Real-time monitoring system and method for coal mine electromechanical operation faults

Abstract

The invention discloses a coal mine electromechanical operation fault real-time monitoring system and method, and the system comprises a multi-sensor module which obtains the multi-source information of the operation state of coal mine electromechanical equipment based on a multi-sensor fusion technology, and a signal processing module carries out the noise filtering processing of the multi-source information, so as to reduce the signal interference in a complex environment. The fault diagnosis module establishes a fault diagnosis model based on the filtered data to realize real-time monitoring of faults; the coal mine electromechanical operation fault real-time monitoring system and method solve the problems that in the prior art, the anti-interference capacity is limited, the reliability of data collection is insufficient, the fault monitoring effect is directly affected, how to reduce the interference degree through reasonable sensor layout and network transmission mechanism design, and in the algorithm level, the real-time monitoring of the coal mine electromechanical operation fault is achieved. The problems that an existing analysis method is limited in noise suppression capacity and the actual application effect of a monitoring system is limited are solved.

Description

Technical Field

[0001] The present invention relates to the field of mine safety technology, and in particular to a real-time monitoring system and method for electromechanical operation failures in a coal mine. Background Art

[0002] In modern coal mines, real-time monitoring of electromechanical equipment is an important measure to ensure the safety of underground operations and improve the reliability of equipment operation. Existing real-time monitoring of electromechanical operation failures in coal mines usually uses sensor technology, data acquisition and processing systems, and advanced analysis algorithms to monitor the operating status of key electromechanical equipment in coal mines. These are aimed at timely detecting potential mechanical failures and performing preventive maintenance, thereby avoiding sudden failures of equipment and improving the continuity and safety of coal mine production.

[0003] Existing monitoring systems face many challenges in their application, especially in the complex electromagnetic environment underground in coal mines. These challenges mainly include: how to ensure that the data collected by sensors is accurate and stable under extreme conditions such as strong electromagnetic interference, large changes in temperature and humidity, and strong mechanical vibrations. The current anti-interference capability is limited, resulting in insufficient reliability of data acquisition, which directly affects the effect of fault monitoring. In addition, due to the wide variety and uneven distribution of underground signal sources, signals are prone to mutual interference. How to reduce the degree of interference through reasonable sensor layout and network transmission mechanism design is also an urgent problem to be solved. At the algorithm level, the existing analysis methods have limited noise suppression capabilities, and it is often difficult to effectively distinguish fault signals from background noise, which affects the accuracy of fault identification. These deficiencies limit the actual application effect of the monitoring system, especially in scenarios requiring high precision and high reliability. Summary of the invention

[0004] The purpose of the present invention is to provide a real-time monitoring system and method for electromechanical operation faults in coal mines, so as to solve the problems that the existing technology has limited anti-interference ability, resulting in insufficient reliability of data collection, directly affecting the effect of fault monitoring, and how to reduce the degree of interference through reasonable sensor layout and network transmission mechanism design; at the algorithm level, the existing analysis method has limited noise suppression ability and limits the actual application effect of the monitoring system.

[0005] To achieve the above object, the present invention provides the following technical solution: a real-time monitoring system for electromechanical operation failures in coal mines, the system comprising:

[0006] Multi-sensor module, used to obtain multi-source information on the operating status of coal mine electromechanical equipment based on multi-sensor fusion technology;

[0007] The signal processing module is connected to the multi-sensor module and is used to perform noise filtering on multi-source information to reduce signal interference in complex environments;

[0008] The fault diagnosis module is connected to the signal processing module and is used to establish a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults.

[0009] Preferably, the multi-sensor module acquires multi-source information on the operating status of coal mine electromechanical equipment based on multi-sensor fusion technology, including:

[0010] Deploy various types of sensors;

[0011] Collect data corresponding to each type of sensor;

[0012] Data standardization conversion;

[0013] Integrate the information from each sensor to form a comprehensive information set;

[0014] The process of data standardization conversion includes:

[0015] The collected raw data are transformed to the same standard interval according to the linear function;

[0016] Set the scale factor of the conversion formula to k and the translation parameter to b, where the transformed normalized data is equal to the original data multiplied by k plus b;

[0017] The condition for applying the formula is that the sensor data is within the valid measurement range;

[0018] Compare and analyze the differences based on standard interval values ​​to reduce the impact of mutual interference between different dimensions.

[0019] Preferably, the step of applying the formula to determine whether the standardized data exceeds a set standard deviation comprises:

[0020] Determine the mean μ and standard deviation σ of the standard data;

[0021] Design the deviation judgment threshold T. If the standardized data x is outside (μ-T×σ,μ+T×σ), it is marked as an abnormal signal;

[0022] Abnormal data records its location and time for easy tracking and investigation;

[0023] Identify the stability and reliability of data based on deviation analysis.

[0024] Preferably, the step of setting the conditional formula for deviation determination further includes:

[0025] Set the formula to determine the logic T value to determine the fault tolerance range;

[0026] If the data deviates significantly from the central value, that is, x>μ+3×σ or x<μ-3×σ, an early warning is triggered;

[0027] If the cumulative number of deviations exceeds the limit within the set time, it will be regarded as a precursor to equipment failure;

[0028] Based on this logical condition, the data of potential fault sources are automatically screened and diagnosed and processed in priority.

[0029] Preferably, the signal processing module performs noise filtering on multi-source information to reduce signal interference in a complex environment, including:

[0030] The frequency component features are extracted by using discrete Fourier transform for the acquired multi-source information;

[0031] Apply wavelet transform to remove large-scale random noise contained in the low-frequency part;

[0032] Use high-pass filters to filter out high-frequency noise and unwanted signal interference;

[0033] Implement bandpass filtering to ensure that the information of the useful signal frequency band is retained and clearly identifiable;

[0034] The processed data is saved and provided to the subsequent analysis module for fault detection.

[0035] Preferably, the signal processing module extracts frequency component features from the acquired multi-source information using discrete Fourier transform, including:

[0036] The frequency characteristics of multi-source information in the operation process of coal mine electromechanical equipment are extracted through discrete Fourier transform, and the time domain signal is converted into a frequency domain signal to facilitate the analysis of the frequency components.

[0037] Preferably, the signal processing module applies wavelet transform to remove large-scale random noise contained in the low-frequency part, including:

[0038] Wavelet transform is used to decompose the signal and filter out large-scale random noise.

[0039] Preferably, the fault diagnosis module establishes a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults, including:

[0040] Apply support vector machine learning model for pre-training based on historical fault data;

[0041] Perform feature extraction and dimension reduction operations on the filtered measured data to improve analysis efficiency;

[0042] Input the data into the trained support vector machine algorithm to determine the health status of the equipment;

[0043] Match online data streams with trained fault models to complete status prediction;

[0044] Model parameters are regularly adjusted to accommodate wear or changes in the mechatronic system to maintain diagnostic accuracy.

[0045] Preferably, the fault diagnosis module establishes a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults, further comprising:

[0046] The support vector machine model is trained based on historical fault data. The data is divided into different categories by finding an optimal hyperplane. Assuming that the input data is x, the corresponding label is y∈{-1,+1}, where -1 indicates a normal state and +1 indicates a fault state. By solving the following optimization problem, an optimal classification hyperplane is found;

[0047] In the data processing after filtering, the signal data needs to be feature extracted first. The features can be statistical features in the time domain and frequency domain;

[0048] The data after feature extraction and dimension reduction will be used as input to the pre-trained SVM model. The SVM will classify and judge based on the input feature vector and output the health status of the device.

[0049] Continuously monitor the operating status of the equipment and input new data into the model for status prediction through real-time data streams, ensuring that the health status of the equipment operation can be monitored and fed back in real time;

[0050] At regular intervals, the support vector machine is retrained based on the latest fault data and equipment operation data, and the grid search method is used to adjust the model's hyperparameters to optimize the classification performance. The penalty parameter C and kernel parameter γ of the SVM are selected through cross-validation.

[0051] A real-time monitoring method for coal mine electromechanical operation failures, using the real-time monitoring system for coal mine electromechanical operation failures, the method comprising:

[0052] S1: Obtain multi-source information on the operating status of coal mine electromechanical equipment through multi-sensor fusion technology;

[0053] S2: Perform signal processing on the acquired multi-source information, including filtering the noise to reduce signal interference in complex environments and ensure the validity and accuracy of the data;

[0054] S3: Based on the multi-source data after filtering, a fault diagnosis model is established to analyze the operating status of coal mine electromechanical equipment, identify and monitor possible equipment failures in real time.

[0055] It can be seen from the above technical solution that the present invention has the following beneficial effects:

[0056] The real-time monitoring system and method for electromechanical operation failures in coal mines obtains multi-source information on the operating status of electromechanical equipment in coal mines based on multi-sensor fusion technology through a multi-sensor module. The signal processing module performs noise filtering on the multi-source information to reduce signal interference in complex environments. The fault diagnosis module establishes a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults, which significantly improves data stability and reliability, improves data acquisition accuracy, and improves fault identification accuracy, ensuring that the system performs well in scenarios with high reliability and high precision requirements, improves production efficiency and safety, and solves the problems that the existing technology has limited anti-interference ability, resulting in insufficient reliability of data acquisition, which directly affects the effect of fault monitoring, how to reduce the degree of interference through reasonable sensor layout and network transmission mechanism design, and at the algorithm level, the existing analysis method has limited noise suppression ability and limits the actual application effect of the monitoring system. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a schematic diagram of the connection of the system modules of the present invention;

[0058] Figure 2 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] like Figure 1 As shown, a real-time monitoring system for electromechanical operation failures in a coal mine is characterized in that the system includes a multi-sensor module, a signal processing module and a fault diagnosis module. The signal processing module is connected to the multi-sensor module, and the fault diagnosis module is connected to the signal processing module. The multi-sensor module is used to obtain multi-source information on the operating status of electromechanical equipment in a coal mine based on multi-sensor fusion technology. The signal processing module is used to perform noise filtering on the multi-source information to reduce signal interference in a complex environment. The fault diagnosis module is used to establish a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults.

[0061] The system obtains the operating data of coal mine electromechanical equipment by arranging multiple sensors. The data captured by various sensors include various status information such as temperature, vibration, and pressure of the equipment. The signal processing module preprocesses these multi-source data and reduces the interference of noise on the signal through filtering technology, thereby ensuring the reliability of the data. The processed data is input into the fault diagnosis module. Based on the historical fault data and operating status data, the system establishes a fault diagnosis model through machine learning methods to monitor and judge the health status of the equipment in real time. The multi-source information of the equipment is obtained through the multi-sensor module. Combined with the noise filtering function of the signal processing module, it can effectively reduce the interference in complex environments and improve the accuracy of monitoring data. Based on the analysis of real-time data and historical data, the fault diagnosis module can timely warn of equipment failures, ensure the normal operation of coal mine electromechanical equipment, and reduce the downtime and maintenance time caused by equipment failures.

[0062] In alternative embodiments, the multi-sensor module may use different types of sensors, such as ultrasonic sensors or infrared sensors, to monitor different equipment operating parameters. The signal processing module may use other filtering techniques, such as Kalman filtering, to enhance the robustness of data processing. In addition, the fault diagnosis module may integrate other machine learning algorithms, such as random forests or neural networks, to improve the accuracy and efficiency of fault identification.

[0063] The multi-sensor module obtains multi-source information on the operating status of coal mine electromechanical equipment based on multi-sensor fusion technology, including:

[0064] Deploy various types of sensors;

[0065] Collect data corresponding to each type of sensor;

[0066] Data standardization conversion;

[0067] Integrate the information from each sensor to form a comprehensive information set;

[0068] The process of data standardization conversion includes:

[0069] The collected raw data are transformed to the same standard interval according to the linear function;

[0070] Set the scale factor of the conversion formula to k and the translation parameter to b, where the transformed normalized data is equal to the original data multiplied by k plus b;

[0071] The condition for applying the formula is that the sensor data is within the valid measurement range;

[0072] Compare and analyze the differences based on standard interval values ​​to reduce the impact of mutual interference between different dimensions.

[0073] This implementation method collects different types of data by arranging various types of sensors, such as temperature sensors to collect temperature information, vibration sensors to collect vibration information, etc. The collected raw data is standardized to ensure the comparability between data of different dimensions. The data is converted to the same standard range through a linear function. After the data is standardized, the system integrates this information to form a comprehensive information set for subsequent processing and analysis. The process of data standardization conversion reduces the interference between different dimensions and improves the accuracy of multi-sensor information fusion. This fusion method can better reflect the operating status of the equipment, improve the real-time performance of the system and the reliability of fault diagnosis.

[0074] Data normalization can be achieved through other mathematical transformation methods, such as logarithmic transformation or Z-score normalization. These methods can also reduce the difference in data dimensions and improve the accuracy of fused data. In addition, the combination and distribution of different sensors can also be adjusted according to actual equipment requirements.

[0075] The steps to apply the formula to determine whether the standardized data exceeds the set standard deviation include:

[0076] Determine the mean μ and standard deviation σ of the standard data;

[0077] Design the deviation judgment threshold T. If the standardized data x is outside (μ-T×σ,μ+T×σ), it is marked as an abnormal signal;

[0078] Abnormal data records its location and time for easy tracking and investigation;

[0079] Identify the stability and reliability of data based on deviation analysis.

[0080] By performing statistical analysis on the standardized data, the system first calculates the mean value μ and standard deviation σ of the data to ensure the overall distribution of the collected data. Then, the threshold T for deviation judgment is set. Once the standardized data x exceeds the range of (μ-T×σ,μ+T×σ), the system marks it as an abnormal signal. The data of the abnormal signal will be recorded, including its location and time, for subsequent investigation. This process helps to quickly identify unstable factors in the operating status of the equipment and improve the reliability analysis of the data. Through real-time monitoring of data deviations, abnormal signals can be identified before potential problems occur in the equipment, reducing the suddenness and unpredictability of equipment failures. In addition, recording the time and location of abnormal data facilitates subsequent tracking and analysis, effectively improving the efficiency of equipment maintenance.

[0081] The deviation judgment formula can use other statistical methods, such as determining the range of abnormal data through quantile analysis. The threshold T can also be dynamically adjusted according to the device type and the nature of the sensor data to make the judgment criteria more flexible. In addition, different data distribution models (such as normal distribution and Poisson distribution) can also be used for deviation analysis in different scenarios.

[0082] The step of setting the conditional formula for deviation determination further includes:

[0083] Set the formula to determine the logic T value to determine the fault tolerance range;

[0084] If the data deviates significantly from the central value, that is, x>μ+3×σ or x<μ-3×σ, an early warning is triggered;

[0085] If the cumulative number of deviations exceeds the limit within the set time, it will be regarded as a precursor to equipment failure;

[0086] Based on this logical condition, the data of potential fault sources are automatically screened and diagnosed and processed in priority.

[0087] Based on the set judgment logic, the system determines the fault tolerance range by adjusting the T value. For example, if the equipment data deviates from the mean value by more than 3 standard deviations (i.e. x>μ+3×σ or x<μ-3×σ), the system will issue an early warning. If the cumulative number of such abnormal conditions exceeds the limit within the set time period, the system will determine that the equipment may be about to fail and mark it as a precursor to failure. At this time, the system will automatically filter out these potential fault source data and give priority to analysis and processing. Through this automated screening and early warning mechanism, the accuracy and timeliness of fault prediction can be effectively improved. The system triggers an early warning when the deviation accumulates beyond the limit, providing an early diagnostic tool that can take preventive measures before equipment failure actually occurs, reducing downtime and maintenance costs.

[0088] Different tolerance ranges can be used to set different T values ​​for different equipment types and usage environments. For example, in some equipment, stricter deviation judgments (such as T = 2 × σ) may be required, while in other cases, looser judgment criteria may be more appropriate. In addition, the early warning logic can be adjusted based on different statistical analysis methods, such as using moving average or exponential smoothing to further improve the flexibility of judgment.

[0089] The signal processing module performs noise filtering on multi-source information to reduce signal interference in complex environments, including:

[0090] The frequency component features are extracted by using discrete Fourier transform for the acquired multi-source information;

[0091] Apply wavelet transform to remove large-scale random noise contained in the low-frequency part;

[0092] Use high-pass filters to filter out high-frequency noise and unwanted signal interference;

[0093] Implement bandpass filtering to ensure that the information of the useful signal frequency band is retained and clearly identifiable;

[0094] The processed data is saved and provided to the subsequent analysis module for fault detection.

[0095] The signal processing module first converts the collected time domain signal into a frequency domain signal through discrete Fourier transform, and extracts the main frequency components from it. Then, the wavelet transform is applied to decompose the signal to remove the random noise in the low-frequency part. In order to further improve the accuracy of the signal, the system uses a high-pass filter to remove high-frequency noise and unnecessary interference signals, and finally retains the useful signal information in the target frequency band through bandpass filtering to ensure the clarity and integrity of the signal data. The processed signal is stored and used for subsequent fault detection and analysis. This signal processing method can effectively reduce the noise interference of multi-source information in complex environments and improve the accuracy and reliability of data. Through the application of multiple filtering technology, useful signals can be extracted, unnecessary noise effects can be reduced, and high-quality data support can be provided for subsequent fault diagnosis modules to ensure the accuracy and stability of real-time monitoring.

[0096] In alternative implementations, the signal processing module may use other filtering algorithms, such as adaptive filtering or Kalman filtering, to enhance noise suppression capabilities in dynamic environments. Different filter combinations can also be adjusted according to the operating characteristics of the device. For example, in some scenarios, a band-stop filter may be more effective than a high-pass filter. In addition, the frequency band range of the filter can be adjusted according to the operating conditions of the device to optimize the filtering effect.

[0097] The signal processing module uses discrete Fourier transform to extract frequency component features from the acquired multi-source information, including:

[0098] The frequency characteristics of multi-source information in the operation process of coal mine electromechanical equipment are extracted by discrete Fourier transform, and the time domain signal is converted into the frequency domain signal to facilitate the analysis of the frequency components. The formula is:

[0099]

[0100] Among them, X(k) is the kth frequency component in the frequency domain, x(n) is the input time domain signal data, N is the total number of sampling points of the signal, n is the data point index in the time domain, j represents the imaginary unit, and k is the index of the frequency component.

[0101] The signal processing module converts the time domain signals in the multi-source information into frequency domain signals through discrete Fourier transform, thereby extracting the key frequency components. In this process, the Fourier transform formula is used to perform frequency analysis on the signal and decompose the complex time domain signal into different frequency components. By analyzing the frequency components of the frequency domain signal, the system can more effectively identify possible fault characteristics during the operation of the equipment. Through discrete Fourier transform, the system can extract frequency characteristic information from complex time domain signals, making the vibration, noise and other problems of the equipment more clearly visible in the frequency domain. This process helps to detect frequency anomalies, which in turn assists the fault diagnosis module in identifying potential equipment problems.

[0102] The discrete Fourier transform can be replaced by the fast Fourier transform to improve computational efficiency, which is particularly suitable for systems with high real-time processing requirements. For nonlinear or non-stationary signals, the short-time Fourier transform or the Hilbert-Huang transform can be used to further enhance the signal analysis capability.

[0103] The signal processing module uses wavelet transform to remove large-scale random noise contained in the low-frequency part, including:

[0104] Wavelet transform is used to decompose the signal and filter out large-scale random noise. The specific formula is:

[0105]

[0106] Among them, W is the coefficient of wavelet transform, x(t) is the original signal, ψ is the mother wavelet function, t is the time, s is the scale factor, and τ is the translation parameter.

[0107] This implementation method decomposes and processes the signal through wavelet transform. Wavelet transform is a tool that can process time domain and frequency domain information at the same time. It can decompose and analyze different frequency components in the signal. Specifically, wavelet transform can decompose the signal into multiple scales (frequencies) and time positions, and use different scales to extract useful signal features while removing large-scale random noise in the low-frequency part. Through this process, the system can improve the clarity of the signal and reduce data distortion caused by noise. Compared with the traditional Fourier transform, wavelet transform can provide better time-frequency localization capabilities, which is particularly effective for processing complex coal mine electromechanical equipment signals. By removing low-frequency noise and retaining key high-frequency components, wavelet transform greatly improves the accuracy of signal processing, which helps the subsequent fault diagnosis module to make more accurate judgments.

[0108] Different types of wavelet functions can be used, such as Haar wavelet, Daubechies wavelet or Symlet wavelet, and the specific choice depends on the characteristics of the signal and the processing requirements. For some application scenarios, it may be necessary to optimize the effect of the wavelet transform by adjusting the scale factor s and the translation parameter τ. In addition, other denoising algorithms, such as empirical mode decomposition, can be combined to enhance the processing capabilities of complex noise environments.

[0109] The fault diagnosis module establishes a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults, including:

[0110] Apply support vector machine learning model for pre-training based on historical fault data;

[0111] Perform feature extraction and dimension reduction operations on the filtered measured data to improve analysis efficiency;

[0112] Input the data into the trained support vector machine algorithm to determine the health status of the equipment;

[0113] Match online data streams with trained fault models to complete status prediction;

[0114] Model parameters are regularly adjusted to accommodate wear or changes in the mechatronic system to maintain diagnostic accuracy.

[0115] The fault diagnosis module first pre-trains the historical fault data through the support vector machine model. During the training process, the SVM learns how to distinguish between normal and faulty equipment based on the input data features. The filtered measured data is extracted for key features and processed through dimensionality reduction techniques (such as principal component analysis) to reduce the computational complexity. The processed data is then input into the SVM model, and the health status of the equipment is judged in real time through the classification ability of the model. The system continuously monitors the online data and matches it with the pre-trained model to ultimately predict the equipment status. In order to adapt to changes in equipment wear or operating status, the system also regularly adjusts the parameters of the SVM model to ensure the accuracy of fault diagnosis. By using the SVM model, the fault diagnosis module can efficiently and accurately judge the operating status of the equipment and predict the health status based on the real-time data stream. The SVM model can process multi-dimensional data and has strong classification capabilities, making it suitable for complex coal mine electromechanical equipment monitoring scenarios. At the same time, the model parameters are regularly adjusted to ensure that the system can adapt to long-term wear and changes in the equipment, improving the reliability and sustainability of the diagnostic results.

[0116] Other machine learning algorithms can be used instead of SVM, such as random forest, neural network or gradient boosting tree, to improve the prediction accuracy of the model or adapt to different data characteristics. In addition, data dimensionality reduction technology can also adopt other methods such as linear discriminant analysis or factor analysis according to the actual data characteristics to optimize processing efficiency.

[0117] The fault diagnosis module establishes a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults, further including:

[0118] The support vector machine model is trained based on historical fault data. The data is divided into different categories by finding an optimal hyperplane. Assuming that the input data is x, the corresponding label is y∈{-1,+1}, where -1 indicates a normal state and +1 indicates a fault state. An optimal classification hyperplane is found by solving the following optimization problem:

[0119]

[0120] Where w is the normal vector of the decision hyperplane, T is the vector transpose sign, b is the bias of the hyperplane, and x i is the i-th input sample, y i is the label of the i-th sample (+1 or -1), i is the sample sequence number. Through pre-training, the model distinguishes normal state from fault state based on historical data. After training, the SVM model can be applied to real-time monitoring;

[0121] In the data processing after filtering, the signal data needs to be feature extracted first. The features can be statistical features in the time domain and frequency domain;

[0122] The data after feature extraction and dimensionality reduction will be used as input and passed into the pre-trained SVM model. The SVM will classify and judge based on the input feature vector and output the health status of the device. The classification discriminant function of the SVM is:

[0123] f(x)=sign(w T x+b),

[0124] Where x is the input feature vector, w is the trained model weight vector, T is the sign of the vector transpose, b is the bias of the model, and the sign function is used to output +1 (indicating failure) or -1 (indicating normal);

[0125] Continuously monitor the operating status of the equipment and input new data into the model for status prediction through real-time data streams, ensuring that the health status of the equipment operation can be monitored and fed back in real time;

[0126] At regular intervals, the support vector machine is retrained based on the latest fault data and equipment operation data, and the grid search method is used to adjust the model's hyperparameters to optimize the classification performance. The penalty parameter C and kernel parameter γ of the SVM are selected through cross-validation.

[0127] The working principle of this fault diagnosis module is based on the support vector machine classification algorithm. The model uses historical normal and fault data through pre-training to find the best hyperplane that can maximize the interval between categories and classify the newly input real-time data into normal or faulty states. After the signal is filtered, the system extracts important features (such as time domain and frequency domain features) and reduces their dimensionality to improve the computational efficiency of the model. These features are then input into the SVM model, and the model determines the equipment status based on the pre-learned classification rules. In order to maintain the accuracy of the diagnostic model, the system is retrained regularly and the model parameters are optimized through the grid search method. The SVM model has efficient classification capabilities when processing complex data and can accurately distinguish between normal and faulty states of equipment. By retraining the model regularly, the system can adapt to wear and changes during the operation of the equipment to ensure the accuracy and reliability of the diagnosis. The use of the SVM algorithm reduces the false alarm rate, and the ability to update and adjust the model in real time ensures a rapid response to equipment failures.

[0128] Other classification algorithms, such as decision trees, random forests, or deep learning models (such as convolutional neural networks) can be used to replace SVM to further improve classification accuracy or adapt to different types of fault data. In addition, feature extraction methods can also be optimized according to data characteristics, such as using Fourier transform to extract frequency domain features or using wavelet transform to extract time-frequency domain features.

[0129] A real-time monitoring method for coal mine electromechanical operation failure is also provided, using the real-time monitoring system for coal mine electromechanical operation failure, the method comprising:

[0130] S1: Obtain multi-source information on the operating status of coal mine electromechanical equipment through multi-sensor fusion technology;

[0131] S2: Perform signal processing on the acquired multi-source information, including filtering the noise to reduce signal interference in complex environments and ensure the validity and accuracy of the data;

[0132] S3: Based on the multi-source data after filtering, a fault diagnosis model is established to analyze the operating status of coal mine electromechanical equipment, identify and monitor possible equipment failures in real time.

[0133] This method obtains multi-source information of coal mine electromechanical equipment through multi-sensor fusion technology. In the signal processing stage, the system performs noise filtering on the data to ensure the accuracy and effectiveness of the data. The processed data is input into the fault diagnosis model. The system analyzes the operating status of the equipment according to the characteristics of these data and issues real-time warnings when abnormalities are found. Through the fault diagnosis model, the system can identify and monitor potential equipment failures, thereby achieving preventive maintenance. This method can effectively improve the fault monitoring efficiency of coal mine electromechanical equipment, reduce the incidence of sudden failures, and improve the safety and stability of equipment operation. Through the application of multi-sensor fusion technology and fault diagnosis models, this method can provide accurate and real-time fault warnings in complex environments, reducing maintenance costs and downtime.

[0134] In alternative implementations, the type or distribution of sensors can be adjusted according to the requirements of specific equipment to optimize data collection. The filtering method in the signal processing link can also be optimized according to the characteristics of environmental interference, such as using a more complex adaptive filtering algorithm. In addition, the fault diagnosis model can use different machine learning or deep learning models according to the different types of coal mine equipment.

[0135] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A real-time monitoring system for coal mine electromechanical operation failures, characterized in that: The system comprises: Multi-sensor module, used to obtain multi-source information on the operating status of coal mine electromechanical equipment based on multi-sensor fusion technology; The signal processing module is connected to the multi-sensor module and is used to perform noise filtering on multi-source information to reduce signal interference in complex environments; The fault diagnosis module is connected to the signal processing module and is used to establish a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults.

2. A real-time monitoring system for coal mine electromechanical operation failure according to claim 1, characterized in that: The multi-sensor module obtains multi-source information on the operating status of coal mine electromechanical equipment based on multi-sensor fusion technology, including: Deploy various types of sensors; Collect data corresponding to each type of sensor; Data standardization conversion; Integrate the information from each sensor to form a comprehensive information set; The process of data standardization conversion includes: The collected raw data are transformed to the same standard interval according to the linear function; Set the scale factor of the conversion formula to k and the translation parameter to b, where the transformed normalized data is equal to the original data multiplied by k plus b; The condition for applying the formula is that the sensor data is within the valid measurement range; Compare and analyze the differences based on standard interval values ​​to reduce the impact of mutual interference between different dimensions.

3. A real-time monitoring system for coal mine electromechanical operation failure according to claim 2, characterized in that: The step of applying the formula to determine whether the standardized data exceeds the set standard deviation includes: Determine the mean μ and standard deviation σ of the standard data; Design the deviation judgment threshold T. If the standardized data x is outside (μ-T×σ,μ+T×σ), it is marked as an abnormal signal; Abnormal data records its location and time for easy tracking and investigation; Identify the stability and reliability of data based on deviation analysis.

4. A real-time monitoring system for coal mine electromechanical operation failure according to claim 3, characterized in that: The step of setting the conditional formula for deviation determination further includes: Set the formula to determine the logic T value to determine the fault tolerance range; If the data deviates significantly from the central value, that is, x>μ+3×σ or x<μ-3×σ, an early warning is triggered; If the cumulative number of deviations exceeds the limit within the set time, it will be regarded as a precursor to equipment failure; Based on this logical condition, the data of potential fault sources are automatically screened and diagnosed and processed in priority.

5. A real-time monitoring system for coal mine electromechanical operation failure according to claim 1, characterized in that: The signal processing module performs noise filtering on multi-source information to reduce signal interference in complex environments, including: The frequency component features are extracted by using discrete Fourier transform for the acquired multi-source information; Apply wavelet transform to remove large-scale random noise contained in the low-frequency part; Use high-pass filters to filter out high-frequency noise and unwanted signal interference; Implement bandpass filtering to ensure that the information of the useful signal frequency band is retained and clearly identifiable; The processed data is saved and provided to the subsequent analysis module for fault detection.

6. A real-time monitoring system for coal mine electromechanical operation failure according to claim 5, characterized in that: The signal processing module extracts frequency component features from the acquired multi-source information using discrete Fourier transform, including: The frequency characteristics of multi-source information in the operation process of coal mine electromechanical equipment are extracted through discrete Fourier transform, and the time domain signal is converted into a frequency domain signal to facilitate the analysis of the frequency components.

7. A real-time monitoring system for coal mine electromechanical operation failure according to claim 5, characterized in that: The signal processing module uses wavelet transform to remove large-scale random noise contained in the low-frequency part, including: Wavelet transform is used to decompose the signal and filter out large-scale random noise.

8. A real-time monitoring system for coal mine electromechanical operation failure according to claim 5, characterized in that: The fault diagnosis module establishes a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults, including: Apply support vector machine learning model for pre-training based on historical fault data; Perform feature extraction and dimension reduction operations on the filtered measured data to improve analysis efficiency; Input the data into the trained support vector machine algorithm to determine the health status of the equipment; Match online data streams with trained fault models to complete status prediction; Model parameters are regularly adjusted to accommodate wear or changes in the mechatronic system to maintain diagnostic accuracy.

9. A real-time monitoring system for coal mine electromechanical operation failure according to claim 8, characterized in that: The fault diagnosis module establishes a fault diagnosis model based on the filtered data to achieve real-time monitoring of faults, further comprising: The support vector machine model is trained based on historical fault data. The data is divided into different categories by finding an optimal hyperplane. Assuming that the input data is x, the corresponding label is y∈{-1,+1}, where -1 indicates a normal state and +1 indicates a fault state. By solving the following optimization problem, an optimal classification hyperplane is found; In the data processing after filtering, the signal data needs to be feature extracted first. The features can be statistical features in the time domain and frequency domain; The data after feature extraction and dimension reduction will be used as input to the pre-trained SVM model. The SVM will classify and judge based on the input feature vector and output the health status of the device. Continuously monitor the operating status of the equipment and input new data into the model for status prediction through real-time data streams, ensuring that the health status of the equipment operation can be monitored and fed back in real time; At regular intervals, the support vector machine is retrained based on the latest fault data and equipment operation data, and the grid search method is used to adjust the model's hyperparameters to optimize the classification performance. The penalty parameter C and kernel parameter γ of the SVM are selected through cross-validation.

10. A real-time monitoring method for coal mine electromechanical operation failures, characterized in that: The real-time monitoring system for coal mine electromechanical operation failure according to any one of claims 1 to 9 is adopted, and the method comprises: S1: Obtain multi-source information on the operating status of coal mine electromechanical equipment through multi-sensor fusion technology; S2: Perform signal processing on the acquired multi-source information, including filtering the noise to reduce signal interference in complex environments and ensure the validity and accuracy of the data; S3: Based on the multi-source data after filtering, a fault diagnosis model is established to analyze the operating status of coal mine electromechanical equipment, identify and monitor possible equipment failures in real time.

Images