water content of the muck in the target environment, the difference between the current plastic limit and water content is calculated;

[0138] generate a current coordinate according to a difference between the current liquid limit and the water content and a difference between the current plastic limit and the water content;

[0139] input the current coordinate into the trained target model, and output a blockage prediction classification label.

[0140] It should be noted that, in this paper, the term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or terminal including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes elements inherent to such process, method, article or terminal. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of other identical elements in the process, method, article or terminal including the element.

[0141] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing relevant hardware (such as a processor, a controller, etc.) through a computer program. The program can be stored in a computer-readable computer-readable storage medium, and the program can include the processes of the above-mentioned method embodiments when executed. The computer-readable storage medium can be a memory, a magnetic disc, an optical disc, etc.

[0142] It should be understood that the application is not limited to the above examples, and those skilled in the art can make improvements or changes according to the above description, and all these improvements and changes should belong to the protection scope of the appended claims of the present application.

Claims

1. A method for predicting a blockage based on a shield caking decision map, characterized by, The blockage prediction method based on the shield mud cake judgment graph comprises: acquiring a target shield mud cake judgment graph, mathematically modeling the target shield mud cake judgment graph, and generating a two-dimensional coordinate space; calculating the area proportion of each risk area in the target shield mud cake judgment graph, generating a plurality of sample points in the two-dimensional coordinate space according to all the area proportions; generating sample point label assignment requirements according to the target shield mud cake judgment graph, dividing a corresponding classification label for each sample point according to the sample point label assignment requirements, generating sample data, training a target model according to the sample data, and obtaining a trained target model; during real-time blockage prediction, acquiring the liquid limit, plastic limit and water content of the muck in a target environment, preprocessing the liquid limit, plastic limit and water content of the muck in the target environment, inputting the preprocessed liquid limit, plastic limit and water content of the muck in the target environment into the trained target model for processing, and outputting a blockage prediction classification label; the calculation of the area proportion of each risk area in the target shield mud cake judgment graph and the generation of a plurality of sample points in the two-dimensional coordinate space according to all the area proportions specifically comprise: calculating the area proportion of each risk area according to the range of each risk area in the target shield mud cake judgment graph; acquiring a preset total number of sample points, and generating a preset number of sample points in the two-dimensional coordinate space by using the scatter point method according to all the area proportions; in the two-dimensional coordinate space, a preset number of sample points are randomly generated by using the scatter point method according to the area proportions, and the coordinate value of each sample point in the two-dimensional coordinate space is generated by a random function, the random function is a standard uniform distribution pseudo-random number generator, and the random number generator includes a true random number generator and a pseudo-random number generator; training according to the sample data, finding model hyperparameters through Bayesian optimization, setting the structure of the model as a framework of 2x16x32x1, inputting the features of the liquid limit-water content and the plastic limit-water content, outputting the blockage prediction classification label, and setting the learning rate to 0.

001.

2. The method of claim 1, wherein, the acquisition of the target shield mud cake judgment graph, the mathematical modeling of the target shield mud cake judgment graph, and the generation of the two-dimensional coordinate space specifically comprise: acquiring a target shield mud cake judgment graph; extracting the numerical range of the difference between the plastic limit and the water content and the numerical range of the difference between the liquid limit and the water content in the target shield mud cake judgment graph; determining the numerical boundary according to the corresponding numerical range, and generating the two-dimensional coordinate space according to the numerical boundary.

3. The method of claim 1, wherein, the generation of sample point label assignment requirements according to the target shield mud cake judgment graph specifically comprises: assigning values to the target shield mud cake judgment graph according to the two-dimensional coordinate space, and constructing a mathematical expression of the boundary of each risk area; calculating the coordinate relationship of points in each risk area according to each mathematical expression, and generating sample point label assignment requirements according to each coordinate relationship.

4. The method of claim 1, wherein, the division of a corresponding classification label for each sample point according to the sample point label assignment requirements, the generation of sample data, the training of a target model according to the sample data, and the obtaining of a trained target model specifically comprise: Obtaining the coordinates of each sample point in the two-dimensional coordinate space, and according to the coordinates of each sample point, dividing a corresponding classification label for each sample point according to the sample point label assignment requirement; According to the classification label and the coordinates of each sample point, generate sample data, train the target model according to the sample data, and obtain the trained target model.

5. The method of claim 4, wherein, According to the classification label and the coordinates of each sample point, generate sample data, train the target model according to the sample data, and obtain the trained target model. According to the classification label and the coordinates of each sample point, generate multiple groups of data, and obtain the sample data by summarizing all the data; Input the coordinates of each group of data in the sample data into the target model, and optimize the target model according to the output of the target model and the classification label of each group of data, and end the training when the training condition is reached, and obtain the trained target model.

6. The method of claim 1, wherein, When performing real-time jam prediction, the liquid limit, plastic limit and water content of the slag soil in the target environment are obtained, and the liquid limit, plastic limit and water content of the slag soil in the target environment are preprocessed and input into the trained target model for processing, and the jam prediction classification label is output, which specifically includes: When performing real-time jam prediction, the liquid limit, plastic limit and water content of the slag soil in the target environment are obtained, and the liquid limit and water content of the slag soil in the target environment are calculated according to the liquid limit and water content of the slag soil in the target environment, and the difference between the current liquid limit and water content is calculated according to the plastic limit and water content of the slag soil in the target environment. According to the current liquid limit and water content difference and the current plastic limit and water content difference, generate the current coordinates; Input the current coordinates into the trained target model, and output the jam prediction classification label.

7. A jam prediction system based on a shield cake building map, characterized by, The jam prediction system based on the shield mud cake judgment graph includes: A two-dimensional coordinate space generation module is used to obtain a target shield mud cake judgment graph, model the target shield mud cake judgment graph, and generate a two-dimensional coordinate space; A sample point generation module is used to calculate the area proportion of each risk area in the target shield mud cake judgment graph, and generate multiple sample points in the two-dimensional coordinate space according to all the area proportions; A training module is used to generate sample point label assignment requirements according to the target shield mud cake judgment graph, divide a corresponding classification label for each sample point according to the sample point label assignment requirements, generate sample data, train a target model according to the sample data, and obtain a trained target model; An application module is used to obtain the liquid limit, plastic limit and water content of the slag soil in the target environment when performing real-time jam prediction, and input the liquid limit, plastic limit and water content of the slag soil in the target environment into the trained target model for processing after preprocessing, and output the jam prediction classification label. The calculation of the area proportion of each risk area in the target shield mud cake judgment graph, and the generation of multiple sample points in the two-dimensional coordinate space according to all the area proportions, specifically includes: According to the range of each risk area in the target shield mud cake judgment graph, calculate the area proportion of each risk area; Obtaining a total number of preset sample points, and generating a preset number of sample points in the two-dimensional coordinate space by using a scatter point method according to all area ratios; In the two-dimensional coordinate space, a preset number of sample points are randomly generated by using a scatter point method according to area ratios, and a coordinate value of each sample point in the two-dimensional coordinate space is generated by a random function, the random function being a standard uniform distribution pseudo-random number generator, wherein the random number generator includes a true random number generator and a pseudo-random number generator; According to the sample data, the model is trained, the model hyperparameters are found through Bayesian optimization, the structure of the model is set as a framework of 2x16x32x1, the input features are liquid limit-water content and plastic limit-water content, the output features are blockage prediction classification labels, and the learning rate is 0.

001.

8. A terminal, characterized by comprising: The terminal comprises a memory, a processor, and a shielded mud cake determination graph-based blockage prediction program stored on the memory and capable of running on the processor, and the shielded mud cake determination graph-based blockage prediction program, when executed by the processor, implements the steps of the shielded mud cake determination graph-based blockage prediction method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a shielded mud cake determination graph-based blockage prediction program, and the shielded mud cake determination graph-based blockage prediction program, when executed by the processor, implements the steps of the shielded mud cake determination graph-based blockage prediction method according to any one of claims 1-6.

Images