Temperature and humidity monitoring system based on cloud computing fusion of Internet of Things

Abstract

The invention discloses a temperature and humidity monitoring system based on Internet of Things cloud computing fusion, which overcomes the defects in the prior art and comprises a sensing layer, a data acquisition system and a temperature and humidity control system, the data acquisition system is composed of a plurality of temperature and humidity acquisition devices, and the temperature and humidity control system controls temperature and humidity according to data of the data acquisition system; the network layer communicates with the gateway platform of the sensing layer, acquires data acquired by the data acquisition system, and transmits the data to the cloud service layer; the cloud service layer comprises a cloud data sharing pool, the cloud data sharing pool stores the data sent by the network layer, and the cloud service layer processes and analyzes the data through a cloud computing platform; and the application layer is in communication connection with the cloud service layer and is used for monitoring the temperature and humidity state, calling data and issuing a control instruction to a temperature and humidity control system.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power equipment, and in particular to a temperature and humidity monitoring system based on the fusion of Internet of Things and cloud computing. Background Art

[0002] Switchgear is a key electrical equipment. In the process of power generation, transmission, distribution and power conversion of the power system, it is mainly responsible for opening, closing, controlling and protecting electrical equipment. Nowadays, switchgear has been widely used in many industries such as power plants, substations, petrochemicals, metallurgy and steel rolling, light industry and textiles, as well as distribution stations and box-type substations in load centers such as factories and mines, urban residential areas, high-rise buildings, and large public buildings. Existing switchgear monitoring usually relies on on-site manual inspections and cannot obtain the operating parameters and environmental data of the equipment in real time. Excessive temperature of the switchgear will lead to accelerated aging of electronic components, shorten their service life, and even directly cause damage; excessive humidity of the switchgear will make the insulation material susceptible to moisture, resulting in a decrease in insulation performance. At the same time, excessive humidity may also cause partial discharge, further damage the insulation material, and increase the risk of equipment failure. Therefore, it is particularly important to obtain the temperature and humidity conditions of the switchgear in real time, and relevant personnel and systems can take corresponding measures in time according to the obtained temperature and humidity conditions. Summary of the invention

[0003] The purpose of the present invention is to overcome the shortcomings of the prior art that the environmental data of the switch cabinet cannot be obtained and analyzed in real time, resulting in the inability to promptly deal with the situation of excessive temperature or excessive humidity in the switch cabinet, increasing the risk of equipment operation, and to provide a temperature and humidity monitoring system based on the integration of Internet of Things and cloud computing.

[0004] The purpose of the present invention is achieved through the following technical solutions: The temperature and humidity monitoring system based on the integration of the Internet of Things and cloud computing is used to monitor the temperature and humidity conditions in the switch cabinet, including: the perception layer, including a data acquisition system composed of multiple temperature and humidity acquisition devices and a temperature and humidity control system. The temperature and humidity control system controls the temperature and humidity according to the data of the data acquisition system; The network layer communicates with the gateway platform of the perception layer, obtains the data collected by the data acquisition system, and transmits the data to the cloud service layer; The cloud service layer includes a cloud data sharing pool, which stores data sent by the network layer. The cloud service layer processes and analyzes data through the cloud computing platform. The application layer communicates with the cloud service layer and is used to monitor the temperature and humidity status, call data, and issue control instructions to the temperature and humidity control system.

[0005] Preferably, the perception layer is specifically an ESP32 monitoring system, which includes an environmental information acquisition module, an ESP32 control module, an OLED display module, an output control module and an alarm module. The ESP32 control module is connected to the environmental information acquisition module, the OLED display module, the output control module and the alarm module at the same time; the ESP32 control module is also provided with an MQTT protocol port, and the ESP32 control module communicates with the network management platform through the MQTT protocol.

[0006] Preferably, the cloud service layer processes and analyzes data through a cloud computing platform, specifically: A high-voltage switchgear temperature and humidity prediction model based on BP neural network is constructed to analyze the influencing factors of temperature and humidity changes inside the switchgear. The number of BP neural network input nodes is 5, including external ambient temperature x1, external ambient humidity x2, atmospheric pressure x3, wind speed x4 and 24h time node x5. The number of BP neural network output nodes is 2, including switchgear internal temperature y1 and switchgear internal humidity y2. The number of hidden layer nodes is determined by trial and error according to the empirical formula. The effect of the high-voltage switchgear temperature and humidity prediction model is compared within the range to determine the optimal number of hidden layer nodes.

[0007] Preferably, the temperature and humidity prediction model of the high-voltage switch cabinet based on the BP neural network is further optimized by a GA algorithm to construct a temperature and humidity prediction model of the high-voltage switch cabinet based on a GA-BP neural network. The construction process of the GA-BP neural network is specifically as follows: Import the temperature and humidity data in the switch cabinet and use the mapminmax function to normalize the data; Determine the GA topology; Coding and population initialization use real number coding method. Each individual in the initial population contains all connection weights and thresholds in the BP neural network. Assuming the chromosome length in the population is L, the chromosome length formula is as follows: L=lm+ln+m+n Among them, l, m, and n are the number of nodes in the input layer, hidden layer, and output layer respectively; Determine the fitness function, substitute the initialized network weights and thresholds into the BP neural network, and the individual fitness value is the absolute value of the difference between the predicted result and the measured result; Genetic evolution operation, using roulette method to determine the parent individuals for crossover, using real number crossover method for crossover operation, generating new offspring individuals by combining the genes of parent individuals; Determine whether the genetic evolution operation has reached the set maximum number of times. If the maximum number of evolutions is reached, GA stops running and the last generation population is obtained; The chromosome with the strongest adaptability in the last generation of the GA is decoded and substituted into the BP neural network to obtain the initial connection weights and thresholds to start training. When the network reaches the set accuracy or the maximum number of iterations, the BP neural network training is completed, and the optimization of the BP neural network is completed. As a preferred method, the temperature and humidity monitoring system based on the integration of IoT and cloud computing also optimizes the GA-BP neural network through the MapReduce algorithm to construct the MR-GA-BP neural network. The process is as follows: Step 1: read the weather environment data outside the cabinet and the temperature and humidity data inside the cabinet stored in the cloud data sharing pool as training sets, and use the MapReduce program framework to split the temperature and humidity prediction training set to obtain k different temperature and humidity prediction data training subsets; where k represents the number of parallel Maps in the cloud cluster; Step 2: Use the Map function program logic to process the temperature and humidity prediction data training subset, that is, use the GA-BP neural network to perform parallel training on the temperature and humidity prediction data training set, that is, use k different GA-BP neural networks to train and predict temperature and humidity; Step 3: The result after the Map operation is transferred to the Reduce stage through the Shuffle stage of the MapReduce program framework, that is, the weight of the predicted value output by each Map operation is determined by the weight calculation method for k different GA-BP predicted values, so as to realize the calculation of the final predicted value of temperature and humidity.

[0008] Preferably, the temperature and humidity acquisition device also includes a constant temperature heating plate, which is used to simulate the heating phenomenon during the actual operation of the switch cabinet; the constant temperature heating plate includes a silicone rubber heating sheet and a temperature controller, and the temperature controller is used to control the heating temperature of the silicone rubber heating sheet.

[0009] Preferably, after the data collection system collects data, it also performs data cleaning on the data. Data cleaning includes processing missing values ​​and outliers, and filling the data using the mean interpolation method. The formula is as follows: In the formula, x i To remove data, x i-1 and x i+1 They are the data before and after the data is removed respectively; After the data is cleaned, the data is normalized using the min-max function. The formula is as follows: In the formula, x n is the normalized data, x is the original data, and x min and x max are the maximum and minimum values ​​of the original data respectively.

[0010] The beneficial effects of the present invention are as follows: the data acquisition system of the present invention can collect temperature and humidity data in the switch cabinet to ensure the timeliness of the data; the cloud service layer can process and analyze a large amount of data quickly and efficiently using the processing power of cloud computing; the application layer allows users to remotely monitor the temperature and humidity status of the switch cabinet without on-site inspections, thereby improving work efficiency. Users can issue control instructions to the temperature and humidity control system through the application layer to achieve remote control. At the same time, the solution of the present invention makes a good prediction of the temperature and humidity of the switch cabinet. Users can take corresponding measures on the switch cabinet in advance based on the prediction results, thereby improving the use efficiency of the switch cabinet and further ensuring the stability of the switch cabinet operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is an overall solution architecture diagram of the present invention; Figure 2 It is a functional block diagram of the ESP32 monitoring system of the present invention; Figure 3 It is a block diagram of the principle of the ESP32 monitoring system of the present invention communicating with other devices; Figure 4 It is a schematic diagram of the temperature and humidity prediction model of high-voltage switchgear based on BP neural network; Figure 5 It is the flow chart of the GA-BP neural network combined deformation prediction model; Figure 6 It is a schematic diagram of the temperature and humidity prediction model of high-voltage switchgear based on MR-GA-BP neural network; Figure 7 It is a Map process diagram of the present invention; Figure 8 It is the Reduce process diagram of the present invention. DETAILED DESCRIPTION

[0012] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0013] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0014] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0015] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0016] Example 1: Figure 1 As shown in the figure, the temperature and humidity monitoring system based on the integration of IoT cloud computing is used to monitor the temperature and humidity in the switch cabinet, including: The perception layer includes a data acquisition system and a temperature and humidity control system composed of multiple temperature and humidity acquisition devices. The temperature and humidity control system controls the temperature and humidity according to the data from the data acquisition system; The network layer communicates with the gateway platform of the perception layer, obtains the data collected by the data acquisition system, and transmits the data to the cloud service layer; The cloud service layer includes a cloud data sharing pool, which stores data sent by the network layer. The cloud service layer processes and analyzes data through the cloud computing platform. The application layer communicates with the cloud service layer and is used to monitor the temperature and humidity status, call data, and issue control instructions to the temperature and humidity control system.

[0017] In this embodiment, the temperature and humidity monitoring system is mainly composed of a cloud computing subsystem and an Internet of Things subsystem. Among them, the Internet of Things subsystem is mainly responsible for on-site information collection and transmission. The switch cabinet Internet of Things subsystem is a wireless sensor network and control system, which is composed of a large number of related nodes. It is responsible for transmitting the collected temperature and humidity data inside the switch cabinet to the gateway via the aggregation node, and the gateway node uploads the data to the cloud computing platform via the Internet. The cloud computing platform in the cloud is mainly responsible for storing and making decisions on the collected temperature and humidity data inside the switch cabinet, and issuing dehumidification instructions to the remote temperature and humidity control system through the Internet based on the results of data processing. Electric field engineering technicians can use wired PC networking, mobile phones or tablet computers and other mobile terminals to access the cloud computing platform. This solution comprehensively analyzes the Internet of Things technology and cloud computing technology. Both have advantages and disadvantages, complement each other, and integrate resources to ensure information collection, data sharing, remote control and collaborative operation of the temperature and humidity monitoring system.

[0018] Specifically, the perception layer includes multiple temperature and humidity collection devices to form data collection and temperature and humidity control systems. Both the collection system and the temperature and humidity control system are controlled by the ESP32 IoT module. Data collection is mainly completed through temperature and humidity sensors to complete the internal temperature and humidity collection of the switch cabinet, and the small meteorological station to complete the external climate collection, including temperature, humidity, atmospheric pressure, wind speed, wind direction and other parameters; temperature and humidity control is mainly achieved by receiving control instructions to control dehumidifiers, heaters, etc. The information perceived by the collection system enters the network layer through the IoT gateway. The mobile communication network, Internet and other transmission networks in the network layer communicate with the IoT gateway platform through the MQTT protocol. The cloud service layer accesses the cloud data sharing pool to achieve data access, data persistence, event processing engine, distributed data processing and basic business support, thereby achieving the goals of resource management, data management, event alarm and data analysis; finally, the mobile terminal APP, Web and open API of the application layer are used to monitor the status, call data and issue control instructions.

[0019] like Figure 2 As shown, the perception layer is specifically an ESP32 monitoring system, which includes an environmental information acquisition module, an ESP32 control module, an OLED display module, an output control module and an alarm module. The ESP32 control module is connected to the environmental information acquisition module, the OLED display module, the output control module and the alarm module at the same time; the ESP32 control module is also provided with an MQTT protocol port, and the ESP32 control module communicates with the network management platform through the MQTT protocol.

[0020] In this embodiment, Figure 3 As shown, the communication process between the ESP32 monitoring system and other devices is as follows: (1) The temperature and humidity sensors collect the temperature and humidity data inside the switch cabinet, and the voltage and current sensors collect the voltage and current flowing through the busbar to calculate its heat generation. The small weather station completes the collection of external climate data, including temperature, humidity, atmospheric pressure, wind speed, wind direction and other parameters; (2) The ESP32 control module sends the collected environmental information to the network layer through the IoT gateway. The mobile communication network, Internet and other transmission networks in the network layer establish communication with the IoT gateway platform through the MQTT protocol and send the data to the cloud computing platform; (3) The cloud computing platform is mainly responsible for storing and processing the collected internal and external data of the switch cabinet. The processing is to send the collected temperature and humidity data to the trained SSA-BP neural network temperature and humidity prediction model, and transmit the predicted data of the next time period and the current data to the client segment for query and monitoring; (4) The staff can view the current information of the switch cabinet, the predicted value in the cabinet, and the historical information through the mobile terminal APP, Web, and open API of the application layer. At the same time, the application layer receives the temperature and humidity data and automatically determines whether there is a risk of condensation. If so, it sends control instructions for the dehumidifier and heater to ESP32. The staff can also send commands through the mobile terminal to start and stop the dehumidifier and heater.

[0021] BP neural network is a multi-layer feedforward neural network, which is trained according to the error back propagation algorithm. The learning process of this algorithm includes forward propagation of information and back propagation of error. During forward propagation, the sample features are input from the input layer, and the samples are processed nonlinearly by each hidden layer to generate an output signal. If the error between the actual output and the expected output is greater than the set value, the back propagation process is entered to back propagate the error signal. Error back propagation is to back propagate the output error through the hidden layer to the output layer layer by layer, obtain the error learning signal of each layer, and finally correct the weight and threshold of each neuron according to the error learning signal to reduce the error to within the expected error.

[0022] By analyzing the influencing factors of temperature and humidity changes inside the switch cabinet, a high-voltage switch cabinet temperature and humidity prediction model based on BP neural network is constructed. Figure 4 As shown: A high-voltage switchgear temperature and humidity prediction model based on BP neural network is constructed to analyze the influencing factors of temperature and humidity changes inside the switchgear. The number of BP neural network input nodes is 5, including external ambient temperature x1, external ambient humidity x2, atmospheric pressure x3, wind speed x4 and 24h time node x5. The number of BP neural network output nodes is 2, including switchgear internal temperature y1 and switchgear internal humidity y2. The number of hidden layer nodes is determined by trial and error according to the empirical formula. The effect of the high-voltage switchgear temperature and humidity prediction model is compared within the range to determine the optimal number of hidden layer nodes.

[0023] The temperature and humidity acquisition equipment also includes a constant temperature heating plate, which is used to simulate the heating phenomenon during the actual operation of the switch cabinet; the constant temperature heating plate includes a silicone rubber heating sheet and a temperature controller, and the temperature controller is used to control the heating temperature of the silicone rubber heating sheet. Since eddy current losses are generated during the operation of high-current and high-voltage switch cabinets, and the eddy current losses generate a large amount of heat, affecting the temperature and humidity inside the cabinet. Therefore, the heating phenomenon caused by eddy current losses in the cabinet needs to be considered when collecting data. However, the heat generated by eddy current losses during the actual operation of the switch cabinet is difficult to measure. This embodiment uses a constant temperature heating plate to simulate the heating phenomenon during the actual operation of the switch cabinet. The silicone rubber heating sheet is placed at the busbar, and its temperature control device is led out of the cabinet by a wire.

[0024] After the data collection system collects data, it also performs data cleaning on the data. Data cleaning includes processing missing values ​​and outliers, and filling the data using the mean interpolation method. The formula is as follows: In the formula, x i To remove data, x i-1 and x i+1 They are the data before and after the data is removed respectively; After the data is cleaned, the data is normalized using the min-max function. The formula is as follows: In the formula, x n is the normalized data, x is the original data, and x min and x max are the maximum and minimum values ​​of the original data respectively.

[0025] Example 2: A temperature and humidity monitoring system based on the integration of the Internet of Things and cloud computing. Its principle and implementation method are basically the same as those of Example 1. The differences are as follows: Figure 5 As shown, the temperature and humidity prediction model of the high-voltage switch cabinet based on the BP neural network is also optimized by the GA algorithm to construct a temperature and humidity prediction model of the high-voltage switch cabinet based on the GA-BP neural network. The construction process of the GA-BP neural network is specifically as follows: Import the temperature and humidity data in the switch cabinet and use the mapminmax function to normalize the data; Determine the GA topology; Coding and population initialization use real number coding method. Each individual in the initial population contains all connection weights and thresholds in the BP neural network. Assuming the chromosome length in the population is L, the chromosome length formula is as follows: L=lm+ln+m+n Among them, l, m, and n are the number of nodes in the input layer, hidden layer, and output layer respectively; Determine the fitness function, substitute the initialized network weights and thresholds into the BP neural network, and the individual fitness value is the absolute value of the difference between the predicted result and the measured result; Genetic evolution operation, using roulette method to determine the parent individuals for crossover, using real number crossover method for crossover operation, generating new offspring individuals by combining the genes of parent individuals; Determine whether the genetic evolution operation has reached the set maximum number of times. If the maximum number of evolutions is reached, GA stops running and the last generation population is obtained; The chromosome with the strongest adaptability in the last generation of the GA is decoded and substituted into the BP neural network to obtain the initial connection weights and thresholds to start training. When the network reaches the set accuracy or the maximum number of iterations, the BP neural network training is completed, and the optimization of the BP neural network is completed. Genetic Algorithm (GA) is an adaptive global optimization probability search algorithm formed by simulating the genetic and evolutionary process of organisms in the natural environment. The basic idea of ​​GA optimization of BP neural network model is: GA search is not limited to one point, so it has the ability of global optimization; GA search uses probability method to search efficiently in the solution space, so it has a faster convergence speed. The initial weights and thresholds of BP neural network are obtained by using these two abilities of GA, so that the model can converge to the global optimal point at a faster speed, making the final prediction result closer to the actual measured value.

[0026] Example 3: A temperature and humidity monitoring system based on the integration of the Internet of Things and cloud computing. Its principle and implementation method are basically the same as those of Example 2. The differences are as follows: Figure 6 As shown in the figure, the temperature and humidity monitoring system based on the integration of IoT and cloud computing also optimizes the GA-BP neural network through the MapReduce algorithm to construct the MR-GA-BP neural network. The process is as follows: Step 1: read the weather environment data outside the cabinet and the temperature and humidity data inside the cabinet stored in the cloud data sharing pool as training sets, and use the MapReduce program framework to split the temperature and humidity prediction training set to obtain k different temperature and humidity prediction data training subsets; where k represents the number of parallel Maps in the cloud cluster; Step 2: Use the Map function program logic to process the temperature and humidity prediction data training subset, that is, use the GA-BP neural network to perform parallel training on the temperature and humidity prediction data training set, that is, use k different GA-BP neural networks to train and predict temperature and humidity; Step 3: The result after the Map operation is transferred to the Reduce stage through the Shuffle stage of the MapReduce program framework, that is, the weight of the predicted value output by each Map operation is determined by the weight calculation method for k different GA-BP predicted values, so as to realize the calculation of the final predicted value of temperature and humidity.

[0027] In view of the randomness of the weight selection of the BP neural network, it is easy to fall into the local minimum. The main method of the genetic optimization of the BP neural network is to use the weight of the BP neural network as the initial population of the genetic algorithm, calculate the fitness value of each individual in the population according to the fitness function, select individuals with high fitness values ​​as the parent individuals of the next generation, cross the parent individuals according to a certain crossover probability, generate new individuals, perform mutation operations on the new individuals, and judge whether the obtained individuals meet the requirements. In this process, if there are many individual data of the selected population, the calculation of fitness is very time-consuming. Therefore, this paper adopts the parallelization of the genetic algorithm. The basic operation is: use the Map function to calculate the error of each individual (the inverse of the error is the fitness function), and then perform the basic operations of the genetic algorithm in the Reduce function.

[0028] MapReduce can implement distributed storage and parallel computing for problems with large data sets, and can be implemented on large-scale clusters, which can greatly improve computing efficiency and save time. In addition, the MapReduce framework shields the underlying implementation details, reducing the difficulty of implementation. MapReduce consists of two parts: Map and Reduce functions. These two functions are relatively simple and can be written by readers according to actual conditions. The task of the MapReduce programming framework is to accept job requests through the job submission node, and then the Mapper and Recucer are sent by the jobtracker process to the idle tasktracker process. The jobtracker process is also responsible for monitoring the execution status of this process. The tasktracker process completes the task after receiving it. The overall process is as follows: (1) Input stage. First, the input data stored in the distributed file system is read. Then, the input data is sliced ​​into a series of data slices, and each data slice is assigned to a Map function. (2) Map phase. The input data slice is regarded as a set of (key, value) key-value pairs, and then the assigned (key, value) key-value pairs are processed by the Map function logic written by the user to generate new (key, value) intermediate process key-value pairs; (3) Shuffle phase. The (key, value) intermediate key value traces are transferred from the Map node to the Reduce node. At the same time, the same (key, value) intermediate key value pairs are merged and sorted. Due to factors such as bandwidth and CPU running speed, the time consumed may be much longer than the running time of the Map function and the Reduce function. (4) CDReduce stage. Execute the Reduce function written by the user and process all intermediate values ​​according to the data processing logic set by the user. The intermediate key or intermediate key chain (list of values) outputs the new (key, value) key-value pair; (5) Output phase: Output the program processing results of the Reduce phase and store them in the distributed file system.

[0029] This embodiment combines the Multi-Agent concept with cloud computing technology to improve the GA-BP model. The MR-GA-BP algorithm introduces the Multi-Agent concept into GA-BP and runs GA-BP through weight balancing. Different nodes of GA-BP have different prediction accuracies. The nodes with higher accuracy have higher weights when calculating the final prediction value of GA-BP. The final prediction value of GA-BP can be expressed as: in, represents the final prediction value of GA-BP; α k The kth GA-BP prediction weight; y k represents the predicted value of the kth GA-BP, and the predicted weight α k It can be solved by gradient ascent method and standard function error E: Among them, t represents the expected value of the input training set of each GA-BP. According to the gradient ascent method, the prediction weight α k for: Where η represents the learning rate, and η=0.1.

[0030] Combining them gives: Among them, α k Through α k ←α k +△α k to update.

[0031] Specifically, the Map process is as follows Figure 7 As shown in the figure, the population individual information of the genetic algorithm is used as the input of the Map function. The individual information is a real number string formed by real number encoding of the BP neural network parameters such as the connection weight between the input layer and the hidden layer, the hidden layer threshold, the connection weight between the hidden layer and the output layer, and the output layer threshold. The map function needs to calculate the error of each individual during training and generate a key-value pair of <individual, error>. The Reduce process is as follows Figure 8As shown, the <individual, error> key-value pair is generated through the output of the Map function, which can be converted into the <individual, fitness value> key-value pair. The Reduce function requires the <individual, fitness value> key-value pair to merge. In order to avoid the situation where the best individuals are eliminated, the individuals with the highest fitness value are first directly transformed into the next generation. Then the roulette algorithm is used to select the individuals to enter the next generation. The selected individuals must be crossovered according to a certain crossover probability. The uniform crossover method is used here. In order to ensure genetic diversity, a mutation process is also required. Since real number coding is used, crossover cannot change all weights, so the mutation probability needs to be appropriately increased.

[0032] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not disclosed in the present application.

[0033] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. The temperature and humidity monitoring system based on the integration of IoT and cloud computing is used to monitor the temperature and humidity in the switch cabinet. Its characteristics are: include: The perception layer includes a data acquisition system and a temperature and humidity control system composed of multiple temperature and humidity acquisition devices. The temperature and humidity control system controls the temperature and humidity according to the data from the data acquisition system; The network layer communicates with the gateway platform of the perception layer, obtains the data collected by the data acquisition system, and transmits the data to the cloud service layer; The cloud service layer includes a cloud data sharing pool, which stores data sent by the network layer. The cloud service layer processes and analyzes data through the cloud computing platform. The application layer communicates with the cloud service layer and is used to monitor the temperature and humidity status, call data, and issue control instructions to the temperature and humidity control system.

2. The temperature and humidity monitoring system based on the integration of Internet of Things and cloud computing according to claim 1 is characterized in that: The perception layer is specifically an ESP32 monitoring system, which includes an environmental information acquisition module, an ESP32 control module, an OLED display module, an output control module and an alarm module. The ESP32 control module is connected to the environmental information acquisition module, the OLED display module, the output control module and the alarm module at the same time; the ESP32 control module is also provided with an MQTT protocol port, and the ESP32 control module communicates with the network management platform through the MQTT protocol.

3. The temperature and humidity monitoring system based on the integration of Internet of Things and cloud computing according to claim 1 is characterized in that: The cloud service layer processes and analyzes data through the cloud computing platform, specifically: A high-voltage switchgear temperature and humidity prediction model based on BP neural network is constructed to analyze the influencing factors of temperature and humidity changes inside the switchgear. The number of BP neural network input nodes is 5, including external ambient temperature x1, external ambient humidity x2, atmospheric pressure x3, wind speed x4 and 24h time node x5. The number of BP neural network output nodes is 2, including switchgear internal temperature y1 and switchgear internal humidity y2. The number of hidden layer nodes is determined by trial and error according to the empirical formula. The effect of the high-voltage switchgear temperature and humidity prediction model is compared within the range to determine the optimal number of hidden layer nodes.

4. The temperature and humidity monitoring system based on the integration of Internet of Things and cloud computing according to claim 3 is characterized in that: The temperature and humidity prediction model of the high-voltage switch cabinet based on the BP neural network is also optimized by the GA algorithm to construct a temperature and humidity prediction model of the high-voltage switch cabinet based on the GA-BP neural network. The construction process of the GA-BP neural network is specifically as follows: Import the temperature and humidity data in the switch cabinet and use the mapminmax function to normalize the data; Determine the GA topology; Coding and population initialization use real number coding method. Each individual in the initial population contains all connection weights and thresholds in the BP neural network. Assuming the chromosome length in the population is L, the chromosome length formula is as follows: L=lm+ln+m+n Among them, l, m, and n are the number of nodes in the input layer, hidden layer, and output layer respectively; Determine the fitness function, substitute the initialized network weights and thresholds into the BP neural network, and the individual fitness value is the absolute value of the difference between the predicted result and the measured result; Genetic evolution operation, using roulette method to determine the parent individuals for crossover, using real number crossover method for crossover operation, generating new offspring individuals by combining the genes of parent individuals; Determine whether the genetic evolution operation has reached the set maximum number of times. If the maximum number of evolutions is reached, GA stops running and the last generation population is obtained; The chromosome with the strongest adaptability in the last generation of the GA is decoded and substituted into the BP neural network to obtain the initial connection weights and thresholds to start training. When the network reaches the set accuracy or the maximum number of iterations, the BP neural network training is completed, and the optimization of the BP neural network is completed.

5. The temperature and humidity monitoring system based on the integration of Internet of Things and cloud computing according to claim 4 is characterized in that: The GA-BP neural network is also optimized by MapReduce algorithm to construct the MR-GA-BP neural network. The process is as follows: Step 1: read the weather environment data outside the cabinet and the temperature and humidity data inside the cabinet stored in the cloud data sharing pool as training sets, and use the MapReduce program framework to split the temperature and humidity prediction training set to obtain k different temperature and humidity prediction data training subsets; where k represents the number of parallel Maps in the cloud cluster; Step 2: Use the Map function program logic to process the temperature and humidity prediction data training subset, that is, use the GA-BP neural network to perform parallel training on the temperature and humidity prediction data training set, that is, use k different GA-BP neural networks to train and predict temperature and humidity; Step 3: The result after the Map operation is transferred to the Reduce stage through the Shuffle stage of the MapReduce program framework, that is, the weight of the predicted value output by each Map operation is determined by the weight calculation method for k different GA-BP predicted values, so as to realize the calculation of the final predicted value of temperature and humidity.

6. The temperature and humidity monitoring system based on the fusion of Internet of Things and cloud computing according to claim 1 is characterized in that: The temperature and humidity acquisition device also includes a constant temperature heating plate, which is used to simulate the heating phenomenon during the actual operation of the switch cabinet; The constant temperature heating plate comprises a silicone rubber heating sheet and a temperature controller, and the temperature controller is used to control the heating temperature of the silicone rubber heating sheet.

7. The temperature and humidity monitoring system based on the fusion of Internet of Things and cloud computing according to claim 1 is characterized in that: After the data collection system collects data, it also performs data cleaning on the data. Data cleaning includes processing missing values ​​and outliers, and filling the data using the mean interpolation method. The formula is as follows: In the formula, x i To remove data, x i-1 and x i+1 They are the data before and after the data is removed respectively; After the data is cleaned, the data is normalized using the min-max function. The formula is as follows: In the formula, x n is the normalized data, x is the original data, and x min and x max are the maximum and minimum values ​​of the original data respectively.

Images