A building energy load prediction device and prediction method thereof

By adopting a variety of installation methods and real-time data update technology in building energy load prediction devices, combined with machine learning models, the problems of uncertainty in prediction results and inconvenient installation are solved, and more accurate and flexible energy load prediction and efficient installation are achieved. .

CN118329120BActive Publication Date: 2024-11-15SUZHOU CONSTR ENG QUALITY TESTING CENT CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410550882.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-07
Publication Date
2024-11-15
Estimated Expiration
2044-05-07

AI Technical Summary

Technical Problem

When facing weather changes and personnel behavior changes, the existing building energy load prediction devices have high uncertainty in the prediction results, and the sensor installation is inconvenient and poor versatility.

Method used

A building energy load prediction device including sensor module, user behavior monitoring component, data processing and analysis module, real-time update and online learning module, decision support module and risk assessment module was designed, using magnets, suction cups and adapters. This installation method combines machine learning and time series analysis models to realize real-time data updates and online learning, adapting to environmental changes and user behavior changes.

Benefits of technology

It improves the accuracy and stability of the prediction results, simplifies the sensor installation process, enhances the universality and installation efficiency of the device, and can effectively compatible with installation requirements in different locations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118329120B_ABST
    Figure CN118329120B_ABST
Patent Text Reader

Abstract

The present invention belongs to the field of energy planning technology, and discloses a building energy load prediction device and a prediction method thereof, including a sensor module, a user behavior monitoring component, a data processing and analysis module, a model prediction module, a real-time update and online learning module, a decision support module and a risk assessment module. The present invention automatically collects and processes the latest sensor data to ensure the continuity and timeliness of the data. At the same time, the prediction model can be adaptively trained and updated according to the new data to ensure the accuracy and reliability of the model. Since changes in user behavior may lead to inconsistencies with historical data, thereby affecting the accuracy of the prediction results, the prediction device adopts advanced algorithms and deep learning technology, which can analyze and model user behavior in real time, so as to better capture changes in user behavior and improve the accuracy and stability of the prediction results.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of energy planning, and specifically relates to a building energy load prediction device and a prediction method thereof. Background Art

[0002] The building energy load forecasting device is a device that uses computer technology, mathematical models and statistical analysis methods to predict and analyze building energy consumption. It can collect data on internal and external building environmental parameters (such as temperature, humidity, light, etc.), building usage and equipment operation, and combine building characteristics and historical energy consumption data to use mathematical models and algorithms to predict and analyze energy consumption in the future. The building energy load forecasting device can provide reliable data support for building energy management, helping building owners and managers optimize building energy consumption, improve energy efficiency, and reduce energy consumption costs.

[0003] However, there are still problems with the building energy load prediction devices currently used. The prediction of energy load is affected by many uncertain factors, such as weather changes, changes in personnel behavior, etc., and the uncertainty factors of sudden situations increase. These uncertain factors are difficult to predict accurately and may lead to uncertainty in the prediction results. At the same time, building energy load prediction is a dynamic process that requires real-time updating of data and models. If the data or model cannot be updated in time, the prediction results may lose accuracy. Building energy load prediction is usually based on historical data for model training and prediction. However, changes in user behavior may lead to inconsistencies with historical data, thereby affecting the accuracy of the prediction results.

[0004] At the same time, the data acquisition of the building energy load prediction device mainly relies on the data input collected by the sensor to provide data support for subsequent predictions. Conventional sensors need to collect different data and need to collect data at various locations in the building, so the sensors need to be installed in various places in the building, such as smooth surfaces or metal surfaces and doors and windows. At this time, due to the limitations of the installation surface, various installation methods are required to achieve installation. The current sensor module is not compatible with the installation of multiple surfaces, resulting in many inconveniences during installation and poor versatility. Summary of the invention

[0005] The purpose of the present invention is to provide a building energy load prediction device and a prediction method thereof to solve the problems raised in the above background technology.

[0006] In order to achieve the above-mentioned purpose, the present invention provides the following technical solutions: a building energy load prediction device, comprising a sensor module, a user behavior monitoring component, a data processing and analysis module, a model prediction module, a real-time update and online learning module, a decision support module and a risk assessment module, the output end of the sensor module and the user behavior monitoring component is electrically connected to the input end of the data processing and analysis module, the output end of the data processing and analysis module is electrically connected to the input end of the model prediction module, the output end of the model prediction module is electrically connected to the input end of the real-time update and online learning module, the output end of the real-time update and online learning module is electrically connected to the input end of the decision support module, and the output end of the decision support module is electrically connected to the input end of the risk assessment module.

[0007] As a further technical solution of the present invention, the sensor module includes a temperature sensor, a humidity sensor and a light sensor. The temperature sensor is used to measure the temperature of the environment inside and outside the building. The humidity sensor is used to obtain the humidity level of the environment inside and outside the building. The light sensor is used to monitor the light intensity inside and outside the building. The user behavior monitoring component includes a human infrared sensor and a surveillance camera. The human infrared sensor is used to detect human thermal radiation in the building and identify the activities and positions of people. The surveillance camera is used to monitor the activities of people in the building in real time.

[0008] As a further technical solution of the present invention, the data processing and analysis module includes data cleaning and preprocessing, feature extraction, and model training and optimization, and the prediction model module includes a machine learning model and a time series analysis model.

[0009] As a further technical solution of the present invention, the real-time update and online learning module includes real-time data update, online learning and iterative optimization, the decision support module includes energy consumption prediction results and suggestions and strategy support, and the risk assessment module includes prediction error analysis and risk assessment report.

[0010] As a further technical solution of the present invention, it includes a mounting seat, a threaded sleeve is fixedly installed in the middle of the top of the mounting seat, a sensor module is provided above the mounting seat, a threaded rod is fixedly connected to the middle of the bottom of the sensor module, the sensor module is threadedly connected to the threaded sleeve through the threaded rod, card slots are opened on the left and right sides of the mounting seat, and a magnet is embedded in the middle of the bottom of the mounting seat.

[0011] As a further technical solution of the present invention, an adapter frame is provided on the outer side surface of the mounting seat, and clamping strips are installed on the left and right sides of the inner side surface of the adapter frame. The adapter frame is movably connected between the clamping strip and the clamping slot. The adapter frame is displaced forward and backward relative to the mounting seat, and a suction cup is detachably installed on the bottom end of the adapter frame.

[0012] As a further technical solution of the present invention, a longitudinal guide rail is detachably installed on one side of the adapter frame, a clamping plate is provided at the bottom of the adapter frame, a guide block is fixedly installed on one side of the clamping plate, the clamping plate is movably connected to the longitudinal guide rail through the guide block, the clamping plate is displaced up and down relative to the adapter frame, the longitudinal guide rail is movably connected to a screw rod inside, a threaded groove is opened in the middle of the guide block and is threadedly connected to the screw rod.

[0013] When installing the sensor module, a variety of installation methods can be selected. When the plane to be installed is made of iron material and has a smooth surface, the device can be directly installed using the magnet on the back of the mounting seat, that is, the magnet on the back of the mounting seat is directly adsorbed to the smooth surface of the iron object to complete the installation of the mounting seat. At the same time, the threaded rod can be selectively rotated to remove the sensor module, and different sensors can be installed on the threaded sleeve. When the installation plane is a smooth non-iron plane, the card strip on the inner side of the adapter frame and the card slot can be movably connected. At this time, the mounting seat can be installed between the adapter frame, and a suction cup can be selectively installed. At this time, the suction cup can be used to complete the adsorption between it and the smooth plane to complete the installation. When the installation plane cannot be magnetically adsorbed or vacuum adsorbed, a longitudinal guide rail can be selectively installed, and the area to be installed is placed between the adapter frame and the splint. By rotating the screw rod, the guide block can be driven to move up and down relative to the longitudinal guide rail. By controlling the splint to approach the direction of the adapter frame, the mounting seat can be clamped to the position to be installed, completing the clamping installation process, and realizing the installation of sensor modules in various environments.

[0014] By utilizing the magnetic adsorption of the magnet, the vacuum adsorption of the suction cup, and the clamping between the adapter and the clamping plate, the sensor can be installed at different locations inside and outside the building, and the installation method can be quickly adjusted only by the adapter. The whole process can be completed quickly and can effectively accommodate installation requirements in different locations, effectively avoiding the problem that traditional devices need to use different installation structures to achieve installation. The overall versatility is strong, which effectively reduces the initial layout time and improves the layout efficiency.

[0015] A prediction method of a building energy load prediction device comprises the following steps:

[0016] S1: Install sensor modules, including temperature sensors, humidity sensors, and light sensors, to regularly obtain real-time data of environmental parameters inside and outside the building, install human infrared sensors and surveillance cameras to monitor user behavior, and integrate the data with environmental parameter data;

[0017] S2: Data cleaning and preprocessing: cleaning the raw data collected by the sensor, removing outliers and noise, and performing data interpolation to ensure data quality and continuity;

[0018] S3: Feature extraction, extracting features related to energy consumption from the cleaned data, including time, temperature, humidity, and light, using statistical methods, including but not limited to other feature extraction techniques such as frequency domain analysis;

[0019] S4: Model training and optimization, using machine learning algorithms including but not limited to neural networks, support vector machines or time series analysis models, training on historical data, optimizing the model using cross-validation, finding the best model parameters and structure, and appropriate feature selection methods;

[0020] S5: Real-time update and online learning: The latest sensor data is obtained regularly or in real time and added to the prediction model to maintain the accuracy and timeliness of the model. Based on the real-time updated data, the prediction model is learned and optimized online to adapt to changes in the environment and user behavior.

[0021] S6: Energy load forecasting, based on the trained model, predicts the energy load in the future according to the current environmental parameters and user behavior, using machine learning models or time series analysis models to make predictions based on the patterns and trends of historical data;

[0022] S7: Decision support: present the energy consumption forecast results to building managers or energy managers in an intuitive and easy-to-understand way, helping them to formulate reasonable energy scheduling and energy-saving measures. Based on the forecast results, corresponding energy-saving suggestions are provided, including but not limited to adjusting the equipment operating temperature range and optimizing equipment scheduling strategies.

[0023] S8: Risk assessment: perform error analysis on the prediction results, evaluate the accuracy of the model, and provide confidence intervals or reliability indicators. Generate a risk assessment report, point out possible risks and uncertainties, and provide corresponding treatment suggestions.

[0024] The beneficial effects of the present invention are as follows:

[0025] 1. The present invention predicts the energy load in the future by real-time monitoring of the environmental parameters inside and outside the building and the user behavior. The building manager or energy manager can formulate reasonable energy scheduling and energy-saving measures according to the prediction results, optimize the equipment scheduling strategy, thereby improving energy efficiency and reducing the negative impact on the environment. The prediction model has real-time update and online learning functions, and can iteratively optimize the model according to the latest sensor data, which enables the prediction model to adapt to environmental changes and changes in user behavior, maintain accuracy and timeliness. At the same time, through online learning, the model can continuously improve its own prediction ability and improve the accuracy and stability of the prediction results. Considering the impact of uncertain factors such as weather changes and changes in personnel behavior on energy load, the prediction device adopts flexible algorithms and models. The structure can effectively model these factors. At the same time, the device also provides risk assessment and treatment suggestions to help building managers or energy managers understand the confidence of the prediction results and provide corresponding response strategies to reduce the impact of uncertainty on energy scheduling and energy-saving measures. The prediction device can automatically collect and process the latest sensor data to ensure the continuity and timeliness of the data. At the same time, the prediction model can be adaptively trained and updated according to the new data to ensure the accuracy and reliability of the model. Since changes in user behavior may lead to inconsistencies with historical data, thereby affecting the accuracy of the prediction results, the prediction device uses advanced algorithms and deep learning technology to analyze and model user behavior in real time, so as to better capture changes in user behavior and improve the accuracy and stability of the prediction results.

[0026] 2. The present invention utilizes the magnetic adsorption of the magnet, the vacuum adsorption of the suction cup, and the clamping between the adapter frame and the clamping plate, so that the sensor can be installed at different positions inside and outside the building, and the installation method can be quickly adjusted only by the adapter frame. The whole process can be completed quickly and can effectively accommodate the installation requirements in different positions, effectively avoiding the problem that traditional devices need to use different installation structures to achieve installation. The overall versatility is strong, which effectively reduces the initial layout time and improves the layout efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the overall process of the present invention;

[0028] Figure 2 It is a schematic diagram of the coordination between the adapter frame and the clamping plate structure of the present invention;

[0029] Figure 3 It is an exploded schematic diagram of the adapter frame and the clamping plate structure of the present invention;

[0030] Figure 4 It is an exploded schematic diagram of the mounting base, the sensor module and the adapter structure of the present invention;

[0031] Figure 5 It is a schematic diagram of the coordination between the adapter frame and the suction cup structure of the present invention;

[0032] Figure 6 It is a schematic diagram of the bottom end structure of the mounting base of the present invention.

[0033] In the figure: 1. Mounting base; 2. Sensor module; 3. Threaded sleeve; 4. Threaded rod; 5. Slot; 6. Adapter frame; 7. Clip strip; 8. Magnet; 9. Suction cup; 10. Clamp; 11. Longitudinal guide rail; 12. Guide block; 13. Screw. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] like Figures 1 to 6 As shown, in an embodiment of the present invention, a building energy load prediction device includes a sensor module, a user behavior monitoring component, a data processing and analysis module, a model prediction module, a real-time update and online learning module, a decision support module and a risk assessment module, the output end of the sensor module and the user behavior monitoring component is electrically connected to the input end of the data processing and analysis module, the output end of the data processing and analysis module is electrically connected to the input end of the model prediction module, the output end of the model prediction module is electrically connected to the input end of the real-time update and online learning module, the output end of the real-time update and online learning module is electrically connected to the input end of the decision support module, and the output end of the decision support module is electrically connected to the input end of the risk assessment module.

[0036] The device can monitor the environmental parameters and user behaviors inside and outside the building in real time, pre-process and extract features of the data, and make predictions through machine learning algorithms and time series analysis models to provide accurate energy load forecasting results. At the same time, the device can update the model in real time and learn online based on the latest data to ensure the timeliness and accuracy of the prediction model. Based on the prediction results of energy load, the device can also provide decision support for building managers and energy managers to help them formulate reasonable energy scheduling strategies and energy-saving measures. The device will generate intuitive and easy-to-understand prediction results and provide corresponding energy-saving suggestions, such as adjusting the operating temperature range of equipment, optimizing equipment scheduling strategies, etc., so as to maximize the use of energy and achieve effective use of energy. In addition, the device also has a risk assessment function, performs error analysis on the prediction results, evaluates the accuracy of the model, and provides confidence intervals or reliability indicators. Through the risk assessment report, the device can point out possible risks and uncertainties, and provide corresponding processing suggestions to help managers make wise decisions.

[0037] like Figure 6 As shown, the sensor module includes a temperature sensor, a humidity sensor and a light sensor. The temperature sensor is used to measure the temperature of the environment inside and outside the building. The humidity sensor is used to obtain the humidity level of the environment inside and outside the building. The light sensor is used to monitor the light intensity inside and outside the building. The user behavior monitoring component includes a human infrared sensor and a surveillance camera. The human infrared sensor is used to detect human thermal radiation in the building and identify the activities and positions of people. The surveillance camera is used to monitor the activities of people in the building in real time.

[0038] Among them, temperature sensors can measure the temperature inside and outside the building, humidity sensors are used to obtain humidity levels, and light sensors monitor light intensity. These sensors provide real-time environmental data to help understand the thermal environment inside the building and provide a basis for comfort control and energy-saving scheduling. The data provided by the sensor module can be used to intelligently adjust the operation of air conditioning, heating and ventilation equipment to provide appropriate indoor temperature and humidity, and achieve a balance between energy saving and comfort. Light sensors monitor the light intensity indoors and outdoors, and can automatically adjust the brightness and opening time of indoor lighting equipment to maximize the use of natural light and save energy. Human infrared sensors and surveillance cameras can detect the activities and positions of people in the building. Through these monitoring, intelligent equipment control can be achieved, such as automatically shutting down equipment when there is no human activity to reduce energy consumption. Surveillance cameras provide real-time video monitoring to help managers detect abnormal situations in a timely manner and take corresponding measures to improve the safety of the building. The data collected by the sensor module and user behavior monitoring components can be used for data analysis to provide decision support for building managers, such as optimizing energy management, improving the indoor environment, and improving building efficiency.

[0039] like Figure 6As shown, the data processing and analysis module includes data cleaning and preprocessing, feature extraction, model training and optimization, and the prediction model module includes machine learning models and time series analysis models.

[0040] The data processing and analysis module can clean and preprocess the raw data collected by the sensor module and the user behavior monitoring component, which includes steps such as removing outliers, filling missing values, and smoothing data to ensure the accuracy and consistency of the data, laying the foundation for subsequent feature extraction and model training. The data processing and analysis module can extract meaningful features from the preprocessed data. These features can include environmental parameters such as temperature, humidity, and light intensity, as well as information such as the frequency and duration of personnel activities. Through feature extraction, the raw data can be converted into input features that can be used by the prediction model. The data processing and analysis module is responsible for training the prediction model using training data. It can select a suitable machine learning model or time series analysis model, and learn the relationship between data through the training process to establish an effective prediction model. In addition, model training also involves parameter adjustment and optimization to improve the prediction accuracy and robustness of the model. The prediction model module uses the trained model to predict future indicators such as building environment and energy consumption, which can help building managers understand future trends and changes and make corresponding optimization decisions. For example, according to the prediction results, the equipment operation strategy and energy supply plan can be adjusted to reduce energy consumption costs or improve comfort.

[0041] like Figure 6 As shown, the real-time update and online learning module includes real-time data update, online learning and iterative optimization, the decision support module includes energy consumption forecast results and suggestions and strategy support, and the risk assessment module includes forecast error analysis and risk assessment report.

[0042] The real-time update and online learning module can update the data collected by the sensor module and the user behavior monitoring component in real time to ensure the timeliness and accuracy of the data. This can keep the prediction model effective in real-time environmental changes and improve the accuracy of the prediction results. The real-time update and online learning module can combine new data with historical data, conduct online learning and iterative optimization, so as to continuously improve the accuracy and robustness of the prediction model, which can enable the prediction model to adapt to data changes more quickly and continue to provide accurate prediction results. The decision support module uses the prediction results of the prediction model to provide building managers with energy consumption prediction results and corresponding suggestions and strategy support, which can help managers make more scientific and accurate decisions, optimize energy management and equipment operation strategies, reduce energy consumption costs, and improve building efficiency. The risk assessment module can analyze the prediction error of the prediction model, evaluate the risk of the prediction results, and generate a corresponding risk assessment report, which can help managers understand the reliability of the prediction results and take corresponding measures.

[0043] like Figures 1 to 5 As shown, it also includes a mounting base 1, a threaded sleeve 3 is fixedly installed in the middle of the top of the mounting base 1, a sensor module 2 is arranged above the mounting base 1, a threaded rod 4 is fixedly connected to the middle of the bottom of the sensor module 2, the sensor module 2 is threadedly connected to the threaded sleeve 3 through the threaded rod 4, a card slot 5 is opened on both sides of the mounting base 1, a magnet 8 is embedded in the middle of the bottom of the mounting base 1, an adapter frame 6 is arranged on the outer side of the mounting base 1, and a card strip 7 is installed on both sides of the inner side of the adapter frame 6, and the adapter frame 6 is connected to the card slot 5 through the card strip 7. The adapter frame 6 is movably connected between the adapter frame 6 and the mounting seat 1. The bottom end of the adapter frame 6 is detachably mounted with a suction cup 9, and one side of the adapter frame 6 is detachably mounted with a longitudinal guide rail 11. A clamping plate 10 is provided below the adapter frame 6, and a guide block 12 is fixedly mounted on one side of the clamping plate 10. The clamping plate 10 is movably connected with the longitudinal guide rail 11 through the guide block 12. The clamping plate 10 is displaced up and down relative to the adapter frame 6. The interior of the longitudinal guide rail 11 is movably connected with a screw rod 13. A threaded groove is provided in the middle of the guide block 12 and is threadedly connected to the screw rod 13.

[0044] When installing the sensor module 2, a variety of installation methods can be selected. When the plane to be installed is made of iron material and has a smooth surface, the device can be directly installed using the magnet 8 on the back of the mounting seat 1, that is, the magnet 8 on the back of the mounting seat 1 is directly adsorbed to the smooth surface of the iron object to complete the installation of the mounting seat 1. At the same time, the threaded rod 4 can be selectively rotated to disassemble the sensor module 2, and different sensors can be installed on the threaded sleeve 3. When the installation plane is a smooth non-iron plane, the mounting seat 1 can be installed by movably connecting the card strip 7 on the inner side of the adapter 6 with the card slot 5. It is installed between the adapter frame 6, and a suction cup 9 is optionally installed. At this time, the suction cup 9 can be used to complete the installation by adsorption between it and the smooth surface. When the installation surface cannot be magnetically adsorbed or vacuum adsorbed, a longitudinal guide rail 11 can be optionally installed, and the area to be installed is placed between the adapter frame 6 and the clamping plate 10. By rotating the screw rod 13, the guide block 12 can be driven to move up and down relative to the longitudinal guide rail 11. By controlling the clamping plate 10 to approach the direction of the adapter frame 6, the mounting seat 1 can be clamped to the position to be installed, completing the clamping installation process and realizing the installation of the sensor module 2 in various environments.

[0045] By utilizing the magnetic adsorption of the magnet 8, the vacuum adsorption of the suction cup 9, and the clamping between the adapter frame 6 and the clamping plate 10, the sensor can be installed at different positions inside and outside the building, and the installation method can be quickly adjusted only through the adapter frame 6. The whole process can be completed quickly and can effectively accommodate the installation requirements in different positions, effectively avoiding the problem that traditional devices need to use different installation structures to achieve installation. The overall versatility is strong, which effectively reduces the initial layout time and improves the layout efficiency.

[0046] A prediction method of a building energy load prediction device comprises the following steps:

[0047] S1: Install sensor modules, including temperature sensors, humidity sensors, and light sensors, to regularly obtain real-time data of environmental parameters inside and outside the building, install human infrared sensors and surveillance cameras to monitor user behavior, and integrate the data with environmental parameter data;

[0048] S2: Data cleaning and preprocessing: cleaning the raw data collected by the sensor, removing outliers and noise, and performing data interpolation to ensure data quality and continuity;

[0049] S3: Feature extraction, extracting features related to energy consumption from the cleaned data, including time, temperature, humidity, and light, using statistical methods, including but not limited to other feature extraction techniques such as frequency domain analysis;

[0050] S4: Model training and optimization, using machine learning algorithms including but not limited to neural networks, support vector machines or time series analysis models, training on historical data, optimizing the model using cross-validation, finding the best model parameters and structure, and appropriate feature selection methods;

[0051] S5: Real-time update and online learning: The latest sensor data is obtained regularly or in real time and added to the prediction model to maintain the accuracy and timeliness of the model. Based on the real-time updated data, the prediction model is learned and optimized online to adapt to changes in the environment and user behavior.

[0052] S6: Energy load forecasting, based on the trained model, predicts the energy load in the future according to the current environmental parameters and user behavior, using machine learning models or time series analysis models to make predictions based on the patterns and trends of historical data;

[0053] S7: Decision support: present the energy consumption forecast results to building managers or energy managers in an intuitive and easy-to-understand way, helping them to formulate reasonable energy scheduling and energy-saving measures. Based on the forecast results, corresponding energy-saving suggestions are provided, including but not limited to adjusting the equipment operating temperature range and optimizing equipment scheduling strategies.

[0054] S8: Risk assessment: perform error analysis on the prediction results, evaluate the accuracy of the model, and provide confidence intervals or reliability indicators. Generate a risk assessment report, point out possible risks and uncertainties, and provide corresponding treatment suggestions.

[0055] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A building energy load prediction device, characterized in that: It includes a sensor module, a user behavior monitoring component, a data processing and analysis module, a model prediction module, a real-time update and online learning module, a decision support module and a risk assessment module, wherein the output end of the sensor module and the user behavior monitoring component is electrically connected to the input end of the data processing and analysis module, the output end of the data processing and analysis module is electrically connected to the input end of the model prediction module, the output end of the model prediction module is electrically connected to the input end of the real-time update and online learning module, the output end of the real-time update and online learning module is electrically connected to the input end of the decision support module, and the output end of the decision support module is electrically connected to the input end of the risk assessment module; The sensor module includes a temperature sensor, a humidity sensor and a light sensor. The temperature sensor is used to measure the temperature of the environment inside and outside the building. The humidity sensor is used to obtain the humidity level of the environment inside and outside the building. The light sensor is used to monitor the light intensity inside and outside the building. The user behavior monitoring component includes a human infrared sensor and a surveillance camera. The human infrared sensor is used to detect human thermal radiation in the building and identify the activities and positions of people. The surveillance camera is used to monitor the activities of people in the building in real time. The data processing and analysis module includes data cleaning and preprocessing, feature extraction, model training and optimization, and the model prediction module includes a machine learning model and a time series analysis model; The real-time update and online learning module includes real-time data update, online learning and iterative optimization, the decision support module includes energy consumption prediction results and suggestions and strategy support, and the risk assessment module includes prediction error analysis and risk assessment report.

2. A building energy load prediction device according to claim 1, characterized in that: The invention comprises a mounting seat (1), a threaded sleeve (3) is fixedly mounted in the middle of the top of the mounting seat (1), a sensor module (2) is arranged above the mounting seat (1), a threaded rod (4) is fixedly connected in the middle of the bottom of the sensor module (2), the sensor module (2) is threadedly connected to the threaded sleeve (3) via the threaded rod (4), a clamping groove (5) is provided on both the left and right sides of the mounting seat (1), and a magnet (8) is embedded in the middle of the bottom of the mounting seat (1).

3. A building energy load prediction device according to claim 2, characterized in that: An adapter frame (6) is provided on the outer side surface of the mounting seat (1), and clamping strips (7) are installed on both left and right sides of the inner side surface of the adapter frame (6). The adapter frame (6) is movably clamped with the clamping strip (7) and the clamping slot (5). The adapter frame (6) is displaced forward and backward relative to the mounting seat (1), and a suction cup (9) is detachably installed at the bottom end of the adapter frame (6).

4. A building energy load prediction device according to claim 3, characterized in that: A longitudinal guide rail (11) is detachably mounted on one side of the adapter frame (6), a clamping plate (10) is provided below the adapter frame (6), a guide block (12) is fixedly mounted on one side of the clamping plate (10), the clamping plate (10) is movably connected to the longitudinal guide rail (11) through the guide block (12), the clamping plate (10) is displaced up and down relative to the adapter frame (6), a screw rod (13) is movably connected inside the longitudinal guide rail (11), a threaded groove is provided in the middle of the guide block (12) and is threadedly connected to the screw rod (13).

5. The prediction method of a building energy load prediction device according to claim 4, characterized in that: The following steps are involved: S1: Install sensor modules, including temperature sensors, humidity sensors, and light sensors, to regularly obtain real-time data of environmental parameters inside and outside the building, install human infrared sensors and surveillance cameras to monitor user behavior, and integrate the data with environmental parameter data; S2: Data cleaning and preprocessing: cleaning the raw data collected by the sensor, removing outliers and noise, and performing data interpolation to ensure data quality and continuity; S3: Feature extraction, extracting features related to energy consumption from the cleaned data, including time, temperature, humidity, and light, using statistical methods, including but not limited to frequency domain analysis techniques; S4: Model training and optimization, using machine learning algorithms including but not limited to neural networks, support vector machines, and time series analysis models to train historical data, optimize the model using cross-validation, and find the best model parameters and structure; S5: Real-time update and online learning: regularly and in real time obtain the latest sensor data and add it to the prediction model to maintain the accuracy and timeliness of the model. Based on the real-time updated data, the prediction model is learned and iteratively optimized online to adapt to changes in the environment and user behavior. S6: Energy load forecasting, based on the trained model, predicts the energy load in the future according to the current environmental parameters and user behavior, using machine learning models and time series analysis models to make predictions based on the patterns and trends of historical data; S7: Decision support: present the energy consumption forecast results to building managers and energy managers in an intuitive and easy-to-understand way, helping them to formulate reasonable energy scheduling and energy-saving measures. Based on the forecast results, corresponding energy-saving suggestions are provided, including but not limited to adjusting the equipment operating temperature range and optimizing equipment scheduling strategies. S8: Risk assessment: perform error analysis on the prediction results, evaluate the accuracy of the model, provide confidence intervals and reliability indicators, generate a risk assessment report, point out the existing risks and uncertainties, and provide corresponding treatment suggestions.