A teaching building intelligent power management system and method

Abstract

The application discloses a kind of teaching building wisdom power management system and method, belong to involve power management field.The teaching building wisdom power management system includes multiple-way relay, measurement and control circuit and cloud server, and cloud server is according to the historical power consumption data of each electrical equipment, to each electrical equipment is time series prediction and energy consumption prediction, and then according to time series prediction and energy consumption prediction generation control instruction and energy-saving strategy, through measurement and control circuit according to control instruction to multiple-way relay is controlled, realize to each electrical equipment in each classroom control.The application takes the various electrical equipment under the teaching building place as the research range, with energy saving and emission reduction as the research purpose, realizes the automatic energy-saving management of each electrical equipment in teaching building.

Description

Technical Field

[0001] This invention relates to the field of power management, and in particular to a smart power management system and method for teaching buildings. Background Technology

[0002] In the field of education, although the numerous electrical appliances in school buildings provide great convenience for students' learning, they also bring about problems such as excessive energy waste, which directly affect the utilization and management of energy, and make energy conservation and emission reduction work in various places particularly important.

[0003] The electrical equipment in the teaching building includes smart sockets, smart meters, smart lighting systems, and a teaching building energy-saving control system. These products or technologies can solve most electricity problems, but they also have corresponding drawbacks. Smart sockets can be remotely and timed, but cannot achieve refined management. Smart meters can display information such as power, current, and voltage, but require relevant knowledge to detect abnormal operation. Smart lighting systems can recognize when people turn them on and adjust the brightness according to the environment, but the number of electrical devices is limited. The teaching building energy-saving control system can be remotely controlled via an app and computer, but requires human monitoring, posing a risk of inadequate supervision.

[0004] In summary, the main drawbacks of existing products and technologies are: ① they require human control; and ② they are limited to powering only one type of electrical appliance. Therefore, there is an urgent need for an intelligent system suitable for managing the power supply of various electrical appliances in teaching buildings to reduce energy waste caused by human factors. Summary of the Invention

[0005] The purpose of this invention is to provide a smart power management system and method for teaching buildings, which can automatically manage energy conservation of various electrical devices in teaching buildings.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] A smart power management system for a teaching building, the power management system comprising: multi-channel relays, measurement and control circuits, and a cloud server;

[0008] The multiple power input terminals of the measurement and control circuit are connected one-to-one with the multiple output terminals of the mains power, and the multiple power output terminals of the measurement and control circuit are connected one-to-one with the multiple power input terminals of the multi-channel relay; the multiple power output terminals of the multi-channel relay are connected one-to-one with the power supply terminal of each electrical device in each classroom.

[0009] The first signal output terminal of the measurement and control circuit is connected to the signal input terminal of the cloud server; the measurement and control circuit is used to collect the voltage and current signals of each electrical device when it is working, and after obtaining the historical power consumption data of each electrical device based on the voltage and current signals, it remotely sends the data to the cloud server.

[0010] The signal output terminal of the cloud server is connected to the signal input terminal of the measurement and control circuit, and the second signal output terminal of the measurement and control circuit is connected to the control terminal of the multi-channel relay. The cloud server is used to perform time-series prediction and energy consumption prediction for each electrical device based on the historical power consumption data of each electrical device, and then generate control commands and energy-saving strategies based on the time-series prediction and energy consumption prediction. The measurement and control circuit controls the multi-channel relay according to the control commands to realize the control of each electrical device in each classroom.

[0011] A smart power management method for teaching buildings includes:

[0012] Collect historical energy consumption data and historical usage duration data of various electrical devices in the teaching building;

[0013] Input the historical energy consumption data of all electrical equipment in the teaching building into the first Transformer model, and output the predicted energy consumption value of each electrical equipment.

[0014] Input the historical usage time data of all electrical equipment in the teaching building into the second Transformer model, and output the predicted power usage time of each electrical equipment;

[0015] Based on users' historical electricity consumption data, user profiles are constructed;

[0016] Input user profiles, predicted electricity usage time, and predicted energy consumption into the recommendation system, and output a recommendation strategy with energy saving as the goal in the usage scenario.

[0017] Based on the recommended strategy and the power consumption patterns of the equipment, an energy-saving strategy is generated.

[0018] Based on the energy-saving strategy, the electrical equipment in the usage scenario is automatically controlled and adjusted.

[0019] Optionally, historical energy consumption data and historical usage time data of various electrical devices in the teaching building are collected, followed by:

[0020] Data cleaning is performed on historical electricity consumption data; the historical electricity consumption data is historical energy consumption data or historical usage duration data.

[0021] All historical electricity consumption data after cleaning of electrical equipment are integrated into a unified dataset, and then merged, deduplicated, and normalized to obtain an integrated dataset.

[0022] Feature selection and dimensionality reduction are performed on the data in the integrated dataset.

[0023] Optionally, historical energy consumption data of all electrical equipment in the teaching building can be input into the first Transformer model to output predicted energy consumption values ​​for each piece of equipment, specifically including:

[0024] A value encoder is used to perform feature mapping on each historical energy consumption data point to extract value information;

[0025] A time encoder is used to encode the time corresponding to each historical energy consumption data to extract time information;

[0026] By combining all extracted value information and time information, the sequence characteristics of energy consumption data can be obtained;

[0027] Position-coded features are obtained by performing positional encoding on the sequence features of energy consumption data.

[0028] The position-encoded features are input into the multi-head attention layer, encoded, and then linearly mapped to obtain the output features of the multi-head attention layer.

[0029] The output features of the multi-head attention layer are input into the first summation and normalization layer to obtain the output features of the first summation and normalization layer.

[0030] The output features of the first summing and normalizing layer are input into the fully connected layer to obtain the output features of the fully connected layer.

[0031] The output features of the fully connected layer are input into the second summation and normalization layer to obtain the output features of the second summation and normalization layer.

[0032] The output features of the second additive normalization layer are input into the encoding layer for feature encoding, and the output is a context vector containing overall information of all historical energy consumption data.

[0033] Based on the context vector and the energy consumption prediction value of the previous time step, the decoding layer is used to perform feature decoding to generate the energy consumption prediction value of the next time step.

[0034] Optionally, based on the energy-saving strategy, the electrical equipment in the usage scenario is automatically controlled and adjusted, and the system further includes:

[0035] Data analysis and statistics are performed on the electricity consumption of electrical equipment to generate electricity consumption reports and energy-saving effect analysis reports.

[0036] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0037] This invention discloses a smart power management system and method for teaching buildings. A cloud server performs time-series and energy consumption predictions for each electrical device based on historical power consumption data. Then, based on these predictions, control commands and energy-saving strategies are generated. Through a measurement and control circuit, multiple relays are controlled according to the control commands, thereby achieving control over the electrical equipment in each classroom. This invention focuses on various electrical devices in teaching buildings, aiming at energy conservation and emission reduction, and realizes automatic energy-saving management of various electrical devices in teaching buildings.

[0038] The cloud server uses a combination of recommendation system and Transformer model to generate energy-saving strategies, and dynamically adjusts the energy-saving strategies according to actual needs. At the same time, it can avoid resource waste or safety hazards caused by human intervention or misoperation, and realize the autonomous learning and self-optimization of the power system. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a schematic diagram of the structure of a smart power management system for a teaching building provided in Embodiment 1 of the present invention;

[0041] Figure 2 This is a schematic diagram of the measurement and control circuit provided in Embodiment 1 of the present invention;

[0042] Figure 3 This is a schematic diagram of the power supply module provided in Embodiment 1 of the present invention;

[0043] Figure 4 This is a flowchart of a smart power management method for a teaching building provided in Embodiment 2 of the present invention. Detailed Implementation

[0044] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0045] This invention takes intelligent power supply systems as the research object, various electrical devices in teaching buildings as the research scope, and energy conservation and emission reduction as the research objective. It explores the design principles, key technologies, and optimization strategies of intelligent power supply systems for powering various electrical devices in teaching buildings.

[0046] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] Example 1

[0048] like Figure 1 As shown in the figure, an embodiment of the present invention provides a smart power management system for a teaching building, including: a multi-channel relay, a measurement and control circuit, and a cloud server.

[0049] The multiple power input terminals of the measurement and control circuit are connected one-to-one with the multiple output terminals of the mains power, and the multiple power output terminals of the measurement and control circuit are connected one-to-one with the multiple power input terminals of the multi-channel relay; the multiple power output terminals of the multi-channel relay are connected one-to-one with the power supply terminal of each electrical device in each classroom.

[0050] The first signal output terminal of the measurement and control circuit is connected to the signal input terminal of the cloud server; the measurement and control circuit is used to collect the voltage and current signals of each electrical device when it is working, and after obtaining the historical power consumption data of each electrical device based on the voltage and current signals, it remotely sends the data to the cloud server.

[0051] The signal output terminal of the cloud server is connected to the signal input terminal of the measurement and control circuit, and the second signal output terminal of the measurement and control circuit is connected to the control terminal of the multi-channel relay. The cloud server is used to perform time-series prediction and energy consumption prediction for each electrical device based on the historical power consumption data of each electrical device, and then generate control commands and energy-saving strategies based on the time-series prediction and energy consumption prediction. The measurement and control circuit controls the multi-channel relay according to the control commands to realize the control of each electrical device in each classroom.

[0052] Reference Figure 1 Each classroom is equipped with an AC input interface, which connects to multiple loads (electrical equipment) and multiple relays. Each classroom's electrical equipment has its own independent power supply line, which is led out from the multiple relays; each output of the mains power is determined by the number of electrical devices in each teaching building.

[0053] One end of the monitoring and control circuit is connected to a multi-channel relay, and the other end is connected to the mains power (220VAC). Each mains power supply passes through the monitoring and control circuit before being connected to the electrical appliance.

[0054] The cloud server deploys a Transformer model and a recommendation system. Based on the database accumulated through sampling, it makes effective predictions and generates a recommendation list. The Transformer model can learn the power consumption patterns of devices based on historical power consumption data, predicting device usage time and energy consumption. At the same time, the recommendation system analyzes historical power consumption behavior data to better understand the power consumption environment, habits, and time periods of devices, in order to generate more personalized recommendation strategies and scenario lists.

[0055] Based on the prediction results, control commands are generated and transmitted to the main control circuit of the measurement and control circuit to control the relay, thereby effectively preventing power waste.

[0056] For example, the power management system further includes a mini-program terminal. The mini-program terminal is connected to a cloud server. The mini-program terminal is used to record students' usage habits of various electrical devices during self-study and stores these usage habits on the cloud server. The cloud server is used to perform time-series prediction and energy consumption prediction for each electrical device based on the usage habits, and generate control instructions and energy-saving strategies based on the time-series prediction and energy consumption prediction. The cloud server is also used to compare the actual usage time of the electrical devices. When the actual usage time exceeds the predicted time, a warning message will be sent to the mini-program terminal to remind students to turn off the electrical devices. If the electrical devices are not turned off within a preset time period after the warning message is sent, a shutdown control instruction is generated. The measurement and control circuit controls multiple relays according to the shutdown control instruction to realize the shutdown of the electrical devices.

[0057] Students can use the mini-program to make reasonable use of various electrical devices in the teaching building, and each student's usage habits for various electrical devices will be stored in the cloud server.

[0058] The mini-program primarily provides an energy-saving mode for students studying in a classroom. The original predictions were based on historical electricity usage patterns (before and after class), which are predictable. However, students' self-study breaks these patterns, causing conflicts with the predictions. The mini-program aims to resolve these conflicts by generating energy-saving strategies based on students' usual self-study habits. For example, students might prefer to study in a corner, only turn on the corner light, and typically leave after a few hours.

[0059] Furthermore, the power management system also includes a web-based interface. The web interface is connected to a cloud server. The cloud server is used to obtain the actual energy consumption of the electrical equipment. If the actual energy consumption exceeds the predicted energy consumption threshold range, an alarm message will be sent to the web interface.

[0060] Administrators can use the web interface to monitor the usage of each classroom in the teaching building so that they can report emergency information in a timely manner.

[0061] The recommendation list is sent to the web / mini-program platform. The recommendation list mainly recommends the best device settings to save electricity. For example, it tells system users what parameters, such as the air conditioner temperature and fan speed, can save electricity at a certain time. This is equivalent to giving users feedback and seeing whether they adopt it.

[0062] like Figure 2 As shown, the measurement and control circuit includes: a multi-channel sampling circuit and a microcontroller.

[0063] The multiple power input terminals of the multi-channel sampling circuit are connected one-to-one with the multiple power output terminals of the mains power supply, and the multiple power output terminals of the multi-channel sampling circuit are connected one-to-one with the multiple power input terminals of the multi-channel relay. The signal output terminal of the multi-channel sampling circuit is connected to the first signal input terminal of the microcontroller. The multi-channel sampling circuit is used to collect voltage and current signals from each electrical device during operation. The first signal output terminal of the microcontroller is connected to the signal input terminal of the cloud server. The microcontroller is used to obtain historical power consumption data of each electrical device based on the voltage and current signals and then remotely send it to the cloud server. The signal output terminal of the cloud server is connected to the second signal input terminal of the microcontroller, and the second signal output terminal of the microcontroller is connected to the control terminal of the multi-channel relay. The microcontroller is used to control the multi-channel relay according to the control instructions from the cloud server, thereby controlling each electrical device in each classroom.

[0064] Each output of the electrical equipment has a unique ID number. This means that the ID numbers of the electrical equipment in each classroom are independent. The voltage and current data for each output are acquired by a pin of the microcontroller through the microcontroller chip's built-in ADC, and the processed signal is then sent to the sampling circuit. In other words, a pin of the microcontroller is also associated with the ID of each power line, thus establishing a one-to-one correspondence. The microcontroller is responsible for converting the voltage and current information into other electrical parameters such as power and energy consumption. After processing the information, the microcontroller uploads it to a cloud server for storage via an IoT module. Simultaneously, the microcontroller also receives instructions from the cloud server and effectively controls the relays accordingly, thereby effectively preventing energy waste.

[0065] In the cloud server, each piece of electrical equipment in each teaching building has a series of parameters under its ID number, including but not limited to the equipment ID, start time, usage duration, and power consumption. In addition, each student also has a unique ID and corresponding data—the student's usage habits (obtained through a mini-program to address irregular student study patterns in the teaching building, avoiding conflicts with predictions trained based on actual weekdays, rest days, holidays, and free time—this is an effective measure). These include, but are not limited to, the equipment ID, start time, and usage duration. During the model's prediction process, it calls the current usage time of a particular piece of electrical equipment from the database and predicts its usage duration. If the usage duration is exceeded, a warning message is sent to the student, reminding them to turn off the equipment. If the student does not respond, the model automatically turns off the equipment. The model can also predict the energy consumption of each piece of electrical equipment. If it exceeds the predicted energy consumption threshold, an alarm message is sent to the administrator. The administrator can promptly receive the pushed alarm message through the web interface and dispatch an electrician to repair the malfunctioning equipment.

[0066] For example, the power management system further includes an Internet of Things (IoT) module and a power supply module. The measurement and control circuit is connected to a cloud server via the IoT module. The power supply module draws power from the mains and supplies power to the measurement and control circuit, the multiple relays, and the IoT module.

[0067] like Figure 3 As shown, the power supply module includes: an isolated step-down rectifier bridge, a BUCK circuit, and a linear regulator. The isolated step-down rectifier bridge is used to isolate and rectify the mains power, outputting DC power to the BUCK circuit. The BUCK circuit is used to step down the DC power. The linear regulator is used to regulate the stepped-down DC power, outputting the regulated DC power to the measurement and control circuit, multiplex relays, and the Internet of Things (IoT) module.

[0068] Figure 3 The area within the dashed box represents the power supply module. Due to the limited number of modules and low power consumption in this system's hardware circuitry, active power factor correction is unnecessary when isolating and rectifying the mains power. A simple rectifier bridge is sufficient to achieve DC output, thus reducing the system's hardware cost to some extent. Considering efficiency, a DC-DC circuit—a BUCK topology—is used after rectification to reduce the voltage to a suitable level for the IoT module and relays. To ensure the microcontroller can achieve high-precision ADC sampling and that the sampling circuit chip operates stably, a low-noise, low-ripple linear regulator is used to provide the operating voltage, avoiding interference from switching noise in the DC-DC circuit.

[0069] To address the energy waste caused by students' improper electricity use and inadequate supervision by administrators in current teaching buildings, a smart power management system based on a recommendation system and Transformer technology to generate energy-saving strategies is proposed. Its innovations are as follows:

[0070] (1) Energy saving strategy: Energy saving strategy is generated by combining recommendation system and Transformer model. The energy saving strategy is dynamically adjusted according to actual needs. At the same time, it can avoid resource waste or safety hazards caused by human intervention or misoperation, and realize the autonomous learning and self-optimization of power system.

[0071] (2) Intelligent control: By equipping an IoT module and providing a mini-program and backend management platform for remote control, it helps managers improve efficiency.

[0072] (3) High flexibility and strong scalability: The system adopts the "platform + hardware" design mode, which can adapt to the needs of electrical equipment in different scenarios.

[0073] (4) Wide equipment capacity: Increase the sample diversity in the model dataset and expand the types of equipment vertically.

[0074] Example 2

[0075] This invention provides a smart power management method for teaching buildings, such as... Figure 4 As shown, it includes:

[0076] Step 1: Collect historical energy consumption data and historical usage time data of each electrical device in the teaching building.

[0077] Data Acquisition:

[0078] Historical power consumption data (including voltage, current, energy consumption, switch status, and usage duration) of various electrical devices in the teaching building (each device in each classroom has a unique ID) is measured using measurement and control circuits and stored in a database as a training set for time series and energy consumption prediction. Simultaneously, usage patterns of these devices are acquired using IoT-enabled devices to provide effective solutions for generating subsequent energy-saving strategies. Energy consumption and usage duration are the targets for prediction.

[0079] Among them, the usage mode is the device configuration, such as the air conditioner's fan speed and target temperature, etc.

[0080] Data processing:

[0081] Due to measurement errors, signal interference, and equipment failures in the collection of electricity consumption data, the data obtained by the server contains a large amount of noisy data. Therefore, it is necessary to clean and integrate the original dataset to generate a high-quality dataset, which will facilitate better modeling, training, and prediction in the future.

[0082] Specifically, the data cleaning stage requires the use of statistics, visualization, and domain knowledge to identify and process outliers, missing values, and duplicate values, such as removing them or replacing them with the mean; the data integration stage requires integrating the different electrical equipment in different classrooms measured by different measuring devices into a unified dataset, and performing merging, deduplication, and normalization; the data transformation stage requires the use of feature selection, dimensionality reduction, and other techniques to transform the raw data into a form more suitable for analysis and modeling.

[0083] Data can be obtained directly from the cloud via the measurement and control circuit and the internal controller of the IoT device, or it can be retrieved from the database.

[0084] Historical energy consumption data and historical usage duration data of various equipment in the teaching building were collected. The data was preprocessed (data cleaning and data integration) to obtain a high-quality, usable dataset, which was used as input data for the feature embedding layer of Transformer prediction.

[0085] Step 2: Input the historical energy consumption data of all electrical equipment in the teaching building into the first Transformer model, and output the predicted energy consumption value of each electrical equipment.

[0086] Compared to LSTM and GRU in the field, using the Transformer model can guarantee the correctness and accuracy of predictions.

[0087] This invention uses the same method for predicting equipment energy consumption and usage duration. Here, we take equipment energy consumption prediction as an example to introduce the equipment energy consumption prediction process.

[0088] Step 2.1: For each historical energy consumption value in the sequence, use the value encoder E v Perform feature mapping to extract value information;

[0089] Step 2.2: For the time information in the sequence, a time encoder E is used. t The time corresponding to each data item in the sequence is encoded to extract the time information. This is because the energy consumption of the teaching building equipment has a strong time correlation. For each energy consumption data item X... iFirst, extract the time information. Assuming the time is 2023-06-01 12:30, the extracted information includes: June 1st, Thursday, 12:30 PM. This can be represented by a numerical vector as [6, 1, 4, 12, 30]. For the minute information, in actual processing, 15 minutes are used as a time interval, that is, 1 hour is divided into 4 segments. Therefore, when inputting the time encoder, this time information should be represented as [6, 1, 4, 12, 2].

[0090] (In addition to electricity consumption data and time information, supplementary features containing other relevant information, such as weather data and seasonal factors, can be introduced. These features can enhance the model's ability to predict energy consumption.)

[0091] Step 2.3: Combine the extracted value information and time information to obtain the sequence feature representation of the energy consumption data as X. seq =E v (X)+E t (X); where E v (X) represents the value information, E t (X) represents time information;

[0092] Step 2.4: Perform position encoding on the above sequence features to obtain position-encoded features. The position encoder uses a position encoding function for calculation, the position encoding function e t Represented as

[0093]

[0094] Where, d m The dimension of the sequence feature is represented by t, which represents the odd or even position in the sequence feature. Odd positions are processed using a cosine function, and even positions are processed using a sine function. i represents the index of the positional coding feature, with values ​​of 0, 1, 2, ...

[0095] Step 2.5: Input the position-encoded feature X into the multi-head attention layer, encode the position-encoded feature, and then perform a linear mapping to obtain the output feature of the multi-head attention layer, represented as Multihead(H) = concat(head1,...,head1). u W 0 Where Multihead represents the output features of the multihead attention layer, H represents the number of attention heads, and W represents the number of attention heads. 0 Represents the weight vector, head1,...,head u This represents the output of u self-attention layers in a multi-head attention layer.

[0096]

[0097] in, W Q W k W V All of these are learnable parameters.

[0098] Step 2.6: Input the output features of the multi-head attention layer into the first summation normalization layer to obtain the output features of the first summation normalization layer.

[0099] Step 2.7: Input the output features of the first summed normalized layer into the fully connected layer to obtain the output features of the fully connected layer.

[0100] Step 2.8: Input the output features of the fully connected layer into the second summation normalization layer to obtain the output features of the second summation normalization layer.

[0101] Step 2.9: Based on historical experience and the number of encoders obtained through repeated parameter tuning during deep learning training, repeat steps 2.5 to 2.8, input the obtained output features into the encoding layer for feature encoding, and output a context vector containing overall information of the input sequence, which is used as the input of the decoder.

[0102] Step 2.10: Input the energy consumption data sequence to be predicted into the decoder for feature decoding. The specific process is the same as steps 2.5 to 2.8. The decoder generates the energy consumption prediction value for each time step in the target sequence step by step based on the context vector output by the encoder and the prediction result of the previous time step.

[0103] Step 3: Input the historical usage time data of all electrical equipment in the teaching building into the second Transformer model, and output the predicted power usage time value of each electrical equipment.

[0104] Step 4: Build a user profile based on the user's historical electricity consumption data.

[0105] Customized User Profiles: In recommendation systems, to provide personalized recommendation services, it is necessary to build user profiles, which involves describing and modeling user characteristics and preferences. This includes basic user information, electricity consumption behavior, preferences, and goals. By collecting and analyzing users' historical electricity consumption data, combined with user feedback and behavioral data, fine-grained user profiles can be constructed.

[0106] Step 5: Input the user profile, predicted electricity usage time, and predicted energy consumption into the recommendation system, and output a recommendation strategy with energy saving as the goal in the usage scenario.

[0107] Applications of Data Mining Technology: Data mining technology plays a crucial role in recommender systems. By using Transformers to mine historical electricity consumption data of users and devices, it is possible to discover correlations, consumption patterns, and characteristics between users and devices. Simultaneously, using Transformers to predict future electricity consumption allows for dynamic optimization of energy-saving strategies.

[0108] Inputs and Outputs: The inputs to the recommendation system include user profile information and electricity consumption data. User profile information includes the user's personal characteristics, electricity consumption preferences, and goals. Electricity consumption data includes indicators such as electricity usage time, electricity consumption, and electricity usage patterns. Based on these inputs, the output of the recommendation system is to recommend device parameter settings according to the usage scenario to achieve energy saving.

[0109] Usage Scenario Recommendations: The output of the recommendation system is customized based on specific usage scenarios. For example, in a teaching scenario, the system can recommend energy-efficient home appliances based on the user's electricity consumption behavior and preferences to reduce energy consumption. Simultaneously, based on the user's daily routine and behavioral patterns, the system can also recommend suitable electricity usage patterns and timed control strategies to further optimize energy efficiency.

[0110] Step 6: Based on the recommended strategy and the power consumption patterns of the equipment, generate an energy-saving strategy.

[0111] Energy-saving strategy generation stage: Based on the output results of the recommendation system model and combined with the power consumption patterns of the equipment, energy-saving strategies are generated, including information such as the switching time of the equipment, usage mode, and aging warning, in order to achieve the best energy-saving effect.

[0112] Step 7: Based on the energy-saving strategy, automatically control and adjust the electrical equipment in the usage scenario.

[0113] Energy-saving strategy execution phase: The generated energy-saving strategy is applied to the scene devices to realize automatic control and adjustment of the devices, thereby achieving intelligence and energy saving.

[0114] Energy-saving effect report generation: Data analysis and statistics on the power consumption of equipment are performed to generate power consumption reports and energy-saving effect analysis reports. Users can provide feedback and monitor the energy-saving effect in real time, helping them to better understand the power consumption and energy-saving effect of the equipment, and further optimize energy-saving strategies. The energy-saving effect report can also provide decision-making basis for managers.

[0115] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Regarding the methods disclosed in the embodiments, since they correspond to the systems disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the system section description.

[0116] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A smart power management system for teaching buildings, characterized in that, The power management system includes: multi-channel relays, measurement and control circuits, and a cloud server; The multiple power input terminals of the measurement and control circuit are connected one-to-one with the multiple output terminals of the mains power, and the multiple power output terminals of the measurement and control circuit are connected one-to-one with the multiple power input terminals of the multi-channel relay; the multiple power output terminals of the multi-channel relay are connected one-to-one with the power supply terminal of each electrical device in each classroom. The first signal output terminal of the measurement and control circuit is connected to the signal input terminal of the cloud server; the measurement and control circuit is used to collect the voltage and current signals of each electrical device when it is working, and after obtaining the historical power consumption data of each electrical device based on the voltage and current signals, it remotely sends the data to the cloud server. The signal output terminal of the cloud server is connected to the signal input terminal of the measurement and control circuit, and the second signal output terminal of the measurement and control circuit is connected to the control terminal of the multi-channel relay. The cloud server is used to perform time-series prediction and energy consumption prediction for each electrical device based on the historical power consumption data of each electrical device, and then generate control commands and energy-saving strategies based on the time-series prediction and energy consumption prediction. The measurement and control circuit controls the multi-channel relay according to the control commands to realize the control of each electrical device in each classroom. A Transformer model and a recommendation system are deployed on the cloud server. Historical energy consumption data of all electrical equipment in the teaching building are input into the first Transformer model, which outputs predicted energy consumption values ​​for each device. Historical usage duration data of all electrical equipment in the teaching building are input into the second Transformer model, which outputs predicted usage duration values ​​for each device. User profiles are constructed based on users' historical electricity consumption data. User profiles, predicted usage duration values, and predicted energy consumption values ​​are input into the recommendation system, which outputs energy-saving recommendation strategies for specific usage scenarios. Based on these recommendation strategies and the electricity consumption patterns of the devices, energy-saving strategies are generated. Based on these energy-saving strategies, the electrical equipment in the usage scenario is automatically controlled and adjusted.

2. The intelligent power management system for teaching buildings according to claim 1, characterized in that, The power management system also includes: a mini-program terminal; The mini-program connects to the cloud server; The mini-program is used to record students' usage habits of various electrical devices during self-study and to store these usage habits on a cloud server. The cloud server is used to perform time-series prediction and energy consumption prediction for each electrical device according to the usage habits, and to generate control instructions and energy-saving strategies based on the time-series prediction and energy consumption prediction. The cloud server is also used to compare the actual usage time of the electrical equipment. When the actual usage time exceeds the predicted time, a warning message will be sent to the mini-program to remind students to turn off the electrical equipment. If the electrical equipment is not turned off within a preset time period after the warning message is sent, a shutdown control command will be generated. The measurement and control circuit will control multiple relays according to the shutdown control command to turn off the electrical equipment.

3. The intelligent power management system for teaching buildings according to claim 1, characterized in that, The power management system also includes: a web interface; Connect the web client to the cloud server; The cloud server is used to obtain the actual energy consumption of electrical equipment. If the actual energy consumption exceeds the threshold range of the predicted energy consumption, an alarm message will be sent to the web terminal.

4. The intelligent power management system for teaching buildings according to claim 1, characterized in that, The measurement and control circuit includes: a multi-channel sampling circuit and a microcontroller; The multiple power input terminals of the multi-channel sampling circuit are connected one-to-one with the multiple power output terminals of the mains power, and the multiple power output terminals of the multi-channel sampling circuit are connected one-to-one with the multiple power input terminals of the multi-channel relay; the signal output terminal of the multi-channel sampling circuit is connected to the first signal input terminal of the microcontroller; the multi-channel sampling circuit is used to collect the voltage and current signals of each electrical device during operation. The first signal output terminal of the microcontroller is connected to the signal input terminal of the cloud server. The microcontroller is used to obtain the historical power consumption data of each electrical device based on the voltage signal and the current signal, and then remotely send it to the cloud server. The signal output terminal of the cloud server is connected to the second signal input terminal of the microcontroller, and the second signal output terminal of the microcontroller is connected to the control terminal of the multi-channel relay. The microcontroller is used to control the multi-channel relay according to the control instructions of the cloud server, so as to control each electrical device in each classroom.

5. The intelligent power management system for teaching buildings according to claim 1, characterized in that, The power management system also includes: an Internet of Things (IoT) module and a power supply module; The measurement and control circuit is connected to the cloud server via an Internet of Things (IoT) module; The power supply module draws power from the mains and supplies it to the measurement and control circuit, multiple relays, and the Internet of Things module.

6. The intelligent power management system for teaching buildings according to claim 5, characterized in that, The power supply module includes: an isolated step-down rectifier bridge, a BUCK circuit, and a linear regulator; The isolated buck rectifier bridge is used to isolate and rectify the mains power and output DC power to the BUCK circuit; The BUCK circuit is used to step down the DC voltage; Linear regulators are used to regulate the stepped-down DC power and output the regulated DC power to measurement and control circuits, multi-channel relays, and Internet of Things (IoT) modules.

7. A method for intelligent power management in a teaching building, characterized in that, include: Collect historical energy consumption data and historical usage duration data of various electrical devices in the teaching building; Input the historical energy consumption data of all electrical equipment in the teaching building into the first Transformer model, and output the predicted energy consumption value of each electrical equipment. Input the historical usage time data of all electrical equipment in the teaching building into the second Transformer model, and output the predicted power usage time of each electrical equipment; Based on users' historical electricity consumption data, user profiles are constructed; Input user profiles, predicted electricity usage time, and predicted energy consumption into the recommendation system, and output a recommendation strategy with energy saving as the goal in the usage scenario. Based on the recommended strategy and the power consumption patterns of the equipment, an energy-saving strategy is generated. Based on the energy-saving strategy, the electrical equipment in the usage scenario is automatically controlled and adjusted.

8. The intelligent power management method for teaching buildings according to claim 7, characterized in that, Collect historical energy consumption data and historical usage time data of various electrical devices in the teaching building, and then include: Data cleaning is performed on historical electricity consumption data; the historical electricity consumption data is historical energy consumption data or historical usage duration data. All historical electricity consumption data after cleaning of electrical equipment are integrated into a unified dataset, and then merged, deduplicated, and normalized to obtain an integrated dataset. Feature selection and dimensionality reduction are performed on the data in the integrated dataset.

9. The intelligent power management method for teaching buildings according to claim 7, characterized in that, The historical energy consumption data of all electrical equipment in the teaching building are input into the first Transformer model, and the predicted energy consumption values ​​of each electrical device are output, including: A value encoder is used to perform feature mapping on each historical energy consumption data point to extract value information; A time encoder is used to encode the time corresponding to each historical energy consumption data to extract time information; By combining all extracted value information and time information, the sequence characteristics of energy consumption data can be obtained; Position-coded features are obtained by performing positional encoding on the sequence features of energy consumption data. The position-encoded features are input into the multi-head attention layer, encoded, and then linearly mapped to obtain the output features of the multi-head attention layer. The output features of the multi-head attention layer are input into the first summation and normalization layer to obtain the output features of the first summation and normalization layer. The output features of the first summing and normalizing layer are input into the fully connected layer to obtain the output features of the fully connected layer. The output features of the fully connected layer are input into the second summation and normalization layer to obtain the output features of the second summation and normalization layer. The output features of the second additive normalization layer are input into the encoding layer for feature encoding, and the output is a context vector containing overall information of all historical energy consumption data. Based on the context vector and the energy consumption prediction value of the previous time step, the decoding layer is used to perform feature decoding to generate the energy consumption prediction value of the next time step.

10. The intelligent power management method for teaching buildings according to claim 7, characterized in that, Based on the energy-saving strategy, the system automatically controls and adjusts the electrical equipment in the usage scenario, and then further includes: Data analysis and statistics are performed on the electricity consumption of electrical equipment to generate electricity consumption reports and energy-saving effect analysis reports.

Images