Method and system for monitoring safe operation of dust remover

Abstract

The invention relates to the technical field of dust removers, and discloses a dust remover safe operation monitoring method and system, and the method comprises the steps: collecting dust concentration data, gas pressure data, equipment operation temperature data and equipment operation vibration data of a dust remover in real time through a sensor, and obtaining a sensor data set; carrying out denoising processing on the sensor data set by adopting a Kalman filtering algorithm, and carrying out normalization processing on data of different sensors to obtain preprocessed data; performing dimension reduction processing on the preprocessed data by adopting a principal component analysis method to obtain feature data, taking the feature data as input of an LSTM-CNN hybrid model, and outputting equipment operation state classification and fault type identification through the LSTM-CNN hybrid model; performing graded early warning according to an output result of the model, and recommending a filter bag replacement period and ash removal strategy adjustment based on Weibull distribution; according to the invention, safe and stable operation of the dust remover is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of dust collectors, and in particular to a method and system for monitoring the safe operation of a dust collector. Background Art

[0002] In the industrial production process, dust collectors, as key environmental protection equipment, are widely used in various dust-generating scenarios, such as the electricity, cement, steel and other industries. Their function is to separate the dust generated in the production process from the gas to achieve the purpose of purifying the gas, protecting the environment and ensuring the normal operation of production equipment. However, during the long-term operation of the dust collector, due to the influence of various complex factors, its safe operation faces many challenges. In the existing technology, the dust collector monitoring method mostly relies on a single sensor, which cannot fully reflect the operating status of the equipment. It relies on preset thresholds and has difficulty in coping with dynamic changes under complex working conditions, which can easily lead to missed faults. Summary of the Invention

[0003] The purpose of the present invention is to solve the above problems and to design a dust collector safe operation monitoring method and system.

[0004] A first aspect of the present invention provides a method for monitoring the safe operation of a dust collector, the method comprising the following steps: Use sensors to collect dust concentration data, gas pressure data, equipment operating temperature data, and equipment operating vibration data of the dust collector in real time to obtain a sensor data set; The Kalman filter algorithm is used to denoise the sensor data set, and the data of different sensors are normalized to obtain preprocessed data; The principal component analysis method is used to reduce the dimension of the preprocessed data to obtain feature data. The feature data is used as the input of the LSTM-CNN hybrid model, and the LSTM-CNN hybrid model outputs the equipment operation status classification and fault type identification; A graded warning is issued based on the output of the model, and filter bag replacement cycles and cleaning strategy adjustments are recommended based on the Weibull distribution.

[0005] Optionally, in a first implementation of the first aspect of the present invention, the denoising process is performed on the sensor data set using a Kalman filter algorithm, and normalization is performed on data from different sensors to obtain preprocessed data, including: Build a corresponding dynamic model for each type of sensor, and predict the state of the sensor data at the current moment based on the dynamic model and the state of the sensor data at the previous moment; Compare the actual measured data with the predicted results, analyze the differences between the two, and revise the predicted values; For different types of sensor data after Kalman filter denoising, normalization operation is performed according to the measurement range to obtain preprocessed data.

[0006] Optionally, in a second implementation of the first aspect of the present invention, the step of performing dimensionality reduction processing on the preprocessed data using principal component analysis to obtain feature data includes: The preprocessed data are organized into a matrix, the covariance matrix is ​​calculated, the eigenvalues ​​and eigenvectors of the covariance matrix are solved, and the data are sorted in descending order of eigenvalues. The first m principal components are selected according to the cumulative contribution rate, the principal component scores are calculated, and the preprocessed data are reduced to m dimensions based on the principal component scores to obtain feature data.

[0007] Optionally, in a third implementation of the first aspect of the present invention, the CNN network in the LSTM-CNN hybrid model consists of 2 convolutional layers and 1 pooling layer, the convolution layer adopts a 3×3 convolution kernel, the activation function adopts a ReLU function, and the pooling layer adopts maximum pooling; the LSTM network contains 2 layers of LSTM units, with 128 neurons in each layer.

[0008] Optionally, in a fourth implementation of the first aspect of the present invention, using the feature data as input to the LSTM-CNN hybrid model and outputting device operating status classification and fault type identification through the LSTM-CNN hybrid model includes: The feature data is arranged in chronological order to form time series data, which is then fed into the CNN network. The convolution layer performs sliding convolution on the time series data to extract local feature patterns in the data, and the convolution output is downsampled through the pooling layer. The local feature sequence extracted by the CNN network is input into the LSTM network. The LSTM unit captures the long-term and short-term dependencies in the local feature sequence through its internal gating mechanism. The temporal features output by the LSTM network are fused with the local features extracted by the CNN to form a comprehensive feature vector containing spatiotemporal feature information; First, the comprehensive feature vector is mapped to a high-dimensional feature space through a fully connected layer, and then the output of the fully connected layer is converted into the probability distribution of each operating status category and fault type through the Softmax activation function. The final equipment operating status classification and fault type identification results are determined based on the maximum probability.

[0009] Optionally, in a fifth implementation of the first aspect of the present invention, performing graded warnings based on the output results of the model and recommending filter bag replacement cycles and cleaning strategy adjustments based on Weibull distribution include: Combined with the dust collector's design parameters, historical operating data, and safety standards, a multi-level warning threshold is established for each fault type and operating state. Based on the equipment operating status classification and fault type identification results output by the LSTM-CNN hybrid model, a weighted calculation is performed to obtain the comprehensive risk value of the equipment's current operating status. The calculated comprehensive risk value is compared with the multi-level warning threshold to determine the warning level corresponding to the current equipment operation status and automatically trigger the corresponding warning measures.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, performing graded warnings based on the output results of the model and recommending filter bag replacement cycles and cleaning strategy adjustments based on Weibull distribution further includes: Obtain historical operating data of the filter bags. Based on the filter bag life data, use the maximum likelihood estimation method to estimate the shape parameters and scale parameters of the Weibull model distribution to determine the distribution characteristics of the filter bag life. Substitute the current filter bag operating time and status parameters into the Weibull distribution model to calculate the failure probability of the filter bag at different time points in the future. Based on the preset failure probability threshold, the remaining life of the filter bag is predicted and the optimal filter bag replacement cycle is generated. When the pressure difference of the dust collector continues to rise and the dust concentration is not effectively reduced, adjust the cleaning strategy.

[0011] A second aspect of the present invention provides a dust collector safe operation monitoring system, the system comprising: The acquisition module is used to use sensors to collect dust concentration data, gas pressure data, equipment operating temperature data, and equipment operating vibration data of the dust collector in real time to obtain a sensor data set; The preprocessing module is used to perform denoising on the sensor data set using the Kalman filter algorithm, and perform normalization on the data of different sensors to obtain preprocessed data; The recognition module is used to perform dimensionality reduction on the preprocessed data using principal component analysis to obtain feature data. The feature data is used as the input of the LSTM-CNN hybrid model, and the LSTM-CNN hybrid model outputs the equipment operation status classification and fault type identification; The early warning module is used to provide graded early warnings based on the output results of the model, and recommend filter bag replacement cycles and cleaning strategy adjustments based on Weibull distribution.

[0012] The third aspect of the present invention provides a dust collector safety operation monitoring device, which includes a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory to enable the dust collector safety operation monitoring device to perform each step of the dust collector safety operation monitoring method as described in any one of the above items.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, implement the various steps of the dust collector safe operation monitoring method as described in any one of the above items.

[0014] In the technical solution provided by the present invention, sensors are used to collect dust concentration data, gas pressure data, equipment operating temperature data and equipment operating vibration data of the dust collector in real time to obtain a sensor data set; a Kalman filter algorithm is used to denoise the sensor data set, and normalization is performed on the data of different sensors to obtain preprocessed data; a principal component analysis method is used to reduce the dimension of the preprocessed data to obtain feature data, and the feature data is used as the input of an LSTM-CNN hybrid model, and the LSTM-CNN hybrid model is used to output the equipment operating status classification and fault type identification; a graded warning is performed according to the output results of the model, and filter bag replacement is recommended based on the Weibull distribution. Adjustment of cycle and cleaning strategy; The present invention can collect data on dust concentration, pressure, temperature, vibration and other aspects in real time through multi-dimensional sensor arrangement, comprehensively reflect the operating status of the dust collector, overcome the problem of single monitoring means in the existing technology, conduct in-depth analysis and processing of the collected data, accurately judge the operating status of the dust collector, timely discover potential safety hazards, improve the accuracy and reliability of monitoring, realize intelligent early warning of safety hazards, and can promptly notify staff to take measures before a fault occurs, effectively reducing the probability of safety accidents, ensuring the safe and stable operation of the dust collector, and improving the continuity and stability of industrial production, thereby improving production efficiency and reducing production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Various other advantages and benefits will become apparent to those skilled in the art by reading the following detailed description of the preferred embodiment.The accompanying drawings are only for the purpose of illustrating the preferred embodiment and are not to be considered as limiting the present invention.

[0016] Figure 1 A flow chart of a dust collector safe operation monitoring method provided by an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a dust collector safety operation monitoring system provided by an embodiment of the present invention; Figure 3A schematic structural diagram of a dust collector safety operation monitoring device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The terms "first," "second," "third," "fourth," and so forth (if any) in the description and claims of the present invention and in the accompanying drawings are used to distinguish similar items and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that shown or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, apparatus, product, or device that includes a series of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, product, or device.

[0018] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 A flow chart of a dust collector safe operation monitoring method provided by an embodiment of the present invention, the method specifically comprises the following steps: Step 101: Use sensors to collect dust concentration data, gas pressure data, equipment operating temperature data, and equipment operating vibration data of the dust collector in real time to obtain a sensor data set; In this embodiment, different types of sensors are deployed at key locations of the dust collector, such as the air inlet, air outlet, filter bag area, ash hopper, and fan: the dust concentration sensor uses a laser scattering dust concentration sensor with a sampling frequency of 1 time per minute. Two sensors are deployed at the air inlet, one at the air outlet, and one at the entrance of each filter bag chamber. These sensors acquire dust concentration data in real time for evaluating dust removal efficiency and detecting dust leakage. Gas pressure sensor: A high-precision piezoresistive pressure sensor with a measuring range of 0-50kPa is used. One is installed at each of the inlet and outlet pipes of the dust collector, and one is installed at the pipe connecting the ash hopper and the filter bag chamber. Pressure data is collected every 10 seconds. The inlet and outlet pressure difference is used to determine the resistance of the dust collector and analyze whether there is a blockage problem. K-type thermocouple temperature sensors are installed at key locations such as the filter bag frame, fan motor housing, and air inlet and outlet pipes. The temperature measurement range is -50°C to 1200°C, and temperature data is collected every 30 seconds to prevent equipment damage or fire risks caused by abnormal temperatures. Piezoelectric vibration sensors are installed on the main structural components of the dust collector box, such as the side panels, brackets, and fan base. The measurement frequency range is 1Hz-10kHz, and vibration data is collected in real time at a sampling frequency of 1000Hz. Mechanical faults such as loose components and wear are identified by analyzing vibration characteristics.

[0019] Step 102: Use the Kalman filter algorithm to perform denoising on the sensor data set, perform normalization on the data of different sensors, and obtain pre-processed data; In this embodiment, a corresponding dynamic model is constructed for each type of sensor, and the state of the sensor data at the current moment is predicted based on the dynamic model and the state of the sensor data at the previous moment; the actual measured data is compared with the predicted results, the difference between the two is analyzed, and the predicted value is corrected; for the different types of sensor data after Kalman filtering and denoising, a normalization operation is performed according to the measurement range to obtain preprocessed data.

[0020] In this embodiment, for the dust concentration sensor, a dynamic model is constructed based on the fluid mechanics characteristics of the gas-solid two-phase flow inside the dust collector. The influence of factors such as the dust input amount at the air inlet, the filtration efficiency of the filter bag, and the air leakage rate of the equipment on the change of dust concentration is considered. The dust concentration at each location at the current moment is predicted by combining the dust concentration state of the air inlet, air outlet, and near the filter bag at the previous moment. For the pressure sensor, a dynamic model is established based on the Bernoulli equation and the continuity equation of gas in the pipeline, taking into account parameters such as gas flow rate, pipeline resistance, and the pressure difference between the inlet and outlet of the equipment. The gas pressure at the inlet and outlet of the dust collector and key structures at the current moment is predicted by combining the pressure data at the previous moment. The dynamic model of the temperature sensor is based on the principles of heat transfer, considering factors such as heat generation, heat dissipation, and heat conduction during the operation of the equipment. The temperature value at the current moment is predicted by combining the temperature state of the key heat-generating components and the air flow channel at the previous moment. The dynamic model of the vibration sensor is based on the principles of mechanical dynamics, considering factors such as the speed, load, and component connection state during equipment operation. The vibration state of the main structural components at the current moment is predicted by combining the vibration data at the previous moment. The actual measurement data collected in real time by various sensors are compared point by point with the results predicted by the dynamic model. Taking dust concentration as an example, if there is a deviation between the actual measured dust concentration at the air inlet and the predicted value, the cause of the deviation is analyzed. It may be that sudden changes in production conditions cause fluctuations in dust generation, or local damage to the filter bag causes dust leakage. Combined with the dust collector's operating history data and real-time operating information, the feedback mechanism of the Kalman filter algorithm is used to adjust the parameters in the prediction model, thereby correcting the predicted value to make the corrected estimate closer to the actual operating status. For pressure, temperature and vibration data, the difference between the actual measurement value and the predicted value is also analyzed. Taking into account potential factors such as pipeline blockage, abnormal heating of equipment, and loose components, the prediction model is optimized to obtain a more accurate current state estimate. For dust concentration data after Kalman filtering and denoising, if its measurement range is 0-1000mg / m³, a linear transformation is used to map the data to the [0,1] interval, so that measurement values ​​of different magnitudes are on a unified scale. For pressure sensor data, if the range is 0-50kPa, each pressure measurement value is converted to the standard interval according to the corresponding normalization formula. Temperature sensor data is normalized based on its measurement range of -50℃-1200℃ to eliminate the impact of the temperature data's magnitude. Vibration sensor data is normalized based on its measurement frequency range using a specific conversion formula.

[0021] Step 103: Use principal component analysis to perform dimensionality reduction on the preprocessed data to obtain feature data, use the feature data as input to the LSTM-CNN hybrid model, and output the equipment operation status classification and fault type identification through the LSTM-CNN hybrid model; In this embodiment, the preprocessed data is organized into a matrix, its covariance matrix is ​​calculated, the eigenvalues ​​and eigenvectors of the covariance matrix are solved, and the data are arranged in descending order of the eigenvalues; the first m principal components are selected according to the cumulative contribution rate, the principal component scores are calculated, and the preprocessed data is reduced to m dimensions based on the principal component scores to obtain characteristic data.

[0022] In this embodiment, the sensor data of different types, after being denoised and normalized by Kalman filtering, are arranged according to time series or sample numbers to form a data matrix of dimension n × p, where n represents the number of samples. For example, if one hour is the sampling period, there are 24 samples of monitoring data for one day. p represents the feature dimension, that is, the number of sensors, such as the total number of different types of sensors such as dust concentration, pressure, temperature, and vibration. Each row of the matrix represents all the eigenvalues ​​of a sample, and each column represents the measurement value of a certain type of sensor in all samples. The data matrix is ​​centered, that is, the mean of the column is subtracted from each element to obtain the centered data matrix. The covariance matrix of the centered data matrix is ​​calculated. The covariance matrix is ​​a p×p symmetric matrix whose elements represent the covariance between the i-th feature and the j-th feature, reflecting the correlation between the two features. The covariance matrix can be used to analyze the degree of correlation between different sensor data. Perform eigendecomposition on the covariance matrix and solve its characteristic equation. By solving this equation, p eigenvalues ​​and corresponding eigenvectors are obtained. Each eigenvalue represents the data variance in the direction of the corresponding eigenvector. The larger the eigenvalue, the more significant the data change in that direction and the more information it contains. Arrange the obtained eigenvalues ​​in descending order, and reorder the corresponding eigenvectors accordingly. Calculate the contribution rate and cumulative contribution rate of each eigenvalue. The contribution rate of a single eigenvalue is the ratio of the eigenvalue to the sum of all eigenvalues, and the cumulative contribution rate is the sum of the contribution rates of the first k eigenvalues. According to actual needs, set a cumulative contribution rate threshold, such as 95%, and select the minimum m values ​​that make the cumulative contribution rate reach or exceed the threshold, i.e., the first m principal components. These m principal components can retain most of the information in the original data while reducing the data dimension from p to m. Multiply the centered data matrix by the matrix composed of the first m selected eigenvectors to obtain the principal component score matrix. Each row in the principal component score matrix represents the score of a sample on the m principal components. These scores constitute the feature data after dimensionality reduction.

[0023] In this embodiment, the CNN network in the LSTM-CNN hybrid model consists of two convolutional layers and one pooling layer. The convolutional layer uses a 3×3 convolution kernel, the activation function uses the ReLU function, and the pooling layer uses maximum pooling. The LSTM network contains two layers of LSTM units, with 128 neurons in each layer.

[0024] In this embodiment, the feature data is used as the input of the LSTM-CNN hybrid model, and the LSTM-CNN hybrid model is used to output the equipment operation status classification and fault type identification, including: arranging the feature data in chronological order to form time series data, inputting the time series data into the CNN network, performing a sliding convolution operation on the time series data through the convolution layer to extract local feature patterns in the data, and downsampling the convolution output through the pooling layer; inputting the local feature sequence extracted by the CNN network into the LSTM network, and the LSTM unit captures the long-term and short-term dependencies in the local feature sequence through the internal gating mechanism; fusing the time series features output by the LSTM network with the local features extracted by the CNN to form a comprehensive feature vector containing spatiotemporal feature information; first mapping the comprehensive feature vector to a high-dimensional feature space through a fully connected layer, and then converting the output of the fully connected layer into a probability distribution of each operation status category and fault type through a Softmax activation function, and determining the final equipment operation status classification and fault type identification results according to the maximum probability.

[0025] Step 104: Perform graded warnings based on the output results of the model, and recommend filter bag replacement cycles and cleaning strategy adjustments based on Weibull distribution.

[0026] In this embodiment, a corresponding multi-level warning threshold is established for each fault type and operating state in combination with the design parameters, historical operating data and safety standards of the dust collector. Based on the equipment operating state classification and fault type identification results output by the LSTM-CNN hybrid model, a comprehensive risk value of the current operating state of the equipment is obtained through weighted calculation. The calculated comprehensive risk value is compared with the multi-level warning threshold to determine the warning level corresponding to the current equipment operating state, and the corresponding warning measures are automatically triggered.

[0027] In this embodiment, the design parameters, historical operating data and industry safety standards of the dust collector are collected, including design parameters such as rated processing air volume, design pressure loss, filter bag material and specifications, etc. The historical operating data includes sensor data during normal operation and various faults, and industry safety standards such as environmental emission limits and equipment safety operating procedures. For each fault type and operating state, the fault type is such as filter bag damage, ash hopper blockage, fan failure, and the operating state is such as normal, slightly abnormal, and seriously abnormal, and the corresponding key monitoring indicators are analyzed, such as dust concentration, inlet and outlet pressure difference, and temperature change rate. Statistical analysis methods are used to determine the distribution range of each indicator under different fault levels, and multi-level warning thresholds are set. For example, for a filter bag damage fault, the inlet and outlet dust concentration difference of 5-10mg / m³ is set as the first warning threshold, 10-20mg / m³ is set as the second warning threshold, and more than 20mg / m³ is set as the third warning threshold. Based on the output of the LSTM-CNN hybrid model, the equipment operating status classification probabilities are calculated, such as 0.7 for normal state, 0.2 for minor anomaly, and 0.1 for major anomaly. Fault type identification results include 0.6 for filter bag damage, 0.3 for ash hopper blockage, and 0.1 for fan failure. A weight coefficient is assigned to each status and fault type based on the severity, frequency, and impact of the fault on production. For example, the weight coefficient for filter bag damage is set to 0.4, for ash hopper blockage to 0.3, and for fan failure to 0.3. The probability values ​​of each status and fault type are multiplied by their corresponding weights and then added together to obtain a comprehensive risk value for the equipment's current operating status. For example, the comprehensive risk value = 0.1 × 0.4 (major anomaly × filter bag damage weight) + 0.3 × 0.3 (minor anomaly × ash hopper blockage weight) + ..., which quantifies the overall risk level of the equipment. The calculated comprehensive risk value is compared with the established multi-level warning thresholds. If the risk value is lower than the first-level warning threshold, the equipment is judged to be in normal operation, and the system only records routine data; if the risk value is between the first-level and second-level warning thresholds, it is judged to be a slight abnormality in the first-level warning, and the system automatically sends a text message to the operation and maintenance personnel's mobile phone, and highlights the abnormal parameters on the monitoring interface; if the risk value is between the second-level and third-level warning thresholds, it is judged to be a moderate abnormality in the second-level warning. In addition to the above measures, the sound and light alarm device in the workshop is triggered, and a preliminary fault diagnosis report is automatically generated; if the risk value exceeds the third-level warning threshold, it is judged to be a serious abnormality in the third-level warning. The system immediately cuts off some non-critical power supplies of the dust collector and starts the backup dust removal equipment. At the same time, it notifies the company's senior management and technical experts through multiple channels such as text messages, emails, and APP push, to realize an automated early warning mechanism with graded response.

[0028] In this embodiment, the historical operation data of the filter bag is obtained, and based on the filter bag life data, the maximum likelihood estimation method is used to estimate the shape parameters and scale parameters of the Weibull model distribution to determine the distribution characteristics of the filter bag life; the current filter bag operation time and state parameters are substituted into the Weibull distribution model, and the failure probability of the filter bag at different time points in the future is calculated. According to the preset failure probability threshold, the remaining life prediction result of the filter bag is determined, and the optimal filter bag replacement cycle recommendation is generated; when the pressure difference of the dust collector continues to rise and the dust concentration is not effectively reduced, the cleaning strategy is adjusted.

[0029] In this embodiment, historical operating data of the filter bags is extracted from the equipment management database, including parameters such as installation date, replacement date, pressure difference change during operation, cleaning frequency, and inlet dust concentration. A filter bag life data set containing multiple samples is constructed. Assuming that the filter bag life follows a Weibull distribution, the maximum likelihood estimation method is used to solve the shape parameter and scale parameter that maximize the likelihood function through an iterative optimization algorithm such as the Newton-Raphson method. After the estimation is completed, the physical meaning of the shape parameter is analyzed: when the shape parameter is greater than 1, the filter bag failure rate increases with time, indicating the presence of aging; when the shape parameter is equal to 1, the failure rate is constant, which is consistent with the characteristics of random failure; when the shape parameter is less than 1, the failure rate decreases with time, which is common in the early failure stage. The scale parameter reflects the average life characteristics of the filter bag. Substitute the current filter bag's operating time and real-time status parameters, such as current pressure difference and dust emission concentration, into the established Weibull distribution model to calculate the failure probability of the filter bag at different time points in the future. Set a failure probability threshold, such as 5% or 10%. When the failure probability exceeds the threshold, the corresponding time point is the remaining life prediction result. Combined with production plans, maintenance costs, and downtime losses, the optimal replacement cycle recommendation is generated: if the predicted remaining life is less than the next planned maintenance cycle, it is recommended to replace it in advance; if the remaining life is long but close to the equipment warranty period, it is recommended to replace it before the warranty expires. At the same time, considering the economic efficiency of filter bag replacement, the total maintenance cost under different replacement cycles is calculated, including the filter bag procurement cost, replacement labor cost, and downtime loss. The solution with the lowest total cost is selected as the final recommendation; Continuously monitor the inlet and outlet pressure difference and dust emission concentration of the dust collector, and establish a time series of the pressure difference change rate and the dust emission concentration change rate. When the pressure difference continues to rise in multiple consecutive monitoring cycles, that is, the pressure difference change rate is >0 and exceeds the preset rising rate threshold, and the dust emission concentration is not effectively reduced, that is, the dust emission concentration change rate is ≥0 or the decline is less than expected, it indicates that there is too much dust on the surface of the filter bag and the cleaning effect is poor. At this time, the system automatically triggers the cleaning strategy adjustment process: first, increase the cleaning frequency and shorten the time interval between two adjacent cleaning operations; if the pressure difference is still not effectively controlled, increase the cleaning intensity, such as increasing the pulse injection pressure or extending the injection time; if the above measures are ineffective, start the deep cleaning mode, including increasing the offline cleaning time, adjusting the cleaning sequence, etc. During the adjustment process, continuously monitor the pressure difference and dust concentration changes to form a closed-loop feedback mechanism to ensure dynamic optimization of the cleaning strategy.

[0030] See also Figure 2 , a schematic diagram of the structure of a dust collector safety operation monitoring system provided by an embodiment of the present invention, the system includes: The acquisition module is used to use sensors to collect dust concentration data, gas pressure data, equipment operating temperature data, and equipment operating vibration data of the dust collector in real time to obtain a sensor data set; The preprocessing module is used to perform denoising on the sensor data set using the Kalman filter algorithm, and perform normalization on the data of different sensors to obtain preprocessed data; The recognition module is used to perform dimensionality reduction on the preprocessed data using principal component analysis to obtain feature data. The feature data is used as the input of the LSTM-CNN hybrid model, and the LSTM-CNN hybrid model outputs the equipment operation status classification and fault type identification; The early warning module is used to provide graded early warnings based on the output results of the model, and recommend filter bag replacement cycles and cleaning strategy adjustments based on Weibull distribution.

[0031] Figure 3 is a schematic structural diagram of a dust collector safety operation monitoring device provided by an embodiment of the present invention. The dust collector safety operation monitoring device 300 may vary significantly due to different configurations or performance, and may include one or more processors (central processing units, CPUs) 310 (for example, one or more processors) and memory 320, and one or more storage media 330 (for example, one or more mass storage devices) storing application programs 333 or data 332. The memory 320 and storage medium 330 may be either short-term storage or persistent storage. The program stored in the storage medium 330 may include one or more modules (not shown), each of which may include a series of instruction operations in the dust collector safety operation monitoring device 300. Furthermore, the processor 310 may be configured to communicate with the storage medium 330, and execute the series of instruction operations in the storage medium 330 on the dust collector safety operation monitoring device 300 to implement the method provided in the above embodiment.

[0032] The dust collector safety operation monitoring device 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating devices 331, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 3 The structure of the dust collector safe operation monitoring device shown does not constitute a limitation on the computer device provided by the present invention, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0033] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions. When the instructions are run on a computer, the computer executes the various steps of the dust collector safety operation monitoring method provided in the above embodiments.

[0034] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices, and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0035] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0036] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A dust collector safe operation monitoring method, characterized in that: The method comprises the following steps: Use sensors to collect dust concentration data, gas pressure data, equipment operating temperature data, and equipment operating vibration data of the dust collector in real time to obtain a sensor data set; The Kalman filter algorithm is used to denoise the sensor data set, and the data of different sensors are normalized to obtain preprocessed data; The principal component analysis method is used to reduce the dimension of the preprocessed data to obtain feature data. The feature data is used as the input of the LSTM-CNN hybrid model, and the LSTM-CNN hybrid model outputs the equipment operation status classification and fault type identification; A graded warning is issued based on the output of the model, and filter bag replacement cycles and cleaning strategy adjustments are recommended based on the Weibull distribution.

2. A dust collector safe operation monitoring method according to claim 1, characterized in that: The Kalman filter algorithm is used to perform denoising on the sensor data set, and normalization is performed on the data of different sensors to obtain preprocessed data, including: Build a corresponding dynamic model for each type of sensor, and predict the state of the sensor data at the current moment based on the dynamic model and the state of the sensor data at the previous moment; Compare the actual measured data with the predicted results, analyze the differences between the two, and revise the predicted values; For different types of sensor data after Kalman filter denoising, normalization operation is performed according to the measurement range to obtain preprocessed data.

3. A dust collector safe operation monitoring method according to claim 1, characterized in that: The principal component analysis method is used to perform dimensionality reduction processing on the pre-processed data to obtain feature data, including: The preprocessed data are organized into a matrix, the covariance matrix is ​​calculated, the eigenvalues ​​and eigenvectors of the covariance matrix are solved, and the data are sorted in descending order of eigenvalues. The first m principal components are selected according to the cumulative contribution rate, the principal component scores are calculated, and the preprocessed data are reduced to m dimensions based on the principal component scores to obtain feature data.

4. A dust collector safe operation monitoring method according to claim 1, characterized in that: The CNN network in the LSTM-CNN hybrid model consists of two convolutional layers and one pooling layer. The convolutional layer uses a 3×3 convolution kernel, the activation function uses the ReLU function, and the pooling layer uses maximum pooling. The LSTM network contains two layers of LSTM units, with 128 neurons in each layer.

5. A dust collector safe operation monitoring method according to claim 4, characterized in that: The feature data is used as the input of the LSTM-CNN hybrid model, and the LSTM-CNN hybrid model outputs the equipment operation status classification and fault type identification, including: The feature data is arranged in chronological order to form time series data, which is then fed into the CNN network. The convolution layer performs sliding convolution on the time series data to extract local feature patterns in the data, and the convolution output is downsampled through the pooling layer. The local feature sequence extracted by the CNN network is input into the LSTM network. The LSTM unit captures the long-term and short-term dependencies in the local feature sequence through its internal gating mechanism. The temporal features output by the LSTM network are fused with the local features extracted by the CNN to form a comprehensive feature vector containing spatiotemporal feature information; First, the comprehensive feature vector is mapped to a high-dimensional feature space through a fully connected layer, and then the output of the fully connected layer is converted into the probability distribution of each operating status category and fault type through the Softmax activation function. The final equipment operating status classification and fault type identification results are determined based on the maximum probability.

6. A dust collector safe operation monitoring method according to claim 1, characterized in that: The output of the model is used to provide graded warnings, and filter bag replacement cycles and cleaning strategy adjustments are recommended based on the Weibull distribution, including: Combined with the dust collector's design parameters, historical operating data, and safety standards, a multi-level warning threshold is established for each fault type and operating state. Based on the equipment operating status classification and fault type identification results output by the LSTM-CNN hybrid model, a weighted calculation is performed to obtain the comprehensive risk value of the equipment's current operating status. The calculated comprehensive risk value is compared with the multi-level warning threshold to determine the warning level corresponding to the current equipment operation status and automatically trigger the corresponding warning measures.

7. A dust collector safe operation monitoring method according to claim 1, characterized in that: The method of providing graded warnings based on the output of the model and recommending filter bag replacement cycles and cleaning strategy adjustments based on Weibull distribution also includes: Obtain historical operating data of the filter bags. Based on the filter bag life data, use the maximum likelihood estimation method to estimate the shape parameters and scale parameters of the Weibull model distribution to determine the distribution characteristics of the filter bag life. Substitute the current filter bag operating time and status parameters into the Weibull distribution model to calculate the failure probability of the filter bag at different time points in the future. Based on the preset failure probability threshold, the remaining life of the filter bag is predicted and the optimal filter bag replacement cycle is generated. When the pressure difference of the dust collector continues to rise and the dust concentration is not effectively reduced, adjust the cleaning strategy.

8. A dust collector safety operation monitoring system, characterized in that: The system includes: The acquisition module is used to use sensors to collect dust concentration data, gas pressure data, equipment operating temperature data, and equipment operating vibration data of the dust collector in real time to obtain a sensor data set; The preprocessing module is used to perform denoising on the sensor data set using the Kalman filter algorithm, and perform normalization on the data of different sensors to obtain preprocessed data; The recognition module is used to perform dimensionality reduction on the preprocessed data using principal component analysis to obtain feature data. The feature data is used as the input of the LSTM-CNN hybrid model, and the LSTM-CNN hybrid model outputs the equipment operation status classification and fault type identification; The early warning module is used to provide graded early warnings based on the output results of the model, and recommend filter bag replacement cycles and cleaning strategy adjustments based on Weibull distribution.

9. A dust collector safety operation monitoring device, characterized in that: The dust collector safe operation monitoring device includes a memory and at least one processor, and instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the dust collector safe operation monitoring device performs each step of the dust collector safe operation monitoring method as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the various steps of the dust collector safe operation monitoring method as described in any one of claims 1 to 7 are implemented.

Images