Method and system for intelligently predicting the effect of plugging the end of a working face based on a network model

Abstract

The application provides a method and system for intelligently predicting the effect of a working face end head leakage stopping based on a network model, which first uses computational fluid dynamics to simulate the effect of a Janus type foam slurry in stopping the end head under different working condition parameters, obtains a goaf oxygen concentration distribution image data, pre-processes the simulated data, and then trains a deep learning model based on the pre-processed data set, so that the trained deep learning model fits the mapping relationship between different working condition parameters and the goaf oxygen concentration distribution characteristics, and can quickly predict the leakage stopping process effect after inputting the actual working face end head leakage stopping material injection pressure, injection flow and injection position and other parameter values. The method provides a working face end head leakage stopping material with leakage stopping and oxygen isolation and cooling and fire extinguishing effects, and on this basis, the network model is combined to timely predict the end head leakage stopping process effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of mine fire prevention and extinguishing, and in particular to a method and system for intelligently predicting the effect of a working face end leakage plugging based on a network model. BACKGROUND

[0002] Coal spontaneous combustion induced fire not only leads to the loss of coal resources, but also interferes with the normal production of the mine and poses a threat to the life safety of miners. In the fully mechanized caving face, end leakage is one of the key factors leading to coal spontaneous combustion in the goaf. Therefore, researchers have developed various end leakage plugging technologies to prevent and control coal spontaneous combustion in the goaf. For example, the prior art introduces a gas bag type rapid leakage plugging device applied to the end of the fully mechanized caving face. The device is made of thermoplastic rubber material, which has good plasticity, high elasticity, heat resistance, compression deformation resistance, and excellent mechanical properties. Although this gas bag can play a role in plugging, it does not have the function of cooling and extinguishing fire. Moreover, since the gas bag itself is flammable, it has limited effect on the prevention and control of coal spontaneous combustion in the goaf of the high-temperature fully mechanized caving face. In addition, the prior art also uses solidified foam to plug the leakage channel at the end of the fully mechanized caving face. In order to further improve the effect of the solidified foam plugging application process, researchers usually use computational fluid dynamics (CFD) to simulate the accumulation and diffusion mode of the solidified foam at the end of the working face in advance to guide the design of the plugging process. However, CFD simulation still has the problems of long calculation period and insufficient timeliness, especially in the case of an emergency fire, it is difficult to provide an effective emergency response scheme in time.

[0003] Therefore, it is urgent to develop a new material that has the dual functions of plugging and oxygen isolation and cooling and fire extinguishing. However, how to intelligently, quickly and efficiently predict the effect of the end leakage plugging process of the new material has become a key technical problem that technicians in the field urgently need to solve. SUMMARY

[0004] The present application aims to provide a method and system for intelligently predicting the effect of an end leakage plugging process based on a network model, to provide a working face end leakage plugging material that has the effects of plugging and oxygen isolation and cooling and fire extinguishing, and on this basis, to intelligently, quickly and efficiently predict the effect of the end leakage plugging process in combination with a network model.

[0005] In a first aspect, the present application provides a method for predicting the effect of an end leakage plugging process based on a network model, which comprises:

[0006] CFD simulation is used to simulate the effect of Janus type foam slurry on sealing the end of fully mechanized caving face under different working condition parameters, and obtain the oxygen concentration distribution image data of the goaf; wherein the Janus type foam slurry is mixed by cement-based slurry, Janus type water-based foam and coagulant;

[0007] Data normalization and data enhancement preprocessing are performed on the data obtained by CFD simulation, and a data set is constructed;

[0008] A deep learning model is trained based on the data set, and a trained deep learning model is obtained; wherein the deep learning model includes an encoder, a discriminator, a decoder and a deep regression neural network model, and the trained deep learning model fits the mapping relationship between different working condition parameters and the oxygen concentration distribution characteristics of the goaf;

[0009] The current actual working condition parameters are input into the trained deep learning model, and the trained deep learning model outputs the oxygen concentration distribution image data of the goaf, which is used to evaluate the effect of the end sealing process.

[0010] In one possible implementation, training a deep learning model based on the data set specifically includes:

[0011] The working condition data set I n is input into the deep regression neural network model, and the output end of the deep regression neural network model is connected to the encoder;

[0012] The goaf oxygen concentration distribution image data set P n is input into the encoder, and the encoder applies a deep convolutional neural network to downsample the data set P n , compresses the data set P n into a reduced dimension data feature as a negative sample, and generates random data with the same dimension as the reduced dimension data feature from a certain random normal distribution as a positive sample, and inputs the positive sample and the negative sample into the discriminator, and the discriminator adjusts the encoder to compress the goaf oxygen concentration distribution characteristics, so that the decoder is infinitely close to the true random normal distribution;

[0013] The encoder inputs the reduced dimension data feature into the decoder, and the decoder applies a transpose convolutional neural network to upsample the reduced dimension data feature to obtain a reconstructed data set, and finally the trained deep learning model outputs the predicted goaf oxygen concentration distribution image.

[0014] In other possible implementations, the encoder adopts an objective function corresponding to the adversarial auto-encoding algorithm as follows:

[0015]

[0016] where x is the data in the real dataset, an array; z is the random data in the random distribution, an array; P is the data distribution probability function; d represents the discriminator, g represents the decoder, min g max d V(d,g) represents the cross-entropy loss of the generator to be minimized in the case that the discriminator maximizes the true, false picture cross-entropy loss V(d,g); represents the expectation of the real data set; represents the expectation of the random distribution data set.

[0017] In another possible implementation, the cement-based slurry is formed by uniformly mixing coal-based solid waste and Portland cement, which are two incombustible inorganic cementitious materials, and water, the mass fraction of the coal-based solid waste in the two inorganic cementitious materials being 29.3-35%, and the mass ratio of water to the two inorganic cementitious materials being 0.5-0.6;

[0018] The Janus type water-based foam is composed of a hydrophilic polymer, a hydrophobic nanoparticle and water; the mass fraction of the hydrophilic polymer in the Janus type water-based foam is 1-3.2%, the mass fraction of the hydrophobic nanoparticle in the Janus type water-based foam is 1.2-2%, and the addition amount of the Janus type water-based foam is 2-2.21 times the volume of the cement-based slurry;

[0019] The accelerator comprises one or more of liquid silicate, powdered silicate and aluminum sulfate, and the addition amount of the accelerator is 2.9-6.1% of the mass of the cement-based slurry.

[0020] In other possible implementations, the hydrophilic polymer comprises one or more of polyvinyl alcohol, xanthan gum and sodium alginate, the hydrophobic nanoparticle comprises one or more of silicon dioxide, titanium dioxide and carbon nanotubes, and the working condition parameters include grouting pressure, grouting flow rate and grouting pipe position.

[0021] In another possible implementation, the data obtained by the CFD simulation is preprocessed by data normalization and data enhancement, including: normalizing the working condition parameters in the CFD simulation by using the z-score normalization method; and performing data enhancement on the goaf oxygen concentration distribution image data by using the Transforms tool of the TorchVision module in PyTorch.

[0022] In another possible implementation, the encoder is composed of 7 layers of convolutional neural network layers, the decoder is composed of 8 layers of transposed convolutional neural network layers, and the discriminator uses 2 layers of convolutional neural network layers for down-sampling and 1 layer of fully connected layer for feature compression.

[0023] In the second aspect, the application further provides a system for predicting the effect of a tip plugging process based on a network model, wherein:

[0024] The simulation data module is used to obtain oxygen concentration distribution image data of a goaf at the end of a fully-mechanized caving face under different working condition parameters by using CFD simulation, wherein the Janus type foam slurry is mixed by a cement-based slurry, a Janus type water-based foam and a coagulant.

[0025] The data set construction module is used to perform data normalization and data enhancement preprocessing on the data obtained by the CFD simulation, and construct a data set.

[0026] The training module is used to train a deep learning model based on the data set, and obtain a trained deep learning model, wherein the deep learning model comprises an encoder, a discriminator, a decoder and a deep regression neural network model, and the trained deep learning model fits a mapping relationship between different working condition parameters and oxygen concentration distribution characteristics of the goaf.

[0027] The model application module is used to input current actual working condition parameters to the trained deep learning model, and the trained deep learning model outputs oxygen concentration distribution image data of the goaf, and the oxygen concentration distribution image data of the goaf is used to evaluate the effect of the end sealing process.

[0028] In a third aspect, the embodiments of the present application further provide a computer readable storage medium, which comprises a computer program, and when the computer program runs on a processor, the processor executes the method of the first aspect.

[0029] In a fourth aspect, the embodiments of the present application provide an electronic device comprising a processor and a memory. The memory is used to store one or more computer programs, and when the one or more computer programs stored in the memory are executed by the processor, the processor can implement the method of any one of the possible embodiments of the first aspect.

[0030] In a fifth aspect, the embodiments of the present application further provide a computer program product, and when the computer program product runs on a processor, the processor executes the method of any one of the possible embodiments of any one of the aspects.

[0031] The method and system for intelligently predicting the effect of the end sealing process based on the network model provided by the embodiments of the present application have the following beneficial effects:

[0032] (1) The Janus type foam slurry is developed, and the material is endowed with the performance advantages of sealing and oxygen isolation and cooling and fire extinguishing.

[0033] (2) The method can intelligently, quickly and efficiently predict the plugging effect of the end of the fully mechanized caving face, and can solve the problems of long simulation calculation period and poor timeliness of CFD simulation, and efficiently deal with the possible emergency fire in the goaf of the fully mechanized caving face.

[0034] (3) A discriminator is constructed in the encoder-decoder process, and the encoder is used as a generator, and the feature mapping result of the oxygen concentration distribution in the goaf and a random normal distribution are trained through the discriminator, so as to promote the oxygen concentration distribution in the goaf to generate a high-quality reconstructed image through the decoder.

[0035] (4) A convolutional autoencoder model is established to realize dimension reduction and reconstruction of the oxygen concentration distribution data, the encoder uses a deep convolutional neural network instead of a linear layer, and the decoder uses a transposed convolutional neural network for upsampling, thereby solving the problems of weak representation ability of linear layers for high-dimensional data, overfitting or underfitting of feature extraction. BRIEF DESCRIPTION OF DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0037] Figure 1 A working face goaf schematic diagram of a "U" shape ventilation system is provided for the embodiments of the present application.

[0038] Figure 2 A method flowchart for intelligently predicting the process effect of plugging the end based on a network model is provided for the embodiments of the present application.

[0039] Figure 3 A structure schematic diagram of a deep learning model is provided for the embodiments of the present application.

[0040] Figure 4 An electronic device structure schematic diagram is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0041] With the research and progress of artificial intelligence technology, artificial intelligence is researched and applied in many fields, for example, common smart home, intelligent customer service, virtual assistant, smart speaker, intelligent marketing, unmanned driving, autonomous driving, robot, intelligent medical treatment, etc. It is believed that with the development of technology, artificial intelligence will be applied in more fields and play an increasingly important role.

[0042] Machine learning is a multi-disciplinary subject, involving probability theory, statistics, approximation theory, convex analysis, algorithmic complexity theory, etc. It is a specialized study of how computers simulate or implement human learning behavior to acquire new knowledge or skills, reorganize existing knowledge structure, and continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental approach to making computers intelligent. Its applications are widespread in various fields of artificial intelligence. Machine learning and deep learning generally include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, etc.

[0043] In the description of embodiments of the present application, the terms used in the following embodiments are only for the purpose of describing the specific embodiments and are not intended to be limiting to the present application. As used in the specification and the appended claims of the present application, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "at least one," "one or more," as used in the following embodiments of the present application, mean one or two or more (including two). The term "and / or" is used to describe the association relationship of the associated objects, which means that there can be three relationships; for example, A and / or B can represent the following cases: A exists alone, A and B exist together, B exists alone, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after it.

[0044] In this specification, the reference to "one embodiment" or "some embodiments" etc. means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present application. Thus, the appearance of the phrases "in one embodiment", "in some embodiments", "in other embodiments", "in yet some embodiments", etc. in various places in the specification is not necessarily all referring to the same embodiment, but means "one or more but not all embodiments", unless otherwise specifically stated. The terms "include", "contain", "have" and their variants mean "including but not limited to", unless otherwise specifically stated. The term "connected" includes direct connection and indirect connection, unless otherwise stated. "First", "second", etc. are only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features.

[0045] In the embodiments of the present application, the word "exemplary" or "for example" is used to mean serving as an example, instance, or illustration, at 99 least with regard to the matters described at that point in the disclosure. Any embodiment or implementation described as "exemplary" or "for example" in the embodiments of the present application should not be construed as preferred or advantageous over other embodiments or implementations. Rather, use of the word "exemplary" or "for example" is intended to present concepts in a concrete manner.

[0046] In view of the problems of the prior art, the present application provides a new plugging material applied to the end of a fully mechanized caving face. The material has the effects of plugging, oxygen isolation, temperature reduction and fire extinguishing, so as to solve the problem of spontaneous combustion of residual coal in the goaf caused by air leakage at the upper and lower ends of the fully mechanized caving face.

[0047] As shown in Figure 1 For the goaf of the "U"-shaped ventilation system, according to the size of air leakage and the possibility of residual coal occurrence, the goaf can be divided into three zones: heat dissipation zone, spontaneous combustion zone (oxidation and temperature rise zone) and asphyxiation zone. According to the previous mining experience, it is known that a certain free space 1 is easy to form at the two ends of the fully mechanized caving face, and then air leakage is intensified, and if not timely plugged, it is easy to cause spontaneous combustion of the goaf, which brings great hidden danger to safety production. The new material provided by the present application is used to plug the free space 1 at the two ends of the fully mechanized caving face.

[0048] The new plugging material provided by the embodiments of the present application is a Janus type foam slurry, which is mixed by a cement-based slurry, a Janus type water-based foam and a coagulant. The interface formed by the hydrophobic nanoparticles and the hydrophilic polymer in the slurry has both hydrophilicity and hydrophobicity, which enhances the stability of the bubbles. In the field of modern science and medicine, Janus type may be used to describe that some molecules, proteins or drugs have two different functional or structural characteristics.

[0049] The Janus type foam slurry is characterized in that: the cement-based slurry is formed by uniformly stirring two incombustible inorganic cementitious materials of coal-based solid waste and Portland cement and water, the coal-based solid waste accounts for 29.3-35% of the mass of the two inorganic cementitious materials, and the mass ratio of water to the two inorganic cementitious materials is 0.5-0.6. The Janus type water-based foam is composed of a hydrophilic polymer, a hydrophobic nanoparticle and water. The hydrophilic polymer includes one or more of polyvinyl alcohol, xanthan gum and sodium alginate, and the hydrophobic nanoparticle includes one or more of silicon dioxide, titanium dioxide and carbon nanotubes. The hydrophilic polymer accounts for 1-3.2% of the mass of the Janus type water-based foam, the hydrophobic nanoparticle accounts for 1.2-2% of the mass of the Janus type water-based foam, and the addition amount of the Janus type water-based foam is 2-2.21 times the volume of the cement-based slurry. The setting accelerator includes one or more of liquid silicate, powdered silicate and aluminum sulfate, and the addition amount of the setting accelerator is 2.9-6.1% of the mass of the cement-based slurry. In addition, the Janus type foam slurry can be converted into a solidified foam within 10-30 seconds, and has the properties of efficient stacking and long-term stability.

[0050] Based on the Janus type foam slurry with the effects of plugging and oxygen isolation and temperature reduction and fire extinguishing, the embodiment further trains a deep learning model of the Janus type foam slurry for plugging the end of the fully mechanized caving face under different working conditions. The main process is: first, CFD simulation is used to obtain oxygen concentration distribution image data of the goaf under different working condition parameters, the image data obtained by simulation is preprocessed, and then a deep learning model is trained based on the preprocessed data set. The trained deep learning model fits the mapping relationship between different working condition data and the oxygen concentration distribution characteristics of the goaf, and can quickly predict the plugging process effect after inputting the actual parameters such as grouting pressure, grouting flow and grouting position.

[0051] As shown in Figure 2 The embodiment of the present application provides a method for intelligently predicting the end plugging process effect based on a network model, and the flow of the method can be executed by an electronic device. The method comprises the following steps:

[0052] S201, CFD simulation is used to obtain the effect of the Janus type foam slurry in plugging the end of the fully mechanized caving face under different working condition parameters, and oxygen concentration distribution image data of the goaf is obtained.

[0053] In this embodiment, the different working condition parameters simulated by CFD involve grouting pressure, grouting flow and grouting pipe position, at least 27 working conditions. CFD simulation: first, based on the actual situation of the working face goaf and the Janus type foam slurry plugging end application process, an equal proportion physical model is established. Second, the physical model is meshed. Finally, the distribution of oxygen concentration in the goaf of the fully mechanized caving face under different working conditions is obtained. Through CFD simulation research, the distribution of oxygen concentration in the goaf of the fully mechanized caving face under different working conditions can be obtained, which provides data support for the training of subsequent deep learning model.

[0054] S202, data normalization and data enhancement preprocessing are performed on the data obtained by CFD simulation, and a data set is constructed.

[0055] Specifically, the first step is data normalization: the working condition parameters in CFD simulation are normalized by, for example, z-score normalization method to adapt to the input requirements of deep regression neural network. The second step is data enhancement: for example, the Transforms tool of TorchVision module in PyTorch is used to enhance the data in order to improve the generalization ability of deep learning model. The third step is data set construction: the input data set of the entire deep learning algorithm model is obtained by combining the normalized and enhanced data. After determining the input data set of the entire deep learning algorithm model, the training set and the validation set are constructed, and the training data and the validation data are loaded by rewriting the Dataset and Dataloader in PyTorch, which provides a data basis for the training and validation of deep learning model.

[0056] S203, a deep learning model is trained based on the preprocessed data set, and a trained deep learning model is obtained. The deep learning model includes an encoder, a discriminator, a decoder and a deep regression neural network model, and the trained deep learning model fits the mapping relationship between different working condition parameters and goaf oxygen concentration distribution characteristics.

[0057] S204, the current actual working condition parameters are input to the trained deep learning model, and the trained deep learning model outputs goaf oxygen concentration distribution image data, which is used to evaluate the effect of end plugging process.

[0058] In combination with the figure, Figure 3 A structure of a deep learning model is shown, which includes two parts, the first part is: deep convolutional adversarial auto-encoding algorithm; the second part is: deep regression neural network. Among them, the preprocessed working condition parameter data set is input to the working condition data set I of the deep regression neural network in the deep learning model Figure 3 n ​10. The image data of the oxygen concentration distribution in the goaf under different working conditions after preprocessing is Figure 3 the oxygen concentration distribution data set P of the goaf input into the encoder 2 n 1.

[0059] From Figure 3 It can be seen that the working condition data set I n 10 is input into the deep regression neural network model 9 and then connected to the encoder 2. The data set P n 1 is input into the encoder 2, and the encoder 2 applies a deep convolutional neural network to downsample the data set P n 1, compressing the data set P n 1 into the reduced dimension data feature 3. On the one hand, the reduced dimension data feature 3 output by the encoder 2 is taken as a negative sample (fake-), and random data of the same dimension as the reduced dimension data feature 3 is generated from a certain random normal distribution 7 as a positive sample (real+), and the positive sample (real+) and the negative sample (fake-) are input into the discriminator 8, which adjusts the encoder 2 to compress the oxygen concentration distribution feature of the goaf, so that the decoder 4 approaches the real random normal distribution 7. On the other hand, the encoder 2 inputs the reduced dimension data feature 3 into the decoder 4, and the decoder 4 applies a transposed convolutional neural network to upsample the reduced dimension data feature 3, obtaining the reconstructed data set 5. Finally, the model finally outputs the predicted oxygen concentration distribution image 6 of the goaf. By applying the above model, different working condition data and oxygen concentration distribution features of the goaf can be fitted. When the actual parameters of the grouting pressure, grouting flow and grouting position are input into the trained network model, the intelligent and rapid prediction of the effect of the high-efficiency air leakage plugging process can be realized.

[0060] Figure 3 In the encoder 2 and the decoder 4, a discriminator 8 is constructed between the encoder 2 and the decoder 4, and the decoder 4 is taken as a generator, and the oxygen concentration distribution feature mapping result of the goaf and a random normal distribution are trained by the discriminator 8, so that the oxygen concentration distribution feature generates a high-quality reconstructed image through the decoder, and the objective function of the generative adversarial coding algorithm model is as follows. In the formula, a convolutional layer is applied for downsampling, a transposed convolutional layer is applied for upsampling, the encoder 2 is composed of 7 convolutional neural network layers, the decoder 4 is composed of 8 transposed convolutional neural network layers, and the discriminator 8 uses 2 convolutional neural network layers for downsampling and 1 fully connected layer for outputting compressed features.

[0061]

[0062] In the formula, x is the data in the real data set, an array; z is the random data in the random distribution, an array; P is the data distribution probability function; d represents the discriminator, g represents the decoder, min g max dV(d,g) represents the cross-entropy loss of the decoder to be minimized by the discriminator maximizing the true, false picture cross-entropy loss V(d,g); represents the real data set expectation; represents the random distribution data set expectation.

[0063] The deep learning model provided in the embodiment mainly performs three processes of data dimension reduction, data regression and data reconstruction. Among them, the model is trained based on the CFD simulation result, the data characteristics of the oxygen concentration distribution result of the goaf of the fully mechanized caving face are fitted, and the reconstruction of the oxygen concentration distribution data of the goaf is realized. By training the data regression model, the distribution characteristics of different working condition data and the oxygen concentration of the goaf are fitted, the deep learning agent model loads the weights of each part of the trained data dimension reduction, data regression and data reconstruction model, realizes the input of the grouting pressure, grouting flow and grouting position parameters, maps the oxygen concentration feature vector of the goaf after air leakage through the deep regression neural network model, and then reconstructs the oxygen concentration distribution image of the goaf through the encoder in the deep convolutional generative adversarial algorithm, so as to realize intelligent and rapid prediction of the high-efficiency air leakage stopping process.

[0064] The beneficial effects of the embodiment over the prior art are as follows: (1) a Janus type foam slurry is developed, which endows the material with the performance advantages of air leakage stopping and oxygen isolation and cooling and fire extinguishing. (2) The method can intelligently, quickly and efficiently predict the air leakage stopping effect at the end of the fully mechanized caving face, and can solve the problems of long simulation calculation period and poor timeliness of CFD simulation, and can efficiently deal with possible emergency fire in the goaf of the fully mechanized caving face. (3) In the encoder-decoder process, a discriminator is constructed, and the encoder is used as a generator, and the oxygen concentration distribution feature mapping result and a random normal distribution are trained through the discriminator, so as to make the oxygen concentration distribution feature of the goaf generate a high-quality reconstructed image through the decoder. (4) A convolutional autoencoder model is established to realize the dimension reduction and reconstruction of the oxygen concentration distribution data, the encoder uses a deep convolutional neural network instead of a linear layer, and the decoder uses a transposed convolutional neural network for upsampling, so as to solve the problems of weak representation ability of linear layers for high-dimensional data, overfitting or underfitting of feature extraction.

[0065] All related contents of each step involved in the above method embodiment can be cited to the function description of the corresponding function module, which will not be repeated here.

[0066] The above system can be executed by a chip or a chip module. Each module / unit contained in each device and product described in the above embodiments can be a software module / unit, or a hardware module / unit, or part of a software module / unit and part of a hardware module / unit.

[0067] In some other embodiments of the present application, the electronic device is as followsFigure 4 As shown, the electronic device can integrate the above system, and the hardware components include one or more processors 401, a memory 402, a display 403, one or more application programs (not shown), and one or more computer programs 404, which can be connected through one or more communication buses 405. The one or more computer programs 404 are stored in the memory 402 and configured to be executed by the one or more processors 401, and the one or more computer programs 404 include instructions.

[0068] The application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a computer to implement the method in the above method embodiments. The specific beneficial effects can be referred to the above method embodiments.

[0069] The application further provides a computer program product, which is executed by a computer to implement the method in the above method embodiments. The specific beneficial effects can be referred to the above method embodiments.

[0070] Through the description of the above embodiments, those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example for illustration, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the above described system, device and unit can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0071] Each functional unit in each embodiment of the embodiments of the application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0072] When the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the embodiments of the application essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor execute all or part of the steps of the methods described in the embodiments of the application. The foregoing storage medium includes: a flash memory, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, and various media that can store program codes.

[0073] The above merely provides the specific implementation of the embodiments of the present application, but the protection scope of the embodiments of the present application is not limited thereto, any change or replacement within the technical scope disclosed by the embodiments of the present application should be covered in the protection scope of the embodiments of the present application. Therefore, the protection scope of the embodiments of the present application should be subject to the protection scope of the claims.

Claims

1. A method for intelligently predicting the effect of plugging the end of a working face based on a network model, characterized in that, The application relates to a method for predicting the oxygen concentration distribution in a goaf based on a deep learning model. The method comprises the following steps: CFD simulation is performed on the effect of a Janus-type foam slurry on plugging the end of a fully-mechanized caving face under different working condition parameters, and oxygen concentration distribution image data of the goaf are obtained; the Janus-type foam slurry is formed by mixing a cement-based slurry, a Janus-type water-based foam and a coagulant; Data normalization and data enhancement preprocessing are performed on the data obtained by the CFD simulation, and a data set is constructed; A deep learning model is trained based on the data set, and a trained deep learning model is obtained; the deep learning model comprises an encoder, a discriminator, a decoder and a deep regression neural network model, and the trained deep learning model fits the mapping relationship between different working condition parameters and the oxygen concentration distribution characteristics of the goaf; Actual working condition parameters are input into the trained deep learning model, and the trained deep learning model outputs oxygen concentration distribution image data of the goaf, which is used for evaluating the effect of a plugging process at the end of the fully-mechanized caving face; The cement-based slurry is formed by uniformly stirring two kinds of incombustible inorganic cementitious materials, namely coal-based solid waste and Portland cement, and water; the coal-based solid waste accounts for 29.3-35% of the mass of the two kinds of inorganic cementitious materials; and the mass ratio of water to the two kinds of inorganic cementitious materials is 0.5-0.6; The Janus-type water-based foam is composed of a hydrophilic polymer, a hydrophobic nanoparticle and water; the hydrophilic polymer accounts for 1-3.2% of the mass of the Janus-type water-based foam; the hydrophobic nanoparticle accounts for 1.2-2% of the mass of the Janus-type water-based foam; and the addition amount of the Janus-type water-based foam is 2-2.21 times the volume of the cement-based slurry; 2. The method of claim 1, wherein, The coagulant comprises one or more of liquid silicate, powdered silicate and aluminum sulfate, and the addition amount of the coagulant is 2.9-6.1% of the mass of the cement-based slurry. inputting a working condition data set to the deep regression neural network model, an output end of the deep regression neural network model being connected to an encoder; Gob oxygen concentration distribution image dataset inputting the encoder, the encoder applying a deep convolutional neural network to the dataset down-sampling the dataset compressing into a reduced dimension data feature as a negative sample; and generating random data of the same dimension as the reduced dimension data feature from a determined random normal distribution as a positive sample, inputting the positive sample and the negative sample together into the discriminator, the discriminator adjusting the encoder to compress the gob oxygen concentration distribution feature, so that the decoder approaches the true random distribution normal distribution infinitely The deep learning model is trained based on the data set, and specifically comprises:

3. The method of claim 1, wherein, The encoder inputs the reduced data features into the decoder, the decoder applies a transposed convolutional neural network to up-sample the reduced data features, obtains a reconstructed data set, and finally the trained deep learning model outputs predicted oxygen concentration distribution image data of the goaf. ； where x is data in the real data set, an array; z is random data in a random distribution, an array; P is a data distribution probability function; d represents a discriminator, and g represents a decoder, represents that the generator wants to minimize this cross-entropy loss under the condition that the discriminator maximizes the cross-entropy loss of real and fake pictures; represents the expectation of the real data set; represents the expectation of the random distribution data set.​ 4. The method of claim 1, wherein, The objective function corresponding to the adversarial coding algorithm of the encoder is as follows:

5. The method according to any one of claims 1 to 4, characterized in that, The hydrophilic polymer comprises one or more of polyvinyl alcohol, xanthan gum and sodium alginate; the hydrophobic nanoparticle comprises one or more of silicon dioxide, titanium dioxide and carbon nanotubes; and the working condition parameters comprise grouting pressure, grouting flow and grouting pipe position. The data normalization and data enhancement preprocessing of the data obtained by the CFD simulation comprises: The working condition parameters in the CFD simulation are normalized by using a z-score normalization method; 6. The method according to any one of claims 1 to 4, characterized in that, The oxygen concentration distribution image data of the goaf are subjected to data enhancement by using the Transforms tool of the TorchVision module in PyTorch.

7. A system for intelligently predicting the effect of plugging a face end based on a network model, characterized by, The encoder is composed of 7 convolutional neural network layers, the decoder is composed of 8 transposed convolutional neural network layers, and the discriminator uses 2 convolutional neural network layers for down-sampling and 1 fully-connected layer for feature compression. The application relates to a method for predicting the oxygen concentration distribution in a goaf based on a deep learning model. The method comprises the following steps: The simulation data module is configured to simulate different working condition parameters by using CFD, and to simulate oxygen concentration distribution image data of a goaf at an end of a fully-mechanized caving face under different working condition parameters by using a Yannus type foam slurry; the Yannus type foam slurry is formed by mixing a cement-based slurry, a Yannus type water-based foam and a coagulant; The data set construction module is configured to perform data normalization and data enhancement preprocessing on the data obtained by CFD simulation, and to construct a data set; The training module is configured to train a deep learning model based on the data set, and to obtain a trained deep learning model; the deep learning model includes an encoder, a discriminator, a decoder and a deep regression neural network model, and the trained deep learning model fits a mapping relationship between different working condition parameters and oxygen concentration distribution characteristics of the goaf; The model application module is configured to input current actual working condition parameters to the trained deep learning model, and the trained deep learning model outputs oxygen concentration distribution image data of the goaf, which is used to evaluate the effect of the end sealing process; The cement-based slurry is formed by uniformly stirring two kinds of incombustible inorganic cementitious materials, i.e., coal-based solid waste and Portland cement, and water, the coal-based solid waste accounts for 29.3-35% of the mass of the two kinds of inorganic cementitious materials, and the mass ratio of water to the two kinds of inorganic cementitious materials is 0.5-0.6; The Yannus type water-based foam is composed of a hydrophilic polymer, a hydrophobic nanoparticle and water; the hydrophilic polymer accounts for 1-3.2% of the mass of the Yannus type water-based foam, the hydrophobic nanoparticle accounts for 1.2-2% of the mass of the Yannus type water-based foam, and the addition amount of the Yannus type water-based foam is 2-2.21 times the volume of the cement-based slurry; The coagulant includes one or more of liquid silicate, powdered silicate and aluminum sulfate, and the addition amount of the coagulant is 2.9-6.1% of the mass of the cement-based slurry.

8. A computer-readable storage medium having stored therein a program, characterized by The program is executed by the processor to implement the method of any one of claims 1-6.

9. An electronic device, comprising: The electronic device includes a memory and a processor, and the memory stores a program executable on the processor; when the program is executed by the processor, the electronic device implements the method of any one of claims 1-6.

Images