Operation and Maintenance Management Method and System for an Internet of Things-Based Air Quality Monitoring Platform

Abstract

The present invention discloses an operation and maintenance management method and system for an air quality monitoring platform based on the Internet of Things, which relates to the technical field of air quality monitoring platform management; the present invention corrects data drift in real time through an environmental parameter compensation algorithm to improve monitoring accuracy; uses a carp population optimization algorithm to obtain an improved air quality prediction network to optimize air quality prediction performance; constructs a hybrid network of LoRaWAN and NB-IoT, dynamically adjusts the acquisition frequency according to the risk level, and reduces the communication load; combines the Internet of Things health score model with the Cox proportional hazards model to predict equipment failures, plans the optimal maintenance path through the TSP algorithm, and improves the operation and maintenance efficiency; dynamically optimizes equipment deployment based on the Gaussian process regression algorithm; the present invention realizes the full-chain closed-loop management of "perception-prediction-transmission-maintenance-optimization", significantly improves data accuracy, communication efficiency and equipment reliability, and provides a low-cost and adaptive solution for the air quality monitoring platform.

Description

Technical Field

[0001] The present invention belongs to the technical field of air quality monitoring platform management. Specifically, it particularly relates to a method and system for operation and maintenance management of an air quality monitoring platform based on the Internet of Things. Background Art

[0002] With the acceleration of the global urbanization process, air quality monitoring has become a core link in environmental protection, public health, and policy-making. Traditional monitoring relies on fixed stations, which have limitations such as coverage blind spots, high costs, and insufficient real-time performance. Although the development of Internet of Things technology has improved the deployment flexibility of monitoring devices, it faces challenges such as sensor data drift, low communication efficiency, complex device maintenance, and insufficient dynamic deployment capabilities. Electrochemical, semiconductor, and other sensors in traditional devices are easily interfered by environmental temperature and humidity, resulting in data distortion. There is a lack of an effective real-time compensation mechanism, which affects the monitoring accuracy.

[0003] Traditional time series prediction models rely on a fixed learning rate and are prone to falling into local optima, resulting in large air quality prediction errors and lagging responses. A single communication protocol cannot adapt to the differentiated network requirements of urban and suburban areas. High-density areas are prone to congestion, and signal coverage in low-density areas is weak. There is a lack of quantitative evaluation of device health status and fault prediction capabilities, relying on manual inspections, with slow repair responses and high costs. The device deployment is fixed and it is difficult to dynamically adjust the position according to pollution diffusion and meteorological changes, resulting in monitoring blind spots or redundancy, affecting data representativeness and cost-effectiveness.

[0004] There are problems in the prior art such as serious sensor data drift, insufficient efficiency of prediction models, waste and uneven coverage of communication resources, passive and inefficient device maintenance, and rigid monitoring networks. Summary of the Invention

[0005] (I) Technical Problems to be Solved

[0006] In view of the problems in the related art, the present invention provides a method for operation and maintenance management of an air quality monitoring platform based on the Internet of Things to overcome the above-mentioned technical problems existing in the prior related art.

[0007] (II) Technical Solutions

[0008] To solve the above technical problems, the present invention is implemented through the following technical solutions:

[0009] S1. Calibrate the air quality data set of the monitoring area collected by the Internet of Things device set to obtain a real-time air quality data set;

[0010] S2. Input the real-time air quality data set into an improved air quality prediction network to obtain a predicted air quality data set;

[0011] S3. Perform risk level judgment based on the real-time air quality data set and the predicted air quality data set to obtain a judgment result;

[0012] According to the judgment result, adjust the classification collection, processing, and transmission rules of the multi-source data of the Internet of Things device set for the monitoring area to obtain an optimized data packet;

[0013] S4. Respectively, through the Internet of Things device health score model and the Cox proportional hazards model combined with the optimized data packet, obtain the real-time Internet of Things device health diversity set and the predicted device failure probability set;

[0014] S5. Generate a maintenance work order according to the real-time Internet of Things device health diversity set combined with the predicted device failure probability set;

[0015] Combine the maintenance work order with the location of the maintenance personnel to generate an optimal maintenance path for the maintenance of Internet of Things devices;

[0016] S6. Use the Gaussian process regression algorithm combined with the Internet of Things device set to obtain an optimal Internet of Things device location set for the redeployment of Internet of Things devices;

[0017] Through the multi-technology integration of sensor calibration, intelligent prediction, hybrid communication, health management, and dynamic deployment, the present invention constructs a complete "perception - prediction - transmission - maintenance - optimization" closed-loop system, solves the problems of data distortion, low efficiency, and passive maintenance in the traditional monitoring system, and provides a high-precision, low-cost, and adaptive solution for the operation and maintenance management of the air quality monitoring platform.

[0018] Preferably, the S1 includes the following steps:

[0019] S11. Collect key location points in the monitoring area to obtain a key location point set;

[0020] Deploy devices integrating three types of sensors, namely laser scattering, electrochemistry, and semiconductor, at the key location points in the key location point set to obtain an Internet of Things device point set;

[0021] S12. Collect air quality data of each monitoring area through the Internet of Things device point set to obtain an air quality data set;

[0022] S13. Collect environmental data of each area to obtain an environmental data set; the environmental data includes environmental temperature and environmental humidity;

[0023] Use the environmental parameter compensation algorithm to calculate the drift compensation amount caused by the environmental data to the air quality data in the air quality data set to obtain a drift compensation amount data set;

[0024] According to the drift compensation amount data set, calibrate the air quality data in the air quality data set to obtain a real-time air quality data set;

[0025] The above steps achieve precise data collection and calibration through multi-stage collaboration; first, Internet of Things devices integrating laser scattering, electrochemical, and semiconductor sensors are deployed at key positions in the monitoring area to form a multi-source data collection network; subsequently, combined with environmental temperature and humidity data, a compensation algorithm is used to calculate the sensor drift amount, dynamically calibrate the original air quality data, eliminate environmental interference, and generate a high-precision real-time data set, providing a reliable input for subsequent analysis.

[0026] Preferably, S2 includes the following steps:

[0027] S21. Construct an LSTM network; use historical air quality data to train the LSTM network. During the training process, an optimization algorithm is combined to find the optimal learning rate of the LSTM network to obtain the optimal solution; use the optimal solution as the learning rate of the LSTM network to obtain an improved air quality prediction network;

[0028] S22. Input the air quality data set into the improved air quality prediction network to obtain a predicted air quality data set;

[0029] The above steps improve the prediction ability by improving the LSTM network; an optimization algorithm is used to dynamically adjust the network learning rate to avoid the training falling into a local optimum; real-time calibrated air quality data is input, and a high-precision prediction result is output to achieve a rapid response and early warning of future pollution trends.

[0030] Preferably, the step of finding the optimal learning rate of the LSTM network by combining an optimization algorithm during the training process in S21 includes the following steps:

[0031] S211. Set the prediction accuracy threshold of the LSTM network as z 1 and the prediction accuracy as z 2; set the maximum number of optimization iterations and the survival space size of the carp population;

[0032] S212. Construct an initial carp population set, and set the scale of the initial carp population set according to the learning rate of the LSTM network;

[0033] The survival space size of each carp population in the initial carp population set is fixed, so the number of carp populations will affect the size of the final adult carp. Too many or too few carp populations will reduce the overall benefit;

[0034] S213. Define the fitness function of the carp population number according to the prediction accuracy threshold z 1 and the prediction accuracy z 2;

[0035] S214. Perform iterative operations on the initial carp population set. The higher the fitness value, the better the quantity of the carp population. During each iteration, calculate the fitness value of the carp population quantity according to the fitness function, update the carp population quantity in descending order of the fitness value, and obtain the best carp population quantity and the global best carp population quantity in the carp population during each iteration.

[0036] S215. Repeat S214. When the maximum optimization iteration number is reached, stop the iteration and use the global best carp quantity as the optimal solution.

[0037] The above steps dynamically optimize the LSTM learning rate through the carp population optimization algorithm. Based on the prediction accuracy threshold and the real-time accuracy, construct a fitness function, and iteratively screen the carp population quantity to balance the population density with the survival space constraint, and finally obtain the global optimal learning rate parameter, avoiding the model training falling into the local optimum and improving the convergence speed and accuracy of air quality prediction.

[0038] Preferably, the said S3 includes the following steps:

[0039] S31. Construct a hybrid network of LoRaWAN and NB-IoT. In the hybrid network, NB-IoT is used to transmit data in the dense urban areas of the monitoring area, and LoRaWAN is used to transmit data in the suburban areas of the monitoring area. The determination conditions of the dense urban areas and the suburban areas are determined according to the device density. Design a dual-mode communication module to support the dynamic switching of the communication protocols of LoRaWAN and NB-IoT.

[0040] Design a dual-mode communication module to support the dynamic switching of the communication protocols of LoRaWAN and NB-IoT.

[0041] S32. Set a risk level threshold set. The risk level threshold is divided according to the air quality. Obtain a risk level data set according to the real-time air quality data set, the predicted air quality data set and the risk level threshold set. The risk level data set contains the risk levels of the real-time air quality and the predicted air quality in each area.

[0042] Obtain a risk level data set according to the real-time air quality data set, the predicted air quality data set and the risk level threshold set. The risk level data set contains the risk levels of the real-time air quality and the predicted air quality in each area.

[0043] S33. Set the corresponding air quality data collection frequency according to the risk level threshold set to obtain a collection frequency set. Collect, correct, compress and add spatio-temporal tags to the data in each area of the risk level data set according to the collection frequency set to obtain an optimized data packet.

[0044] Collect, correct, compress and add spatio-temporal tags to the data in each area of the risk level data set according to the collection frequency set to obtain an optimized data packet.

[0045] The optimized data packet contains the air quality data, the environmental data and the IoT device status data in each monitoring area.

[0046] Transmit the optimized data packets through the hybrid networking of LoRaWAN and NB-IoT;

[0047] The above steps optimize the efficiency of data collection and transmission through hybrid communication and dynamic regulation; divide dense urban areas and suburbs based on device density to construct a dual-mode dynamic switching network; divide risk levels according to air quality and match differential collection frequencies, perform real-time correction, compression and add spatio-temporal tags to the data, generate optimized data packets containing air quality, environmental parameters and device status, and transmit them on demand through hybrid networking to reduce network load and ensure the priority of data in key areas.

[0048] Preferably, S4 includes the following steps:

[0049] S41. Collect historical Internet of Things device maintenance record data; based on the historical Internet of Things device maintenance record data, construct an Internet of Things device health score model;

[0050] S42. Extract the signal-to-noise ratio, mean time between failures and calibration score of each Internet of Things device in the real-time Internet of Things device set from the optimized data packet to obtain real-time Internet of Things device data;

[0051] Input the real-time Internet of Things device data into the Internet of Things device health score model to obtain a real-time Internet of Things device health diversity set;

[0052] S43. Extract the Internet of Things device status data set and the environmental data set from the optimized data packet; extract the features of the Internet of Things device status data set and the environmental data set to obtain an Internet of Things device status feature matrix;

[0053] S44. Input each Internet of Things device status feature in the Internet of Things device status feature matrix into the Cox proportional hazards model, calculate the failure probability of each Internet of Things device to obtain a predicted device failure probability set;

[0054] The above steps achieve proactive operation and maintenance through device health assessment and fault prediction; construct a health score model based on historical maintenance data to calculate the device health status in real time; combine the device status feature matrix and the Cox model to predict the failure probability, provide a quantitative basis for maintenance decisions, and reduce the risk of device downtime.

[0055] Preferably, S5 includes the following steps:

[0056] S51. Set the device health score threshold and the failure probability threshold; detect the real-time device health diversity set and the predicted device failure probability set, mark the Internet of Things devices with real-time device health score ≥ device health score threshold or predicted device failure probability ≥ failure probability threshold as devices to be repaired, label the priority of the devices to be repaired to obtain a maintenance work order;

[0057] S52. Set the order assignment threshold; collect the locations of maintenance personnel to obtain the set of maintenance personnel locations;

[0058] Take the priority of the maintenance equipment and the order assignment threshold as constraint conditions, and use the TSP algorithm to match the set of maintenance personnel locations with the locations of the IoT devices in the maintenance work orders to obtain the optimal maintenance path;

[0059] Travel to the location of the equipment to be repaired along the optimal maintenance path;

[0060] The above steps improve the operation and maintenance efficiency through intelligent scheduling; screen high-risk devices based on the health score and the failure probability threshold to generate priority work orders, use the TSP algorithm to plan the optimal path in combination with the locations of maintenance personnel, realize the dynamic optimization of the maintenance order and route of multiple devices, and reduce the response time and labor costs.

[0061] Preferably, the S6 includes the following steps:

[0062] S61. Extract the pollution diffusion data and meteorological parameters in the optimization data packet, and combine the terrain elevation data, predicted air quality data, and real air quality data to obtain the data affecting the devices;

[0063] S62. Collect the location data of each IoT device in the IoT device set to obtain the real-time IoT device location set;

[0064] S63. Through the Gaussian process regression algorithm, combine the IoT device location set and the data affecting the devices to obtain the optimal IoT device location set;

[0065] S64. Use drones to redeploy the IoT devices according to the locations in the optimal IoT device location set;

[0066] The above steps improve the monitoring network efficiency through dynamic optimization; based on the pollution diffusion model, meteorological parameters, and terrain elevation data, combine the difference between predicted and measured air quality, and use the Gaussian process regression algorithm to obtain the optimal IoT device location set; use the optimal IoT device location set and perform adaptive redeployment by drones to eliminate monitoring blind spots and reduce operation and maintenance costs.

[0067] Preferably, the S63 includes the following steps:

[0068] S631. Set the threshold of the movement range of the IoT device; set the constraint conditions according to the communication limit, terrain limit, and the threshold of the movement range of the IoT device;

[0069] S632. Set the objective function to minimize the monitoring deviation and minimize the network deployment cost;

[0070] S633. Use the Gaussian process regression algorithm to combine the set of Internet of Things device locations, constraints, and objective function to obtain the optimal set of Internet of Things device locations;

[0071] The above steps optimize the device layout through Gaussian process regression; set communication and terrain constraints, construct a bi-objective function of monitoring deviation and deployment cost, and combine pollution diffusion and device location data to solve for the optimal set of Internet of Things device locations.

[0072] An operation and maintenance management system for an Internet of Things-based air quality monitoring platform, used to implement the above-mentioned operation and maintenance management method for an Internet of Things-based air quality monitoring platform, including a data collection and correction module, an air quality prediction module, a data collection and transmission optimization module, a device health management module, an intelligent maintenance scheduling module, and a dynamic deployment optimization module;

[0073] The data collection and correction module is used to collect air quality data through Internet of Things devices equipped with multiple types of sensors, and perform data correction in combination with an environmental parameter compensation algorithm; by real-time monitoring of environmental temperature and humidity, calculate the sensor data drift compensation amount using a preset formula, and generate a high-precision real-time air quality data set to ensure the accuracy of subsequent analysis;

[0074] The air quality prediction module constructs an air quality prediction model based on an improved LSTM network, and dynamically adjusts the network learning rate in combination with the carp population optimization algorithm. Train and optimize the model parameters through historical data to achieve accurate prediction of future air quality, and output a prediction data set to support risk warning and decision-making;

[0075] The data collection and transmission optimization module is used to design a hybrid network of LoRaWAN and NB-IoT, dynamically switch communication protocols according to device density, and improve data transmission efficiency; adjust the data collection frequency through risk level division, add spatio-temporal tags and compress to generate optimized data packets, and achieve on-demand classification transmission to reduce network load;

[0076] The device health management module is based on a comprehensive Internet of Things device health sub-model and a Cox proportional hazards model to evaluate the real-time health status of the device and predict the failure probability; calculate the health score through parameters such as signal-to-noise ratio and mean time between failures, and analyze the failure risk in combination with the device status feature matrix to provide a quantitative basis for maintaining Internet of Things devices;

[0077] The intelligent maintenance scheduling module is used to generate priority maintenance work orders according to the device health score threshold and failure probability, integrate the location data of maintenance personnel, and use the TSP algorithm to plan the optimal path; through dynamic dispatching and path matching, improve the maintenance response efficiency and reduce the device downtime;

[0078] The dynamic deployment optimization module adopts the Gaussian process regression algorithm, combines pollution diffusion, meteorological and terrain data to optimize the layout of Internet of Things devices; sets a dual-objective function of monitoring deviation and deployment cost, and implements device redeployment through drones to achieve adaptive adjustment of the monitoring network and efficient utilization of resources.

[0079] (III) Beneficial effects

[0080] The present invention has the following beneficial effects:

[0081] Through the multi-technology integration of sensor calibration, intelligent prediction, hybrid communication, health management, and dynamic deployment, the present invention constructs a complete "perception - prediction - transmission - maintenance - optimization" closed-loop system, solves problems such as data distortion, low efficiency, and passive maintenance in traditional monitoring systems, and provides a high-precision, low-cost, and adaptive solution for the operation and maintenance management of air quality monitoring platforms.

[0082] The present invention improves data accuracy and reliability; corrects sensor drift in real time through the environmental parameter compensation algorithm, combines multi-source sensor fusion, significantly reduces the interference of temperature and humidity on sensitive components such as electrochemistry, ensures the accuracy of air quality data, and provides a credible basis for pollution control; optimizes the performance of the prediction model; adopts the carp population optimization algorithm to dynamically adjust the learning rate of the LSTM network, and

[0083] selects the optimal parameters through the fitness function, avoids the model training falling into local optimum, improves the prediction accuracy and convergence speed, and realizes the efficient early warning of future air quality.

[0084] The present invention enhances communication efficiency and adaptability; by designing a hybrid network of LoRaWAN and NB-IoT, dynamically switches protocols according to device density, combines risk levels to adjust data collection frequency, reduces network load, ensures the transmission priority of key data, and reduces communication costs.

[0085] The present invention realizes intelligent operation and maintenance of devices. It quantifies the device status through the Internet of Things health sub-model, combines the Cox proportional hazards model to predict the failure probability, actively identifies high-risk devices and generates repair work orders; uses the TSP algorithm to plan the optimal path, reduces the response time and movement cost of maintenance personnel, and improves the operation and maintenance efficiency.

[0086] The present invention realizes the dynamic optimization of the monitoring network layout; based on the Gaussian process regression algorithm, balances the monitoring deviation and deployment cost with the objective function, combines data such as pollution diffusion and terrain elevation, and dynamically redeploys devices through drones to eliminate monitoring blind spots, improve the flexibility of network coverage and resource utilization rate, and reduce long-term operation and maintenance expenses.

[0087] Of course, it is not necessary for any product implementing the present invention to achieve all the above-mentioned advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0089] Figure 1 It is a schematic flowchart of the operation and maintenance management method for the air quality monitoring platform based on the Internet of Things of the present invention;

[0090] Figure 2 It is a schematic diagram of the modules of the operation and maintenance management system for the air quality monitoring platform based on the Internet of Things of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0091] The technical solutions in the embodiments of the invention will be clearly and completely described below with reference to the drawings in the embodiments of the invention. Obviously, the described embodiments are only a part of the embodiments of the invention, rather than all the embodiments. Based on the embodiments of the invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the invention.

[0092] In the description of the present invention, it should be understood that the terms "opening", "upper", "lower", "top", "middle", "inner", etc. indicating the orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, rather than indicating or implying that the components or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be construed as a limitation of the invention.

[0093] Embodiment 1:

[0094] Please refer to Figure 1 , the present invention discloses an operation and maintenance management method for an air quality monitoring platform based on the Internet of Things, including the following steps:

[0095] S1. Calibrate the air quality data set of the monitoring area collected through the Internet of Things device set to obtain a real-time air quality data set;

[0096] The S1 includes the following steps:

[0097] S11. Collect key location points in the monitoring area to obtain a key location point set; deploy devices integrating three types of sensors, namely laser scattering, electrochemistry, and semiconductor, at the key location points in the key location point set to obtain an Internet of Things device point set;

[0098] S12. Collect the air quality data of each monitoring area through the Internet of Things device point set to obtain an air quality data set;

[0099] S13. Collect the environmental data of each area to obtain an environmental data set; the environmental data includes environmental temperature and environmental humidity.

[0100] Use an environmental parameter compensation algorithm to calculate the drift compensation amount caused by the environmental data to the air quality data in the air quality data set, and obtain a drift compensation amount data set; the formula of the environmental parameter compensation algorithm is as follows.

[0101] ;

[0102] Wherein, a represents the drift compensation amount caused by the environmental data to the air quality data in the air quality data set; b 1, b 2 and b 3 respectively represent preset coefficients for the electrochemical sensors in the device; T represents the environmental temperature, RH represents the environmental humidity;

[0103] Calibrate the air quality data in the air quality data set according to the drift compensation amount data set to obtain a real-time air quality data set.

[0104] S2. Input the real-time air quality data set into an improved air quality prediction network to obtain a predicted air quality data set.

[0105] The S2 includes the following steps:

[0106] S21. Construct an LSTM network, and set the time step of the input layer of the LSTM network and the number of features of the historical air quality data; set the hidden layer of the LSTM network to 1 layer, and the number of hidden units to 32; set the structure of the output layer to 1 fully connected neuron, and the output content is the air quality prediction data at a future moment.

[0107] Use the historical air quality data to train the LSTM network. During the training process, combine an optimization algorithm to find the optimal learning rate of the LSTM network to obtain an optimal solution; use the optimal solution as the learning rate of the LSTM network to obtain an improved air quality prediction network.

[0108] The steps of combining an optimization algorithm to find the optimal learning rate of the LSTM network to obtain an optimal solution during the training process in S21 include the following steps:

[0109] S211. Set the prediction accuracy threshold of the LSTM network to z 1 and the prediction accuracy to z 2; set the maximum number of optimization iterations and the survival space size of the carp population.

[0110] S212. Construct an initial carp population set. According to the learning rate of the LSTM network, set the scale of the initial carp population set to v , then the initial carp population set is represented as , where u i represents the i -th carp population in the initial carp population set;

[0111] The living space size of each carp population in the initial carp population set is fixed. Then, the number of carp populations will affect the size of the final adult carps. Too many or too few carp populations will reduce the overall benefit;

[0112] S213. According to the prediction accuracy threshold z 1 and the prediction accuracy z 2, define the fitness function of the carp population number. The fitness function formula is as follows,

[0113] ;

[0114] Z represents the fitness function, l represents the bias of the fitness function, which is used to adjust the output of the fitness function;

[0115] S214. Perform iterative operations on the initial carp population set. The higher the fitness value, the better the number of carp populations; in each round of iteration, calculate the fitness value of the carp population number according to the fitness function, update the carp population number from high to low according to the fitness value, and obtain the best carp population number and the global best carp population number in the carp population in each round of iteration;

[0116] S215. Repeat S214. When the maximum optimization iteration number is reached, stop the iteration, and take the global best carp number as the optimal solution;

[0117] S22. Input the air quality data set into the improved air quality prediction network to obtain the predicted air quality data set;

[0118] S3. Perform risk level judgment according to the real-time air quality data set and the predicted air quality data set to obtain the judgment result;

[0119] According to the judgment result, adjust the classification collection, processing and transmission rules of the multi-source data of the Internet of Things device set for the monitoring area to obtain the optimized data packet;

[0120] The S3 includes the following steps:

[0121] S31. Construct a hybrid network of LoRaWAN and NB-IoT; in the hybrid network, NB-IoT is used to transmit data in dense urban areas in the monitoring area, and LoRaWAN is used to transmit data in suburban areas in the monitoring area; the determination conditions for dense urban areas and suburban areas are determined according to the device density;

[0122] Design a dual-mode communication module that supports dynamic protocol switching between LoRaWAN and NB-IoT for communication;

[0123] S32. Set a risk level threshold set; the risk level thresholds are divided according to air quality; for example, if PM2.5 > 150 μg / m for 1 hour, the risk level is divided into level one; if PM2.5 is 75 - 150 μg / m, the risk level is divided into level two; if PM2.5 < 75 μg / m, the risk level is divided into level three;

[0124] Based on the real-time air quality data set, the predicted air quality data set, and the risk level threshold set, obtain a risk level data set; the risk level data set contains the risk levels of the real-time air quality and predicted air quality in each area;

[0125] S33. According to the risk level threshold set, set the corresponding air quality data collection frequency to obtain a collection frequency set; for example, the collection frequency for level three is 0.001 hz, for level two is 0.05 hz, and for level one is 1 hz;

[0126] Collect, correct, compress, and add spatio-temporal tags to the data in each area in the risk level data set according to the collection frequency set to obtain an optimized data packet; the optimized data packet contains the air quality data, environmental data, and IoT device status data in each monitoring area;

[0127] Transmit the optimized data packet through the hybrid network of LoRaWAN and NB-IoT;

[0128] S4. Respectively, through the IoT device health score model and the Cox proportional hazards model combined with the optimized data packet, obtain a real-time IoT device health diversity set and a predicted device failure probability set;

[0129] S4 includes the following steps:

[0130] S41. Collect historical IoT device repair record data; based on the historical IoT device repair record data, construct an IoT device health score model; the formula of the IoT device health score model is as follows,

[0131] ;

[0132] Among them, H represents the IoT device health score, 、MTBF and Ca respectively represent the signal-to-noise ratio, mean time between failures, and calibration score of the Internet of Things device, p 1, p 2, and p 3 respectively represent the weight coefficients of the set signal-to-noise ratio, mean time between failures, and calibration score, satisfying p 1 + p 2 + p 3 = 1;

[0133] S42. Extract the signal-to-noise ratio, mean time between failures, and calibration score of each Internet of Things device in the real-time Internet of Things device set from the optimized data packet to obtain real-time Internet of Things device data;

[0134] Input the real-time Internet of Things device data into the Internet of Things device health score model to obtain a real-time Internet of Things device health diversity set;

[0135] S43. Extract the Internet of Things device status data set and the environmental data set from the optimized data packet; extract the features of the Internet of Things device status data set and the environmental data set to obtain an Internet of Things device status feature matrix;

[0136] S44. Input each Internet of Things device status feature in the Internet of Things device status feature matrix into the Cox proportional hazards model to calculate the failure probability of each Internet of Things device, obtaining a set of predicted device failure probabilities; the formula of the Cox proportional hazards model is as follows,

[0137] ;

[0138] where, λ represents the failure probability of the Internet of Things device, η represents the baseline intensity function, and exp represents the exponential function, x 1, x 2... x n respectively represent the n status features of the Internet of Things device, α 1, α 2... α n respectively represent x 1, x 2... x n the influence degree coefficients on the failure probability;

[0139] S5. Generate a maintenance work order according to the real-time Internet of Things device health diversity set in combination with the set of predicted device failure probabilities;

[0140] Generate an optimal maintenance path in combination with the maintenance work order and the location of the maintenance personnel for the maintenance of the Internet of Things device;

[0141] S5 includes the following steps:

[0142] S51. Set the device health score threshold and the failure probability threshold; detect the real-time device health diversity set and the predicted device failure probability set, mark the IoT devices with real-time device health score ≥ device health score threshold or predicted device failure probability ≥ failure probability threshold as devices to be repaired, label the priority of the devices to be repaired, and obtain a repair work order;

[0143] S52. Set the dispatching threshold; collect the locations of maintenance personnel to obtain a maintenance personnel location set;

[0144] Take the priority of the repaired devices and the dispatching threshold as constraint conditions, and match the maintenance personnel location set with the locations of the IoT devices in the repair work order through the TSP algorithm to obtain the optimal repair path;

[0145] Go to the location of the device to be repaired through the optimal repair path;

[0146] S6. Use the Gaussian process regression algorithm combined with the IoT device set to obtain an optimal IoT device location set for IoT device redeployment;

[0147] S6 includes the following steps:

[0148] S61. Extract the pollution diffusion data and meteorological parameters in the optimization data packet, and combine with the terrain elevation data, predicted air quality data, and real air quality data to obtain device impact data;

[0149] S62. Collect the location data of each IoT device in the IoT device set to obtain a real-time IoT device location set;

[0150] S63. Through the Gaussian process regression algorithm, combine the IoT device location set and the device impact data to obtain an optimal IoT device location set;

[0151] S63 includes the following steps:

[0152] S631. Set the IoT device movement range threshold; set constraint conditions according to communication limitations, terrain limitations, and the IoT device movement range threshold;

[0153] S632. Set the objective function to minimize the monitoring deviation and minimize the network deployment cost; the objective function formula is as follows,

[0154] ;

[0155] Where, f represents the total number of monitoring areas, e1 represents the predicted pollution concentration data of each monitoring area in the predicted air quality data, e 2 represents the actual pollution concentration data of each monitoring area in the actual air quality data, g represents the network deployment cost, d represents the set network deployment cost coefficient;

[0156] S633. Use the Gaussian process regression algorithm to combine the set of IoT device locations, constraints, and objective function to obtain the optimal set of IoT device locations;

[0157] S64. Use drones to redeploy IoT devices according to the locations in the optimal set of IoT device locations.

[0158] Embodiment 2:

[0159] Please refer to Figure 2 , an operation and maintenance management system for an IoT-based air quality monitoring platform, used to implement the above-mentioned operation and maintenance management method for an IoT-based air quality monitoring platform, including a data collection and correction module, an air quality prediction module, a data collection and transmission optimization module, a device health management module, an intelligent maintenance scheduling module, and a dynamic deployment optimization module;

[0160] The data collection and correction module is used to collect air quality data through IoT devices deploying multiple types of sensors and perform data correction in combination with an environmental parameter compensation algorithm; by real-time monitoring of environmental temperature and humidity, calculate the sensor data drift compensation amount using a preset formula, and generate a high-precision real-time air quality data set to ensure the accuracy of subsequent analysis;

[0161] The air quality prediction module constructs an air quality prediction model based on an improved LSTM network and dynamically adjusts the network learning rate in combination with the carp population optimization algorithm. Train and optimize the model parameters through historical data to achieve accurate prediction of future air quality, and output a prediction data set to support risk warning and decision-making;

[0162] The data collection and transmission optimization module is used to design a LoRaWAN and NB-IoT hybrid network, dynamically switch communication protocols according to device density, and improve data transmission efficiency; adjust the data collection frequency through risk level division, add spatio-temporal tags and compress to generate optimized data packets, and achieve on-demand classified transmission to reduce network load;

[0163] The device health management module is based on a comprehensive IoT device health sub-model and a Cox proportional hazards model to evaluate the real-time health status of devices and predict the failure probability; calculate the health score through parameters such as signal-to-noise ratio and mean time between failures, and analyze the failure risk in combination with the device status feature matrix to provide a quantitative basis for maintaining IoT devices;

[0164] The intelligent maintenance scheduling module is used to generate priority maintenance work orders according to the equipment health score threshold and failure probability, integrate the location data of maintenance personnel, and use the TSP algorithm to plan the optimal path; through dynamic order assignment and path matching, improve the maintenance response efficiency and reduce the equipment downtime;

[0165] The dynamic deployment optimization module adopts the Gaussian process regression algorithm, combines pollution diffusion, meteorological and terrain data to optimize the layout of Internet of Things devices; sets a double-objective function of monitoring deviation and deployment cost, and implements equipment redeployment through drones to achieve the adaptive adjustment of the monitoring network and the efficient use of resources.

[0166] In the description of this specification, the descriptions referring to the terms "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0167] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of this specification. These embodiments are selected and specifically described in this specification to better explain the principle and practical application of the invention, so that those skilled in the art can understand and utilize the invention well.

Claims

1. A method for operation and maintenance management of an air quality monitoring platform based on the Internet of Things, characterized in that, It includes the following steps: S1. Collect air quality data of each monitoring area through the Internet of Things device set to obtain an air quality data set; collect environmental data of each area to obtain an environmental data set; the environmental data includes environmental temperature and environmental humidity; use an environmental parameter compensation algorithm to calculate the drift compensation amount caused by the environmental data to the air quality data in the air quality data set to obtain a drift compensation amount data set; calibrate the air quality data in the air quality data set according to the drift compensation amount data set to obtain a real-time air quality data set; S2. Construct an LSTM network; Use historical air quality data to train the LSTM network. During the training process, dynamically adjust the learning rate of the LSTM network in combination with the carp population optimization algorithm to find the optimal learning rate of the LSTM network and obtain the optimal solution; use the optimal solution as the learning rate of the LSTM network to obtain an improved air quality prediction network; input the real-time air quality data set into the improved air quality prediction network to obtain a predicted air quality data set; S3. Construct a hybrid network of LoRaWAN and NB-IoT; in the hybrid network, NB-IoT is used to transmit data in densely populated urban areas in the monitoring area, and LoRaWAN is used to transmit data in suburban areas in the monitoring area; the determination conditions for densely populated urban areas and suburban areas are determined according to the device density; Design a dual-mode communication module to support protocol dynamic switching between LoRaWAN and NB-IoT communication; Judge the risk level according to the real-time air quality data set and the predicted air quality data set to obtain a judgment result; According to the judgment result, adjust the classification collection, processing and transmission rules of multi-source data of the Internet of Things device set for the monitoring area to obtain an optimized data packet; Transmit the optimized data packet through the hybrid network of LoRaWAN and NB-IoT; S4. Respectively, through the Internet of Things device health sub-model and the Cox proportional hazards model combined with the optimized data packet, obtain a real-time Internet of Things device health diversity set and a predicted device failure probability set; S5. Generate a maintenance work order according to the real-time Internet of Things device health diversity set combined with the predicted device failure probability set; Combine the maintenance work order and the location of the maintenance personnel, and generate an optimal maintenance path through the TSP algorithm for Internet of Things device maintenance; S6. Use the Gaussian process regression algorithm combined with the Internet of Things device set to obtain an optimal Internet of Things device location set for Internet of Things device redeployment.

2. The operation and maintenance management method of the air quality monitoring platform based on the Internet of Things according to claim 1, characterized in that The S1 includes the following steps: S11. Collect key location points in the monitoring area to obtain a key location point set; Deploy devices integrating three types of sensors, namely laser scattering, electrochemistry, and semiconductor, at the key location points in the key location point set to obtain an Internet of Things device point set; S12. Collect air quality data of each monitoring area through the Internet of Things device point set to obtain an air quality data set; S13. Collect environmental data of each area to obtain an environmental data set; the environmental data includes environmental temperature and environmental humidity; Use an environmental parameter compensation algorithm to calculate the drift compensation amount caused by the environmental data to the air quality data in the air quality data set to obtain a drift compensation amount data set; Calibrate the air quality data in the air quality dataset according to the drift compensation amount dataset to obtain a real-time air quality dataset.

3. The method for operation and maintenance management of the air quality monitoring platform based on the Internet of Things according to claim 1, characterized in that The S2 includes the following steps: S21. Construct an LSTM network; use historical air quality data to train the LSTM network. During the training process, combine an optimization algorithm to find the optimal learning rate of the LSTM network to obtain an optimal solution; use the optimal solution as the learning rate of the LSTM network to obtain an improved air quality prediction network. S22. Input the air quality dataset into the improved air quality prediction network to obtain a predicted air quality dataset.

4. The method for operation and maintenance management of the air quality monitoring platform based on the Internet of Things according to claim 3, wherein The step of combining an optimization algorithm to find the optimal learning rate of the LSTM network and obtain an optimal solution during the training process in S21 includes the following steps: S211. Set the prediction accuracy threshold of the LSTM network to z 1 and the prediction accuracy to z 2; set the maximum number of optimization iterations and set the size of the living space of the carp population; S212. Construct an initial carp population set, and set the scale of the initial carp population set according to the learning rate of the LSTM network. S213. Define the fitness function of the carp population according to the predicted accuracy threshold z 1 and the predicted accuracy z 2. S214. Perform iterative operations on the initial carp population set; during each round of iteration, calculate the fitness value of the carp population according to the fitness function, update the carp population according to the fitness value from high to low, and obtain the best carp population and the global best carp population in the carp population during each round of iteration. S215. Repeat S214. When the maximum optimization iteration number is reached, stop the iteration, and use the global best carp population as the optimal solution.

5. The operation and maintenance management method of the air quality monitoring platform based on the Internet of Things according to claim 1, characterized in that, The S3 includes the following steps: S31. Construct a hybrid network of LoRaWAN and NB-IoT; in the hybrid network, NB-IoT is used to transmit data in dense urban areas in the monitoring area, and LoRaWAN is used to transmit data in suburban areas in the monitoring area; the determination conditions for dense urban areas and suburban areas are determined according to the device density. Design a dual-mode communication module to support protocol dynamic switching between LoRaWAN and NB-IoT communication. S32. Set a risk level threshold set; the risk level threshold is divided according to air quality. Obtain a risk level dataset according to the real-time air quality dataset, the predicted air quality dataset, and the risk level threshold set; the risk level dataset contains the risk levels of the real-time air quality and the predicted air quality in each area. S33. Set corresponding air quality data collection frequencies according to the risk level threshold set to obtain a collection frequency set. Collect, correct, compress, and add spatio-temporal tags to the data in each area in the risk level dataset according to the collection frequency set to obtain an optimized data packet. The optimized data packet contains air quality data, environmental data, and IoT device status data in each monitoring area. Transmit the optimized data packet through the hybrid network of LoRaWAN and NB-IoT.

6. The operation and maintenance management method of the air quality monitoring platform based on the Internet of Things according to claim 1, characterized in that The S4 includes the following steps: S41. Collect historical IoT device maintenance record data; based on the historical IoT device maintenance record data, construct an IoT device health score model. S42. Extract the signal-to-noise ratio, mean time between failures, and calibration score of each IoT device in the real-time IoT device set from the optimized data packet to obtain real-time IoT device data. Input the real-time Internet of Things (IoT) device data into the IoT device health score model to obtain the real-time IoT device health score set; S43. Extract the IoT device status data set and the environmental data set from the optimized data packet; extract the features of the IoT device status data set and the environmental data set to obtain the IoT device status feature matrix; S44. Input each IoT device status feature in the IoT device status feature matrix into the Cox proportional hazards model, calculate the failure probability of each IoT device, and obtain the predicted device failure probability set.

7. The operation and maintenance management method of the air quality monitoring platform based on the Internet of Things according to claim 6, characterized in that The S5 includes the following steps: S51. Set the device health score threshold and the failure probability threshold; detect the real-time device health score set and the predicted device failure probability set, mark the IoT devices with real-time device health score ≥ device health score threshold or predicted device failure probability ≥ failure probability threshold as devices to be repaired, and label the priority of the devices to be repaired to obtain the maintenance work order; S52. Set the dispatch threshold; collect the locations of the maintenance personnel to obtain the maintenance personnel location set; Take the priority of the maintenance device and the dispatch threshold as the constraint conditions, and match the maintenance personnel location set with the locations of the IoT devices in the maintenance work order through the TSP algorithm to obtain the optimal maintenance path; Go to the location of the device to be repaired through the optimal maintenance path.

8. The operation and maintenance management method of the air quality monitoring platform based on the Internet of Things according to claim 1, characterized in that The S6 includes the following steps: S61. Extract the pollution diffusion data and meteorological parameters from the optimized data packet, and combine the terrain elevation data, the predicted air quality data, and the real air quality data to obtain the data affecting the device; S62. Collect the location data of each IoT device in the IoT device set to obtain the real-time IoT device location set; S63. Through the Gaussian process regression algorithm, combine the IoT device location set and the data affecting the device to obtain the optimal IoT device location set; S64. Use the drone to redeploy the IoT devices according to the locations in the optimal IoT device location set.

9. The operation and maintenance management method of the air quality monitoring platform based on the Internet of Things according to claim 8, characterized in that, The S63 includes the following steps: S631. Set the IoT device movement range threshold; set the constraint conditions according to the communication limit, the terrain limit, and the IoT device movement range threshold; S632. Set the objective function to minimize the monitoring deviation and minimize the network deployment cost; S633. Use the Gaussian process regression algorithm to combine the IoT device location set, the constraint conditions, and the objective function to obtain the optimal IoT device location set.

10. A safety and environmental protection system based on the collaborative treatment of multiple pollutants, characterized in that, Implement the operation and maintenance management method of the air quality monitoring platform based on the Internet of Things as described in any one of claims 1-9, including a data collection and correction module, an air quality prediction module, a data collection and transmission optimization module, a device health management module, an intelligent maintenance scheduling module, and a dynamic deployment optimization module.

Images