Heading machine simulation method and device, electronic equipment, readable storage medium and chip

Abstract

The embodiment of the invention provides a tunneling machine simulation method and device, electronic equipment, a readable storage medium and a chip, and the tunneling machine simulation method comprises the steps: obtaining equipment parameters in first time; obtaining a diagnosis training model and an initial simulation model; inputting the equipment parameters in the first time into a diagnosis training model to obtain initial parameters; inputting the initial parameters into the initial simulation model, and determining a first simulation model; acquiring equipment parameters in second time; inputting the equipment parameters in the second time to a first simulation model for error analysis, and determining first error parameters; optimizing the first simulation model based on the first error parameter, and determining a second simulation model; acquiring equipment parameters in third time; the equipment parameters in the third time are input into the second simulation model, action prediction parameters are obtained, and the action prediction parameters are related to the actual movement of the intelligent heading machine. According to the scheme, the tunneling machine simulation model is planned in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel boring machines, and in particular to a tunnel boring machine simulation method, device, electronic equipment, readable storage medium and chip. Background Art

[0002] At present, when the intelligent tunnel boring machine is cutting underground, the digital twin model currently used for the intelligent tunnel boring machine collects data through sensors or IoT devices to provide data for the simulation model. However, in the digital twin simulation process, since the operation of the tunnel boring machine is difficult to accurately capture, and based on the complexity of the underground environment, the data analysis results are difficult to achieve real-time and accurate levels, which will lead to deviations between the simulation model and the actual performance, and the deviation cannot be corrected quickly in real time, affecting the correctness and effectiveness of the tunnel boring machine data status detection. Summary of the invention

[0003] The purpose of the embodiments of the present invention is to provide a tunnel boring machine simulation method, device, electronic device, readable storage medium and chip, which can solve the problem that there is a deviation between the simulation model and the actual performance of the intelligent tunnel boring machine and the deviation cannot be corrected quickly in real time.

[0004] In view of this, an embodiment of the first aspect of the present invention provides a tunnel boring machine simulation method.

[0005] An embodiment of the second aspect of the present invention provides a tunnel boring machine simulation device.

[0006] An embodiment of the third aspect of the present invention provides an intelligent tunneling machine.

[0007] An embodiment of a fourth aspect of the present invention provides an electronic device.

[0008] An embodiment of a fifth aspect of the present invention provides a readable storage medium.

[0009] An embodiment of the sixth aspect of the present invention provides a chip.

[0010] In order to achieve the above-mentioned objectives, an embodiment of the first aspect of the present invention provides a tunnel boring machine simulation method for an intelligent tunnel boring machine, the intelligent tunnel boring machine comprising a fuselage and a cutting head movably connected to the fuselage, a binocular camera and at least one pressure sensor being provided on the fuselage and / or the cutting head, the tunnel boring machine simulation method comprising: obtaining equipment parameters within a first time, the equipment parameters comprising load parameters corresponding to the pressure sensor and spatial motion parameters corresponding to the binocular camera; obtaining a diagnostic training model and an initial simulation model; inputting the equipment parameters within the first time into the diagnostic training model to obtain initial parameters; inputting the initial parameters into the initial simulation model to determine a first simulation model; obtaining equipment parameters within a second time; inputting the spatial motion parameters within the second time into the first simulation model for error analysis to determine a first error parameter; optimizing the first simulation model based on the first error parameter to determine a second simulation model; obtaining equipment parameters within a third time; inputting the equipment parameters within the third time into the second simulation model to obtain action prediction parameters, which are related to the actual movement of the intelligent tunnel boring machine.

[0011] According to the tunnel boring machine simulation method proposed in the present invention, an intelligent tunnel boring machine is applied to underground cutting operations. The intelligent tunnel boring machine includes a cutting head for cutting and crushing underground rocks and a fuselage movably connected to the cutting head. The rotation of the cutting head is controlled by the fuselage of the intelligent tunnel boring machine. The setting positions of the pressure sensor and the binocular camera can be on at least one of the cutting head and the fuselage. Specifically, the pressure sensor can be set separately on the cutting head or the fuselage, and can also be set on the cutting head and the fuselage at the same time. The binocular camera can be set separately on the cutting head or the fuselage, and can also be set on the cutting head and the fuselage at the same time. On this basis, when the intelligent tunnel boring machine enters the underground environment and starts cutting, the load parameters collected from the pressure sensor include pressure, impact load, impact angle and other parameter data, and the spatial motion parameters collected from the binocular camera include the speed, running angle and spatial image data of the intelligent tunnel boring machine; according to the collected pressure, impact load, impact angle, speed of the intelligent tunnel boring machine, running angle of the intelligent tunnel boring machine and spatial image data, the diagnostic training model is used to clean and preprocess the data to obtain the parameter data after noise removal and preprocessing, and the parameter data is used as a training set for training, and the first simulation model is obtained in combination with the initial simulation model. The first simulation model can perform digital twin simulation of the underground working environment of the intelligent tunnel boring machine. Through the load parameters collected by the pressure sensor and the spatial motion parameters collected by the binocular camera, a real-time simulation model of the underground cutting operation of the intelligent tunnel boring machine is constructed. The spatial orientation information of the intelligent tunnel boring machine, the pressure borne by the intelligent tunnel boring machine, the impact load, the operating angle and the speed are updated in real time. By displaying it on the operation interface, the enterprise operator can interact with the enterprise operator so that the enterprise operator can observe the working status of the intelligent tunnel boring machine in real time and predict the motion parameters of the intelligent tunnel boring machine in the underground working, and control the next operation of the intelligent tunnel boring machine, that is, the motion speed and direction. After the intelligent tunnel boring machine has been operating underground for a period of time, that is, the second time, the first simulation model is optimized using the diagnostic training model, with the purpose of reducing the difference between the intelligent tunnel boring machine parameters output by the first simulation model and the load parameters and spatial motion parameters corresponding to the actual intelligent tunnel boring machine, and optimizing the simulation model. Specifically, the diagnostic training model will compare the load parameters and spatial motion parameters collected by the sensor and binocular camera in the second time with the parameters of the intelligent tunnel boring machine corresponding to the second time output by the first simulation model, and determine the error value between the real parameters of the intelligent tunnel boring machine and the parameters output by the simulation model by combining the point cloud data model and the entropy cross loss function, which is the first error parameter; the first simulation model is corrected through the first error parameter to determine the second simulation model, that is, to obtain the final simulation model.In the third time, that is, after the intelligent tunnel boring machine starts to work normally underground for a period of time, the parameters such as the equipment operation angle, speed, and stress state uploaded in real time by the pressure sensor and binocular camera are analyzed through the second simulation model, and the equipment and environmental morphology are fitted to calculate the spatial position of the intelligent tunnel boring machine in the underground environment. The enterprise operator can see the current spatial position information and working status of the intelligent tunnel boring machine on the operation interface, and plan the optimal action route of the intelligent tunnel boring machine in a short time according to the optimized simulation model, determine the action prediction parameters, and control the intelligent tunnel boring machine to move in the underground working environment according to the action prediction parameters. It can be understood that through the tunnel boring machine simulation method, according to the pressure sensor and binocular camera acquisition parameters combined with the simulation model for optimization, the tunnel boring machine simulation model can be planned in real time, the simulation error can be reduced, the calculation accuracy and the authenticity of the simulation model can be improved, and the simulation efficiency can be improved through deep learning, and real-time three-dimensional simulation data monitoring can be completed.

[0012] Among them, there is no limitation on the type of intelligent tunnel boring machine. The tunnel boring machine can include an open tunnel boring machine and a shield tunnel boring machine, or it can be other types of tunnel boring machines. As long as sensors and binocular cameras can be configured on the tunnel boring machine, the tunnel boring machine simulation method of the present invention can be implemented.

[0013] In the above technical solution, the equipment parameters within the first time are input into the diagnostic training model to obtain initial parameters, including: obtaining a server terminal; uploading load parameters and spatial motion parameters to the server terminal and determining them as a first data set; inputting the first data set into the diagnostic training model for characterization processing to obtain initial parameters.

[0014] In this technical solution, the pressure sensor and binocular camera collect load parameters and spatial motion parameters and upload them to the server terminal for storage, so that the parameters of the intelligent tunnel boring machine operating underground can be saved in real time, improving operation safety and data security. The server terminal is a server for enterprise operators in the system to perform digital twin simulation of the intelligent tunnel boring machine; and at the same time, the load parameters and spatial motion parameters are input as the first data set into the diagnostic training model for characterization processing, data cleaning is performed, redundant noise data in the data is removed, initial parameters are obtained, and the first simulation model is determined based on the initial parameters.

[0015] In the above technical solution, the device parameters within the second time are input into the first simulation model for error analysis to determine the first error parameters, including: uploading the device parameters within the second time to the server terminal; de-discrete the device parameters within the server terminal according to the diagnostic training model to obtain a second data set; input the second data set into the first simulation model to obtain a second prediction parameter; and compare the device parameters and the second prediction parameters according to the diagnostic training model to determine the first error parameter.

[0016] In this technical solution, after the intelligent tunnel boring machine has been operating in an underground working environment for a period of time, that is, after the second time, the load parameters and spatial motion parameters collected by the pressure sensor and the binocular camera are uploaded to the server terminal for storage, and the diagnostic training model performs discretization processing on the load parameters and spatial motion parameters, and after data cleaning and preprocessing, the second data set is input into the first simulation model as the second data set, and the result of the digital twin simulation of the first simulation model is output. The second data set is used as a parameter sample, and the simulation result output by the first simulation model is used as a prediction parameter and input into the diagnostic training model. The diagnostic training model combines the entropy cross loss function and the ten-fold cross validation algorithm to perform a differential comparison on the parameter sample and the prediction parameter, and determines the first error parameter, which is the difference between the simulation result output by the first simulation model and the equipment parameter obtained in the actual work of the intelligent tunnel boring machine. Differential comparison is performed by the entropy cross loss function and the ten-fold cross validation algorithm to reduce the simulation error and improve the calculation accuracy and the authenticity of the simulation model.

[0017] In the above technical solution, the first simulation model is optimized based on the first error parameter to determine the second simulation model, including: obtaining historical simulation model data, the historical simulation model data is obtained from the server terminal; the historical simulation model data is arranged in time sequence and input into the first simulation model to obtain at least one historical prediction parameter; when the difference between the historical simulation model data and the historical prediction parameter is less than the first error parameter, the second simulation model is determined; otherwise, the historical prediction parameter is inserted into the end of the historical simulation model data to update the historical simulation model data.

[0018] In this technical solution, the first simulation model is optimized by the first error parameter and the historical simulation model data to determine the final simulation model, i.e., the second simulation model. Specifically, the server terminal determines the historical simulation model data according to the device parameters collected by the pressure sensor and the binocular camera in the first time and the second time, and the parameters of the simulation output of the first simulation model; the device parameters and the model output parameters in the historical simulation model data are input into the first simulation model in a time series arrangement to obtain at least one output result, the output result is a historical prediction parameter, the historical prediction parameter and the historical simulation model data corresponding to the same time are input into the diagnosis training model, the difference between the historical prediction parameter and the historical simulation model data corresponding to the same time is determined by the entropy cross loss function and the ten-fold cross validation algorithm, the difference is compared with the first error parameter, when the difference is less than the first error parameter, it means that the first simulation model is optimized and qualified, and the second simulation model corresponding to the first error parameter is determined; otherwise, the output result of the first simulation model is used as the historical simulation model data, arranged in time series, and inserted into the end of the time series arrangement, and optimized again until the difference is less than the first error parameter. Through deep learning, the simulation efficiency is improved and the calculation time and cost are reduced.

[0019] In the above technical solution, the equipment parameters within the third time are input into the second simulation model to obtain action prediction parameters, which are related to the actual movement of the intelligent tunnel boring machine, including: determining the spatial motion parameters and load parameters corresponding to the equipment parameters within the third time; determining the position parameters and operating angle of the intelligent tunnel boring machine corresponding to the third time according to the spatial motion parameters, and determining the impact parameters corresponding to the third time according to the load parameters; inputting the equipment parameters within the third time into the second simulation model to obtain action prediction parameters; determining the spatial coordinates and predicted operating angle corresponding to the action prediction parameters; controlling the intelligent tunnel boring machine to change the operating angle to the predicted operating angle, move to the position corresponding to the spatial coordinates, and determine the impact parameters corresponding to the spatial coordinates.

[0020] In this technical solution, when the simulation model is optimized and the intelligent tunnel boring machine starts to operate normally in the underground environment, that is, after the third time, the collected equipment parameters are input into the second simulation model through the pressure sensor and the binocular camera, and the enterprise operator can monitor the digital twin real scene simulation of the intelligent tunnel boring machine working underground in real time through the operation interface. In addition, the second simulation model outputs the action prediction parameters corresponding to the intelligent tunnel boring machine in a short time, including the spatial coordinates of the destination and the operating angle of the intelligent tunnel boring machine. The enterprise operator can adjust the action parameters of the tunnel boring machine according to the action prediction parameters to move it to the destination position corresponding to the action prediction parameters, and observe the impact parameters at the spatial coordinates through the pressure sensor, so as to improve the work efficiency and work quality of underground excavation while ensuring the excavation accuracy.

[0021] An embodiment of the second aspect of the present invention provides a tunnel boring machine simulation device, which includes: an acquisition module, used to acquire equipment parameters within a first time, the equipment parameters including load parameters corresponding to a pressure sensor and spatial motion parameters corresponding to a binocular camera; acquiring a diagnostic training model and an initial simulation model; acquiring equipment parameters within a second time; acquiring equipment parameters within a third time; a determination module, used to input the equipment parameters within the first time into the diagnostic training model to obtain initial parameters; input the initial parameters into the initial simulation model to determine a first simulation model; an error determination module, used to input the equipment parameters within the second time into the first simulation model for error analysis to determine a first error parameter; a model optimization module, used to optimize the first simulation model based on the first error parameter to determine a second simulation model; a control module, used to input the equipment parameters within the third time into the second simulation model to obtain action prediction parameters, which are related to the actual movement of the intelligent tunnel boring machine.

[0022] An embodiment of the third aspect of the present invention provides an intelligent tunnel boring machine, which includes: a fuselage and a cutting head movably connected to the fuselage, the fuselage and / or the cutting head are provided with a binocular camera and at least one pressure sensor and a tunnel boring machine simulation device provided by the second aspect.

[0023] An embodiment of the fourth aspect of the present invention provides an electronic device, which includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, and when the program or instruction is executed by the processor, the steps of the method of the first aspect are implemented.

[0024] An embodiment of the fifth aspect of the present invention provides a readable storage medium, on which a program or instruction is stored, and when the program or instruction is executed by a processor, the steps of the method of the first aspect are implemented.

[0025] An embodiment of the sixth aspect of the present invention provides a chip, the chip includes a processor and a communication interface, the communication interface and the processor are coupled, and the processor is used to run programs or instructions to implement the steps of the method of the first aspect.

[0026] Additional aspects and advantages of the present invention will become apparent from the following description or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A schematic diagram of a process for generating a simulation model according to an embodiment of the present application is shown;

[0028] Figure 2 A schematic diagram of a process of a tunnel boring machine simulation method according to an embodiment of the present application is shown;

[0029] Figure 3 A partial schematic diagram of a method for simulating a roadheader according to an embodiment of the present application is shown;

[0030] Figure 4 A partial schematic diagram of a method for simulating a roadheader according to an embodiment of the present application is shown;

[0031] Figure 5 A partial schematic diagram of a method for simulating a roadheader according to an embodiment of the present application is shown;

[0032] Figure 6 A partial schematic diagram of a method for simulating a roadheader according to an embodiment of the present application is shown;

[0033] Figure 7 A schematic structural block diagram of a tunnel boring machine simulation device according to an embodiment of the present application is shown;

[0034] Figure 8 A schematic structural block diagram of an electronic device according to an embodiment of the present application is shown;

[0035] Fig. 9 A schematic structural block diagram of an intelligent tunnel boring machine according to an embodiment of the present application is shown.

[0036] in, Figures 7 to 9 The corresponding relationship between the reference numerals and the component names is as follows:

[0037] 900: tunnel boring machine simulation device; 902: acquisition module; 904: determination module; 906: error determination module; 908: model optimization module; 910: control module; 1000: electronic equipment; 1109: memory; 1110: processor; 3000: intelligent tunnel boring machine; 3002: fuselage; 3004: cutting head; 3100: binocular camera; 3102: pressure sensor. DETAILED DESCRIPTION

[0038] In order to more clearly understand the above-mentioned purposes, features and advantages of the embodiments of the present invention, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings and specific implementation methods. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0039] In the following description, many specific details are set forth to facilitate a full understanding of the present application. However, the embodiments of the present invention may also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present application is not limited to the specific embodiments disclosed below.

[0040] The following is combined with Figures 1 to 9, the tunnel boring machine simulation method, device, electronic device, readable storage medium and chip provided in the embodiments of the present application are described in detail through specific embodiments and their application scenarios.

[0041] This embodiment provides a tunnel boring machine simulation method for Fig. 9 The intelligent tunnel boring machine 3000 shown in the figure includes a body 3002 and a cutting head 3004 movably connected to the body 3002. The body 3002 and / or the cutting head 3004 are provided with a binocular camera 3100 and at least one pressure sensor 3102. Figure 2 As shown, the tunnel boring machine simulation method includes:

[0042] Step S100: Obtaining device parameters within the first time;

[0043] Step S102: obtaining a diagnostic training model and an initial simulation model;

[0044] Step S104: inputting the equipment parameters within the first time into the diagnosis training model to obtain initial parameters;

[0045] Step S106: inputting initial parameters into the initial simulation model to determine the first simulation model;

[0046] Step S108: Obtaining device parameters within the second time;

[0047] Step S110: inputting the spatial motion parameters within the second time into the first simulation model for error analysis to determine the first error parameter;

[0048] Step S112: optimizing the first simulation model based on the first error parameter to determine the second simulation model;

[0049] Step S114: obtaining device parameters within the third time period;

[0050] Step S116: input the equipment parameters within the third time into the second simulation model to obtain action prediction parameters, which are related to the actual movement of the intelligent tunnel boring machine.

[0051] According to the tunnel boring machine simulation method proposed in the present invention, an intelligent tunnel boring machine is applied to underground cutting operations. The intelligent tunnel boring machine includes a cutting head for cutting and crushing underground rocks and a fuselage movably connected to the cutting head. The rotation of the cutting head is controlled by the fuselage of the intelligent tunnel boring machine. The setting positions of the pressure sensor and the binocular camera can be on at least one of the cutting head and the fuselage. Specifically, the pressure sensor can be set separately on the cutting head or the fuselage, and can also be set on the cutting head and the fuselage at the same time. The binocular camera can be set separately on the cutting head or the fuselage, and can also be set on the cutting head and the fuselage at the same time. On this basis, when the intelligent tunnel boring machine enters the underground environment and starts cutting, the load parameters collected from the pressure sensor include pressure, impact load, impact angle and other parameter data, and the spatial motion parameters collected from the binocular camera include the speed, running angle and spatial image data of the intelligent tunnel boring machine; according to the collected pressure, impact load, impact angle, speed of the intelligent tunnel boring machine, running angle of the intelligent tunnel boring machine and spatial image data, the diagnostic training model is used to clean and preprocess the data to obtain the parameter data after noise removal and preprocessing, and the parameter data is used as a training set for training, and the first simulation model is obtained in combination with the initial simulation model. The first simulation model can perform digital twin simulation of the underground working environment of the intelligent tunnel boring machine. Through the load parameters collected by the pressure sensor and the spatial motion parameters collected by the binocular camera, a real-time simulation model of the underground cutting operation of the intelligent tunnel boring machine is constructed. The spatial orientation information of the intelligent tunnel boring machine, the pressure borne by the intelligent tunnel boring machine, the impact load, the operating angle and the speed are updated in real time. By displaying it on the operation interface, the enterprise operator can interact with the enterprise operator so that the enterprise operator can observe the working status of the intelligent tunnel boring machine in real time and predict the motion parameters of the intelligent tunnel boring machine in the underground working, and control the next operation of the intelligent tunnel boring machine, that is, the motion speed and direction. After the intelligent tunnel boring machine has been operating underground for a period of time, that is, the second time, the first simulation model is optimized using the diagnostic training model, with the purpose of reducing the difference between the intelligent tunnel boring machine parameters output by the first simulation model and the load parameters and spatial motion parameters corresponding to the actual intelligent tunnel boring machine, and optimizing the simulation model. Specifically, the diagnostic training model will compare the load parameters and spatial motion parameters collected by the sensor and binocular camera in the second time with the parameters of the intelligent tunnel boring machine corresponding to the second time output by the first simulation model, and determine the error value between the real parameters of the intelligent tunnel boring machine and the parameters output by the simulation model by combining the point cloud data model and the entropy cross loss function, which is the first error parameter; the first simulation model is corrected through the first error parameter to determine the second simulation model, that is, to obtain the final simulation model.In the third time, that is, after the intelligent tunnel boring machine starts to work normally underground for a period of time, the parameters such as the equipment operation angle, speed, and stress state uploaded in real time by the pressure sensor and binocular camera are analyzed through the second simulation model, and the equipment and environmental morphology are fitted to calculate the spatial position of the intelligent tunnel boring machine in the underground environment. The enterprise operator can see the current spatial position information and working status of the intelligent tunnel boring machine on the operation interface, and plan the optimal action route of the intelligent tunnel boring machine in a short time according to the optimized simulation model, determine the action prediction parameters, and control the intelligent tunnel boring machine to move in the underground working environment according to the action prediction parameters. It can be understood that through the tunnel boring machine simulation method, according to the pressure sensor and binocular camera acquisition parameters combined with the simulation model for optimization, the tunnel boring machine simulation model can be planned in real time, the simulation error can be reduced, the calculation accuracy and the authenticity of the simulation model can be improved, and the simulation efficiency can be improved through deep learning, and real-time three-dimensional simulation data monitoring can be completed.

[0052] Among them, there is no limitation on the type of intelligent tunnel boring machine. The tunnel boring machine can include an open tunnel boring machine and a shield tunnel boring machine, or it can be other types of tunnel boring machines. As long as sensors and binocular cameras can be configured on the tunnel boring machine, the tunnel boring machine simulation method of the present invention can be implemented.

[0053] In one embodiment, Figure 3 As shown, the equipment parameters in the first time are input into the diagnosis training model to obtain initial parameters, including:

[0054] Step S1042: Obtain a server terminal;

[0055] Step S1044: uploading the load parameters and the spatial motion parameters to the server terminal and determining them as the first data set;

[0056] Step S1046: input the first data set into the diagnosis training model for characterization processing to obtain initial parameters.

[0057] In this technical solution, the pressure sensor and binocular camera collect load parameters and spatial motion parameters and upload them to the server terminal for storage, so that the parameters of the intelligent tunnel boring machine operating underground can be saved in real time, improving operation safety and data security. The server terminal is a server for enterprise operators in the system to perform digital twin simulation of the intelligent tunnel boring machine; and at the same time, the load parameters and spatial motion parameters are input as the first data set into the diagnostic training model for characterization processing, data cleaning is performed, redundant noise data in the data is removed, initial parameters are obtained, and the first simulation model is determined based on the initial parameters.

[0058] In one embodiment, Figure 4As shown, the device parameters in the second time are input into the first simulation model for error analysis to determine the first error parameters, including:

[0059] Step S1102: uploading the device parameters within the second time to the server terminal;

[0060] Step S1104: performing a de-discrete point process on the device parameters in the server terminal according to the diagnosis training model to obtain a second data set;

[0061] Step S1106: inputting the second data set into the first simulation model to obtain a second prediction parameter;

[0062] Step S1108: performing a difference comparison between the device parameter and the second prediction parameter according to the diagnosis training model to determine a first error parameter.

[0063] In this technical solution, after the intelligent tunnel boring machine has been operating in an underground working environment for a period of time, that is, after the second time, the load parameters and spatial motion parameters collected by the pressure sensor and the binocular camera are uploaded to the server terminal for storage, and the diagnostic training model performs discretization processing on the load parameters and spatial motion parameters, and after data cleaning and preprocessing, the second data set is input into the first simulation model as the second data set, and the result of the digital twin simulation of the first simulation model is output, and the second data set is used as a parameter sample, and the simulation result output by the first simulation model is used as a prediction parameter, which is input into the diagnostic training model. The diagnostic training model combines the entropy cross loss function (Cross-EntropyLoss, CE) and the 10-fold cross validation (10-fold Cross Validation) algorithm to perform differential comparison on the parameter sample and the prediction parameter, and determine the first error parameter, which is the difference between the simulation result output by the first simulation model and the equipment parameter obtained in the actual work of the intelligent tunnel boring machine. Differential comparison is performed through the entropy cross loss function and the 10-fold cross validation algorithm to reduce the simulation error and improve the calculation accuracy and the authenticity of the simulation model.

[0064] Among them, the entropy cross loss function is used to reduce the difference between the real distribution and the predicted distribution, thereby improving the accuracy of the prediction, and adding a penalty term to it. The penalty term is proportional to the parameter sample in the model parameters, that is, it is proportional to the device parameters collected by the binocular camera and the pressure sensor. The purpose of adding the penalty term is to simplify the model. By limiting the value range and size of the model parameters, the model will not be overly complex on the training data, and the generalization ability on unknown data can be improved. In addition, by adding the penalty term, the model can be promoted to automatically select features that have a significant impact on the prediction results, ignore irrelevant features or noise features, and further improve the accuracy of the simulation model prediction. The ten-fold cross validation algorithm divides the collected device data into ten parts, and takes turns to use nine of them as training data and one as test data for testing. Each test will obtain the corresponding accuracy, and the average value of the correct value of the results of ten tests is used as an estimate of the accuracy of the algorithm.

[0065] It can be understood that by using the entropy cross loss function and the ten-fold cross validation algorithm for double verification, the accuracy of model prediction and calculation accuracy can be effectively improved.

[0066] In one embodiment, Figure 5 As shown, the first simulation model is optimized based on the first error parameter to determine the second simulation model, including:

[0067] Step S1122: Acquire historical simulation model data, where the historical simulation model data is acquired from the server terminal;

[0068] Step S1124: Arranging the historical simulation model data in time sequence and inputting the data into the first simulation model to obtain at least one historical prediction parameter;

[0069] Step S1126: when the difference between the historical simulation model data and the historical prediction parameter is less than the first error parameter, determining the second simulation model;

[0070] Step S1128: Otherwise, insert the historical prediction parameters into the end of the historical simulation model data to update the historical simulation model data.

[0071] In this technical solution, the first simulation model is optimized by the first error parameter and the historical simulation model data to determine the final simulation model, i.e., the second simulation model. Specifically, the server terminal determines the historical simulation model data according to the device parameters collected by the pressure sensor and the binocular camera in the first time and the second time, and the parameters of the simulation output of the first simulation model; the device parameters and the model output parameters in the historical simulation model data are input into the first simulation model in a time series arrangement to obtain at least one output result, the output result is a historical prediction parameter, the historical prediction parameter and the historical simulation model data corresponding to the same time are input into the diagnosis training model, the difference between the historical prediction parameter and the historical simulation model data corresponding to the same time is determined by the entropy cross loss function and the ten-fold cross validation algorithm, the difference is compared with the first error parameter, when the difference is less than the first error parameter, it means that the first simulation model is optimized and qualified, and the second simulation model corresponding to the first error parameter is determined; otherwise, the output result of the first simulation model is used as the historical simulation model data, arranged in time series, and inserted into the end of the time series arrangement, and optimized again until the difference is less than the first error parameter. Through deep learning, the simulation efficiency is improved and the calculation time and cost are reduced.

[0072] In one embodiment, Figure 6 As shown, the equipment parameters within the third time are input into the second simulation model to obtain action prediction parameters, which are related to the guidance of the intelligent tunnel boring machine and include:

[0073] Step S1162: determining the spatial motion parameters and load parameters corresponding to the equipment parameters within the third time;

[0074] Step S1164: determining the position parameter and the operating angle of the intelligent tunnel boring machine corresponding to the third time according to the spatial motion parameter, and determining the impact parameter corresponding to the third time according to the load parameter;

[0075] Step S1166: inputting the device parameters within the third time into the second simulation model to obtain action prediction parameters;

[0076] Step S1168: Determine the spatial coordinates and predicted operating angle corresponding to the action prediction parameters;

[0077] Step S1170: Control the intelligent tunnel boring machine to change the operating angle to the predicted operating angle, move to the corresponding position of the spatial coordinates, and determine the impact parameters corresponding to the spatial coordinates.

[0078] In this technical solution, when the simulation model is optimized and the intelligent tunnel boring machine starts to operate normally in the underground environment, that is, after the third time, the collected equipment parameters are input into the second simulation model through the pressure sensor and the binocular camera, and the enterprise operator can monitor the digital twin real scene simulation of the intelligent tunnel boring machine working underground in real time through the operation interface. In addition, the second simulation model outputs the action prediction parameters corresponding to the intelligent tunnel boring machine in a short time, including the spatial coordinates of the destination and the operating angle of the intelligent tunnel boring machine. The enterprise operator can adjust the action parameters of the tunnel boring machine according to the action prediction parameters to move it to the destination position corresponding to the action prediction parameters, and observe the impact parameters at the spatial coordinates through the pressure sensor, so as to improve the work efficiency and work quality of underground excavation while ensuring the excavation accuracy.

[0079] In a specific embodiment, Figure 1 As shown, the optimization process of the simulation model (i.e., the second simulation model) includes:

[0080] Step S600: the tunnel boring machine equipment starts running; Step S602: the binocular camera acquires a digital image; Step S604: uploading parameters such as running angle, speed, and force; Step S606: uploading to the server; Step S608: uploading to the database; Step S610: transmitting to the data cleaning and preprocessing module; Step S612: transmitting to the diagnostic model training module; Step S6122: transmitting to the data analysis and comparison module; Step S6124: constructing a point cloud data model; Step S6126: de-discrete processing; Step S6128: linearly fitting the model morphology and calculating the spatial position and direction; Step S614: transmitting to the result output module; Step S616: judging during operation; Step S618: outputting the simulation model without error.

[0081] Specifically, when the intelligent tunnel boring machine equipment starts to operate in the underground environment, the server will first receive the digital image, operating angle, speed, force and other parameters transmitted from the binocular camera and at least one sensor, and upload the binocular camera digital image and operating angle, speed, force and other parameters to the database for storage; after receiving the data parameters from the binocular camera and the sensor, the server corrects the parameter analysis results through the data cleaning and preprocessing module to generate a simulation model, which is the initial real-time simulation model; the initial real-time simulation model is trained through the diagnostic model training module, and the diagnostic model training module is responsible for training the training set through migration learning, wherein the parameters in the training set include the binocular camera digital image, operating angle, speed, and force parameters; the equipment operation status parameters and simulation model parameters are generated through the result output module. Among them, the diagnostic model training module will transfer the training set to the data analysis and comparison module, and upload the physical quantity parameter information such as the operating angle, speed and force of the intelligent tunnel boring machine equipment in real time according to the sensor, and cooperate with the image data collected by the binocular camera to build a point cloud data model. The model is de-discrete through data analysis, and the model morphology is linearly fitted to calculate the spatial position direction. The spatial position direction is the spatial coordinate and movement direction of the intelligent tunnel boring machine in the underground working environment. It is compared with the parameters output in the result output module. When there is no error, the simulation model (i.e., the second simulation model) is output. When there is an error, the errors in the parameters and models are corrected, and returned to the data cleaning and preprocessing module to regenerate the simulation model, thereby determining the current optimal simulation model. The process of correcting the errors in the parameters and models includes:

[0082] Step S700: Obtain historical simulation model data; Step S702: Upload to the server; Step S704: Arrange in time sequence; Step S706: Convolutional neural network training; Step S708: Output simulation model; Step S710: Cross-loss function optimization model of penalty term; Step S712: Verify simulation results by ten-fold cross validation algorithm; Step S714: If qualified, determine the simulation model; Step S716: If unqualified, return to the simulation model.

[0083] Historical simulation model data is obtained from the server, and the historical simulation model data includes data parameters transmitted by the binocular camera and sensors uploaded by the server, as well as parameters output by the historical simulation model; the historical simulation model data is arranged in time series through the diagnostic model training module, and the training set is trained based on migration learning using a convolutional neural network with residual network (Residual Network, ResNet) and dense convolutional network (Dense Convolutional Network, DenseNet) as basic units, and a sensitivity analysis is performed on the parameters; the simulation model output by the result output module is optimized, that is, the model is optimized using the entropy cross loss function of the penalty term, and the entropy cross loss function is used to reduce the difference between the real distribution and the predicted distribution, thereby improving the accuracy of the prediction results. The optimized simulation model is optimized for the second time, and the simulation results are verified by a ten-fold cross-validation algorithm. The differences in the simulation results are corrected by combining the entropy cross loss function and the ten-fold cross-validation algorithm to improve the accuracy of the algorithm and reduce the deviation of the prediction results. The accuracy threshold is set by the operator. When the simulation results after the ten-fold cross-validation algorithm meet the accuracy threshold requirements, the simulation model is judged to meet the qualification requirements and the simulation model (i.e., the second simulation model) is output; otherwise, the simulation model is judged to be unqualified, the prediction parameters output by the simulation model are inserted into the end of the time series arrangement, and the convolutional neural network training is performed again until the simulation model (i.e., the second simulation model) is output.

[0084] Optionally, the simulation model (i.e., the second simulation model) generation process includes: testing the underground environment and tunnel boring machine equipment parameters; simulating the underground working state of the tunnel boring machine and pre-outputting the simulation model; real-time data-driven optimization and planning the optimal simulation model.

[0085] Optionally, testing the underground environment and tunnel boring machine equipment parameters includes: using a Hopkinson strut to measure the angle, distance, speed, force and other physical quantity parameter information between the underground tunnel boring machine equipment and the surrounding environment in space, as well as the operation law of the equipment under different impact loads, impact frequencies and impact angles;

[0086] Optionally, the underground working state of the tunnel boring machine is simulated, and the pre-output simulation model includes: collecting digital images of the tunnel boring machine equipment through sensors and binocular cameras on the intelligent tunnel boring machine, and storing them on the server; a data cleaning and preprocessing module is responsible for correcting the model data analysis results and generating a simulation model; a diagnostic model training module is responsible for training the training set using a convolutional neural network based on transfer learning with residual networks and dense convolutional networks as basic units; a result output module is responsible for outputting model state parameters, and outputting the optimal simulation model through a data-driven correction optimization process.

[0087] Optionally, real-time data-driven optimization and planning of the optimal simulation model include: real-time uploading of parameters such as equipment operating angle, speed, stress state, etc., and performing data analysis on the model to remove discrete points and perform linear fitting of the equipment and environmental morphology, thereby calculating the spatial orientation of the tunnel boring machine equipment, and optimizing the equipment operating parameter errors in the working state of the tunnel boring machine according to the function formula to determine the current optimal simulation model.

[0088] It can be understood that by collecting and analyzing the operating rules of the equipment under impact load, impact frequency, and impact angle, the simulation model can predict the optimal action parameters, reduce the equipment loss of the intelligent tunnel boring machine during the underground cutting process, and improve the equipment operating life.

[0089] like Figure 7 As shown, an embodiment of the present invention provides a tunnel boring machine simulation device 900, which includes: an acquisition module 902, used to acquire equipment parameters within a first time, the equipment parameters including load parameters corresponding to a pressure sensor and spatial motion parameters corresponding to a binocular camera; acquiring a diagnostic training model and an initial simulation model; acquiring equipment parameters within a second time; acquiring equipment parameters within a third time; a determination module 904, used to input the equipment parameters within the first time into the diagnostic training model to obtain initial parameters; input the initial parameters into the initial simulation model to determine a first simulation model; an error determination module 906, used to input the equipment parameters within the second time into the first simulation model for error analysis to determine a first error parameter; a model optimization module 908, used to optimize the first simulation model based on the first error parameter to determine a second simulation model; a control module 910, used to input the equipment parameters within the third time into the second simulation model to obtain action prediction parameters, which are related to the actual movement of the intelligent tunnel boring machine.

[0090] like Figure 8 As shown, an embodiment of the present application also provides an electronic device 1000, including a processor 1110, a memory 1109, and a program or instruction stored in the memory 1109 and executable on the processor 1110. When the program or instruction is executed by the processor 1110, each process of the embodiment of the above-mentioned tunnel boring machine simulation method is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0091] Optionally, processor 1110 is used to obtain equipment parameters within a first time, the equipment parameters including load parameters corresponding to the pressure sensor and spatial motion parameters corresponding to the binocular camera; obtain a diagnostic training model and an initial simulation model; input the equipment parameters within the first time into the diagnostic training model to obtain initial parameters; input the initial parameters into the initial simulation model to determine a first simulation model; obtain equipment parameters within a second time; input the spatial motion parameters within the second time into the first simulation model for error analysis to determine a first error parameter; optimize the first simulation model based on the first error parameter to determine a second simulation model; obtain equipment parameters within a third time; input the equipment parameters within the third time into the second simulation model to obtain action prediction parameters, which are related to the actual movement of the intelligent tunnel boring machine.

[0092] Optionally, the processor 1110 is further configured to obtain a server terminal; upload the load parameters and the spatial motion parameters to the server terminal, and determine them as a first data set.

[0093] Optionally, the processor 1110 is further configured to perform a difference comparison between the device parameter and the second prediction parameter according to the diagnosis training model to determine a first error parameter.

[0094] Optionally, the processor 1110 is also used to obtain historical simulation model data, which is obtained from a server terminal; the historical simulation model data is arranged in time sequence and input into the first simulation model to obtain at least one historical prediction parameter; when the difference between the historical simulation model data and the historical prediction parameter is less than a first error parameter, the second simulation model is determined; otherwise, the historical prediction parameter is inserted into the end of the historical simulation model data to update the historical simulation model data.

[0095] Optionally, the processor 1110 is also used to control the intelligent tunnel boring machine to change the operating angle to the predicted operating angle, move to a position corresponding to the spatial coordinates, and determine the impact parameters corresponding to the spatial coordinates.

[0096] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, each process of the above-mentioned tunnel boring machine simulation method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0097] The processor is the processor in the electronic device in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0098] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface and the processor are coupled, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned tunnel boring machine simulation method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0099] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0100] According to the tunnel boring machine simulation method, device, electronic device, readable storage medium and chip provided by the present invention, the tunnel boring machine simulation model can be planned in real time according to the parameters collected by the pressure sensor and the binocular camera in combination with the simulation model for optimization, and the simulation error can be reduced, and the calculation accuracy and the authenticity of the simulation model can be improved. Through deep learning, the simulation efficiency can be improved, and real-time three-dimensional simulation data monitoring can be realized.

[0101] In the present invention, the terms "first", "second", and "third" are used for descriptive purposes only and should not be understood as indicating or implying relative importance; the term "plurality" refers to two or more, unless otherwise clearly defined. The terms "installed", "connected", "connected", "fixed", etc. should be understood in a broad sense. For example, "connected" can be a fixed connection, a detachable connection, or an integral connection; "connected" can be a direct connection or an indirect connection through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0102] In the description of the present invention, it is necessary to understand that the directions or positional relationships indicated by terms such as “upper”, “lower”, “left”, “right”, “front” and “back” are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or unit referred to must have a specific direction, be constructed and operated in a specific orientation, and therefore, cannot be understood as a limitation on the present invention.

[0103] In the description of this specification, the description of the terms "one embodiment", "some embodiments", "specific embodiments", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0104] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A tunnel boring machine simulation method, characterized in that: For an intelligent tunnel boring machine, the intelligent tunnel boring machine comprises a fuselage and a cutting head movably connected to the fuselage, the fuselage and / or the cutting head are provided with a binocular camera and at least one pressure sensor, and the tunnel boring machine simulation method comprises: Acquire device parameters within a first time, wherein the device parameters include load parameters corresponding to the pressure sensor and spatial motion parameters corresponding to the binocular camera; Obtaining a diagnostic training model and an initial simulation model; Inputting the equipment parameters within the first time period into the diagnostic training model to obtain initial parameters; Inputting the initial parameters into the initial simulation model to determine a first simulation model; Obtaining the device parameters within a second time period; Inputting the equipment parameters within the second time into the first simulation model for error analysis to determine a first error parameter; Optimizing the first simulation model based on the first error parameter to determine a second simulation model; Obtaining the device parameters within a third time; The equipment parameters within the third time are input into the second simulation model to obtain action prediction parameters, and the action prediction parameters are related to the actual movement of the intelligent tunnel boring machine.

2. The tunnel boring machine simulation method according to claim 1, characterized in that: The inputting the equipment parameters within the first time into the diagnosis training model to obtain initial parameters comprises: Get the server terminal; Uploading the load parameter and the spatial motion parameter to the server terminal to determine them as a first data set; The first data set is input into the diagnosis training model for characterization processing to obtain initial parameters.

3. The tunnel boring machine simulation method according to claim 2, characterized in that: The step of inputting the device parameters within the second time into the first simulation model for error analysis to determine the first error parameter includes: uploading the device parameters within the second time to the server terminal; De-discrete the device parameters in the server terminal according to the diagnostic training model to obtain a second data set; inputting the second data set into the first simulation model to obtain second prediction parameters; The device parameter and the second prediction parameter are compared for differences according to the diagnosis training model to determine a first error parameter.

4. The tunnel boring machine simulation method according to claim 3, characterized in that: The step of optimizing the first simulation model based on the first error parameter to determine a second simulation model includes: Acquiring historical simulation model data, wherein the historical simulation model data is acquired from the server terminal; Arranging the historical simulation model data in time sequence and inputting them into the first simulation model to obtain at least one historical prediction parameter; When the difference between the historical simulation model data and the historical prediction parameter is less than the first error parameter, determining a second simulation model; Otherwise, the historical prediction parameters are inserted into the end of the historical simulation model data to update the historical simulation model data.

5. The tunnel boring machine simulation method according to claim 1, characterized in that: The step of inputting the equipment parameters within the third time into the second simulation model to obtain action prediction parameters, wherein the action prediction parameters are related to the actual movement of the intelligent tunnel boring machine, includes: determining spatial motion parameters and load parameters corresponding to the device parameters within the third time; Determine the position parameter and the running angle of the intelligent tunnel boring machine corresponding to the third time according to the spatial motion parameter, and determine the impact parameter corresponding to the third time according to the load parameter; Inputting the device parameters within the third time into the second simulation model to obtain action prediction parameters; Determining the spatial coordinates and predicted operating angles corresponding to the action prediction parameters; The intelligent tunnel boring machine is controlled to change the operating angle to the predicted operating angle, move to the position corresponding to the spatial coordinate, and determine the impact parameters corresponding to the spatial coordinate.

6. A tunnel boring machine simulation device, characterized in that: include: An acquisition module, used for acquiring device parameters within a first time, wherein the device parameters include load parameters corresponding to the pressure sensor and spatial motion parameters corresponding to the binocular camera; Acquire a diagnostic training model and an initial simulation model; acquire the device parameters within a second period of time; acquire the device parameters within a third period of time; A determination module, used for inputting the equipment parameters within the first time into the diagnosis training model to obtain initial parameters; Inputting the initial parameters into the initial simulation model to determine a first simulation model; an error determination module, configured to input the device parameters within the second time into the first simulation model for error analysis to determine a first error parameter; A model optimization module, configured to optimize the first simulation model based on the first error parameter to determine a second simulation model; A control module is used to input the equipment parameters within the third time into the second simulation model to obtain action prediction parameters, which are related to the actual movement of the intelligent tunnel boring machine.

7. An intelligent tunnel boring machine, characterized in that: include: A fuselage and a cutting head movably connected to the fuselage, wherein the fuselage and / or the cutting head are provided with a binocular camera and at least one pressure sensor; The tunnel boring machine simulation device as claimed in claim 6.

8. An electronic device, characterized in that: The method comprises a processor, a memory and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements the steps of the tunnel boring machine simulation method as described in any one of claims 1 to 5.

9. A readable storage medium, characterized in that: The readable storage medium stores a program or instruction, and when the program or instruction is executed by the processor, the steps of the tunnel boring machine simulation method according to any one of claims 1 to 5 are implemented.

10. A chip, characterized in that: The chip includes a processor and a communication interface, the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the steps of the tunnel boring machine simulation method according to any one of claims 1 to 5.

Images