Remote intelligent quality control system based on unattended operation

Abstract

The invention relates to the field of quality control systems, and particularly discloses an unattended remote intelligent quality control system, which comprises a hardware system, a software program system and an air monitoring instrument for monitoring air quality, the software program system comprises a data acquisition module; an intelligent prediction module; a self-learning optimization module; a virtual calibration simulation module; and a multi-source data fusion module. According to the remote intelligent quality control system based on unattended operation, all intelligent modules in a software program system work cooperatively, for example, a calibration control module can automatically generate a control instruction according to results of an intelligent prediction module and a virtual calibration simulation module and send the control instruction to a PLC; according to the invention, the automatic process from data acquisition, analysis, decision making to execution in the calibration process is realized, the dependence on manual operation is greatly reduced, and errors and unstable conditions caused by human factors are reduced.

Description

Technical Field

[0001] The present invention relates to the field of quality control systems, and in particular to an unattended remote intelligent quality control system. Background Art

[0002] Real-time monitoring and precise management of ambient air quality have become one of the core contents of environmental protection work. In this context, the requirements for air quality monitoring technology are becoming increasingly stringent, and efforts are being made to comprehensively and accurately grasp the air quality status and its changing trends through scientific and efficient monitoring methods. At present, the pollutants that are routinely concerned in ambient air quality monitoring mainly include sulfur dioxide (SO 2 ), nitrogen dioxide (NO 2 ), ozone (O 3 ) and carbon monoxide (CO), these four gases are listed as key monitoring targets due to their significant impact on human health and the environment.

[0003] However, the traditional gas monitoring calibration system has obvious limitations. The existing calibration process relies on equipment such as standard gas, zero gas generator and dynamic calibrator. Although these devices can ensure the accuracy of monitoring data to a certain extent, their operation is complicated and highly dependent on manual labor. Operation and maintenance personnel need to visit the monitoring site regularly for on-site operations, including replacing standard gas, starting calibration procedures, monitoring the calibration process and manually recording calibration results. This process is not only time-consuming and labor-intensive, increasing operation and maintenance costs, but also susceptible to human factors, which may lead to instability or errors in calibration results. In addition, the low frequency of traditional calibration methods also limits the timeliness of monitoring data. When air quality changes rapidly or sudden pollution incidents occur, it is difficult to reflect the real situation in a timely manner, which affects the rapid response and effective governance of pollution sources by environmental management departments.

[0004] Therefore, it is necessary to propose an unmanned remote intelligent quality control system to solve the above problems. Summary of the invention

[0005] The main purpose of the present invention is to provide an unmanned remote intelligent quality control system, which can effectively solve the problems in the background technology.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] A remote intelligent quality control system based on unattended operation, comprising a hardware system, a software program system and an air monitoring instrument for monitoring air quality;

[0008] The hardware system includes a three-way solenoid valve and a zero gas generator, a standard gas, and a dynamic calibrator connected to the three-way solenoid valve through a pipeline, wherein the gas outlet of the dynamic calibrator is connected to the gas inlet of the air monitoring instrument;

[0009] It also includes a PLC electrically connected to the three-way solenoid valve and the dynamic calibrator, wherein the PLC is used to control the opening and closing of the three-way solenoid valve, transmit control instructions to the dynamic calibrator, and receive feedback data from the dynamic calibrator;

[0010] The software program system is connected to the PLC signal, and the software program system includes:

[0011] Data acquisition module, used to collect hardware system data, including but not limited to the data collected by air monitoring instruments and the working parameter data of dynamic calibrators;

[0012] Intelligent prediction module, based on big data analysis and deep learning algorithms, is used to integrate historical calibration data, environmental condition data, and real-time operation data of air monitoring instruments to predict the drift trend and optimal calibration time of air monitoring instruments;

[0013] Self-learning optimization module, which is used to analyze calibration data, environmental condition data and air monitoring instrument performance parameters during and after the calibration process, and automatically adjust the calibration procedure parameters and calibration limits;

[0014] The virtual calibration simulation module is used to build a virtual calibration environment and air monitoring instrument model using existing data before the calibration operation, simulate the effects of different calibration schemes, and evaluate and select the best calibration strategy;

[0015] Multi-source data fusion module, used to integrate multi-source heterogeneous data, including but not limited to data from air monitoring instruments, operating status of hardware systems, and environmental condition data;

[0016] User interaction module, used to build a human-computer interaction interface, visualize the overall operation status of the system, and receive operation instructions from operation and maintenance personnel;

[0017] The calibration control module is used to generate accurate control instructions and send them to the PLC based on the calibration prediction information of the intelligent prediction module, the best calibration solution provided by the virtual calibration simulation module, and the calibration parameters set by the user through the user interaction module.

[0018] Preferably, a data processing and storage module is used to centrally process, analyze and store the data collected by the data collection module.

[0019] Preferably, the software program system further comprises a communication module, and the software program system is connected to the PLC signal via the communication module;

[0020] The communication module is also used for data transmission and communication connection with external devices, and the external devices are used to collect environmental condition data, which include but are not limited to environmental meteorological data and surrounding pollution source distribution data. The external devices are but not limited to servers, meteorological data collection terminals, and pollution source monitoring equipment.

[0021] Preferably, the software program system also includes an alarm and log module, which is used to monitor the system operation status in real time, including but not limited to the hardware system, calibration result abnormalities, and communication interruptions, and when an abnormal event occurs, an alarm message is immediately generated and an alarm signal is sent to the operation and maintenance personnel.

[0022] Preferably, the standard gas is used to store SO 2 、NO 2 , O 3 , CO, and provide a benchmark reference for the calibration of air monitoring instruments to determine the accuracy and deviation of air monitoring instrument measurements;

[0023] The zero gas generator is used to generate SO-free gas by physical or chemical methods. 2 、NO 2 , O 3 , CO, and is used for zero point calibration of air monitoring instruments to eliminate background interference and drift and ensure the accuracy of the measurement starting point.

[0024] Preferably, the implementation steps of the intelligent prediction module include:

[0025] S1.1: Extract historical calibration data from the data processing and storage module, obtain real-time environmental meteorological data and surrounding pollution source distribution data through the communication module, and receive real-time operation data of air monitoring instruments;

[0026] S1.2: Eliminate erroneous or missing data and normalize data;

[0027] S1.3: Calculate the change range within a specific period of time for environmental meteorological data;

[0028] Estimate the potential impact of pollution sources on monitoring instruments based on their distribution;

[0029] Extracting dispersion indicators such as recent measurements from air monitoring instrument operating data;

[0030] S1.4: Use long short-term memory network to build the model and divide the data set into training set, validation set and test set;

[0031] S1.5: Input the latest feature data into the trained model to predict the drift trend of the monitoring instrument;

[0032] S1.6: Based on the prediction results and the preset calibration accuracy threshold, the optimal calibration time point is determined and pushed to the calibration control module.

[0033] Preferably, in step S1.1, the historical calibration data includes but is not limited to calibration time, concentration, deviation, environmental parameters and air monitoring instrument parameters.

[0034] Preferably, the steps of the data processing and storage module when performing centralized data processing include but are not limited to cleaning, organizing and archiving the data.

[0035] Preferably, the self-learning optimization module includes the following implementation steps:

[0036] S2.1: Obtain calibration data and dynamic calibrator parameters from the data processing and storage module, collect environmental condition data and air monitoring instrument performance parameters, and store them in association with each other;

[0037] S2.2: Analyze the calibration data in depth, count the deviation index and calibration success rate of different concentration calibration points, classify and group the calibration data according to the environmental and air monitoring instrument parameters, analyze their characteristic changes, and construct machine learning models such as multivariate linear regression or decision trees with calibration data statistics, environmental and instrument parameters as inputs and calibration procedure key parameters and limit adjustment values ​​as outputs;

[0038] S2.3: Divide the historical calibration data into training sets and test sets, use the training set to train the model, aiming to minimize the mean square error, adjust the model parameters through algorithms such as gradient descent, and use the test set to verify the model accuracy and generalization ability. If the accuracy and generalization ability of the model are not up to standard, optimize the model by adding features, adjusting complexity or changing algorithms, and retrain and verify.

[0039] S2.4: After the calibration is completed, the actual data is input into the model to obtain the optimized recommended values ​​of the key parameters and limits of the calibration procedure, generate update instructions and send them to the calibration control module, and record the data. The model is regularly retrained and updated based on the new calibration data to cope with the dynamic changes in instrument performance and environment and ensure the stability of monitoring accuracy.

[0040] Preferably, in step S2.1, calibration data including but not limited to standard gas concentration, measurement value, calibration deviation, gas flow rate and mixing ratio setting information of the dynamic calibrator 4 are obtained from the data processing and storage module.

[0041] Compared with the prior art, the present invention provides an unmanned remote intelligent quality control system, which has the following beneficial effects:

[0042] 1. The unattended remote intelligent quality control system works together through the various intelligent modules in the software program system. For example, the calibration control module can automatically generate control instructions based on the results of the intelligent prediction module and the virtual calibration simulation module and send them to the PLC, realizing the automation process of the calibration process from data collection, analysis, decision-making to execution, greatly reducing the dependence on manual operation, reducing errors and instability caused by human factors, and improving the accuracy and reliability of calibration. There is no need for operation and maintenance personnel to frequently go to the monitoring site for on-site calibration operations. For example, the self-learning optimization module can automatically adjust the calibration parameters according to instrument performance and environmental changes, reducing unnecessary manual intervention and on-site maintenance times, thereby saving labor costs, transportation costs and time costs, and improving operation and maintenance efficiency.

[0043] 2. This unmanned remote intelligent quality control system can predict the drift trend and optimal calibration time of the monitoring instrument in advance through the intelligent prediction module combined with the rich data provided by the multi-source data fusion module, thereby achieving more accurate and timely calibration arrangements. Compared with traditional low-frequency calibration methods, it can respond quickly when the air quality changes, ensuring that the monitoring data can timely and accurately reflect the real situation of the environment, which helps environmental management departments to quickly grasp the pollution dynamics and take timely control measures, thereby improving the monitoring and response capabilities to environmental changes.

[0044] 3. This unmanned remote intelligent quality control system uses a virtual calibration simulation module to simulate the effects of different schemes and screen the best strategy before actual calibration. The self-learning optimization module intelligently optimizes the calibration program parameters and limits according to the calibration data and environmental conditions throughout the life cycle of the instrument. The combination of the two ensures that the calibration strategy is always in the optimal state, which can effectively respond to factors such as instrument aging and environmental changes, and continuously maintain stable and reliable monitoring accuracy, providing high-quality data support for environmental monitoring, which is conducive to more accurate assessment of air quality and pollution source conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the structure of the hardware system of the present invention;

[0046] Figure 2 It is a structural diagram of the software program system of the present invention. DETAILED DESCRIPTION

[0047] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0048] An unattended remote intelligent quality control system includes a hardware system, a software program system, and an air monitoring instrument for monitoring air quality. The hardware system includes a three-way solenoid valve and a zero gas generator connected to the three-way solenoid valve through a pipeline, a standard gas, and a dynamic calibrator. The standard gas is specifically a standard gas storage tank, which is used to store SO 2 、NO 2 , O 3 , CO, and provide a reference for the calibration of air monitoring instruments to determine the accuracy and deviation of air monitoring instrument measurements. Zero gas generators are used to generate air that does not contain SO by physical or chemical methods. 2 、NO 2 , O 3 , CO, and is used for zero-point calibration of air monitoring instruments to eliminate background interference and drift and ensure the accuracy of the measurement starting point. The outlet of the dynamic calibrator is connected to the inlet of the air monitoring instrument. The dynamic calibrator accurately adjusts the flow rate and mixing ratio of the standard gas and the zero gas according to the system settings, simulates the gaseous pollutant environment with a variety of concentration gradients, and fully calibrates the air monitoring instrument to ensure its accuracy and reliability under different measurement ranges and working conditions. At the same time, it can feedback parameter information such as gas flow and pressure during the calibration process;

[0049] In order to realize the control of the hardware system, a PLC electrically connected to the three-way solenoid valve and the dynamic calibrator is also included. The PLC is used to control the opening and closing of the three-way solenoid valve and transmit control instructions to the dynamic calibrator and receive feedback data from the dynamic calibrator. The software program system is connected to the PLC signal;

[0050] Furthermore, the software program system includes:

[0051] Data acquisition module, used to collect hardware system data, including but not limited to the data collected by air monitoring instruments and the working parameter data of dynamic calibrators;

[0052] The intelligent prediction module, based on big data analysis and deep learning algorithms, integrates historical calibration data, environmental meteorological data (temperature, humidity, air pressure, etc.), surrounding pollution source distribution data, and real-time operation data of air monitoring instruments to predict the drift trend and optimal calibration time of air monitoring instruments, plan calibration tasks in advance, avoid monitoring data deviation caused by air monitoring instrument drift, and ensure the continuity and accuracy of data quality;

[0053] The self-learning optimization module deeply analyzes the calibration data, environmental condition data and air monitoring instrument performance parameters during and after the calibration process, automatically adjusts the key parameters of the calibration procedure (including the gas mixing ratio of the dynamic calibrator, the calibration time interval, etc.) and the calibration limit, and intelligently optimizes the quality control standards according to the aging degree and performance changes of the air monitoring instrument to ensure stable and reliable monitoring accuracy throughout the life cycle of the air monitoring instrument;

[0054] The virtual calibration simulation module is used to build a virtual calibration environment and air monitoring instrument model using existing data before calibration, simulate the effects of different calibration schemes, evaluate and screen the best calibration strategy, and compare the actual and virtual results after calibration, verify and optimize the virtual model, improve the model prediction accuracy and reliability, and provide more accurate guidance for subsequent calibration;

[0055] The multi-source data fusion module is used to integrate heterogeneous data from multiple sources, such as air monitoring instruments, hardware system status sensors, environmental meteorological stations, surrounding pollution source monitoring equipment, and geographic information systems, and to unify the format, clean, pre-process, and analyze the association of the data, so as to explore the potential relationships and rules between the data, provide a comprehensive and accurate data foundation for each functional module of the system, and support intelligent decision-making and calibration operations;

[0056] It should be noted that the above-mentioned hardware system status sensor refers to real-time perception of various parameters of each hardware device in the hardware system, including but not limited to temperature, voltage, current, frequency, pressure, flow, etc., so as to timely understand whether the working status of the hardware device is normal, including but not limited to voltage sensors, current sensors, temperature sensors, etc.

[0057] The user interaction module is used to build a friendly human-computer interaction interface to visualize the overall operation status of the system, including the connection and working status of the hardware devices in the hardware system, the real-time curve and historical trend of the monitoring data, the progress and results of the calibration task, etc. The operation and maintenance personnel can conveniently set the system parameters, control the calibration task, view detailed reports and data records through this interface, and realize remote monitoring and management of the system;

[0058] The calibration control module generates precise control instructions and sends them to the PLC based on the calibration prediction information of the intelligent prediction module, the optimal calibration solution provided by the virtual calibration simulation module, and the calibration parameters set by the user through the user interaction and visualization module. The PLC controls the three-way solenoid valve, dynamic calibrator and other hardware devices to perform the calibration operation, and monitors the progress and status of the calibration process to ensure that the calibration procedure proceeds smoothly according to the predetermined process.

[0059] The data processing and storage module centrally processes, analyzes and stores the monitoring data, calibration data and hardware equipment status data from the data acquisition module, including data cleaning, organization and archiving, calculation of calibration error, uncertainty and other parameters, judging whether the calibration results are qualified according to preset rules, and storing the processed data in the database. At the same time, it provides data query and call services for other modules, and supports data backtracking analysis and report generation.

[0060] The software program system also includes a communication module, which is responsible for data transmission and communication connection between the software program system and PLC and other external devices (including servers, meteorological data acquisition terminals, pollution source monitoring equipment, etc.). It ensures stable and efficient data interaction, including hardware device status information and calibration data uploaded by PLC to the software program system, control instructions sent by the software program system to PLC, and transmission of meteorological data, pollution source data, etc. between other external devices.

[0061] The software program system also includes an alarm and log module, which monitors the system operation status in real time, including hardware equipment failure, abnormal calibration results, communication interruption, etc. When an abnormal event occurs, an alarm message is immediately generated, and an alarm signal is sent to the operation and maintenance personnel in a variety of ways (including sound and light alarms, SMS notifications, email notifications, etc.). At the same time, various event information during the system operation is recorded in detail, including operation logs, fault logs, calibration logs, etc., to provide a basis for system maintenance, troubleshooting and data analysis.

[0062] The implementation steps of the intelligent prediction module include:

[0063] Step 1: Obtain historical calibration data from the data processing and storage module, including the time of each calibration, calibration point concentration, measurement deviation, environmental meteorological data at the time, and related parameters of the monitoring instrument; receive the current temperature, humidity, air pressure and other meteorological data from the environmental meteorological data acquisition terminal in real time through the communication module; receive surrounding pollution source distribution data from external data sources (environmental protection departments or related monitoring networks), including the location of surrounding factories, types of pollutants emitted and approximate emissions; obtain current operating data from air monitoring instruments in real time, including the fluctuation range of measurement values, response time, etc.

[0064] Step 2: Clean all types of collected data and remove data points that are obviously wrong or missing. If a meteorological data collection point suddenly shows abnormally high or low values ​​that exceed the reasonable range, it will be corrected or marked as invalid data; standardize the data so that data from different sources and different magnitudes can be analyzed on the same scale, including temperature and humidity in environmental meteorological data, which can be normalized to a specific interval (between 0 and 1). For the distribution data of surrounding pollution sources, if it is the emission of different pollutants, it is normalized according to the weight of its impact on the monitoring instrument; the processed data is sorted according to the time series to construct a multidimensional time series data set, in which the data at each time point contains information in multiple dimensions such as calibration-related data, environmental data, pollution source data, and instrument operation data.

[0065] Step 3: Extract meaningful features from multidimensional time series data and calculate the rate of change of environmental meteorological data, including the average rate of change of temperature over the past period of time, the fluctuation range of humidity, etc. These features will affect the performance and drift of monitoring instruments; based on the distribution data of surrounding pollution sources, calculate the distance between the location of the air monitoring instrument and the main pollution source, the pollutant concentration gradient under the influence of wind direction and other features to evaluate the potential impact of pollution sources on the instrument; for the operating data of air monitoring instruments, extract statistical features such as the standard deviation and coefficient of variation of recent measurements to reflect the stability of instrument measurements; construct some time-based features, such as the time of the day, the day of the week, the season of the year, etc., because the operating environment and performance of air monitoring instruments may be different at different times; combine the extracted features into feature vectors, with each time point corresponding to a feature vector, for subsequent model training and prediction.

[0066] Step 4: Use the long short-term memory network (LSTM) in the deep learning algorithm to build a prediction model, divide the preprocessed and feature-engineered data set into a training set, a validation set, and a test set, and divide them according to the proportion (70% training set, 20% validation set, 10% test set); use the training set to train the LSTM model, and continuously adjust the weights and biases of the model through the back propagation algorithm during the training process to minimize the loss function between the predicted value and the actual value. The loss function can use the mean square error (MSE); during the training process, use the validation set to evaluate the model, monitor the model's loss value and accuracy on the validation set, and stop training when the performance of the model on the validation set no longer improves to prevent overfitting.

[0067] Step 5: Input the latest feature vector into the trained LSTM model to obtain the drift trend prediction result of the monitoring instrument; determine the optimal calibration time point based on the predicted drift trend and the set calibration accuracy requirements, and predict that the measurement deviation will exceed the acceptable threshold at a certain time point in the future. This time point is used as the optimal calibration time point, and the relevant information is sent to the calibration control module to arrange the calibration task in advance.

[0068] The self-learning optimization module includes the following implementation steps:

[0069] Step 1: The self-learning optimization module first obtains calibration-related core data from the data processing and storage module, including calibration data such as standard gas concentration, measurement value, and calibration deviation, as well as the gas flow and mixing ratio setting parameters of the dynamic calibrator. At the same time, it collects environmental condition data, including temperature, humidity, air pressure, etc., as well as air monitoring instrument performance parameters, including but not limited to response time, usage time, etc. These different sources and different types of data are associated and integrated according to specific logic and identification, laying the foundation for subsequent in-depth analysis, so that the data can form an organic whole to explore the hidden laws and connections.

[0070] Step 2: After the data integration is completed, conduct a multi-dimensional statistical analysis on the calibration data. Calculate the deviation indicators of calibration points of different concentrations, where the average deviation intuitively presents the average deviation between the measured value and the true value, and the standard deviation accurately describes the discrete characteristics of the deviation. At the same time, the calibration success rate is clarified. Furthermore, the calibration data is finely classified and grouped according to the environmental condition data and instrument performance parameters. In different temperature ranges, humidity intervals, and instrument usage periods, the unique characteristics and change trends of the calibration data are deeply explored, so as to accurately understand the specific impact mechanism of various factors on the calibration results, and provide strong data feature support for building accurate models.

[0071] Step 3: Based on the previously extracted data features, a machine learning model is constructed. A multivariate linear regression model or a decision tree model can be used. The calibration data statistical indicators and environmental and air monitoring instrument parameters are used as input variables. The key parameters of the calibration procedure (dynamic calibrator gas mixing ratio adjustment) and the calibration limit adjustment value are set as output targets. The historical calibration data is divided into a training set and a test set at a ratio of 70% and 30%. The model is trained using the training set data, guided by error functions such as mean square error, and the model parameters are continuously optimized with the help of algorithms such as gradient descent, so that the model prediction value is as close as possible to the actual historical adjustment value. After the training is completed, the accuracy and generalization ability of the model are tested using the test set, and the model performance is evaluated based on the verification results. If the model does not meet the expected standards, the model is repeatedly adjusted and optimized by increasing the feature dimension, adjusting the model complexity, or replacing the model algorithm, until the model performance meets the requirements and can accurately predict the calibration parameter adjustment value.

[0072] Step 4: After each calibration process is completed, the actual data involved in this calibration, including calibration data, environmental condition data, and air monitoring instrument performance parameters, are input into the fully trained model. After calculation, the model generates optimization recommendations for key parameters and calibration limits of the calibration procedure. The slut model determines that the current environment is high temperature and the instrument has been used for a long time. The specific standard gas flow ratio in the dynamic calibrator is appropriately increased and the calibration limit is reasonably relaxed to improve the calibration accuracy. The calibration procedure update instruction is generated based on these optimization recommendations and passed to the calibration control module so that the updated parameter settings can be enabled during the next calibration operation. At the same time, the data and operation process related to this optimization are recorded in detail in the log library to provide a detailed basis for subsequent traceability analysis. With the continuous advancement of calibration practice, new data continues to accumulate. The module will re-divide the training set and test set according to the preset cycle (specific time interval or calibration number threshold), and conduct the model training and verification process again, and continuously update the model parameters to ensure that the model can closely fit the dynamic changes of instrument performance and environmental conditions, and always provide accurate and effective parameter adjustment strategies for calibration work, so as to ensure that the monitoring instrument maintains stable and reliable monitoring accuracy throughout its life cycle.

[0073] The implementation steps of the virtual calibration simulation module are as follows:

[0074] Step 1: Obtain historical calibration data from the data processing and storage module, including the standard gas concentration settings under different calibration schemes, the flow and mixing ratio parameters of the dynamic calibrator, the corresponding measurement results of the air monitoring instrument, and the environmental conditions data at the time (including temperature, humidity, air pressure, etc.). Collect detailed technical specifications and performance parameters of the air monitoring instrument, including the measurement principle, sensor characteristics, response curve, range, accuracy index and other information of the air monitoring instrument. These data will help build a more accurate instrument model. Obtain real-time data of the current environment and relevant information about the upcoming calibration, such as the preparation of the standard gas, environmental prediction data at the expected calibration time, etc., to provide a basis for building a virtual calibration environment.

[0075] Step 2: Based on the technical principles and performance parameters of the air monitoring instrument, a virtual monitoring instrument model is constructed using a mathematical model or a physical model. If the air monitoring instrument is based on an electrochemical sensor, a mathematical equation model describing the relationship between the electrochemical reaction process of the sensor and the measurement signal is established; for instruments based on optical principles, a model of the relationship between light absorption, scattering and concentration measurement is constructed. In the model, the nonlinear characteristics of the instrument, the influence of environmental factors such as temperature and humidity on the measurement are considered, and the parameters in the model are determined through historical data and technical information. A virtual calibration environment model is constructed based on environmental data and calibration equipment information. The physical properties of the standard gas under different environmental conditions are simulated (including gas diffusion rate, the influence of humidity on gas concentration, etc.), as well as the gas output characteristics of the dynamic calibrator under different settings (including flow stability, mixing uniformity, etc.). The model should be able to generate a virtual calibration gas environment based on the input calibration scheme parameters (including standard gas concentration sequence, flow setting, etc.).

[0076] Step 3: Based on the constructed virtual calibration environment and air monitoring instrument model, design a variety of calibration schemes for simulation. The calibration scheme includes different standard gas concentration combinations, different gas flow and mixing ratio settings, different calibration time intervals, and different environmental condition simulations. Set the standard gas concentration from low to high, combined with different ambient temperature and humidity conditions, and simulate the measurement process and calibration results of the monitoring instrument under these schemes. For each simulated calibration scheme, calculate the evaluation indicators, including calibration error (deviation between the measured value and the standard value), calibration accuracy (repeatability of multiple measurements), uncertainty (measurement uncertainty caused by environmental factors and fluctuations in the characteristics of air monitoring instruments), etc. These indicators are used to evaluate the advantages and disadvantages of different calibration schemes and determine the calibration strategy that performs best under different conditions. This includes which standard gas concentration sequence and flow setting can minimize the calibration error and achieve the highest accuracy under specific ambient temperature and humidity.

[0077] Step 4: When performing the actual calibration operation, follow the selected calibration scheme and record the actual calibration data, including the measured values ​​of the monitoring instrument, the actual parameter settings of the dynamic calibrator, the actual measured values ​​of the environmental conditions, etc. Compare and analyze the actual calibration results with the results of the previous virtual simulation. Calculate the difference index between the actual results and the virtual results. If the difference is large, analyze the reasons, including actual factors not taken into account by the virtual model (sudden instrument failure, environmental interference sources not included in the model, etc.), inaccurate model parameters, etc.

[0078] Step 5: Based on the comparative analysis of the actual calibration and virtual results, optimize and adjust the virtual calibration environment model and the air monitoring instrument model. If the difference is caused by inaccurate model parameters, use the actual calibration data and environmental data to re-estimate the model parameters, and adjust the response curve parameters or environmental impact coefficients in the instrument model by least squares fitting and other methods. If the model structure is imperfect and does not take into account a certain special instrument failure mode or environmental factor influencing mechanism, improve and expand the model structure. Add fault diagnosis modules or more complex environmental interaction terms to the instrument model to improve the model's ability to simulate actual conditions. Apply the optimized model to the simulation and evaluation of the next calibration plan, and continuously repeat the above process. As the actual calibration data accumulates, the prediction accuracy and reliability of the virtual model are gradually improved, providing more and more accurate guidance for subsequent calibration operations.

[0079] Working principle:

[0080] When the system is in use, the PLC performs self-inspection on hardware devices such as three-way solenoid valves and dynamic calibrators, checks the connection status and functional integrity of the equipment, and feeds back the self-inspection results to the user interaction module of the software program system through the communication link for display. At the same time, the multi-source data fusion module is started to collect and integrate initial environmental meteorological data, air monitoring instrument status data and other multi-source information, store them in the data processing and storage module, and provide basic data for other software modules. During the data collection and analysis stage, the air monitoring instrument continuously collects gaseous pollutant concentration data in the ambient air samples. The data collection module obtains and transmits it to the data processing and storage module in real time. The multi-source data fusion module simultaneously collects environmental meteorological data, surrounding pollution source data and other information. The intelligent prediction module uses the deep learning algorithm to predict the drift trend and calibration requirement time point of the monitoring instrument based on the fused data and the hardware equipment status data. When the intelligent prediction module predicts the calibration requirement or reaches the preset calibration time interval, it sends a calibration start instruction to the calibration control module. Before starting the calibration, the calibration control module notifies the virtual calibration simulation module to build a virtual calibration environment and instrument model, simulate different calibration schemes and determine the best scheme. The virtual calibration simulation module performs simulation calculations based on the data provided by the multi-source data fusion module and the historical calibration data in the data processing and storage module, and feeds back the best calibration scheme to the calibration control module. According to the solution provided by the virtual calibration simulation module, the three-way solenoid valve is controlled by PLC to switch the gas path, and the zero gas or standard gas is introduced into the dynamic calibrator. The dynamic calibrator accurately adjusts the gas flow and mixing ratio according to the set parameters, generates the mixed gas for calibration and transmits it to the air monitoring instrument. During the entire calibration process, the data acquisition module continuously collects the measurement data of the air monitoring instrument and the working parameter data of the dynamic calibrator, and transmits them to the data processing and storage module for storage and analysis. The data processing and storage module processes and analyzes the calibration data, calculates the measurement deviation, uncertainty and other parameters, and compares the calibration results with the preset calibration limits. The self-learning optimization module calculates the calibration results, environmental conditions and instrument performance according to the calibration results. It can automatically adjust the key parameters of the calibration program and the calibration limits, and feed back the optimization results to the calibration control module for application in subsequent calibration. At the same time, the virtual calibration simulation module compares the actual calibration results with the virtual simulation results, verifies and optimizes the virtual model, and improves its prediction accuracy. The user interaction and visualization module displays the calibration results, system operation status, hardware equipment status and other information to the operation and maintenance personnel in the form of intuitive charts and reports. The operation and maintenance personnel can view detailed data, generate calibration reports, set system parameters, and start or pause calibration tasks through this interface. At the same time, the alarm and log module records important events, fault information and calibration operation records during the system operation for subsequent tracing and analysis.

[0081] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. A remote intelligent quality control system based on unattended operation, characterized in that: Includes hardware systems, software program systems, and air monitoring instruments for monitoring air quality; The hardware system includes a three-way solenoid valve and a zero gas generator, a standard gas, and a dynamic calibrator connected to the three-way solenoid valve through a pipeline, wherein the gas outlet of the dynamic calibrator is connected to the gas inlet of the air monitoring instrument; It also includes a PLC electrically connected to the three-way solenoid valve and the dynamic calibrator, wherein the PLC is used to control the opening and closing of the three-way solenoid valve, transmit control instructions to the dynamic calibrator, and receive feedback data from the dynamic calibrator; The software program system is connected to the PLC signal, and the software program system includes: Data acquisition module, used to collect hardware system data, including but not limited to the data collected by air monitoring instruments and the working parameter data of dynamic calibrators; Intelligent prediction module, based on big data analysis and deep learning algorithms, is used to integrate historical calibration data, environmental condition data, and real-time operation data of air monitoring instruments to predict the drift trend and optimal calibration time of air monitoring instruments; Self-learning optimization module, which is used to analyze calibration data, environmental condition data and air monitoring instrument performance parameters during and after the calibration process, and automatically adjust the calibration procedure parameters and calibration limits; The virtual calibration simulation module is used to build a virtual calibration environment and air monitoring instrument model using existing data before the calibration operation, simulate the effects of different calibration schemes, and evaluate and select the best calibration strategy; Multi-source data fusion module, used to integrate multi-source heterogeneous data, including but not limited to data from air monitoring instruments, operating status of hardware systems, and environmental condition data; User interaction module, used to build a human-computer interaction interface, visualize the overall operation status of the system, and receive operation instructions from operation and maintenance personnel; The calibration control module is used to generate accurate control instructions and send them to the PLC based on the calibration prediction information of the intelligent prediction module, the best calibration solution provided by the virtual calibration simulation module, and the calibration parameters set by the user through the user interaction module.

2. The unmanned remote intelligent quality control system according to claim 1, characterized in that: The software program system further comprises a communication module, and the software program system is connected to the PLC signal via the communication module; The communication module is also used for data transmission and communication connection with external devices, and the external devices are used to collect environmental condition data, which include but are not limited to environmental meteorological data and surrounding pollution source distribution data. The external devices are but not limited to servers, meteorological data collection terminals, and pollution source monitoring equipment.

3. The unmanned remote intelligent quality control system according to claim 2 is characterized in that: The software program system also includes a data processing and storage module, which is used to centrally process, analyze and store the data collected by the data collection module.

4. The unmanned remote intelligent quality control system according to claim 3 is characterized in that: The software program system also includes an alarm and log module, which is used to monitor the system operation status in real time, including but not limited to the hardware system, calibration result abnormalities, and communication interruptions, and when an abnormal event occurs, an alarm message is immediately generated and an alarm signal is sent to the operation and maintenance personnel.

5. The unmanned remote intelligent quality control system according to claim 1 is characterized in that: The standard gas is used to store SO2, NO2, O3, CO, and provide a reference for the calibration of air monitoring instruments to determine the accuracy and deviation of air monitoring instrument measurements; The zero gas generator is used to generate pure gas without SO2, NO2, O3, and CO by physical or chemical methods, and is used for zero point calibration of air monitoring instruments to eliminate background interference and drift and ensure the accuracy of the measurement starting point.

6. The unmanned remote intelligent quality control system according to claim 3 is characterized in that: The implementation steps of the intelligent prediction module include: S1.1: Extract historical calibration data from the data processing and storage module, obtain real-time environmental meteorological data and surrounding pollution source distribution data through the communication module, and receive real-time operation data of air monitoring instruments; S1.2: Eliminate erroneous or missing data and normalize data; S1.3: Calculate the change range within a specific period of time for environmental meteorological data; Estimate the potential impact of pollution sources on monitoring instruments based on their distribution; Extracting dispersion indicators such as recent measurements from air monitoring instrument operating data; S1.4: Use long short-term memory network to build the model and divide the data set into training set, validation set and test set; S1.5: Input the latest feature data into the trained model to predict the drift trend of the monitoring instrument; S1.6: Based on the prediction results and the preset calibration accuracy threshold, the optimal calibration time point is determined and pushed to the calibration control module.

7. The unmanned remote intelligent quality control system according to claim 6 is characterized in that: In step S1.1, the historical calibration data includes but is not limited to calibration time, concentration, deviation, environmental parameters and air monitoring instrument parameters.

8. The unmanned remote intelligent quality control system according to claim 3 is characterized in that: The steps of the data processing and storage module for centralized data processing include but are not limited to cleaning, organizing and archiving the data.

9. The unmanned remote intelligent quality control system according to claim 3 is characterized in that: The self-learning optimization module includes the following implementation steps: S2.1: Obtain calibration data and dynamic calibrator parameters from the data processing and storage module, collect environmental condition data and air monitoring instrument performance parameters, and store them in association with each other; S2.2: Analyze the calibration data in depth, count the deviation index and calibration success rate of different concentration calibration points, classify and group the calibration data according to the parameters of the environment and air monitoring instruments, and analyze their characteristic changes. Using calibration data statistics, environmental and instrument parameters as inputs and calibration procedure key parameters and limit adjustment values ​​as outputs, machine learning models such as multivariate linear regression or decision trees are constructed; S2.3: Divide the historical calibration data into training sets and test sets, use the training set to train the model, aiming to minimize the mean square error, adjust the model parameters through algorithms such as gradient descent, and use the test set to verify the model accuracy and generalization ability. If the accuracy and generalization ability of the model are not up to standard, optimize the model by adding features, adjusting complexity or changing algorithms, and retrain and verify. S2.4: After the calibration is completed, the actual data is input into the model to obtain the optimized recommended values ​​of the key parameters and limits of the calibration procedure, generate update instructions and send them to the calibration control module, and record the data. The model is regularly retrained and updated based on the new calibration data to cope with the dynamic changes in instrument performance and environment and ensure the stability of monitoring accuracy.

10. The unmanned remote intelligent quality control system according to claim 9, characterized in that: In step S2.1, the calibration data obtained from the data processing and storage module includes but is not limited to standard gas concentration, measurement value, calibration deviation, gas flow rate of the dynamic calibrator and mixing ratio setting information.

Images