Real-time detection method and system for activated carbon adsorption device based on deep learning

Abstract

The present invention relates to the technical field of activated carbon adsorption device detection, and in particular to a real-time detection method and system for activated carbon adsorption devices based on deep learning, which can reduce costs and improve overall operational efficiency by optimizing management processes; the method comprises: based on set time nodes, collecting air inlet VOCs concentrations and air outlet VOCs concentrations of the activated carbon adsorption device during operation; the air inlet VOCs concentrations and the air outlet VOCs concentrations correspond one to one; arranging multiple air inlet VOCs concentrations and multiple air outlet VOCs concentrations collected at different time nodes in time series to generate a VOCs processing feature set for the adsorption device; inputting the VOCs processing feature set for the adsorption device into a pre-trained adsorption efficiency evaluation model to obtain an adsorption efficiency stability evaluation index for the activated carbon adsorption device; in response to the adsorption efficiency stability evaluation index not exceeding a preset stability evaluation threshold, an abnormal adsorption efficiency alarm for the activated carbon adsorption device is issued.

Description

Technical Field

[0001] The present invention relates to the technical field of activated carbon adsorption device detection, and in particular to a real-time detection method and system for activated carbon adsorption devices based on deep learning. Background Art

[0002] With the acceleration of industrialization, the emission of volatile organic compounds (VOCs) has received increasing attention. VOCs not only pollute the environment, but may also have adverse effects on human health. In order to effectively control the emission of VOCs, many companies and institutions use activated carbon adsorption devices as one of the treatment methods. Activated carbon is widely used in VOCs treatment facilities due to its efficient physical adsorption performance and relatively simple operation and maintenance.

[0003] However, existing detection methods for activated carbon adsorption devices often rely on only a few key parameters (such as inlet and outlet concentrations) for evaluation, ignoring other important indicators, lacking effective monitoring methods, and unable to monitor its adsorption efficiency and operating status in real time, making it difficult for companies to conduct effective management interventions in a timely manner, and there is a risk of unorganized VOCs emissions or even excessive emissions. In addition, the operating and maintenance costs of activated carbon adsorption devices are high. If they are not managed in a refined manner, they will increase corporate costs, such as the cost of purchasing new activated carbon and disposing of waste activated carbon. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a real-time detection method and system for an activated carbon adsorption device based on deep learning, which can reduce costs and improve overall operational efficiency by optimizing management processes.

[0005] In a first aspect, the present invention provides a real-time detection method for an activated carbon adsorption device based on deep learning, the method comprising:

[0006] Based on the set time node, the VOCs concentration at the air inlet and the VOCs concentration at the air outlet of the activated carbon adsorption device during operation are collected; the VOCs concentration at the air inlet and the VOCs concentration at the air outlet correspond one to one;

[0007] Arrange multiple air inlet VOCs concentrations and multiple air outlet VOCs concentrations collected at different time points in time sequence to generate a VOCs treatment feature set for the adsorption device;

[0008] The VOCs treatment feature set of the adsorption device is input into the pre-trained adsorption efficiency evaluation model to obtain the adsorption efficiency stability evaluation index of the activated carbon adsorption device;

[0009] In response to the adsorption efficiency stability evaluation index not exceeding the preset stability evaluation threshold, an abnormal adsorption efficiency alarm of the activated carbon adsorption device is issued;

[0010] In response to the adsorption efficiency stability evaluation index exceeding the preset stability evaluation threshold, the operation data information and the operation environment data information of the activated carbon adsorption device during operation are collected in real time; the operation data information includes the air inlet pressure, the air outlet pressure, the air inlet gas flow, the air outlet gas flow and the device temperature; the operation environment data information includes the ambient humidity, the ambient temperature and the ambient air pressure;

[0011] Inputting the operation work data information into a pre-trained device operation status evaluation model to obtain a first operation status evaluation index;

[0012] Using the operating environment data information as an index tag, index traversal is performed in a pre-built operating state environment impact space to obtain real-time environment correction parameters;

[0013] Performing environmental impact correction processing on the first operating state evaluation index using the real-time environmental correction parameter to obtain a second operating state evaluation index;

[0014] In response to the second operating state evaluation index not exceeding the preset state evaluation threshold, an abnormal operating state alarm of the activated carbon adsorption device is issued; in response to the second operating state evaluation index exceeding the preset state evaluation threshold, no action is taken.

[0015] Furthermore, the mathematical calculation formula of the device operation status evaluation model is:

[0016]

[0017] in, I 1 represents the first operating status evaluation index, w i Indicates i The weight coefficient of the parameter, X i Indicates i The actual value of the parameter, n Indicates the number of parameters.

[0018] Furthermore, the structure of the device operation status assessment model includes:

[0019] The input layer is used to receive the operation data information collected in real time from the activated carbon adsorption device;

[0020] The feature extraction layer selects the features that have a key impact on the evaluation of the operating status according to the operating mechanism of the activated carbon adsorption device;

[0021] The hidden layer uses a neural network structure of multi-layer perceptron, convolutional neural network, and recurrent neural network to learn and represent the extracted features;

[0022] The output layer is used to calculate the operating status evaluation index of the activated carbon adsorption device according to the output of the hidden layer.

[0023] Furthermore, the method for constructing the adsorption efficiency evaluation model includes:

[0024] Collect data on the activated carbon adsorption unit under different operating conditions;

[0025] Clean the collected data, remove noise and outliers, and standardize and normalize the data;

[0026] Selecting a deep learning model as the basic architecture of the adsorption efficiency evaluation model, wherein the deep learning model includes a recurrent neural network, a long short-term memory network, and a gated recurrent unit;

[0027] Divide the preprocessed data into training set, validation set and test set;

[0028] Use the training set data to train the model and adjust the model parameters through the back propagation algorithm to minimize the prediction error;

[0029] Use the validation set data to tune the model and select the optimal model structure and parameter combination;

[0030] Use test set data to evaluate the performance of the model to ensure the effectiveness and reliability of the model in practical applications;

[0031] Adjust the model structure and parameters based on the evaluation results, including increasing the number of network layers, changing the activation function, and adjusting the learning rate;

[0032] The trained model is deployed to the actual activated carbon adsorption device detection system.

[0033] Furthermore, factors affecting the setting of the stability evaluation threshold include industry standards, equipment performance, historical operating data, operating environment factors and cost considerations.

[0034] Furthermore, the method for obtaining the VOCs processing feature set of the adsorption device includes:

[0035] Arrange the VOCs concentration data at the air inlet and outlet at each time point;

[0036] Arrange the data in chronological order to form a time series data;

[0037] Standardize and normalize time series data; eliminate the impact of dimension and make the data within the same scale range;

[0038] The processed time series data are combined into a feature set, namely the adsorption device VOCs processing feature set.

[0039] Furthermore, the abnormal operating status alarm of the activated carbon adsorption device includes sound alarm, light signal alarm, display screen alarm, remote notification, automatic record report, and visual alarm map.

[0040] On the other hand, the present application also provides a real-time detection system for an activated carbon adsorption device based on deep learning, the system comprising:

[0041] A data acquisition module, based on a set time node, collects the VOCs concentration at the air inlet and the VOCs concentration at the air outlet of the activated carbon adsorption device during operation; the VOCs concentration at the air inlet and the VOCs concentration at the air outlet correspond one to one;

[0042] The feature set generation module arranges the VOCs concentrations of multiple air inlets and multiple air outlets collected at different time points in time sequence to generate a VOCs treatment feature set for the adsorption device;

[0043] The adsorption efficiency evaluation module inputs the VOCs processing feature set of the adsorption device into the pre-trained adsorption efficiency evaluation model to obtain the adsorption efficiency stability evaluation index of the activated carbon adsorption device;

[0044] an abnormal alarm module, in response to the adsorption efficiency stability evaluation index not exceeding a preset stability evaluation threshold, to issue an abnormal adsorption efficiency alarm for the activated carbon adsorption device;

[0045] A refined data acquisition module, in response to the adsorption efficiency stability evaluation index exceeding a preset stability evaluation threshold, collects in real time the operating data information and operating environment data information of the activated carbon adsorption device during operation; the operating data information includes the air inlet pressure, the air outlet pressure, the air inlet gas flow, the air outlet gas flow and the device temperature; the operating environment data information includes the ambient humidity, the ambient temperature and the ambient air pressure;

[0046] The device operation status evaluation module inputs the operation work data information into a pre-trained device operation status evaluation model to obtain a first operation status evaluation index;

[0047] The environmental impact correction module uses the operating environment data information as an index tag, performs index traversal in a pre-built operating state environmental impact space, and obtains a real-time environmental correction parameter; uses the real-time environmental correction parameter to perform environmental impact correction processing on the first operating state evaluation index to obtain a second operating state evaluation index;

[0048] The comprehensive judgment module, in response to the second operating state evaluation index not exceeding the preset state evaluation threshold, issues an abnormal operating state alarm for the activated carbon adsorption device; in response to the second operating state evaluation index exceeding the preset state evaluation threshold, takes no action.

[0049] In a third aspect, the present application provides an electronic device, comprising a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory, and the processor are connected via the bus, and the computer program, when executed by the processor, implements the steps of any one of the above methods.

[0050] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps in any one of the above-mentioned methods when executed by a processor.

[0051] Compared with the prior art, the present invention has the following beneficial effects: the method not only monitors the key parameter of the inlet and outlet VOCs concentration of the activated carbon adsorption device, but also covers the operation data and the operation environment data; this comprehensive and multi-dimensional monitoring method can more accurately reflect the actual operation status and adsorption efficiency of the device, thereby improving the accuracy and reliability of the evaluation;

[0052] By setting time nodes for data collection and inputting them into the deep learning model for analysis in real time, this method can achieve real-time monitoring and dynamic evaluation of the operating status of the activated carbon adsorption device. This real-time feature enables enterprises to identify potential problems in a timely manner and take corresponding management measures, effectively reducing the risk of VOCs unorganized emissions and excessive emissions.

[0053] This method uses a deep learning model to evaluate adsorption efficiency and operating status and automatically trigger abnormal alarms; it reduces the burden of manual monitoring, improves work efficiency, and reduces the possibility of human error;

[0054] By introducing operating environment data information and using real-time environmental correction parameters to correct the operating status evaluation index, the impact of different environmental conditions on the operating status of the device can be considered, improving the environmental adaptability of the evaluation results; making the evaluation results more accurate, which helps enterprises to formulate more scientific and reasonable maintenance plans and management strategies under different environmental conditions;

[0055] This method provides enterprises with a means of refined management through real-time monitoring and intelligent evaluation. Enterprises can carry out targeted maintenance and optimization of activated carbon adsorption devices based on the evaluation results, reduce operating and maintenance costs, and improve the efficiency and life of the device.

[0056] In summary, the real-time detection method of activated carbon adsorption devices based on deep learning provides a solution with high automation, strong intelligence and wide coverage. It can not only help enterprises meet environmental compliance requirements, but also reduce costs and improve overall operational efficiency by optimizing management processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 is a flow chart of the present invention;

[0058] Figure 2 This is a structural diagram of the real-time detection system of activated carbon adsorption device based on deep learning. DETAILED DESCRIPTION

[0059] In the description of this application, those skilled in the art should know that this application can be implemented as a method, an apparatus, an electronic device, and a computer-readable storage medium. Therefore, this application can be specifically implemented in the following forms: complete hardware, complete software (including firmware, resident software, microcode, etc.), a combination of hardware and software. In addition, in some embodiments, this application can also be implemented in the form of a computer program product in one or more computer-readable storage media, and the computer-readable storage medium contains computer program code.

[0060] The above-mentioned computer-readable storage medium may adopt any combination of one or more computer-readable storage media. Computer-readable storage media include: electrical, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or devices, or any combination of the above. More specific examples of computer-readable storage media include: portable computer disks, hard disks, random access memories, read-only memories, erasable programmable read-only memories, flash memories, optical fibers, optical disc read-only memories, optical storage devices, magnetic storage devices, or any combination of the above. In the present application, computer-readable storage media can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or device.

[0061] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws.

[0062] The present application describes the provided methods, devices, and electronic devices through flowcharts and / or block diagrams.

[0063] It should be understood that each box in the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer-readable program instructions. These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine, and these computer-readable program instructions are executed by a computer or other programmable data processing device to produce a device that implements the functions / operations specified by the boxes in the flowchart and / or block diagram.

[0064] These computer-readable program instructions may also be stored in a computer-readable storage medium that enables a computer or other programmable data processing device to work in a specific manner. In this way, the instructions stored in the computer-readable storage medium produce an instruction device product including functions / operations specified in the blocks in the flowchart and / or block diagram.

[0065] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device, so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby enabling the instructions executed on the computer or other programmable data processing apparatus to provide a process for implementing the functions / operations specified in the blocks in the flowchart and / or block diagram.

[0066] The present application is described below in conjunction with the drawings in the present application.

[0067] Embodiment 1: Figure 1 As shown, the real-time detection method of the activated carbon adsorption device based on deep learning of the present invention specifically includes the following steps:

[0068] S1. Based on the set time node, collect the VOCs concentration at the air inlet and the VOCs concentration at the air outlet of the activated carbon adsorption device during operation; the VOCs concentration at the air inlet and the VOCs concentration at the air outlet correspond one to one;

[0069] The method for collecting the VOCs concentration at the air inlet and the VOCs concentration at the air outlet of the activated carbon adsorption device during operation is as follows:

[0070] According to actual needs, set an appropriate data collection time interval to ensure the continuity and real-time nature of data collection;

[0071] Use professional VOCs concentration detection equipment or sensors to perform real-time concentration measurement at the air inlet of the activated carbon adsorption device; record the air inlet VOCs concentration value at each time point to ensure the accuracy and reliability of the data;

[0072] Also use professional VOCs concentration detection equipment or sensors to measure the concentration in real time at the outlet of the activated carbon adsorption device; record the outlet VOCs concentration value at each time point and correspond it to the inlet concentration value one by one for subsequent analysis;

[0073] Ensure the one-to-one correspondence between the VOCs concentration data at the inlet and outlet at the time node, which is the key to evaluating the adsorption efficiency of activated carbon;

[0074] The basis for setting the time node includes:

[0075] Equipment operation cycle: The activated carbon adsorption device will have a fixed operation cycle. The setting of the time node should take these operation cycles into consideration to ensure that there are enough data points in each cycle to reflect the operating status of the device;

[0076] VOCs emission characteristics: VOCs emission characteristics vary in different industries and processes; the time nodes should be set based on these characteristics so that data collection can be intensified during emission peaks or large fluctuations, thereby more accurately evaluating the adsorption efficiency of the device;

[0077] Monitoring accuracy requirements: Set time nodes according to the accuracy and real-time requirements of VOCs emission monitoring of enterprises or institutions; for occasions that require high-precision real-time monitoring, the time nodes need to be set shorter; for occasions that do not require high real-time performance, the time interval can be appropriately extended;

[0078] System resource limitation: Considering the system resources required for data collection, storage and processing, the setting of time nodes should reasonably balance the relationship between monitoring requirements and system resources; avoid overly intensive data collection that leads to system resource tension or waste;

[0079] Historical data analysis: Determine the appropriate time node by analyzing historical data on the operation of the activated carbon adsorption device.

[0080] In this step, by setting an appropriate data collection time interval, continuous monitoring of key parameters of the activated carbon adsorption device during operation can be ensured, thereby providing real-time and accurate data support, laying a solid foundation for subsequent analysis and evaluation; the one-to-one correspondence between the VOCs concentration data at the inlet and outlet is the key to evaluating the activated carbon adsorption efficiency; this step ensures the scientificity and accuracy of the evaluation process by accurately measuring and recording these data, which helps to promptly discover and deal with abnormal adsorption efficiency; taking into account the differences in VOCs emission characteristics in different industries and process processes, this step allows time nodes to be set according to actual needs, thereby encrypting data collection during emission peaks or when fluctuations are large, thereby improving the pertinence and effectiveness of monitoring; when setting time nodes, this step comprehensively considers the limitations of system resources to avoid resource waste or system strain caused by overly intensive data collection; this helps to ensure that the system is running smoothly. While ensuring monitoring needs, it optimizes resource utilization efficiency; by continuously and in real time monitoring the operating status of the activated carbon adsorption device, this step provides strong support for the company's refined management; the company can adjust the operating strategy in time according to the monitoring data, optimize the operating parameters, thereby reducing VOCs emissions, reducing environmental pollution, and reducing operating and maintenance costs; when it is found that the adsorption efficiency stability evaluation index does not exceed the preset stability evaluation threshold, this step can promptly alarm and prompt the company to take corresponding measures to deal with it; it helps the company to respond quickly to abnormal situations, prevent the occurrence of unorganized or excessive emissions of VOCs, and ensure environmental safety and human health; this step S1 provides strong support for subsequent adsorption efficiency evaluation and operating status monitoring by scientifically and reasonably setting time nodes and accurately collecting key parameter data during the operation of the activated carbon adsorption device, which helps to improve the company's environmental management level and emergency response capabilities.

[0081] S2, arranging the VOCs concentrations at multiple air inlets and multiple air outlets collected at different time points in time sequence to generate a VOCs processing feature set for the adsorption device;

[0082] The method for obtaining the VOCs processing feature set of the adsorption device includes:

[0083] The VOCs concentration data of the air inlet and the air outlet at each time point are sorted out; since these data are collected at the same time node, it can be ensured that the VOCs concentration data of each pair of air inlet and air outlet are matched with each other;

[0084] Arrange these data in chronological order to form a time series data; each set of VOCs concentration data at the air inlet and outlet will be sorted according to the time they were collected, retaining the time correlation between the data;

[0085] Considering that the values ​​of VOCs concentrations in different time periods may vary greatly, in order to enable the model to better understand and process these data, the time series data are standardized and normalized; the influence of dimensions is eliminated and the data are within the same scale range;

[0086] The processed time series data are combined into a feature set, namely, the adsorption device VOCs processing feature set; the adsorption device VOCs processing feature set contains the inlet and outlet VOCs concentration information within all acquisition time periods, as well as other auxiliary features that may be extracted.

[0087] In this step, the VOCs concentration data of the air inlet and outlet at each time point are sorted and matched to ensure the integrity and consistency of the data; the sorted data are arranged in chronological order to form time series data, which retains the time correlation between the data and helps to more accurately evaluate the adsorption efficiency and stability of the adsorption device; the time series data are standardized and normalized to eliminate the influence of the dimension and keep the data within the same scale range, avoiding the interference of different dimensions and scales on data analysis and model training; the processed time series data are combined into a feature set, namely the VOCs processing feature set of the adsorption device; this feature set not only contains the VOCs concentration information of the air inlet and outlet during all the collection time periods, but also Extracting other auxiliary features as needed will help to more comprehensively evaluate the performance of the adsorption device; the generated VOCs processing feature set of the adsorption device provides an efficient data basis for subsequent adsorption efficiency evaluation and operation status monitoring; by inputting these data into the pre-trained model, the adsorption efficiency stability evaluation index and the operation status evaluation index can be quickly obtained, thereby helping enterprises to detect problems in a timely manner and conduct effective management interventions; based on real-time and accurate adsorption device performance evaluation results, enterprises can formulate operation and maintenance strategies more scientifically, optimize resource allocation, and reduce operating costs; at the same time, by promptly detecting and handling abnormal adsorption efficiency and abnormal operation status, enterprises can also effectively avoid the risks of VOCs unorganized emissions and excessive emissions, and improve the level of environmental protection management.

[0088] S3, inputting the VOCs processing feature set of the adsorption device into the pre-trained adsorption efficiency evaluation model to obtain the adsorption efficiency stability evaluation index of the activated carbon adsorption device;

[0089] The VOCs treatment feature set of the adsorption device is used as input data and input into the pre-trained adsorption efficiency evaluation model;

[0090] The model automatically extracts key features from the input data, including the trend of concentration changes, fluctuation amplitude, periodicity, etc.

[0091] Based on the extracted features, the model calculates the adsorption efficiency stability evaluation index of the activated carbon adsorption device; this index is a numerical value used to quantify the stability of the adsorption efficiency;

[0092] The model outputs a specific value, namely the adsorption efficiency stability evaluation index; this index will be used in the subsequent judgment and decision-making process;

[0093] The method for constructing the adsorption efficiency evaluation model includes:

[0094] Collect data from a large number of activated carbon adsorption devices under different operating conditions, including VOCs concentrations at the inlet and outlet, timestamps, and other relevant parameters;

[0095] Clean the collected data to remove noise and outliers; standardize and normalize the data to ensure that data of different dimensions can be fairly compared in the same model;

[0096] Selecting a deep learning model suitable for processing time series data as the basic architecture of the adsorption efficiency evaluation model, the deep learning model includes a recurrent neural network, a long short-term memory network and a gated recurrent unit;

[0097] Divide the preprocessed data into training set, validation set and test set;

[0098] Use the training set data to train the model and adjust the model parameters through the back propagation algorithm to minimize the prediction error;

[0099] Use the validation set data to tune the model and select the optimal model structure and parameter combination;

[0100] Use test set data to evaluate the performance of the model to ensure the effectiveness and reliability of the model in practical applications;

[0101] Use precision, recall, and F1 scores to evaluate the classification performance of the model; analyze the types and causes of model errors and identify potential areas for improvement;

[0102] Adjust the model structure and parameters based on the evaluation results, including increasing the number of network layers, changing the activation function, and adjusting the learning rate; introduce regularization technology to prevent overfitting;

[0103] The trained model is deployed to the actual activated carbon adsorption device detection system.

[0104] In this step, the pre-trained deep learning model can automatically extract key information from the complex VOCs treatment feature set; the deep learning model performs well in processing this type of time series data and can significantly improve the accuracy of the evaluation; integrating the model into the actual activated carbon adsorption device detection system can realize real-time evaluation and monitoring of the adsorption efficiency; this helps to timely discover the instability of the adsorption efficiency and take corresponding measures to intervene to ensure that the treatment effect always remains at the expected level; through the obtained adsorption efficiency stability evaluation index, the factors affecting the stability of the adsorption efficiency can be further analyzed, and the operation strategy of the activated carbon adsorption device can be optimized, thereby improving the operation efficiency and economic benefits of the equipment; by collecting and processing a large amount of operation data of the activated carbon adsorption device, not only rich training materials are provided for the evaluation model, but also these data are fully utilized; the implementation of this step promotes the intelligent process of activated carbon adsorption device management; through automated and intelligent evaluation and monitoring methods, the need for manual intervention is reduced, the operating cost is reduced, and the efficiency and accuracy of management are improved; this step significantly improves the accuracy and real-time performance of the adsorption efficiency evaluation of the activated carbon adsorption device by introducing a deep learning model and a real-time evaluation mechanism, providing strong support for optimizing the operation strategy and improving management efficiency.

[0105] S4, in response to the adsorption efficiency stability evaluation index not exceeding the preset stability evaluation threshold, an abnormal adsorption efficiency alarm of the activated carbon adsorption device is issued;

[0106] The factors affecting the setting of the stability evaluation threshold include:

[0107] Industry standards: The stability evaluation threshold needs to meet the relevant national and local standards for VOCs emissions; the standards usually stipulate the limits of VOCs emissions and the minimum efficiency standards that adsorption devices should meet; therefore, the threshold setting should be at least not lower than the minimum standards required by these regulations;

[0108] Equipment performance: Activated carbon adsorption devices of different models, specifications and brands have different performances; therefore, when setting the stability evaluation threshold, the performance parameters of the specific equipment need to be considered to ensure that the threshold can accurately reflect the actual operating status of the equipment;

[0109] Historical operation data: By analyzing the historical operation data of the activated carbon adsorption device, we can understand the stability and change trend of the adsorption efficiency of the equipment under different working conditions; these data provide an important basis for setting a reasonable stability evaluation threshold;

[0110] Operating environment factors: The operating efficiency of the activated carbon adsorption device is affected by a variety of operating conditions and environmental factors. Changes in operating environment factors will cause fluctuations in adsorption efficiency. Therefore, when setting the stability evaluation threshold, the influence of these factors needs to be considered, and the false alarm rate may be reduced by introducing environmental correction parameters.

[0111] Cost considerations: Enterprises also need to consider their own needs and costs when setting stability evaluation thresholds. On the one hand, too low a threshold may lead to frequent alarms and unnecessary maintenance costs. On the other hand, too high a threshold may not be able to detect potential problems in a timely manner, increasing the risk of excessive emissions. Therefore, enterprises need to set reasonable thresholds based on their actual conditions and needs while meeting regulatory requirements.

[0112] In this step, by real-time monitoring and evaluating the stability of adsorption efficiency, the system can detect potential problems and issue alarms in the first place; it helps enterprises to take timely intervention measures to avoid excessive emissions and unorganized emissions of VOCs, protect the environment and human health; the timely abnormal alarm mechanism helps to reduce the risk of VOCs emissions caused by equipment failure or improper operation; it helps enterprises to comply with national and local environmental protection laws and regulations and avoid legal risks and fines due to illegal emissions; by reasonably setting the stability evaluation threshold, enterprises can balance the alarm frequency and maintenance costs according to their actual situation and needs without meeting regulatory requirements; too low thresholds may lead to frequent alarms and unnecessary maintenance costs, while too high thresholds may not be able to detect potential problems in time; therefore, reasonable threshold settings help optimize operating costs and improve economic benefits; the automated abnormal alarm mechanism reduces the burden of manual monitoring and improves management efficiency; managers can focus more on other important tasks while relying on the system to monitor and evaluate the operating status of the activated carbon adsorption device in real time; by collecting and analyzing alarm data and subsequent treatment measures, enterprises can continuously improve and optimize the operation and management of the activated carbon adsorption device to further improve the VOCs treatment effect.

[0113] S5, in response to the adsorption efficiency stability evaluation index exceeding the preset stability evaluation threshold, real-time collection of operation data information and operation environment data information of the activated carbon adsorption device during operation; the operation data information includes air inlet pressure, air outlet pressure, air inlet gas flow, air outlet gas flow and device temperature; the operation environment data information includes ambient humidity, ambient temperature and ambient air pressure;

[0114] Operation work data information collection:

[0115] The inlet pressure is monitored in real time by a pressure sensor installed at the inlet. The pressure change can reflect the state of gas flow.

[0116] The outlet pressure is monitored by a pressure sensor; the pressure difference between the inlet and outlet helps to understand the pressure loss of the gas when it passes through the activated carbon layer;

[0117] The gas flow rate at the air inlet uses a flow meter to measure the volume of gas entering the activated carbon adsorption device per unit time;

[0118] The gas flow at the outlet is monitored using a flow meter to monitor the gas flow leaving the device; the comparison of the inlet and outlet gas flows can reveal whether there are blockages or other problems inside the device;

[0119] The device temperature monitors the temperature inside the activated carbon adsorption device through a temperature sensor; temperature changes may affect the adsorption performance of activated carbon, so real-time monitoring is required;

[0120] Operation environment data information collection:

[0121] Ambient humidity: A humidity sensor is used to monitor the humidity of the environment around the device; humidity has a certain impact on the adsorption efficiency of VOCs, so it needs to be recorded;

[0122] Ambient temperature The ambient temperature is monitored by a temperature sensor; changes in ambient temperature may affect the adsorption efficiency and physical properties of the device;

[0123] Ambient Air Pressure uses an air pressure sensor to monitor the ambient atmospheric pressure; changes in air pressure may affect the density and flow rate of the gas, which in turn affects the adsorption efficiency.

[0124] In this step, by collecting operating data information and operating environment data information in real time, the operating status of the activated carbon adsorption device and its impact on environmental factors can be fully and accurately understood; once any abnormality or potential problem is found, it can respond immediately to avoid the unorganized emission and excessive emission of VOCs and ensure environmental safety; real-time monitoring of the pressure, gas flow rate and device temperature at the air inlet and outlet helps to understand the flow state and adsorption efficiency of the gas in the device; by comparing the pressure difference and gas flow rate at the inlet and outlet, blockage, leakage and other problems inside the device can be discovered in time, so that corresponding maintenance measures can be taken to improve the operating efficiency and stability of the device; real-time monitoring of environmental humidity, temperature and air pressure enables the operation of the activated carbon adsorption device to be better Adapt to changes in the external environment; by recording and analyzing these environmental data, the operating parameters of the device can be adjusted and optimized, its adaptability to different environmental conditions can be improved, and the stability and efficiency of adsorption efficiency can be ensured; through real-time monitoring and data analysis, potential problems can be discovered and resolved in a timely manner to avoid the expansion of problems leading to more serious consequences and higher maintenance costs; at the same time, based on the refined management of data, the replacement cycle of activated carbon and the disposal of waste activated carbon can be optimized, and the operating and maintenance costs can be reduced; the implementation of this step relies on advanced sensor technology and data processing technology, making the detection and management of activated carbon adsorption devices more intelligent; through automated collection and real-time analysis of data, remote monitoring and intelligent management of the device can be achieved, improving management efficiency and accuracy.

[0125] S6, inputting the operation data information into a pre-trained device operation status evaluation model to obtain a first operation status evaluation index;

[0126] The structure of the device operation status assessment model includes:

[0127] The input layer is used to receive the operating data information collected in real time from the activated carbon adsorption device, including the inlet pressure, outlet pressure, inlet gas flow, outlet gas flow and device temperature; the input data is cleaned, normalized and standardized to ensure the stability and accuracy of the model training;

[0128] The feature extraction layer selects the features that have a key impact on the evaluation of the operating status according to the operating mechanism of the activated carbon adsorption device; through linear transformation, nonlinear transformation and feature combination, it further extracts features that are helpful in distinguishing different operating states;

[0129] The hidden layer uses a neural network structure of multi-layer perceptron, convolutional neural network, and recurrent neural network to learn and represent the extracted features; nonlinear activation functions are introduced between neural network layers to enhance the nonlinear expression and learning capabilities of the model; to prevent the model from overfitting, a dropout mechanism can be introduced after certain layers to randomly discard the outputs of some neurons;

[0130] The output layer is used to calculate the operating status evaluation index of the activated carbon adsorption device according to the output of the hidden layer; the index is usually a continuous value or a classification label, which is used to indicate whether the operating status of the device is normal;

[0131] The mathematical calculation formula of the device operation status evaluation model is:

[0132]

[0133] in, I 1 represents the first operating status evaluation index, w i Indicates i The weight coefficient of the parameter, X i Indicates i The actual value of the parameter, n Indicates the number of parameters.

[0134] In this step, the model can receive and process the operating data information collected from the activated carbon adsorption device in real time, including key parameters such as inlet pressure, outlet pressure, inlet gas flow, outlet gas flow, and device temperature; through advanced machine learning or deep learning algorithms, the model can accurately evaluate the current operating status of the device and provide timely and reliable operating status evaluation indexes; in the feature extraction layer, the model not only considers the influence of a single parameter, but also further extracts comprehensive features that help distinguish different operating states through linear transformation, nonlinear transformation, and feature combination; this enables the model to more comprehensively reflect the operating status of the device and improve the accuracy and reliability of the evaluation; the introduction of the device operating status evaluation model realizes the intelligent and automated monitoring of the operating status of the activated carbon adsorption device; enterprises do not need to rely on manual inspections and experience judgments Real-time knowledge of the operating status of the device greatly improves management efficiency and response speed; by real-time monitoring and evaluation of the operating status of the device, enterprises can promptly discover potential problems and anomalies, and take corresponding measures to intervene and adjust; this helps prevent the unorganized emission and excessive emission of VOCs, reduce the risk of environmental pollution, and also helps to extend the service life of the device and reduce operating costs; the structure and parameters of the model can be customized and adjusted according to specific application scenarios and needs; with the continuous development of activated carbon adsorption device technology and the continuous accumulation of operating data, the model can be continuously optimized and updated to adapt to new operating environments and changes in conditions; step S6 introduces a device operating status evaluation model to achieve real-time, accurate, comprehensive, intelligent and automated monitoring and evaluation of the operating status of the activated carbon adsorption device, providing enterprises with strong technical support and decision-making basis.

[0135] S7, using the operating environment data information as an index tag, performing index traversal in a pre-built operating state environment impact space to obtain real-time environment correction parameters;

[0136] The method for constructing the operating state environment impact space includes:

[0137] Collect a large amount of operating status data of activated carbon adsorption devices operating under different environmental conditions; these data include operating data information of the device under different humidity, temperature and air pressure and the corresponding device operating status evaluation;

[0138] Clean the collected data, remove invalid and erroneous data records, and ensure the accuracy and consistency of the data; standardize the collected data so that data of different dimensions can be compared on the same scale;

[0139] Group historical data according to different intervals of environmental parameters to form multiple subsets; divide the ambient temperature into several intervals and put the corresponding data records into the corresponding intervals;

[0140] Use multidimensional indexing technology to build indexes for the grouped data and insert index nodes, each of which represents a combination of feature intervals;

[0141] Insert the preprocessed data into the corresponding index node according to the feature interval; each index node contains all historical data records in the interval and stores the corresponding environmental correction parameters;

[0142] For each index node, the change trend of the device operation state within the interval is calculated as the environment correction parameter; and the calculated environment correction parameter is stored in the corresponding index node.

[0143] In this step, by constructing the operating status environmental impact space, the impact of different environmental parameters on the operating status of the device is taken into account, so that the evaluation results can more accurately reflect the performance of the device under the current environment; the use of multidimensional indexing technology can realize rapid traversal of the operating status environmental impact space, so as to quickly find the data points or data areas that best match the current environmental parameters; once the real-time environmental data information is obtained, the index traversal can be immediately performed to obtain the corresponding environmental correction parameters, and the device operating status evaluation index can be corrected in real time to ensure the timeliness of the evaluation results; by real-time monitoring and evaluation of the operating status of the device, the enterprise can promptly discover and deal with potential problems and avoid the risks of unorganized emissions and excessive emissions. This will reduce fines and reputation losses caused by illegal emissions; based on accurate operating status assessments, companies can more reasonably arrange activated carbon replacement cycles and waste activated carbon disposal plans, reducing unnecessary waste and cost expenditures; provide reliable data support for the company's management decisions, help managers understand the performance of the device in different environments, and formulate more scientific and reasonable operation and maintenance strategies; through real-time monitoring and evaluation of the operating status of the device, potential operating risks can be discovered in advance, providing a time window for the company to take preventive measures; the real-time environmental correction parameter acquisition process in step S7, by constructing and operating the environmental impact space of the status, realizes accurate evaluation and real-time correction of the operating status of the activated carbon adsorption device.

[0144] S8, using the real-time environmental correction parameter to perform environmental impact correction processing on the first operating status evaluation index to obtain a second operating status evaluation index;

[0145] The method for obtaining the second operating status evaluation index includes:

[0146] The first operating state evaluation index is corrected using the real-time environmental correction parameter; the specific method of the correction process depends on the nature of the environmental correction parameter, and generally involves the following common correction methods:

[0147] Additive correction: If the correction parameter represents an increment or decrement to the current state, it can be directly added to the first operating state evaluation index;

[0148] Multiplication correction: If the correction parameter represents a proportional factor or a multiple, it can be multiplied by the first operating state evaluation index;

[0149] Composite correction: Sometimes it may be necessary to combine multiple correction methods, such as performing a proportional correction first and then an incremental correction;

[0150] The specific steps for the correction process are as follows:

[0151] Determine whether to use additive correction, multiplicative correction or composite correction according to the nature of the environmental correction parameters;

[0152] Correcting the first operating status evaluation index according to the determined correction method;

[0153] If additive correction is adopted, the correction parameter is directly added to the first operating state evaluation index;

[0154] If multiplication correction is adopted, the first operation status evaluation index is multiplied by the correction parameter;

[0155] If composite correction is used, multiplication correction is performed first, followed by addition correction;

[0156] After correction processing, the final second operating state evaluation index is obtained.

[0157] In this step, by introducing real-time environmental correction parameters, the first operating status evaluation index is corrected for environmental impact, which can more accurately reflect the actual operating status of the activated carbon adsorption device under the current environment; this correction takes into account the impact of dynamic changes in environmental factors on the performance of the device, thereby avoiding the deviation that may be caused by a single evaluation indicator; real-time collection of environmental data and immediate correction processing enable the second operating status evaluation index to instantly reflect the operating status of the device, which helps enterprises to take management measures in a timely manner to avoid unorganized emissions or excessive emissions of VOCs; through refined environmental correction and operating status evaluation, enterprises can more accurately grasp the operating status of the activated carbon adsorption device, thereby formulating more scientific and reasonable maintenance plans and cost control strategies; this helps to reduce operating and maintenance costs and improve resource utilization efficiency; by introducing multiple correction methods and strict correction steps, the accuracy and reliability of the correction process are ensured; this systematic correction method helps to reduce false alarms and missed alarms and improve the stability and reliability of the entire detection system; step S8 significantly improves the evaluation accuracy and real-time performance of the operating status of the activated carbon adsorption device by introducing real-time environmental correction parameters and refined correction processing, supports the refined management of enterprises, improves the reliability and stability of the system, and promotes the development of technological innovation.

[0158] S9, in response to the second operation status evaluation index not exceeding the preset status evaluation threshold, an abnormal operation status alarm of the activated carbon adsorption device is issued; in response to the second operation status evaluation index exceeding the preset status evaluation threshold, no action is taken;

[0159] Comparing the second operating status evaluation index with a preset status evaluation threshold;

[0160] If the second operation status evaluation index does not exceed the preset status evaluation threshold, it means that the operation status of the activated carbon adsorption device is abnormal or is about to be abnormal, and attention should be paid;

[0161] If the second operation status evaluation index exceeds the preset status evaluation threshold, it means that the activated carbon adsorption device is currently in a normal operation state and no additional intervention is required;

[0162] When the second operation status evaluation index does not exceed the preset threshold, the system should trigger an abnormal alarm mechanism;

[0163] The abnormal operation state alarm mode of the activated carbon adsorption device includes:

[0164] Sound alarm: A high-decibel alarm sound is emitted through a buzzer or speaker on the device to attract the attention of on-site personnel; the sound alarm usually has different tones and frequencies to distinguish different types of abnormalities;

[0165] Light signal alarm: Use light signal devices such as LED lights or flashing lights to display specific colors or flashing patterns on the device or in the control room to indicate abnormal conditions;

[0166] Display screen alarm: Detailed alarm information is displayed on the display screen in the control room or near the device, including abnormality type, occurrence time, current status evaluation index, etc.; providing more comprehensive information to help maintenance personnel quickly understand the problem;

[0167] Remote notification: Send alarm information to designated maintenance personnel or management personnel through remote communication methods such as SMS, email or mobile applications; ensure that relevant personnel can receive alarm notifications in time even if they are not on site;

[0168] Automatic recording and reporting: The system automatically records the time, type, processing results and other information of each alarm and generates corresponding reports; these records can be used for subsequent data analysis, performance evaluation and formulation of improvement measures;

[0169] Visual alarm map: In large factories or parks, a visual alarm map can be used to display the operating status of each activated carbon adsorption device; when an abnormality occurs in a device, the corresponding position on the map is displayed with a special mark to facilitate quick positioning and processing.

[0170] In this step, by real-time monitoring and evaluation of the operating status of the activated carbon adsorption device, potential operating anomalies can be quickly discovered, ensuring that the company can take timely intervention measures to prevent the unorganized emission or excessive emission of VOCs, thereby protecting the environment and human health; this step not only takes into account the stability of the adsorption efficiency, but also combines the operating work data information and the operating environment data information to conduct a comprehensive evaluation of the operating status of the device; this multi-dimensional monitoring method improves the accuracy and reliability of detection; when the operating status of the activated carbon adsorption device is detected to be abnormal, it can trigger a variety of alarm methods, including sound alarms, light signal alarms, display screen alarms, remote notifications, etc., to ensure that relevant personnel can quickly receive the alarm information and take corresponding treatment measures according to the actual situation; through fine Digital management helps enterprises to promptly discover and deal with operating problems of activated carbon adsorption devices, avoiding production interruptions and additional costs caused by equipment failures; at the same time, by optimizing operating parameters and reducing unnecessary maintenance operations, it can further reduce the operating costs of enterprises; the automatic recording and reporting function provides enterprises with rich historical data and statistical analysis reports, which helps enterprises better understand the operating status and performance of equipment, and provides strong support for subsequent decision-making and optimization and improvement; in large factories or parks, the introduction of visual alarm maps makes the operating status monitoring of activated carbon adsorption devices more intuitive and convenient; through special marks on the map, enterprises can quickly locate abnormal devices and take corresponding treatment measures, thereby improving management efficiency and response speed.

[0171] Embodiment 2: Figure 2 As shown, the real-time detection system of activated carbon adsorption device based on deep learning of the present invention specifically includes the following modules;

[0172] A data acquisition module, based on a set time node, collects the VOCs concentration at the air inlet and the VOCs concentration at the air outlet of the activated carbon adsorption device during operation; the VOCs concentration at the air inlet and the VOCs concentration at the air outlet correspond one to one;

[0173] The feature set generation module arranges the VOCs concentrations of multiple air inlets and multiple air outlets collected at different time points in time sequence to generate a VOCs treatment feature set for the adsorption device;

[0174] The adsorption efficiency evaluation module inputs the VOCs processing feature set of the adsorption device into the pre-trained adsorption efficiency evaluation model to obtain the adsorption efficiency stability evaluation index of the activated carbon adsorption device;

[0175] an abnormal alarm module, in response to the adsorption efficiency stability evaluation index not exceeding a preset stability evaluation threshold, to issue an abnormal adsorption efficiency alarm for the activated carbon adsorption device;

[0176] A refined data acquisition module, in response to the adsorption efficiency stability evaluation index exceeding a preset stability evaluation threshold, collects in real time the operating data information and operating environment data information of the activated carbon adsorption device during operation; the operating data information includes the air inlet pressure, the air outlet pressure, the air inlet gas flow, the air outlet gas flow and the device temperature; the operating environment data information includes the ambient humidity, the ambient temperature and the ambient air pressure;

[0177] The device operation status evaluation module inputs the operation work data information into a pre-trained device operation status evaluation model to obtain a first operation status evaluation index;

[0178] The environmental impact correction module uses the operating environment data information as an index tag, performs index traversal in a pre-built operating state environmental impact space, and obtains a real-time environmental correction parameter; uses the real-time environmental correction parameter to perform environmental impact correction processing on the first operating state evaluation index to obtain a second operating state evaluation index;

[0179] The comprehensive judgment module, in response to the second operating state evaluation index not exceeding the preset state evaluation threshold, issues an abnormal operating state alarm for the activated carbon adsorption device; in response to the second operating state evaluation index exceeding the preset state evaluation threshold, takes no action.

[0180] The system can collect key operating data of the activated carbon adsorption device in real time, including inlet and outlet VOCs concentrations, operating data, and operating environment data, and conduct real-time analysis and evaluation through deep learning models, thereby achieving real-time monitoring of adsorption efficiency and device operating status. This real-time monitoring capability enables enterprises to promptly identify potential problems and reduce the risk of VOCs unorganized emissions and excessive emissions.

[0181] The system not only focuses on the traditional inlet and outlet concentration parameters, but also introduces more operating data and environmental data to build a more comprehensive evaluation system; this multi-parameter comprehensive evaluation method improves the accuracy of the evaluation, allowing the system to more accurately reflect the actual operating conditions of the activated carbon adsorption device;

[0182] The system can automatically determine whether the adsorption efficiency and device operating status are abnormal based on the preset thresholds, and trigger the corresponding alarm mechanism; this intelligent alarm function enables enterprises to respond quickly to abnormal situations, take effective measures to intervene, and reduce VOCs emission risks and maintenance costs;

[0183] The system takes into account the impact of environmental factors on the operating status of the device, and corrects the operating status evaluation index through the environmental impact correction module, thereby improving the accuracy and reliability of the evaluation results; this correction mechanism enables the system to more accurately reflect the actual operating status of the device in different environments;

[0184] The system provides enterprises with refined management tools through real-time monitoring and intelligent evaluation. Enterprises can formulate more scientific and reasonable maintenance plans and management strategies based on the data and analysis results provided by the system, reduce operating and maintenance costs, and improve the efficiency and life of activated carbon adsorption devices. The modular design of the system makes it scalable and customizable. Enterprises can expand the functions of the system or customize it according to their own needs and actual conditions to meet the VOCs emission control needs in different scenarios.

[0185] In summary, the real-time detection system of activated carbon adsorption device based on deep learning can help improve the environmental management level of enterprises, and also help reduce operating costs.

[0186] The various variations and specific embodiments of the real-time detection method for an activated carbon adsorption device based on deep learning in the aforementioned embodiment 1 are also applicable to the real-time detection system for an activated carbon adsorption device based on deep learning in this embodiment. Through the aforementioned detailed description of the real-time detection method for an activated carbon adsorption device based on deep learning, those skilled in the art can clearly understand the implementation method of the real-time detection system for an activated carbon adsorption device based on deep learning in this embodiment. Therefore, for the sake of brevity of the specification, it will not be described in detail here.

[0187] In addition, the present application also provides an electronic device, including a bus, a transceiver, a memory, a processor, and a computer program stored in the memory and executable on the processor. The transceiver, the memory, and the processor are respectively connected via a bus. When the computer program is executed by the processor, each process of the above-mentioned method for controlling output data is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0188] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A real-time detection method for activated carbon adsorption device based on deep learning, characterized in that: The method comprises: Based on the set time node, the VOCs concentration at the air inlet and the VOCs concentration at the air outlet of the activated carbon adsorption device during operation are collected; the VOCs concentration at the air inlet and the VOCs concentration at the air outlet correspond one to one; Arrange multiple air inlet VOCs concentrations and multiple air outlet VOCs concentrations collected at different time points in time sequence to generate a VOCs treatment feature set for the adsorption device; The VOCs treatment feature set of the adsorption device is input into the pre-trained adsorption efficiency evaluation model to obtain the adsorption efficiency stability evaluation index of the activated carbon adsorption device; In response to the adsorption efficiency stability evaluation index not exceeding the preset stability evaluation threshold, an abnormal adsorption efficiency alarm of the activated carbon adsorption device is issued; In response to the adsorption efficiency stability evaluation index exceeding the preset stability evaluation threshold, the operation data information and the operation environment data information of the activated carbon adsorption device during operation are collected in real time; the operation data information includes the air inlet pressure, the air outlet pressure, the air inlet gas flow, the air outlet gas flow and the device temperature; the operation environment data information includes the ambient humidity, the ambient temperature and the ambient air pressure; Inputting the operation work data information into a pre-trained device operation status evaluation model to obtain a first operation status evaluation index; Using the operating environment data information as an index tag, index traversal is performed in a pre-built operating state environment impact space to obtain real-time environment correction parameters; Performing environmental impact correction processing on the first operating state evaluation index using the real-time environmental correction parameter to obtain a second operating state evaluation index; In response to the second operating state evaluation index not exceeding the preset state evaluation threshold, an abnormal operating state alarm of the activated carbon adsorption device is issued; in response to the second operating state evaluation index exceeding the preset state evaluation threshold, no action is taken.

2. The real-time detection method for activated carbon adsorption device based on deep learning according to claim 1, characterized in that: The mathematical calculation formula of the device operation status evaluation model is: in, I 1 represents the first operating status evaluation index, w i Indicates i The weight coefficient of the parameter, X i Indicates i The actual value of the parameter, n Indicates the number of parameters.

3. The real-time detection method for activated carbon adsorption device based on deep learning according to claim 2, characterized in that: The structure of the device operation status assessment model includes: The input layer is used to receive the operation data information collected in real time from the activated carbon adsorption device; The feature extraction layer selects the features that have a key impact on the evaluation of the operating status according to the operating mechanism of the activated carbon adsorption device; The hidden layer uses a neural network structure of multi-layer perceptron, convolutional neural network, and recurrent neural network to learn and represent the extracted features; The output layer is used to calculate the operating status evaluation index of the activated carbon adsorption device according to the output of the hidden layer.

4. The real-time detection method for activated carbon adsorption device based on deep learning according to claim 1, characterized in that: The method for constructing the adsorption efficiency evaluation model includes: Collect data on the activated carbon adsorption unit under different operating conditions; Clean the collected data, remove noise and outliers, and standardize and normalize the data; Selecting a deep learning model as the basic architecture of the adsorption efficiency evaluation model, wherein the deep learning model includes a recurrent neural network, a long short-term memory network, and a gated recurrent unit; Divide the preprocessed data into training set, validation set and test set; Use the training set data to train the model and adjust the model parameters through the back propagation algorithm to minimize the prediction error; Use the validation set data to tune the model and select the optimal model structure and parameter combination; Use test set data to evaluate the performance of the model to ensure the effectiveness and reliability of the model in practical applications; Adjust the model structure and parameters based on the evaluation results, including increasing the number of network layers, changing the activation function, and adjusting the learning rate; The trained model is deployed to the actual activated carbon adsorption device detection system.

5. The real-time detection method for activated carbon adsorption device based on deep learning according to claim 1, characterized in that: The factors affecting the setting of the stability evaluation threshold include industry standards, equipment performance, historical operation data, operating environment factors and cost considerations.

6. The real-time detection method for activated carbon adsorption device based on deep learning according to claim 1, characterized in that: The method for obtaining the VOCs processing feature set of the adsorption device includes: Arrange the VOCs concentration data at the air inlet and outlet at each time point; Arrange the data in chronological order to form a time series data; Standardize and normalize time series data; eliminate the impact of dimension and make the data within the same scale range; The processed time series data are combined into a feature set, namely the adsorption device VOCs processing feature set.

7. The real-time detection method for activated carbon adsorption device based on deep learning according to claim 1, characterized in that: The abnormal operating status alarm of the activated carbon adsorption device includes sound alarm, light signal alarm, display screen alarm, remote notification, automatic record report, and visual alarm map.

8. A real-time detection system for activated carbon adsorption device based on deep learning, characterized in that: The system comprises: A data acquisition module, based on a set time node, collects the VOCs concentration at the air inlet and the VOCs concentration at the air outlet of the activated carbon adsorption device during operation; the VOCs concentration at the air inlet and the VOCs concentration at the air outlet correspond one to one; The feature set generation module arranges the VOCs concentrations of multiple air inlets and multiple air outlets collected at different time points in time sequence to generate a VOCs treatment feature set for the adsorption device; The adsorption efficiency evaluation module inputs the VOCs processing feature set of the adsorption device into the pre-trained adsorption efficiency evaluation model to obtain the adsorption efficiency stability evaluation index of the activated carbon adsorption device; an abnormal alarm module, in response to the adsorption efficiency stability evaluation index not exceeding a preset stability evaluation threshold, to issue an abnormal adsorption efficiency alarm for the activated carbon adsorption device; A refined data acquisition module, in response to the adsorption efficiency stability evaluation index exceeding a preset stability evaluation threshold, collects in real time the operating data information and operating environment data information of the activated carbon adsorption device during operation; the operating data information includes the air inlet pressure, the air outlet pressure, the air inlet gas flow, the air outlet gas flow and the device temperature; the operating environment data information includes the ambient humidity, the ambient temperature and the ambient air pressure; The device operation status evaluation module inputs the operation work data information into a pre-trained device operation status evaluation model to obtain a first operation status evaluation index; The environmental impact correction module uses the operating environment data information as an index tag, performs index traversal in a pre-built operating state environmental impact space, and obtains a real-time environmental correction parameter; uses the real-time environmental correction parameter to perform environmental impact correction processing on the first operating state evaluation index to obtain a second operating state evaluation index; The comprehensive judgment module, in response to the second operating state evaluation index not exceeding the preset state evaluation threshold, issues an abnormal operating state alarm for the activated carbon adsorption device; in response to the second operating state evaluation index exceeding the preset state evaluation threshold, takes no action.

9. An electronic device for real-time detection of activated carbon adsorption device based on deep learning, comprising a bus, a transceiver, a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the transceiver, the memory and the processor are connected via the bus, characterized in that: When the computer program is executed by the processor, the steps in the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps in the method according to any one of claims 1 to 7 are implemented.

Images