Building equipment intelligent control method and system based on Internet of Things

Abstract

The invention discloses a building equipment intelligent control method and system based on the Internet of Things, and relates to the field of intelligent control. A building equipment intelligent control system based on the Internet of Things comprises a building modeling module, an initial information acquisition module, an information standard processing module, a data real-time analysis module, a prediction analysis module and an intelligent control module. According to the invention, the building is modeled to obtain the intelligent building model, so that more detailed and dynamic display and control of building equipment are realized; and acquiring a real-time control instruction and a prediction control instruction of the area sample of the building as a control period, calculating the dynamic feasibility of the control period, and selecting the optimal control period according to the dynamic feasibility of the control period, thereby realizing intelligent dynamic control of building setting.

Description

Technical Field

[0001] The present invention relates to the field of intelligent control, and in particular to an intelligent control method and system for building equipment based on the Internet of Things. Background Art

[0002] With the continuous advancement of science and technology and the acceleration of urbanization, the demand for intelligent and automated modern building management is increasing. Traditional building equipment control mainly relies on manual operation and management, which is not only inefficient, but also prone to human misoperation and management blind spots. Traditional building equipment control has many limitations in data collection, transmission and processing, and it is difficult to meet the efficient management needs of modern intelligent buildings.

[0003] The development of Internet of Things technology has provided a new solution for intelligent building management. Through Internet of Things technology, real-time monitoring and intelligent control of various equipment in the building can be achieved, greatly improving the management efficiency and security of the system. Although Internet of Things technology has made certain progress in building management, there are still some shortcomings in the existing technical solutions. The system compatibility is poor, and there is a lack of unified standards between equipment and systems of different manufacturers, resulting in poor system compatibility, low integration and interoperability. The level of intelligent control is low. The existing system is not intelligent enough in energy management. The equipment is controlled according to a fixed mode, and the data cannot be fully utilized for energy consumption optimization and intelligent control. There is a lack of predictive ability, and it is impossible to predict the building situation based on the changes in data for pre-control and early abnormal alarm. Summary of the invention

[0004] In response to the above problems, the present invention provides an intelligent control method and system for building equipment based on the Internet of Things. Based on the Internet of Things technology and combined with artificial intelligence and big data, intelligent control of building equipment is achieved, promoting the development of building management towards intelligence, automation and efficiency.

[0005] An intelligent control method for building equipment based on the Internet of Things, comprising: S1. Obtain building information, take each area of ​​the building as an area sample, and use the building information to model the building to obtain a smart building model; S2. Obtain data from various sources in the building, use the obtained data as initial information, and obtain the number of people in each area sample; S3. Use a standardized processing method on the initial information to obtain standard information; S4. The standard information of the regional sample and the number of personnel data are combined into an independent feature vector of the regional sample, and the independent feature vector of the regional sample is sent to the real-time data analysis model for analysis to obtain a real-time control instruction of the regional sample; S5. The independent feature vector and the real-time control instruction are combined into a prediction feature vector, and the prediction feature vector of a single regional sample and the prediction feature vector of the regional sample connected thereto are combined into an integrated prediction sample, and the integrated prediction sample is sent to the prediction analysis model for analysis to obtain the prediction control instruction of the regional sample, and the prediction control instruction includes the control instruction of three time periods; S6. Control the equipment within the regional sample according to the current real-time control instructions and predictive control instructions of the regional sample, use the real-time control instructions and predictive control instructions of the regional sample as the control cycle, calculate the dynamic feasibility of the control cycle, and select the optimal control cycle according to the dynamic feasibility of the control cycle.

[0006] As a preferred technical solution of the present invention, the dynamic feasibility of the control cycle is calculated, and the optimal control cycle is selected according to the dynamic feasibility of the control cycle; the specific steps of the dynamic feasibility calculation method are: A1. Obtain standard environmental data V of real-time control instructions during the control cycle b , and compare it with the standard environmental data V of the previous period a Perform weighted difference calculation to obtain the first feasibility f1, the calculation formula is:

[0007] in the formula represents the parameters of the standard environmental data of the i-th previous period, Represents the weight of the i-th parameter, which is used to achieve the standardization of different parameter data and the regulation of parameter importance. , where Q i1 and Q i2 They represent the weight value of data volume standardization and the weight value of parameter importance respectively; A2. Obtain the predictive control instruction 1 and standard environment data V in the control cycle b , and sent to the environmental change prediction database to match the prediction standard environmental data V pa , the predicted control instruction 2 and the predicted standard environment data V pa Send to the environmental change prediction database to match the prediction standard environmental data V pb , the standard environmental data V b With the predicted standard environmental data V pa Perform weighted difference calculation to obtain the second feasibility f2, and use the predicted standard environmental data V p1 With the predicted standard environmental data V pb Perform weighted difference calculation to obtain the third feasibility f3; A3. If the currently used control instruction is a real-time control instruction, calculate the dynamic feasibility of the three adjacent control cycles; The dynamic feasibility of the current control cycle is F B =β1f1+β2f2+β3f3; The dynamic feasibility of the previous control cycle is F A =β1f 11 +β1f 12 +β2f2, where f 11 and f 12 They represent the first first feasibility and the second first feasibility of the previous control cycle respectively, and are calculated by weighted difference of adjacent standard environmental data measured twice; The dynamic feasibility of the next control cycle is F C =2*(β2f2+β3f3) / 3, where f2 is obtained by taking the real-time control instruction of the next control cycle as the prediction control instruction 1 through the second feasibility calculation method in step A2, and f3 is obtained by taking the prediction control instruction 1 of the next control cycle as the prediction control instruction 2 through the third feasibility calculation method in step A2; In the formula, β1 is the weight of the first feasibility, β1:β2:β3=3:2:1, and the sum of the weights of each feasibility in the formula is 1; A4. By comparing the dynamic feasibility F A 、F B and F C The value of the control cycle with the largest dynamic feasibility is selected to control the equipment of the regional sample. If the comparison result is F B , then the next control instruction is its corresponding prediction control instruction 1. If the comparison result is F A , then the next control instruction is its corresponding prediction control instruction 2. If the comparison result is F C , then the next control instruction is the corresponding current control instruction.

[0008] As a preferred technical solution of the present invention, it also includes the calculation of the control instructions used being the prediction control instruction 1 and the prediction control instruction 2; When the control instruction adopted is the predictive control instruction 1, the dynamic feasibility of the three adjacent control cycles is calculated; The dynamic feasibility of the current control cycle is F B =β1f 11 +β1f 12 +β2f2; The dynamic feasibility of the previous control cycle is F A =β1f 11 +β1f 12 +β1f 13 ; where f 13It is the third first feasibility of the previous control cycle, which is calculated by weighted difference between adjacent standard environmental data measured twice; The dynamic feasibility of the next control cycle is F C =β1f1+β2f2+β3f3; If the comparison result is F B , then the next control instruction is its corresponding prediction control instruction 2. If the comparison result is F A , then the next control instruction is its corresponding prediction control instruction 3. If the comparison result is F C , then the next control instruction is its corresponding predicted control instruction 1.

[0009] When the control instruction adopted is the predictive control instruction 2, the dynamic credibility of the current control cycle and the next two control cycles is calculated; The dynamic feasibility of the current control cycle is F B =β1f 11 +β1f 12 +β1f 13 ; The dynamic feasibility of the next control cycle is F C =β1f 11 +β1f 12 +β2f2; The dynamic feasibility of the second control cycle is F D =β1f 11 +β1f 12 +β2f2; If the comparison result is F B , then the next control instruction is its corresponding prediction control instruction 3. If the comparison result is F C , then the next control instruction is its corresponding prediction control instruction 2. If the comparison result is F D , then the next control instruction is its corresponding predicted control instruction 1.

[0010] As a preferred technical solution of the present invention, the building information is used to model the building to obtain a smart building model; The acquired building information includes the spatial information of the building, the distribution of sensor equipment, and the distribution of building equipment. The constructed smart building model includes regional samples, the adjacent connectivity relationship between regional samples, the equipment and sensor conditions within the regional samples, and the standard environmental data of the regional samples.

[0011] As a preferred technical solution of the present invention, data from various sources in the building are obtained; The acquired data include temperature sensor data, smoke sensor data, current sensor data, wind speed sensor data, imaging equipment data and equipment operation data. The number of personnel data in each area sample is acquired through the imaging equipment data. As a preferred technical solution of the present invention, it also includes data transmission through wired and wireless methods, wherein the data is first transmitted to a distributed data receiver in a building, and then transmitted from the distributed data receiver to a data processing server for processing.

[0012] As a preferred technical solution of the present invention, standard information is obtained; sensors and devices of regional samples are recorded, data types of different sensors and devices in the regional samples are obtained, and different types of data are converted into standard data through standardized processing methods.

[0013] As a preferred technical solution of the present invention, it also includes a real-time data analysis model, which is trained based on the SVM model. The target of the model is the standard control instructions in the standard control instruction library, which is obtained by training independent feature vectors of several regional samples.

[0014] As a preferred technical solution of the present invention, it also includes a prediction analysis model, which is trained based on a recurrent neural network. The target of the model is three prediction control instructions in the standard control instruction library, and is obtained by training multiple prediction feature vectors in integrated prediction samples.

[0015] An intelligent control system for building equipment based on the Internet of Things, comprising: The building modeling module is used to obtain the information of the building, take each area of ​​the building as an area sample, and use the building information to model the building to obtain a smart building model; The initial information collection module is used to obtain data from various sources in the building, and the obtained data is used as initial information to obtain the number of people in each area sample; An information standard processing module is used to use a standardized processing method on the initial information to obtain standard information; The real-time data analysis module is used to combine the standard information of the regional sample and the personnel quantity data into an independent feature vector of the regional sample, send the independent feature vector of the regional sample into the real-time data analysis model for analysis, and obtain the real-time control instruction of the regional sample; The prediction analysis module is used to combine the independent feature vector and the real-time control instruction into a prediction feature vector, and combine the prediction feature vector of a single regional sample and the prediction feature vector of the regional samples connected thereto into an integrated prediction sample, and send the integrated prediction sample to the prediction analysis model for analysis to obtain the prediction control instruction of the regional sample, which includes the control instructions of two time periods; The intelligent control module is used to control the equipment in the regional sample according to the current real-time control instructions and predictive control instructions of the regional sample, take the real-time control instructions and predictive control instructions of the regional sample as the control cycle, calculate the dynamic feasibility of the control cycle, and select the optimal control cycle according to the dynamic feasibility of the control cycle.

[0016] The present invention has the following advantages: 1. The present invention obtains a smart building model by modeling the building, thereby realizing a more detailed and dynamic display and control of the building equipment; and obtains the real-time control instructions and predictive control instructions of the building's regional samples as the control cycle, calculates the dynamic feasibility of the control cycle, and selects the optimal control cycle according to the dynamic feasibility of the control cycle, thereby realizing intelligent dynamic control of the building equipment.

[0017] 2. The present invention realizes standard processing of equipment data from different manufacturers by using a standardized processing method for initial information, and reduces the cost of intelligent control of building equipment by combining actual wireless and wired methods to transmit data.

[0018] 3. The present invention obtains real-time control instructions by analyzing the independent feature vectors of regional samples using a real-time data analysis model, and obtains real-time control instructions and predictive control instructions of regional samples by analyzing the integrated prediction samples using a predictive analysis model. The real-time control instructions effectively ensure the feasibility of the control instructions. The predictive control instructions effectively predict the changes in the sample area environment through multiple independent feature vectors of the sample area and the connected sample areas, and make predictive control instructions based on the changes, thereby improving the predictive capability of intelligent control of building equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a schematic diagram of the structure of an Internet of Things-based building equipment intelligent control system adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0021] Embodiment 1, a building equipment intelligent control method based on the Internet of Things, comprising: S1. Obtain building information, take each area of ​​the building as an area sample, and use the building information to model the building to obtain a smart building model; The acquired building information includes the spatial information of the building, the distribution of sensor equipment, and the distribution of building equipment. The constructed smart building model includes regional samples, the adjacent connectivity relationship between regional samples, the equipment and sensor conditions within the regional samples, and the standard environmental data of the regional samples.

[0022] Equipment includes air conditioning equipment, fire extinguishing equipment, lighting equipment, power supply equipment and exhaust equipment; S2. Obtain data from various sources in the building, use the obtained data as initial information, and obtain the number of people in each area sample; The acquired data includes temperature sensor data, smoke sensor data, current sensor data, wind speed sensor data, imaging device data and equipment operation data. The imaging device data is used to acquire the number of personnel in each area sample. Specifically, in the implementation process, the number of personnel in the image is acquired by acquiring the imaging device data and using the image recognition algorithm, and the specific number of personnel in each area sample is acquired based on the entry and exit of the personnel. Data is transmitted via wired and wireless methods, depending on the cost of the transmission method. The cost of wired and wireless transmission is calculated, and the transmission method is determined by comparing the costs. Data is first transmitted to the distributed data receivers in the building, and then transmitted to the data processing server for processing.

[0023] S3. Use a standardized processing method on the initial information to obtain standard information; The sensors and devices of the regional samples are recorded, the data types of different sensors and devices in the regional samples are obtained, and different types of data are converted into standard data through standardized processing methods.

[0024] This ensures that data from devices from different sources can be processed together, reducing the cost of purchasing equipment and being able to adapt to old equipment. When adding new equipment, the corresponding data type needs to be added to the smart building model; S4. The standard information of the regional sample and the number of personnel data are combined into an independent feature vector of the regional sample, and the independent feature vector of the regional sample is sent to the real-time data analysis model for analysis to obtain a real-time control instruction of the regional sample; The data real-time analysis model is trained based on the SVM model. The model targets the standard control instructions in the standard control instruction library and is trained through independent feature vectors of several regional samples. SVM (Support Vector Machine) is a supervised learning model widely used in the field of machine learning, mainly used for classification problems. In this method, the SVM model is selected as the basis for training the real-time analysis model to improve the efficiency of data classification and improve the classification performance of the model when the amount of data is small; S5. The independent feature vector and the real-time control instruction are combined into a prediction feature vector, and the prediction feature vector of a single regional sample and the prediction feature vector of the regional sample connected thereto are combined into an integrated prediction sample, and the integrated prediction sample is sent to the prediction analysis model for analysis to obtain the prediction control instruction of the regional sample, and the prediction control instruction includes the control instruction of three time periods; The prediction analysis model is trained based on a recurrent neural network. The model targets three prediction control instructions in the standard control instruction library and is trained by integrating multiple prediction feature vectors in the prediction samples. The prediction analysis model adopted by the method includes 1 integrated input layer, 8 hidden layers and 1 integrated output layer; the integrated input layer is used to receive multiple independent feature vectors of the input integrated prediction samples; the 8 hidden layers are used to analyze the connection between the multiple independent feature vectors of the input, the first 5 layers are used to evaluate the connection between the multiple independent feature vectors, the 6th layer is used to obtain the prediction control instruction 1 according to the connection relationship of the first 5 layers, the 7th layer is used to obtain the prediction control instruction 2 according to the connection relationship of the first 6 layers and the prediction control instruction 1, and the 8th layer is used to obtain the firearm prediction control instruction 3 according to the connection relationship of the first 7 layers and the prediction control instruction 2; the integrated output layer is used to output the prediction control instruction; In a recurrent neural network, the states of different layers are represented by time steps, and the dependencies in the data are obtained by propagating the recurrent connections between different layers in the time steps. The hidden layer state is determined by the hidden state of the previous time step and the current input. The specific formula is:

[0025] The state of the output layer is determined by the current hidden state. The specific formula is:

[0026] in represents the hidden state at the time step, Indicates the current input. Indicates the current output, represents the weight of the input layer, Represents the state quantity of the input layer; S6. Control the equipment within the regional sample according to the current real-time control instructions and predictive control instructions of the regional sample, use the real-time control instructions and predictive control instructions of the regional sample as the control cycle, calculate the dynamic feasibility of the control cycle, and select the optimal control cycle according to the dynamic feasibility of the control cycle.

[0027] Calculate the dynamic feasibility of the control cycle, and select the optimal control cycle according to the dynamic feasibility of the control cycle; the specific steps of the dynamic feasibility calculation method are: A1. Obtain standard environmental data V of real-time control instructions during the control cycle b , and compare it with the standard environmental data V of the previous period a Perform weighted difference calculation to obtain the first feasibility f1, the calculation formula is:

[0028] in the formula represents the parameters of the standard environmental data of the i-th previous period, Represents the weight of the i-th parameter, which is used to achieve the standardization of different parameter data and the regulation of parameter importance. , where Q i1 and Q i2 They represent the weight value of data volume standardization and the weight value of parameter importance respectively; A2. Obtain the predictive control instruction 1 and standard environment data V in the control cycle b , and sent to the environmental change prediction database to match the prediction standard environmental data V pa , the predicted control instruction 2 and the predicted standard environment data V pa Send to the environmental change prediction database to match the prediction standard environmental data V pb , the standard environmental data V b With the predicted standard environmental data V pa Perform weighted difference calculation to obtain the second feasibility f2, and use the predicted standard environmental data V p1 With the predicted standard environmental data V pb Perform weighted difference calculation to obtain the third feasibility f3; A3. If the currently used control instruction is a real-time control instruction, calculate the dynamic feasibility of the three adjacent control cycles; The dynamic feasibility of the current control cycle is F B =β1f1+β2f2+β3f3; The dynamic feasibility of the previous control cycle is F A =β1f 11 +β1f 12 +β2f2, where f 11 and f 12 They represent the first first feasibility and the second first feasibility of the previous control cycle respectively, and are calculated by weighted difference of adjacent standard environmental data measured twice; The dynamic feasibility of the next control cycle is F C=2*(β2f2+β3f3) / 3, where f2 is obtained by taking the real-time control instruction of the next control cycle as the prediction control instruction 1 through the second feasibility calculation method in step A2, and f3 is obtained by taking the prediction control instruction 1 of the next control cycle as the prediction control instruction 2 through the third feasibility calculation method in step A2; In the formula, β1 is the weight of the first feasibility, β1:β2:β3=3:2:1, and the sum of the weights of the feasibility in the formula is 1. In this way, the weights of different subsequent predictive control instructions are reduced layer by layer, and the weights of the control instructions are controlled according to the actual situation, so as to achieve the purpose of dynamically adjusting the control instructions. When the control instruction adopted is the predictive control instruction 1, the dynamic feasibility of the three adjacent control cycles is calculated; The dynamic feasibility of the current control cycle is F B =β1f 11 +β1f 12 +β2f2; The dynamic feasibility of the previous control cycle is F A =β1f 11 +β1f 12 +β1f 13 ; where f 13 It is the third first feasibility of the previous control cycle, which is calculated by weighted difference between adjacent standard environmental data measured twice; The dynamic feasibility of the next control cycle is F C =β1f1+β2f2+β3f3; If the comparison result is F B , then the next control instruction is its corresponding prediction control instruction 2. If the comparison result is F A , then the next control instruction is its corresponding prediction control instruction 3. If the comparison result is F C , then the next control instruction is its corresponding predicted control instruction 1.

[0029] When the control instruction adopted is the predictive control instruction 2, the dynamic credibility of the current control cycle and the next two control cycles is calculated; The dynamic feasibility of the current control cycle is F B =β1f 11 +β1f 12 +β1f 13 ; The dynamic feasibility of the next control cycle is F C =β1f 11 +β1f 12 +β2f2; The dynamic feasibility of the second control cycle is F D =β1f 11+β1f 12 +β2f2; If the comparison result is F B , then the next control instruction is its corresponding prediction control instruction 3. If the comparison result is F C , then the next control instruction is its corresponding prediction control instruction 2. If the comparison result is F D , then the next control instruction is its corresponding prediction control instruction 1; A4. By comparing the dynamic feasibility F A 、F B and F C The value of the control cycle with the largest dynamic feasibility is selected to control the equipment of the regional sample. If the comparison result is F B , then the next control instruction is its corresponding prediction control instruction 1. If the comparison result is F A , then the next control instruction is its corresponding prediction control instruction 2. If the comparison result is F C , then the next control instruction is the corresponding current control instruction.

[0030] Example 2, an intelligent control system for building equipment based on the Internet of Things, see Figure 1 As shown, including: The building modeling module is used to obtain the information of the building, take each area of ​​the building as an area sample, and use the building information to model the building to obtain a smart building model; The initial information collection module is used to obtain data from various sources in the building, and the obtained data is used as initial information to obtain the number of people in each area sample; An information standard processing module is used to use a standardized processing method on the initial information to obtain standard information; The real-time data analysis module is used to combine the standard information of the regional sample and the personnel quantity data into an independent feature vector of the regional sample, send the independent feature vector of the regional sample into the real-time data analysis model for analysis, and obtain the real-time control instruction of the regional sample; The prediction analysis module is used to combine the independent feature vector and the real-time control instruction into a prediction feature vector, and combine the prediction feature vector of a single regional sample and the prediction feature vector of the regional samples connected thereto into an integrated prediction sample, and send the integrated prediction sample to the prediction analysis model for analysis to obtain the prediction control instruction of the regional sample, which includes the control instructions of two time periods; The intelligent control module is used to control the equipment in the regional sample according to the current real-time control instructions and predictive control instructions of the regional sample, take the real-time control instructions and predictive control instructions of the regional sample as the control cycle, calculate the dynamic feasibility of the control cycle, and select the optimal control cycle according to the dynamic feasibility of the control cycle.

[0031] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A building equipment intelligent control method based on the Internet of Things, characterized in that: include: S1. Obtain building information, take each area of ​​the building as an area sample, and use the building information to model the building to obtain a smart building model; S2. Obtain data from various sources in the building, use the obtained data as initial information, and obtain the number of people in each area sample; S3. Use a standardized processing method on the initial information to obtain standard information; S4. The standard information of the regional sample and the number of personnel data are combined into an independent feature vector of the regional sample, and the independent feature vector of the regional sample is sent to the real-time data analysis model for analysis to obtain a real-time control instruction of the regional sample; S5. The independent feature vector and the real-time control instruction are combined into a prediction feature vector, and the prediction feature vector of a single regional sample and the prediction feature vector of the regional sample connected thereto are combined into an integrated prediction sample, and the integrated prediction sample is sent to the prediction analysis model for analysis to obtain the prediction control instruction of the regional sample, and the prediction control instruction includes the control instruction of three time periods; S6. Control the equipment within the regional sample according to the current real-time control instructions and predictive control instructions of the regional sample, use the real-time control instructions and predictive control instructions of the regional sample as the control cycle, calculate the dynamic feasibility of the control cycle, and select the optimal control cycle according to the dynamic feasibility of the control cycle.

2. The method for intelligent control of building equipment based on the Internet of Things according to claim 1, characterized in that: Calculate the dynamic feasibility of the control cycle, and select the optimal control cycle according to the dynamic feasibility of the control cycle; the specific steps of the dynamic feasibility calculation method are: A1. Obtain standard environmental data V of real-time control instructions during the control cycle b , and compare it with the standard environmental data V of the previous period a Perform weighted difference calculation to obtain the first feasibility f1, the calculation formula is: in the formula represents the parameters of the standard environmental data of the i-th previous period, Represents the weight of the i-th parameter, which is used to achieve the standardization of different parameter data and the regulation of parameter importance. , where Q i1 and Q i2 They represent the weight value of data volume standardization and the weight value of parameter importance respectively; A2. Obtain the predictive control instruction 1 and standard environment data V in the control cycle b , and sent to the environmental change prediction database to match the prediction standard environmental data V pa , the predicted control instruction 2 and the predicted standard environment data V pa Send to the environmental change prediction database to match the prediction standard environmental data V pb , the standard environmental data V b With the predicted standard environmental data V pa Perform weighted difference calculation to obtain the second feasibility f2, and use the predicted standard environmental data V p1 With the predicted standard environmental data V pb Perform weighted difference calculation to obtain the third feasibility f3; A3. If the currently used control instruction is a real-time control instruction, calculate the dynamic feasibility of the three adjacent control cycles; The dynamic feasibility of the current control cycle is F B =β1f1+β2f2+β3f3; The dynamic feasibility of the previous control cycle is F A =β1f 11 +β1f 12 +β2f2, where f 11 and f 12 They represent the first first feasibility and the second first feasibility of the previous control cycle respectively, and are calculated by weighted difference of adjacent standard environmental data measured twice; The dynamic feasibility of the next control cycle is F C =2*(β2f2+β3f3) / 3, where f2 is obtained by taking the real-time control instruction of the next control cycle as the prediction control instruction 1 through the second feasibility calculation method in step A2, and f3 is obtained by taking the prediction control instruction 1 of the next control cycle as the prediction control instruction 2 through the third feasibility calculation method in step A2; In the formula, β1 is the weight of the first feasibility, β1:β2:β3=3:2:1, and the sum of the weights of each feasibility in the formula is 1; A4. By comparing the dynamic feasibility F A 、F B and F C The value of the control cycle with the largest dynamic feasibility is selected to control the equipment of the regional sample. If the comparison result is F B , then the next control instruction is its corresponding prediction control instruction 1. If the comparison result is F A , then the next control instruction is its corresponding prediction control instruction 2. If the comparison result is F C , then the next control instruction is the corresponding current control instruction.

3. The method for intelligent control of building equipment based on the Internet of Things according to claim 2, characterized in that: It also includes the calculation conditions when the control instructions used are prediction control instruction 1 and prediction control instruction 2; When the control instruction adopted is the predictive control instruction 1, the dynamic feasibility of the three adjacent control cycles is calculated; The dynamic feasibility of the current control cycle is F B =β1f 11 +β1f 12 +β2f2; The dynamic feasibility of the previous control cycle is F A =β1f 11 +β1f 12 +β1f 13 ; where f 13 It is the third first feasibility of the previous control cycle, which is calculated by weighted difference between adjacent standard environmental data measured twice; The dynamic feasibility of the next control cycle is F C =β1f1+β2f2+β3f3; If the comparison result is F B , then the next control instruction is its corresponding prediction control instruction 2. If the comparison result is F A , then the next control instruction is its corresponding prediction control instruction 3. If the comparison result is F C , then the next control instruction is its corresponding predicted control instruction 1. When the control instruction adopted is the predictive control instruction 2, the dynamic credibility of the current control cycle and the next two control cycles is calculated; The dynamic feasibility of the current control cycle is F B =β1f 11 +β1f 12 +β1f 13 ; The dynamic feasibility of the next control cycle is F C =β1f 11 +β1f 12 +β2f2; The dynamic feasibility of the second control cycle is F D =β1f 11 +β1f 12 +β2f2; If the comparison result is F B , then the next control instruction is its corresponding prediction control instruction 3. If the comparison result is F C , then the next control instruction is its corresponding prediction control instruction 2. If the comparison result is F D , then the next control instruction is its corresponding predicted control instruction 1.

4. The method for intelligent control of building equipment based on the Internet of Things according to claim 1, characterized in that: Use the building information to model the building to obtain a smart building model; The acquired building information includes the spatial information of the building, the distribution of sensor equipment, and the distribution of building equipment. The constructed smart building model includes regional samples, the adjacent connectivity relationship between regional samples, the equipment and sensor conditions within the regional samples, and the standard environmental data of the regional samples.

5. The method for intelligent control of building equipment based on the Internet of Things according to claim 1, characterized in that: Obtain data from various sources in the building; The acquired data include temperature sensor data, smoke sensor data, current sensor data, wind speed sensor data, imaging equipment data and equipment operation data. The number of personnel data in each area sample is acquired through the imaging equipment data.

6. The method for intelligent control of building equipment based on the Internet of Things according to claim 5, characterized in that: It also includes data transmission through wired and wireless methods. The data is first transmitted to a distributed data receiver in a building, and then transmitted from the distributed data receiver to a data processing server for processing.

7. The method for intelligent control of building equipment based on the Internet of Things according to claim 1, characterized in that: Obtain standard information; record the sensors and devices of the regional samples, obtain the data types of different sensors and devices in the regional samples, and convert different types of data into standard data through standardized processing methods.

8. The method for intelligent control of building equipment based on the Internet of Things according to claim 1, characterized in that: It also includes a real-time data analysis model, which is trained based on the SVM model. The target of the model is the standard control instructions in the standard control instruction library, which is obtained through independent feature vector training of samples from several regions.

9. The method for intelligent control of building equipment based on the Internet of Things according to claim 1, characterized in that: It also includes a predictive analysis model, which is trained based on a recurrent neural network. The model targets three predictive control instructions in the standard control instruction library and is trained by integrating multiple predictive feature vectors in the prediction samples.

10. An intelligent control system for building equipment based on the Internet of Things, characterized in that: The system applies an intelligent control method for building equipment based on the Internet of Things as described in any one of claims 1 to 10, including: The building modeling module is used to obtain the information of the building, take each area of ​​the building as an area sample, and use the building information to model the building to obtain a smart building model; The initial information collection module is used to obtain data from various sources in the building, and the obtained data is used as initial information to obtain the number of people in each area sample; An information standard processing module is used to use a standardized processing method on the initial information to obtain standard information; The real-time data analysis module is used to combine the standard information of the regional sample and the personnel quantity data into an independent feature vector of the regional sample, send the independent feature vector of the regional sample into the real-time data analysis model for analysis, and obtain the real-time control instruction of the regional sample; The prediction analysis module is used to combine the independent feature vector and the real-time control instruction into a prediction feature vector, and combine the prediction feature vector of a single regional sample and the prediction feature vector of the regional samples connected thereto into an integrated prediction sample, and send the integrated prediction sample to the prediction analysis model for analysis to obtain the prediction control instruction of the regional sample, which includes the control instructions of two time periods; The intelligent control module is used to control the equipment in the regional sample according to the current real-time control instructions and predictive control instructions of the regional sample, take the real-time control instructions and predictive control instructions of the regional sample as the control cycle, calculate the dynamic feasibility of the control cycle, and select the optimal control cycle according to the dynamic feasibility of the control cycle.

Images