Intelligent mine dangerous state prediction method and device based on digital twin

Abstract

The present invention discloses a method and device for predicting dangerous conditions in intelligent mines based on digital twins, comprising: arranging sound sensors in various key areas of the mine and preprocessing the collected sound signal data to achieve comprehensive monitoring and data collection of the mine environment; and using feature extraction technology and a digital twin network to accurately extract the characteristics of the sound signal and establish a digital twin model, so as to more accurately simulate and predict dangerous conditions in the mine environment; and at the same time, using historical data and dangerous event records, combined with a digital twin model to predict sound signal data, more accurately analyzing the propagation path of the dangerous condition and the areas and equipment that may be affected, thereby improving the accuracy and reliability of the prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twin technology and mine safety management, and in particular to a method and device for predicting dangerous states of smart mines based on digital twins. Background Art

[0002] Digital twin technology is an emerging intelligent modeling method that aims to accurately simulate and predict the behavior and performance of physical entities through digital models. Digital twin technology combines traditional simulation modeling with modern data-driven methods to achieve dynamic simulation and prediction of physical systems, providing support for intelligent decision-making in smart factories, smart cities and other fields.

[0003] In the field of mine safety management, the prediction and monitoring of dangerous conditions has always been an extremely challenging task. Traditional mine safety management mainly relies on manual inspections and fixed sensor monitoring, which has problems such as blind spots in monitoring and lack of real-time performance.

[0004] There are several major problems with traditional mine safety management: 1. Insufficient sensor layout: The sensor layout of traditional mine safety management systems is usually limited to fixed positions, which makes it impossible to fully cover all key areas of the mine, and the comprehensiveness and accuracy of monitoring data are limited; 2. Difficulty in feature extraction: Traditional sound signal feature extraction methods are affected by factors such as noise and interference. The extracted features are not accurate enough to reflect the actual situation of the sound signal, affecting the accuracy of subsequent modeling and prediction; 3. Inaccurate model modeling: Traditional modeling methods have limited modeling capabilities for complex mining environments and changeable dangerous conditions, and it is difficult to accurately capture the complex relationship between sound signals and dangerous conditions; 4. Low prediction accuracy: Traditional prediction methods are usually based on simple rules or empirical models, and have limited prediction accuracy for complex mining environments and dangerous conditions, and are prone to underreporting or false alarms. Summary of the invention

[0005] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0006] In view of the above existing problems, the present invention is proposed.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: collecting and preprocessing the sound signal data of the sound sensors arranged in each key area of ​​the mine;

[0008] The sound features of the preprocessed sound signal data are extracted using feature extraction technology, and the sound features are input into the digital twin network for modeling to form a digital twin model based on a smart mine;

[0009] The digital twin model is trained using a convolutional neural network to learn the correlation between the sound features and the dangerous state;

[0010] The digital twin model after training is used to predict the sound signal data to be predicted to determine whether it is in a dangerous state; wherein:

[0011] When the confidence of the prediction result is greater than the set threshold, the value is 1, indicating that there is a dangerous state;

[0012] When the confidence of the prediction result is ≤ the set threshold, the value is 0, indicating that there is no dangerous state.

[0013] As a preferred solution of the method for predicting dangerous states of intelligent mines based on digital twins described in the present invention, the sound signal data includes the sound of mechanical equipment operation, the sound of geological activities and environmental noise;

[0014] The key areas include mining operation areas, equipment concentration areas, and rock stability monitoring areas.

[0015] As a preferred solution of the method for predicting dangerous states of intelligent mines based on digital twins described in the present invention, the preprocessing removes the interference of environmental noise on the sound signal through a filtering algorithm, and then normalizes the sound signal to ensure that the data is within the same dimensional range. The filtering algorithm is used to remove the interference of environmental noise on the sound signal, which can be expressed by the following formula:

[0016] Filtered Sound(t) = F(Sound(t),Noise(t))

[0017] Among them, Filtered Sound (t) represents the value of the sound signal processed by the filtering algorithm at time t. represents the value of the original sound signal at time t, represents the value of the ambient noise at time t, A function representing a filtering algorithm, which accepts the original sound signal and the ambient noise signal as parameters and returns the filtered sound signal.

[0018] As a preferred solution of the method for predicting dangerous states of intelligent mines based on digital twins described in the present invention, the sound characteristics include spectrum characteristics, energy characteristics and time domain characteristics;

[0019] The spectrum feature, the energy feature, and the time domain feature are extracted and expressed by the following formulas respectively:

[0020]

[0021]

[0022]

[0023] Where (j) represents the frequency of the sound signal The spectrum characteristics under are obtained through Fourier transform. (T) represents the energy characteristics of the sound signal in the time period T, which is the square integral of the sound signal. ( ) indicates the sound signal at time point The time domain feature is the convolution integral of the sound signal and the unit impulse function.

[0024] The extracted sound features are then subjected to dimensionality reduction processing to reduce the data dimension.

[0025] As a preferred solution of the method for predicting dangerous states of smart mines based on digital twins described in the present invention, the sound features are input into the digital twin network for modeling to form a digital twin model based on smart mines, and its mathematical expression formula is as follows:

[0026]

[0027] in, represents the sound features extracted from X, X represents the preprocessed sound signal data, M represents the model of the digital twin network, and Y represents the output of the digital twin network, that is, the prediction result of the mine status.

[0028] As a preferred solution of the method for predicting dangerous states of intelligent mines based on digital twins described in the present invention, a convolutional neural network is used to train the digital twin model to learn the correlation between sound features and dangerous states, which is expressed by the following mathematical expression formula:

[0029]

[0030] in, represents the sound characteristics, B represents the danger characteristics, Represents the extraction and learning of sound features by convolutional neural network. is the representation function of the dangerous state, is a Gaussian function used to measure The distance from B, is the standard deviation of the Gaussian function, which is used to control the range and strength of the correlation. Indicates the correlation between the sound characteristics and the dangerous state;

[0031] The value range of is (0,1). The closer the value is to 1, the higher the correlation between the sound feature and the dangerous state;

[0032] when The closer the value is to 0, the lower the correlation between the sound feature and the dangerous state.

[0033] As a preferred solution of the method for predicting dangerous states of intelligent mines based on digital twins described in the present invention, the digital twin model after training is used to predict the sound signal data to be predicted to determine whether it is in a dangerous state, including:

[0034]

[0035]

[0036] Among them, D is a binary variable, indicating whether a dangerous state exists, 1 means it exists, and 0 means it does not exist. represents the set threshold, P represents the confidence of the prediction result, n represents the sample size, and c represents the probability of the existence of a dangerous state.

[0037] As a preferred solution of the intelligent mine dangerous state prediction device based on digital twin described in the present invention, it includes:

[0038] one or more processors;

[0039] A memory stores operable instructions, which, when executed by the one or more processors, cause the one or more processors to perform operations, including the process of the aforementioned smart mine dangerous state prediction method based on digital twins.

[0040] Beneficial effects of the present invention: The present invention realizes comprehensive monitoring and data collection of the mine environment by arranging sound sensors in various key areas of the mine and preprocessing the collected sound signal data; and with the help of feature extraction technology and digital twin network, it can accurately extract the characteristics of the sound signal and establish a digital twin model, so as to more accurately simulate and predict the dangerous conditions in the mine environment; at the same time, by using historical data and dangerous event records, combined with the digital twin model to predict the sound signal data, it can more accurately analyze the propagation path of the dangerous condition and the areas and equipment that may be affected, thereby improving the accuracy and reliability of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0042] Figure 1 This is a schematic flow chart of the method for predicting dangerous conditions in a smart mine based on digital twins shown in the present invention. DETAILED DESCRIPTION

[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0044] Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without making any creative work should fall within the scope of protection of the present invention.

[0045] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0046] Example 1

[0047] According to an embodiment of the present invention, Figure 1 The flowchart shown is a method for predicting dangerous conditions in a smart mine based on digital twins, which specifically includes the following steps:

[0048] S1: Collect and pre-process the sound signal data of the sound sensors arranged in each key area of ​​the mine. Among them, what needs to be explained in this step is:

[0049] Preprocessing uses a filtering algorithm to remove the interference of environmental noise on the sound signal, and then normalizes it to ensure that the data is within the same dimensional range. The filtering algorithm is used to remove the interference of environmental noise on the sound signal, which can be expressed by the following formula:

[0050] Filtered Sound(t) = F(Sound(t),Noise(t))

[0051] Among them, Filtered Sound (t) represents the value of the sound signal processed by the filtering algorithm at time t. represents the value of the original sound signal at time t, represents the value of the ambient noise at time t, A function representing a filtering algorithm, which accepts the original sound signal and the ambient noise signal as parameters and returns the filtered sound signal.

[0052] In an optional embodiment, the sound signal data includes the sound of mechanical equipment operation, the sound of geological activities, and environmental noise.

[0053] As an example, the operating sound of mechanical equipment: the working sound of mechanical equipment inside the mine (such as drilling rigs, transportation equipment), such as roar, friction, and collision.

[0054] As an example, geological activity sounds: sounds produced by rock cracks, geological changes, etc., such as rock movement, collapse, and earthquakes.

[0055] As an example, environmental noise: noise generated by human activities, natural environment, etc., such as wind and water.

[0056] In an optional embodiment, the key areas include mining operation areas, equipment concentration areas, and rock stability monitoring areas.

[0057] As an example, the mining operation area of ​​a mine: such as the excavation working face and the transportation channel.

[0058] As an example, equipment concentration area: an area with dense equipment, such as a mine workshop or machinery equipment installation area.

[0059] As an example, rock stability monitoring areas: areas where rock stability is poor or prone to rock collapse, areas where the sound of geological activity needs to be monitored.

[0060] In an optional implementation, the sound signal data also includes time domain data, frequency domain data, energy data and spectrogram data.

[0061] As an example, the time domain data is the waveform data of the sound signal, the frequency domain data is the spectrum data of the sound signal, the energy data is the energy distribution data of the sound signal, and the spectrogram data is the time-frequency spectrum image data of the sound signal.

[0062] The above normalization process can be implemented by existing technologies, and this step will not be described in detail.

[0063] Preferably, the embodiment of the present invention inputs the original sound signal and the environmental noise signal into a filtering algorithm, and obtains a filtered sound signal after algorithm processing, thereby effectively removing the interference of the environmental noise on the sound signal and improving the quality and accuracy of the sound signal.

[0064] Preferably, this step realizes comprehensive monitoring and data collection of the mine environment, providing sufficient data support for subsequent dangerous state prediction.

[0065] S2: Use feature extraction technology to extract the sound features of the preprocessed sound signal data, and input the sound features into the digital twin network for modeling to form a digital twin model based on smart mines. Among them, this step also needs to be explained:

[0066] Sound characteristics include spectrum characteristics, energy characteristics and time domain characteristics;

[0067] The spectrum features, energy features and time domain features are extracted and expressed by the following formulas respectively:

[0068]

[0069]

[0070]

[0071] Where (j) represents the frequency of the sound signal The spectrum characteristics under are obtained through Fourier transform. (T) represents the energy characteristics of the sound signal in the time period T, which is the square integral of the sound signal. ( ) indicates the sound signal at time point The time domain feature is the convolution integral of the sound signal and the unit impulse function.

[0072] The extracted sound features are then processed for dimensionality reduction to reduce the data dimension.

[0073] The above-mentioned dimensionality reduction process can be implemented by existing technologies, and this step will not be described in detail.

[0074] Furthermore, the sound features are input into the digital twin network for modeling to form a digital twin model based on the smart mine. The mathematical expression formula is as follows:

[0075]

[0076] in, represents the sound features extracted from X, X represents the preprocessed sound signal data, M represents the model of the digital twin network, and Y represents the output of the digital twin network, that is, the prediction result of the mine status.

[0077] In a preferred embodiment, in addition to the sound signal data, mine environment data such as temperature, humidity, and vibration are also considered. Consideration of these data can provide more comprehensive mine status information, which helps the digital twin model to more accurately understand the operating status of the mine, thereby improving prediction accuracy and reliability; for example, temperature and humidity data can be used to detect whether the equipment is in an overheated state, and vibration data can be used to determine whether the equipment is operating abnormally.

[0078] In a preferred embodiment, by applying historical sound signal data and its corresponding dangerous state records to model training and verification, the digital twin model can learn and identify the correlation between different sound characteristics and dangerous states. By analyzing historical data, the model can learn the dangerous state patterns under the specific environment of the mine, thereby improving the accuracy and reliability of the prediction.

[0079] As an example, historical data (such as historical sound signal data) can be obtained by querying a database in a mine management system. The query method can be implemented by means in the prior art, which will not be described in detail in this embodiment.

[0080] In a preferred embodiment, a labeled data set (such as the correspondence between sound features and actual dangerous states) is prepared for supervised learning, that is, the model learns to predict dangerous states by observing input features and corresponding labels, thereby improving the model's prediction ability.

[0081] Preferably, this step converts complex sound signals into a digital model by constructing a digital twin model that comprehensively considers a variety of environmental data and historical information, so that the mine environmental status can be characterized and analyzed in a more accurate and easier to handle manner, and the dangerous status of the mine can be accurately predicted, so that mine managers can promptly discover potential safety hazards and take corresponding measures to ensure safe production in the mine.

[0082] S3: Use convolutional neural network to train the digital twin model to learn the correlation between sound features and dangerous states. Among them, what needs to be explained in detail in this step is that the correlation between the learned sound features and dangerous states is expressed by the following mathematical expression formula:

[0083]

[0084] in, represents the sound characteristics, B represents the danger characteristics, Represents the extraction and learning of sound features by convolutional neural network. is the representation function of the dangerous state, is a Gaussian function used to measure The distance from B, is the standard deviation of the Gaussian function, which is used to control the range and strength of the correlation. Indicates the correlation between the sound characteristics and the dangerous state;

[0085] The value range of is (0,1). The closer the value is to 1, the higher the correlation between the sound feature and the dangerous state;

[0086] when The closer the value is to 0, the lower the correlation between the sound feature and the dangerous state.

[0087] Preferably, the Gaussian function is used in the above formula to weight the distance between the sound feature and the dangerous state, taking into account the similarity between each other, which can better reflect the correlation between the sound feature and the dangerous state, and the parameters of the Gaussian function can be adjusted according to actual conditions to make the model more suitable for different application scenarios.

[0088] Preferably, through the training process, the model can better understand the correlation between the sound features and the dangerous state, thereby improving the accuracy and reliability of the prediction.

[0089] S4: Use the trained digital twin model to predict the sound signal data to determine whether it is in a dangerous state. The following points need to be explained in this step:

[0090]

[0091]

[0092] Among them, D is a binary variable, indicating whether a dangerous state exists, 1 means it exists, and 0 means it does not exist. represents the set threshold, P represents the confidence of the prediction result, n represents the sample size, and c represents the probability of the existence of a dangerous state;

[0093] When the confidence level of the prediction result P> the set threshold When , D is 1, indicating a dangerous state;

[0094] When the confidence level of the prediction result P ≤ the set threshold When , D is 0, indicating that there is no dangerous state.

[0095] As a preferred implementation, the above formula can quickly and accurately determine whether a dangerous state exists by comparing the confidence of the prediction result with the set threshold, thereby providing an effective decision-making basis for safety management and risk control.

[0096] As a preferred implementation, the threshold in the above formula can be adjusted according to actual conditions and needs, so that risk judgment is flexible and customizable. Different application scenarios can adopt different thresholds to meet different security management needs. This embodiment does not make a unique limitation on it.

[0097] Preferably, this embodiment timely discovers dangerous conditions in the mine environment through the prediction process, provides an important basis for taking corresponding safety measures, and ensures the safety of mine workers.

[0098] Preferably, the embodiments of the present invention can be used to predict areas and equipment that may be affected, help prevent the occurrence and spread of dangerous events, ensure the safety of personnel and equipment, and effectively reduce the damage and impact of dangerous events on the system by taking preventive measures in a timely manner.

[0099] It should be noted that other aspects disclosed in the embodiments of the present invention also propose a smart mine dangerous state prediction device based on digital twins, including: one or more processors and a memory.

[0100] The memory is used to store operable instructions, which, when executed by the one or more processors, cause the one or more processors to perform operations, including the process of the method for predicting dangerous conditions of intelligent mines based on digital twins of the aforementioned embodiment, especially Figure 1 The process of the method shown.

[0101] The intelligent mine danger state prediction device disclosed in the embodiment of the present invention also includes a computer-readable medium storing software, wherein the software includes instructions that can be executed by one or more computers, and the instructions enable one or more computers to perform operations through such execution, and the operations include the process of the intelligent mine danger state prediction method based on digital twins of the aforementioned embodiment, especially Figure 1 The process of the method shown.

[0102] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for predicting dangerous conditions in intelligent mines based on digital twins, characterized in that: include: Collect and pre-process the sound signal data from the sound sensors placed in key areas of the mine; The preprocessing removes the interference of environmental noise on the sound signal through a filtering algorithm, and then normalizes it to ensure that the data is within the same dimension range. The filtering algorithm is used to remove the interference of environmental noise on the sound signal, which can be expressed by the following formula: Filtered Sound(t)=F(Sound(t),Noise(t)) Wherein, Filtered Sound(t) represents the value of the sound signal processed by the filtering algorithm at time t, Sound(t) represents the value of the original sound signal at time t, Noise(t) represents the value of the ambient noise at time t, and F(·) represents the function of the filtering algorithm, which accepts the original sound signal and the ambient noise signal as parameters and returns the filtered sound signal; The sound features of the preprocessed sound signal data are extracted using feature extraction technology, and the sound features are input into the digital twin network for modeling to form a digital twin model based on the smart mine; its mathematical expression formula is as follows: M(f(X))=Y Among them, f(X) represents the sound features extracted from X, X represents the preprocessed sound signal data, M represents the model of the digital twin network, and Y represents the output of the digital twin network, that is, the prediction result of the mine status; The sound features include spectrum features, energy features and time domain features; The spectrum feature, the energy feature and the time domain feature are extracted and expressed by the following formulas respectively: Among them, (j) represents the spectral characteristics of the sound signal at frequency J, which is obtained by Fourier transform, (T) represents the energy characteristics of the sound signal in time period T, which is the square integral of the sound signal, It represents the time domain characteristics of the sound signal at time point t0, which is the convolution integral of the sound signal and the unit impulse function; Then, the extracted sound features are processed for dimensionality reduction to reduce the data dimension; The digital twin model is trained using a convolutional neural network to learn the correlation between the sound features and the dangerous state; it is expressed by the following mathematical expression formula: Among them, f(X) represents the sound feature, B represents the danger feature, F(f(X)) represents the extraction and learning of the sound feature by the convolutional neural network, and g(B) is the representation function of the danger state. is a Gaussian function used to measure the distance between f(X) and B, σ is the standard deviation of the Gaussian function used to control the range and strength of the correlation, and R(f(X), B) represents the correlation between the sound feature and the dangerous state; The value range of R(f(X), B) is (0, 1). When the value of R(f(X), B) is closer to 1, it means that the correlation between the sound feature and the dangerous state is higher; When the value of R(f(X), B) is closer to 0, it means that the correlation between the sound feature and the dangerous state is lower; The digital twin model after training is used to predict the sound signal data to be predicted to determine whether it is in a dangerous state; including: Among them, D is a binary variable, indicating whether a dangerous state exists, 1 indicates existence, 0 indicates non-existence, τ indicates the set threshold, P indicates the confidence of the prediction result, n indicates the sample size, and c indicates the probability of the existence of a dangerous state; When the confidence of the prediction result is greater than the set threshold, the value is 1, indicating that there is a dangerous state; When the confidence of the prediction result is ≤ the set threshold, the value is 0, indicating that there is no dangerous state.

2. The method for predicting dangerous conditions of intelligent mines based on digital twins according to claim 1 is characterized in that: The sound signal data includes the sound of mechanical equipment operation and the sound of geological activities; The key areas include mining operation areas, equipment concentration areas, and rock stability monitoring areas.

3. A smart mine danger state prediction device based on digital twin, characterized in that: include: one or more processors; A memory storing operable instructions, wherein when the instructions are executed by the one or more processors, the one or more processors perform operations, wherein the operations include the process of the method for predicting dangerous conditions in a smart mine based on digital twins as described in any one of claims 1 to 2.

Images