Multi-source geological information in-situ intelligent sensing system of micro-nano geotechnical robot

Abstract

The invention discloses a multi-source geological information in-situ intelligent sensing system of a micro-nano geotechnical robot. The system comprises a micro-nano rock-soil robot cluster, an energy supply unit, electromagnetic modified micro-nano particles, a main control unit and an injection-production system. The micro-nano rock-soil robot is composed of a micro-nano sensor, a power supply, a communication module and a control module, and the sensor can be a temperature sensor, a ground stress sensor, a fluid pressure sensor, a microbial sensor, an ion concentration sensor and a pH value sensor according to monitoring requirements so as to realize monitoring and acquisition of multi-source geological information. And the main control unit processes the monitoring data by using a machine learning technology and comprehensively reconstructs underground information. The system can realize multi-dimensional monitoring of geological information, and promotes research and application in the fields of deep engineering construction, underground resource mining and environmental evaluation.

Description

Technical Field

[0001] The present invention belongs to the technical fields of geological information monitoring and machine learning, and particularly relates to a multi-source geological information in-situ intelligent perception system based on micro-nano geotechnical robots. Background Art

[0002] The deep earth contains rich energy and mineral resources. Therefore, carrying out deep earth exploration is of great strategic significance for ensuring national energy resource security and improving the level of basic scientific research. However, the deep earth environment is complex and changeable, and its evolution process is controlled by the multi-field coupling of the mechanical field, temperature field, seepage field and chemical field.

[0003] Existing geophysical exploration technologies are difficult to obtain multi-source geological information and cannot meet the requirements of in-situ real-time monitoring. Although the method of laying boreholes and borehole sensors can partially solve this problem, the measurement range is limited. In this case, deploying wireless sensors underground to form a wireless sensor network provides a new way for multi-source, in-situ, and real-time geological information detection. However, at present, the contradiction between the sensor size and function integration has not been effectively solved, and the high cost also urgently demands sensor recovery technology.

[0004] Therefore, there is an urgent need for a system based on micro-nano geotechnical robots for realizing multi-source geological information monitoring. Summary of the Invention

[0005] Based on the above technical problems, the present invention provides a multi-source geological information in-situ intelligent perception system for micro-nano geotechnical robots, which is used to solve the problems of limited size and difficult recovery of in-situ measurement devices in the prior art, realize in-situ, real-time, and multi-source monitoring of geological information, and achieve intelligent analysis of monitoring data through machine learning means, so as to accurately reproduce the three-dimensional changes of multiple fields in the target monitoring area.

[0006] The technical solution provided by the present invention is as follows:

[0007] In the first aspect, a multi-source geological information in-situ intelligent perception system for micro-nano geotechnical robots is provided, including a micro-nano geotechnical robot cluster, an energy supply unit, electromagnetic modified micro-nano particles, a main control unit, and an injection and production system;

[0008] The micro-nano geotechnical robot cluster is used to identify the changes in geological information in the target monitoring area;

[0009] The injection and production system is used to inject and recover the micro-nano geotechnical robot cluster and the electromagnetic modified micro-nano particles;

[0010] The energy supply unit is used to generate an electric field in the target monitoring area to supply energy to the micro-nano geotechnical robot cluster;

[0011] The electromagnetic modified micro-nano particles are used to improve the conductivity, magnetic permeability and dielectric constant of the crack environment where the micro-nano geotechnical robots are located, improve the energy transmission efficiency of the power supply unit to the micro-nano geotechnical robot cluster, and be used as tracers and proppants to determine the crack distribution pattern and prevent crack closure;

[0012] The main control unit is used to perform data analysis and three-dimensional modeling on the recovered micro-nano geotechnical robots, construct a three-dimensional visualization model of the multi-physical field according to the modeling results, and control the working behavior of the power supply device in real time.

[0013] In a possible implementation manner, the micro-nano geotechnical robots in the micro-nano geotechnical robot cluster include a housing and the following components disposed in the housing:

[0014] Micro-nano sensors, which are used to monitor the changes in geological information in the target monitoring area;

[0015] A power supply, which is used to supply electric energy;

[0016] A communication module, including a transmitting antenna and a receiving antenna, which is used to realize information exchange with the main control unit;

[0017] A control module, which is used to control the working states of the micro-nano sensors and the communication module, store the collected geological information data and the corresponding time, and perform signal conditioning and data compensation on the micro-nano sensors.

[0018] Furthermore, the micro-nano sensors include inertial sensors and monitoring sensors;

[0019] The monitoring sensors include temperature sensors, stress sensors, fluid pressure sensors, microbial sensors, ion concentration sensors and pH sensors.

[0020] Furthermore, the power supply is an electret, which is charged by the excitation of the power supply unit.

[0021] Furthermore, the housing includes a waterproof layer from the inside to the outside and a shell layer with surface modification treatment; the waterproof layer is made of a polymer coating or silica gel; the method for obtaining the shell layer with surface modification treatment is: mixing the micro-nano geotechnical robots, negatively charged graphene and a coupling agent at high speed to increase the electrostatic repulsion between the micro-nano geotechnical robots.

[0022] In a possible implementation manner, the injection-production system includes an injection well, a production well, a fluid pump, a filter and a separator;

[0023] The injection well and the production well respectively inject and produce the micro-nano geotechnical robot cluster and the electromagnetic modified micro-nano particles through the fluid pump;

[0024] The filter and separator are used to filter and separate the extracted micro-nano geotechnical robot clusters, electromagnetically modified micro-nano particles, fluids, and solid impurities.

[0025] In a possible implementation manner, the power supply unit uses a DC power supply or an AC power supply. The two poles of the power supply are respectively placed between the formations in the target monitoring area, and an electric field is excited in the target monitoring area at intervals less than the limit working duration of the micro-nano geotechnical robots to supply power to the micro-nano geotechnical robot clusters.

[0026] In a possible implementation manner, the electromagnetically modified micro-nano particles are obtained by laying a conductive coating on the surface of ceramic particles.

[0027] In a possible implementation manner, the main control unit includes a data denoising module, a data filling module, and a three-dimensional visualization module;

[0028] The data denoising module is used to select effective components from different scale information components and reconstruct them to achieve noise reduction by using the time-domain denoising method;

[0029] The data filling module is used to make up for the data non-uniformity and structural missing caused by the failure, loss, and uneven distribution of micro-nano geotechnical robots by using machine learning methods, and completely reconstruct the three-dimensional point cloud model of the target monitoring area;

[0030] The three-dimensional visualization module relies on the reconstructed three-dimensional point cloud model to visually display the spatial distribution and time variation of underground multi-source geological information.

[0031] Furthermore, the data denoising module is used to perform data denoising by using the empirical mode decomposition and wavelet transform methods;

[0032] The data filling module is used to perform data filling by using the GAIN data filling algorithm based on the generative adversarial network; the data filling method includes:

[0033] Generate a binary mask matrix for the original data matrix to mark known values and missing values;

[0034] Process the known values in the original data matrix with random noise variables to obtain a first data matrix, and fill in the missing values in the original data matrix to obtain a second data matrix;

[0035] Construct a generator and a discriminator model, perform adversarial training, and obtain an optimized generator; where the generator is used to receive the first data matrix and generate predictions for the missing values; the discriminator is used to receive the second data and the hint matrix, and learn to identify the actual observed data and the data filled by the generator;

[0036] Use the optimized generator to fill in the missing values of the original data to obtain a complete data set.

[0037] The beneficial effects of the present invention are as follows:

[0038] The multi-source geological information in-situ intelligent perception system based on micro-nano geotechnical robots provided by the present invention injects and extracts micro-nano geotechnical robots into the ground in a cycle, records the movement trajectories of the micro-nano geotechnical robots in the target monitoring area and the collected geological information. When the micro-nano geotechnical robots work underground, they only collect sensor data. After being extracted, they perform near-field communication with the nano-geotechnical robots through the main control unit to transmit the data, and then output the geological information data of the target monitoring area after data processing. Therefore, this system can greatly reduce the communication requirements and thresholds, can adopt smaller communication modules, and further realize the miniaturization of the robots.

[0039] The system of the present invention effectively overcomes the problem of limited size of underground sensors by specializing the functions of micro-nano geotechnical robots, and realizes the effective recovery and recycling of micro-nano geotechnical robots, greatly reducing the engineering cost. This system can reproduce the geological conditions of the target monitoring area in all directions and from multiple angles, and can provide effective guidance for deep engineering construction, underground resource exploitation and environmental assessment according to the monitoring results. Description of the Drawings

[0040] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0041] Figure 1 is a schematic structural diagram of a multi-source geological information in-situ intelligent perception system based on micro-nano geotechnical robots;

[0042] Figure 2 is a schematic composition diagram of a micro-nano geotechnical robot;

[0043] Figure 3 is a working flow chart of the data noise reduction module of the main control unit of the multi-source geological information in-situ intelligent perception system of the micro-nano geotechnical robot in the embodiment of the present application;

[0044] Figure 4 is an algorithm structure diagram of the data filling module of the main control unit of the multi-source geological information in-situ intelligent perception system of the micro-nano geotechnical robot in the embodiment of the present application;

[0045] Figure 5 is a working schematic diagram of the multi-source geological information in-situ intelligent perception system of the micro-nano geotechnical robot in the embodiment of the present application;

[0046] Figure 6 It is a flow chart for the construction and operation of an in-situ intelligent perception system for multi-source geological information of a micro-nano geotechnical robot. Specific implementation manners

[0047] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0048] As Figure 1 shown, a multi-source geological information in-situ intelligent perception system for a micro-nano geotechnical robot provided by the present invention includes a micro-nano geotechnical robot cluster 11, an injection and extraction system 12, an energy supply unit 13, electromagnetic modified micro-nano particles 14, and a main control unit 15;

[0049] The micro-nano geotechnical robot cluster 11 is used to identify changes in geological information in the target monitoring area;

[0050] The injection and extraction system 12 is used to inject and recover the micro-nano geotechnical robot cluster and the electromagnetic modified micro-nano particles;

[0051] The energy supply unit 13 is used to generate an electric field in the target monitoring area to supply energy to the micro-nano geotechnical robot cluster;

[0052] The electromagnetic modified micro-nano particles 14 are used to improve the conductivity, magnetic permeability and dielectric constant of the crack environment where the micro-nano geotechnical robot is located, improve the energy transmission efficiency of the energy supply unit to the micro-nano geotechnical robot cluster, and serve as a tracer and a proppant to determine the crack distribution pattern and prevent crack closure;

[0053] The main control unit 15 is used to perform data analysis and three-dimensional modeling on the recovered micro-nano geotechnical robots, construct a three-dimensional visualization model of a multi-physical field according to the modeling results, and real-time control the working behavior of the energy supply device.

[0054] In a possible implementation manner, referring to Figure 2 , the micro-nano geotechnical robots in the micro-nano geotechnical robot cluster 11 include a housing and the following disposed in the housing:

[0055] Micro-nano sensors, used to monitor changes in geological information in the target monitoring area;

[0056] A power source, used to supply electric energy;

[0057] The communication module, including a transmitting antenna and a receiving antenna, is used to realize information exchange with the main control unit;

[0058] The control module is used to control the working states of the micro-nano sensor and the communication module, store the collected geological information data and the corresponding time, and perform signal conditioning and data compensation on the micro-nano sensor.

[0059] Furthermore, the micro-nano sensor includes an inertial sensor and a monitoring sensor;

[0060] The monitoring sensor includes a temperature sensor, a stress sensor, a fluid pressure sensor, a microbial sensor, an ion concentration sensor, and a pH sensor.

[0061] Even further, the inertial sensor includes an accelerometer and a gyroscope, which are used to record the running trajectory of the micro-nano geotechnical robot.

[0062] It can be understood that the monitoring sensor is configured according to specific monitoring requirements, and thus the size of the micro-nano geotechnical robot can be ensured to meet the limitation of the crack size by highly specializing the functions.

[0063] Furthermore, the power supply is an electret, which is charged by the excitation of the energy supply unit.

[0064] Specifically, the material of the electret is silicon dioxide.

[0065] Furthermore, the outer shell includes a waterproof layer from the inside to the outside and a shell layer with surface modification treatment; the waterproof layer is made of a polymer coating or silica gel; the method for obtaining the shell layer with surface modification treatment is: mixing the outer shell, negatively charged graphene, and a coupling agent at high speed to increase the electrostatic repulsion force between micro-nano geotechnical robots, prevent agglomeration, and prevent adsorption by the negatively charged rock fracture surface.

[0066] Specifically, the polymer coating includes polyimide, parylene, and epoxy resin.

[0067] Specifically, the shell layer is alumina ceramic or silicon nitride ceramic.

[0068] Specifically, the coupling agent is sodium dodecyl sulfate or sodium dodecylbenzenesulfonate.

[0069] It should be noted that considering that the inertial sensor is prone to error accumulation during operation, which affects the discrimination of the movement trajectory of the micro-nano geotechnical robot, a trajectory correction algorithm is preset in the control unit 15.

[0070] To ensure the accurate positioning and long-term operation of the micro-nano geotechnical robot, the inertial sensor and the control module are always in working condition, and the monitoring sensor works at a certain time interval according to the preset value of the control module. According to the preset time interval, the ultimate working duration of the micro-nano geotechnical robot is measured in an indoor simulation environment.

[0071] The communication module of the nano-robot described in this application is only used to realize the information exchange with the main control unit. Under underground conditions, data transmission and exchange do not occur between nano-robots. After the nano-robots are extracted to the surface through the extraction of the injection-production system 12 and then extracted through the extraction well, they are read by the external main control unit 15. Therefore, there are no strict restrictions on the antenna, and magnetic induction coils can be used for communication through magnetic field excitation, or radio frequency signals can be used for communication.

[0072] In a possible implementation manner, the injection-production system includes an injection well, an extraction well, a fluid pump, a filter, and a separator;

[0073] The injection well and the extraction well respectively inject and extract the micro-nano geotechnical robot cluster and the electromagnetic modified micro-nano particles through the fluid pump;

[0074] The filter and the separator are used to filter and separate the extracted micro-nano geotechnical robot cluster, electromagnetic modified micro-nano particles, fluid, and solid impurities.

[0075] It can be understood that after data transmission, the separated micro-nano geotechnical robot cluster can be re-injected underground with the electromagnetic modified micro-nano particles after passing through the work performance inspection and maintenance procedures, reducing the monitoring cost.

[0076] It should be noted that before injecting the micro-nano geotechnical robot cluster and the electromagnetic modified micro-nano particles, artificial fracture formation methods such as hydraulic fracturing and supercritical carbon dioxide fracturing should be used to form a migration channel. The electromagnetic modified micro-nano particles 14 should be injected underground before the micro-nano geotechnical robot cluster 11. When the concentration of the electromagnetic modified micro-nano particles recovered by the extraction well is basically unchanged, it is considered that a good electromagnetic signal transmission environment has been formed in the underground migration channel, and then the micro-nano geotechnical robot cluster is injected.

[0077] In a possible implementation manner, the power supply unit uses a DC power supply or an AC power supply. The two poles of the power supply are respectively placed between the formations of the target monitoring area, and an electric field is excited in the target monitoring area at an interval less than the ultimate working duration of the micro-nano geotechnical robot to supply power to the micro-nano geotechnical robot cluster.

[0078] Specifically, the two poles of the power supply are respectively placed in the wellbore walls of the injection well and the extraction well located in the formations of the target monitoring area.

[0079] In a possible implementation manner, the electromagnetic modified micro-nano particles are obtained by laying a conductive coating on the surface of ceramic particles.

[0080] Specifically, the conductive coating is zinc aluminum oxide or polyaniline.

[0081] In a possible implementation manner, the main control unit includes a data denoising module, a data filling module, and a three-dimensional visualization module;

[0082] The data denoising module is used to select effective components from different scale information components and reconstruct them to achieve denoising by using the time-domain denoising method;

[0083] The data filling module is used to make up for the data non-uniformity and structural missing caused by the failure, loss, and uneven distribution of the micro-nano geotechnical robot by using the machine learning method, and completely reconstruct the three-dimensional point cloud model of the target monitoring area;

[0084] The three-dimensional visualization module relies on the reconstructed three-dimensional point cloud model to visually display the spatial distribution and time variation of underground multi-source geological information.

[0085] Furthermore, the time-domain denoising module includes empirical mode decomposition and wavelet transform methods, and the working process is as Figure 3 shown. The specific denoising method is:

[0086] Perform empirical mode decomposition on the monitoring signal of the micro-nano geotechnical robot to obtain a finite number of intrinsic mode functions (IMF) y j ;

[0087] Determine the low-frequency IMF component and high-frequency IMF component according to the noise energy mutation principle, perform wavelet threshold denoising and signal reconstruction on the high-frequency IMF component, and finally output the denoised signal to achieve the extraction of effective signals.

[0088] Furthermore, the data filling module is used to perform data filling by using the GAIN data filling algorithm based on the generative adversarial network, and the algorithm structure is as Figure 4 shown.

[0089] Specifically, the method of data filling includes:

[0090] Generate a binary mask matrix for the original data matrix to mark known values and missing values;

[0091] Process the known values in the original data matrix with random noise variables to obtain a first data matrix, and fill in the missing values in the original data matrix to obtain a second data matrix;

[0092] Build a generator and a discriminator model, conduct adversarial training, and obtain an optimized generator; the generator is used to receive the first data matrix and generate predictions for missing values; the discriminator is used to receive the second data and hint matrix, and learn to identify actual observed data and data filled by the generator.

[0093] Use the optimized generator to fill in the missing values of the original data to obtain a complete data set.

[0094] Furthermore, the method for reconstructing the three-dimensional point cloud model of the target monitoring area is: perform Lagrangian interpolation processing on the obtained complete data set in time and space, and draw a three-dimensional cloud map of the monitoring data.

[0095] See Figure 5 , which is a schematic diagram of the monitoring work of the multi-source geological information in-situ intelligent perception system based on the micro-nano geotechnical robot provided by the present invention. Figure 5 In, the micro-nano geotechnical robot cluster circulates and injects the target monitoring area through the injection-production system to collect geological information and transmits the data to the main control unit for processing. The energy supply unit excites an electric field in the target monitoring area under the control of the main control unit and with the support of electromagnetic modified micro-nano particles.

[0096] Specifically, the process of constructing the multi-source geological information in-situ intelligent perception system of the micro-nano geotechnical robot is as Figure 6 shown, including the following steps:

[0097] S601: Determine the target monitoring area according to the previous geophysical exploration data;

[0098] S602: Set up the injection-production system, the main control unit and the energy supply unit;

[0099] S603: Inject fracturing fluid and electromagnetic modified micro-nano particles into the target monitoring formation until the recovery concentration of the production well remains unchanged;

[0100] S604: Inject micro-nano geotechnical robots and electromagnetic modified micro-nano particles into the target monitoring formation;

[0101] S605: The production well recovers micro-nano geotechnical robots and electromagnetic modified micro-nano particles;

[0102] S606: The main control unit receives the data and circulates and injects micro-nano geotechnical robots and electromagnetic modified nano-particles;

[0103] S607: Intelligent analysis to form a three-dimensional multi-field visualization model of the underground environment.

[0104] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system is characterized by: It includes a micro-nano geotechnical robot cluster, an energy supply unit, electromagnetically modified micro-nano particles, a main control unit, and an injection-production system; The micro-nano geotechnical robot cluster is used to identify changes in geological information in the target monitoring area; The injection and production system is used to inject and recover the micro-nano geotechnical robot cluster and electromagnetically modified micro-nano particles; The energy supply unit is used to excite an electric field in the target monitoring area to supply energy to the micro-nano geotechnical robot cluster; The electromagnetically modified micro-nano particles are used to improve the conductivity, magnetic permeability and dielectric constant of the fracture environment in which the micro-nano geotechnical robot is located, to improve the energy transmission efficiency of the energy supply unit to the micro-nano geotechnical robot cluster, and to serve as tracers and proppants to determine the fracture distribution morphology and prevent fracture closure; The main control unit is used to perform data analysis and three-dimensional modeling on the recovered micro-nano geotechnical robot, construct a multi-physical field three-dimensional visualization model according to the modeling results, and control the working behavior of the energy supply device in real time.

2. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system according to claim 1 is characterized in that: The micro-nano geotechnical robot in the micro-nano geotechnical robot cluster includes a shell and: Micro-nano sensors are used to monitor changes in geological information in the target monitoring area; A power source for supplying electrical energy; A communication module, including a transmitting antenna and a receiving antenna, is used to realize information exchange with the main control unit; The control module is used to control the working states of the micro-nano sensor and the communication module, store the collected geological information data and the corresponding time, and perform signal conditioning and data compensation on the micro-nano sensor.

3. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system according to claim 2 is characterized in that: The micro-nano sensor includes an inertial sensor and a monitoring sensor; The monitoring sensors include a temperature sensor, a stress sensor, a fluid pressure sensor, a microbial sensor, an ion concentration sensor and a pH sensor.

4. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system according to claim 2 is characterized in that: The power source is an electret, which is charged by the excitation of the energy supply unit.

5. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system according to claim 2 is characterized in that: The shell includes a waterproof layer from the inside to the outside and a shell layer with surface modification; the waterproof layer is made of a polymer coating or silica gel; the method for obtaining the shell layer with surface modification is: high-speed rotating and mixing the micro-nano geotechnical robot, negatively charged graphene and a coupling agent to enhance the electrostatic repulsion between the micro-nano geotechnical robots.

6. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system according to claim 1 is characterized in that: The injection-production system includes an injection well, a extraction well, a fluid pump, a filter and a separator; The injection well and the extraction well are used to inject and extract the micro-nano geotechnical robot cluster and the electromagnetically modified micro-nano particles respectively through fluid pumps; The filter and separator are used for filtering and separating the extracted micro-nano rock and soil robot cluster, electromagnetically modified micro-nano particles, fluid and solid impurities.

7. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system according to claim 1 is characterized in that: The energy supply unit adopts a DC power supply or an AC power supply, and the two poles of the power supply are respectively placed between the strata in the target monitoring area. The electric field is excited in the target monitoring area at an interval less than the maximum working time of the micro-nano geotechnical robot to supply energy to the micro-nano geotechnical robot cluster.

8. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system according to claim 1 is characterized in that: The electromagnetically modified micro-nano particles are obtained by laying a conductive coating on the surface of ceramic particles.

9. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system according to claim 1 is characterized in that: The main control unit includes a data noise reduction module, a data filling module and a three-dimensional visualization module; The data denoising module is used to select effective components from information components of different scales and reconstruct them to achieve denoising by adopting the time domain denoising method; The data filling module is used to use machine learning methods to make up for the data unevenness and structural loss caused by the failure, loss, and uneven distribution of the micro-nano geotechnical robot, and completely reconstruct the three-dimensional point cloud model of the target monitoring area; The 3D visualization module relies on the reconstructed 3D point cloud model to visualize the spatial distribution and temporal changes of underground multi-source geological information.

10. The micro-nano geotechnical robot multi-source geological information in-situ intelligent perception system according to claim 9 is characterized in that: The data denoising module is used to perform data denoising using empirical mode decomposition and wavelet transform methods; The data filling module is used to perform data filling using a GAIN data filling algorithm based on a generative adversarial network; The data filling method includes: Generate a binary mask matrix for the original data matrix to mark known values ​​and missing values; Using random noise variables to process known values ​​in the original data matrix to obtain a first data matrix, and filling missing values ​​in the original data matrix to obtain a second data matrix; Construct generator and discriminator models, perform adversarial training, and obtain an optimized generator; the generator is used to receive the first data matrix and generate predictions for missing values; the discriminator is used to receive the second data and prompt matrix, and learn to identify actual observed data and generator-filled data; The optimized generator is used to fill in the missing values ​​of the original data to obtain a complete data set.

Images