Artificial Intelligence-Based Integrated Intelligent Control Method, Device, and Medium for Building Energy Consumption

Abstract

This application relates to the fields of data processing and artificial intelligence, and provides a method, device, and medium for comprehensive intelligent control of building energy consumption based on artificial intelligence. The method includes: acquiring historical energy consumption data, historical personnel distribution data, historical environmental data, current energy consumption data, current environmental data, and current personnel distribution data corresponding to the building to be tested; constructing a target energy consumption state analysis model based on the historical energy consumption data, historical personnel distribution data, and historical environmental data; performing energy consumption state analysis based on the target energy consumption state analysis model, according to the current energy consumption data, current environmental data, and current personnel distribution data, to obtain target energy consumption state data; and generating an intelligent energy consumption control strategy corresponding to the building to be tested based on the target energy consumption state data. This can yield a more accurate and comprehensive intelligent energy consumption control strategy, which is beneficial to improving the accuracy of the intelligent energy consumption control process for the building to be tested.

Description

Technical Field

[0001] This application relates to the fields of data processing and artificial intelligence technology, specifically to a method, device, and medium for comprehensive intelligent control of building energy consumption based on artificial intelligence. Background Technology

[0002] Existing building energy control systems generally suffer from problems such as crude control strategies and insufficient real-time response capabilities. For example, traditional solutions often rely on fixed thresholds or manual experience to set equipment operating parameters. Air conditioning systems, for instance, only start and stop based on preset temperatures, failing to dynamically adapt to real-time variables such as changes in occupancy density and fluctuations in the outdoor environment. This not only results in both energy waste and decreased comfort but also fails to achieve accurate and comprehensive energy control. Therefore, improving the accuracy of intelligent energy control in buildings has become an urgent problem to be solved. Summary of the Invention

[0003] This application provides an artificial intelligence-based method, device, and medium for comprehensive intelligent control of building energy consumption. It can obtain more accurate target energy consumption status data through the constructed target energy consumption status analysis model, thereby generating a more accurate and comprehensive intelligent energy consumption control strategy and improving the accuracy of the intelligent energy consumption control process for the building to be tested.

[0004] The first aspect of this application provides a comprehensive intelligent control method for building energy consumption based on artificial intelligence, the method comprising:

[0005] Obtain historical energy consumption data, historical personnel distribution data, historical environmental data, current energy consumption data, current environmental data, and current personnel distribution data for the building to be tested;

[0006] Based on historical energy consumption data, historical personnel distribution data, and historical environmental data, a target energy consumption status analysis model is constructed.

[0007] Based on the target energy consumption status analysis model, energy consumption status analysis is performed according to current energy consumption-related data, current environmental-related data, and current personnel distribution data to obtain target energy consumption status data;

[0008] Based on the target energy consumption status data, generate intelligent energy consumption control strategies for the buildings to be monitored.

[0009] A second aspect of this application provides an artificial intelligence-based integrated intelligent control device for building energy consumption, the device comprising:

[0010] The acquisition unit is used to acquire historical energy consumption data, historical personnel distribution data, historical environmental data, current energy consumption data, current environmental data, and current personnel distribution data for the building to be tested.

[0011] The first processing unit is used to construct a target energy consumption status analysis model based on the historical energy consumption data, the historical personnel distribution data, and the historical environmental data.

[0012] The second processing unit is used to perform energy consumption status analysis based on the target energy consumption status analysis model, according to the current energy consumption related data, the current environment related data, and the current personnel distribution data, to obtain target energy consumption status data;

[0013] The third processing unit is used to generate an intelligent energy consumption control strategy for the building to be detected based on the target energy consumption status data.

[0014] A third aspect of this application provides a terminal including a processor, an input device, an output device, and a memory, wherein the processor, input device, output device, and memory are interconnected, wherein the memory is used to store a computer program, the computer program including program instructions, and the processor is configured to invoke the program instructions to execute the step instructions as described in the first aspect of this application.

[0015] A fourth aspect of this application provides a computer-readable storage medium storing a computer program for electronic data interchange, wherein the computer program causes a computer to perform some or all of the steps described in the first aspect of this application.

[0016] A fifth aspect of this application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in the first aspect of this application. The computer program product may be a software installation package.

[0017] Implementing the embodiments of this application has the following beneficial effects:

[0018] By acquiring historical energy consumption data, historical personnel distribution data, historical environmental data, current energy consumption data, current environmental data, and current personnel distribution data for the building under test, a target energy consumption status analysis model can be constructed based on the historical energy consumption data, historical personnel distribution data, and historical environmental data. Then, based on this target energy consumption status analysis model, energy consumption status analysis can be performed using the current energy consumption data, current environmental data, and current personnel distribution data to obtain target energy consumption status data. Furthermore, based on this target energy consumption status data, an intelligent energy consumption control strategy for the building under test can be generated. This leads to a more accurate and comprehensive intelligent energy consumption control strategy, improving the accuracy of the intelligent energy consumption control process for the building under test. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This application provides a flowchart illustrating an artificial intelligence-based integrated intelligent control method for building energy consumption.

[0021] Figure 2 This is a schematic diagram of the structure of a terminal provided in an embodiment of this application;

[0022] Figure 3 This application provides a schematic diagram of the structure of an artificial intelligence-based integrated intelligent control device for building energy consumption. Detailed Implementation

[0023] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0025] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application can be combined with other embodiments.

[0026] To better understand the AI-based intelligent building energy consumption control method provided in this application, a brief introduction to existing AI-based intelligent building energy consumption control methods is given below. In existing solutions, at the data utilization level, most energy consumption control systems only collect total energy consumption data, lacking multi-dimensional monitoring of sub-items of energy consumption (such as lighting, elevators, and air conditioning) and environmental parameters (such as light intensity and carbon dioxide concentration), making it difficult to accurately locate abnormal energy consumption nodes. Furthermore, existing models often employ static rule engines, which cannot learn complex energy consumption patterns from historical data. For example, they cannot identify the differentiated demands of "peak office hours on weekday mornings" and "low load on weekend nights," causing control strategies to lag behind actual scenario changes.

[0027] Furthermore, weak anomaly detection capabilities are another key pain point. Traditional solutions rely on single threshold alarms, which are susceptible to data noise and cannot distinguish between anomalies such as equipment malfunctions and operational errors, resulting in high false alarm rates and high troubleshooting costs. With the increasing demand for building intelligence, the limitations of existing technologies in dynamic modeling, multi-source data fusion, and real-time decision-making are becoming increasingly apparent. There is an urgent need to introduce refined control methods based on artificial intelligence to achieve both energy consumption optimization and improved management efficiency.

[0028] To address the aforementioned issues, this application provides an artificial intelligence-based comprehensive intelligent control method for building energy consumption. This method can construct a target energy consumption state analysis model based on historical energy consumption data, historical personnel distribution data, and historical environmental data. This model allows for the acquisition of more accurate target energy consumption state data, which in turn enables the generation of more accurate and comprehensive intelligent energy consumption control strategies. This improves the accuracy of the intelligent energy consumption control process for the building under test.

[0029] Please see Figure 1 , Figure 1 This application provides a flowchart illustrating an artificial intelligence-based intelligent control method for building energy consumption. Figure 1 As shown, the method includes:

[0030] S10: Obtain historical energy consumption data, historical personnel distribution data, historical environmental data, current energy consumption data, current environmental data, and current personnel distribution data for the building to be tested.

[0031] The building to be monitored can be understood as any building requiring comprehensive intelligent control of building energy consumption based on artificial intelligence; this application does not impose any restrictions on this. Historical energy consumption data can be understood as the energy consumption data of the building to be monitored over a historical period. This historical energy consumption data may include, but is not limited to, real-time values, cumulative values, peak / valley values, and energy consumption curves of various energy consumption types such as electricity, gas, and water. For example, historical energy consumption data may include the hourly office electricity consumption data of office building A throughout 2024; this application does not impose any restrictions on this. By obtaining historical energy consumption data, it can be used in subsequent steps to train energy consumption prediction models, mine historical energy consumption patterns (such as weekday / weekend differences, seasonal fluctuations), and analyze the correlation between energy consumption and human activities and environmental factors (such as the relationship between air conditioning energy consumption and outdoor temperature) in subsequent steps.

[0032] Historical personnel distribution data can include information such as the number of people, duration of stay, and movement trajectories in various areas of the building under test during historical time periods. Specifically, this historical personnel distribution data can be collected and obtained through access control systems, cameras (combined with computer vision), etc. For example, historical personnel distribution data can include a heat map of personnel density distribution on the third floor of an office building at 14:00 on a weekday. By obtaining historical personnel distribution data, densely populated areas and time periods can be identified (e.g., meeting rooms are often occupied from 9:00 to 11:00), and energy consumption peaks (e.g., increased air conditioning load during this period) can be further correlated in subsequent steps, which is beneficial for optimizing energy allocation strategies (e.g., automatically reducing lighting / air conditioning power in unoccupied areas). This application does not impose any limitations on this.

[0033] Historical environmental data can be understood as external environmental data related to the energy consumption of the building under test, which may include, but is not limited to, temperature, humidity, light intensity, and wind speed. For example, historical environmental data may include daily maximum temperature data, humidity data, and wind speed data for City B in the summer of 2024. By acquiring historical environmental data, the impact of environmental factors on energy consumption can be analyzed (e.g., high outdoor temperatures leading to a 30% increase in air conditioning energy consumption), and this data can be used as input features in subsequent steps during model training to improve the accuracy of energy consumption prediction (e.g., optimizing control strategies by combining weather forecast data).

[0034] Current energy consumption data can be understood as the real-time energy consumption data of the building under test at the current moment or within a recent time period. Optionally, current energy consumption data can be collected and obtained in real time through devices such as smart meters and sensors. For example, current energy consumption data may include the total real-time electricity consumption of the entire building at 10:00 AM on May 16, 2025. By obtaining current energy consumption data, energy consumption status can be monitored in real time, abnormal fluctuations can be identified (such as a sudden surge in electricity consumption in a certain area), and immediate feedback can be provided for dynamic control in subsequent steps (such as automatically switching energy supply modes when overload is detected).

[0035] Current environmental data can be understood as external environmental parameters at the current moment, such as real-time temperature, humidity, light intensity, and weather conditions (sunny / rainy / snowy). For example, current environmental data may include a current outdoor temperature of 32°C and humidity of 65%. By acquiring current environmental data, the operating parameters of the equipment can be adjusted in subsequent steps in conjunction with the real-time environment (such as automatically dimming indoor lighting when there is sufficient outdoor light). This further enables the trained model to determine more suitable control strategies based on current environmental data, such as triggering emergency control strategies during heavy rain or automatically activating backup power monitoring. This application does not impose any limitations on this.

[0036] Current personnel distribution data can be understood as the real-time distribution of personnel in various areas of the building to be monitored. Optionally, this current personnel distribution data can be acquired and obtained in real time through IoT devices (such as Bluetooth beacons or infrared sensors). For example, this current personnel distribution data could include the number of people in office area A at 14:30 on May 16, 2025, while meeting room B is empty. By acquiring current personnel distribution data, the operating status of public facilities can be dynamically adjusted (such as automatically turning off lights and air conditioning in unoccupied meeting rooms), which is beneficial for optimizing energy allocation priorities (such as prioritizing heating and cooling supply in densely populated areas).

[0037] S20: Construct a target energy consumption status analysis model based on the historical energy consumption data, the historical personnel distribution data, and the historical environmental data.

[0038] The target energy consumption state analysis model can be understood as a model built based on artificial intelligence algorithms (such as machine learning and deep learning) to analyze energy consumption patterns and establish a mapping relationship between energy consumption and factors such as personnel and the environment. Specifically, by inputting historical energy consumption data, personnel distribution characteristics, and environmental characteristics (such as temperature and timestamps), a target energy consumption prediction module for energy consumption prediction processing, a target energy consumption pattern classification module for energy consumption pattern classification processing, and a target energy consumption anomaly detection module for energy consumption anomaly detection processing can be constructed. These three modules can then be further integrated to obtain the target energy consumption state analysis model capable of performing energy consumption state analysis.

[0039] By constructing this target energy consumption status analysis model, we can obtain target energy consumption status data to achieve a reasonable energy consumption range for the current state (such as predicting the reasonable energy consumption value of air conditioning based on the current personnel and environment), and to classify energy consumption patterns and identify abnormal energy consumption patterns (such as judging abnormality when the actual energy consumption exceeds the prediction range). This is beneficial for generating intelligent energy consumption control strategies for the buildings to be monitored based on the target energy consumption status data.

[0040] S30: Based on the target energy consumption status analysis model, perform energy consumption status analysis according to the current energy consumption related data, the current environment related data, and the current personnel distribution data to obtain target energy consumption status data.

[0041] After obtaining the target energy consumption status analysis model, the model can be used to perform real-time calculations on the current data (i.e., the current energy consumption-related data, the current environment-related data, and the current personnel distribution data) to determine whether the building's energy consumption is in a normal state, such as quantifying energy efficiency, predicting energy consumption, classifying energy consumption patterns, and identifying abnormal indicators.

[0042] The target energy consumption status data may include, but is not limited to, predicted energy consumption values, such as the current reasonable energy consumption level predicted based on the target energy consumption status analysis model (e.g., the air conditioning system is predicted to consume 100kW currently), or the predicted future energy consumption level (e.g., the air conditioning system is predicted to consume 150kW in the future); actual-prediction deviation, such as the difference between actual energy consumption and predicted value (e.g., actual consumption is 120kW, predicted consumption is 100kW, deviation +20%); energy consumption mode category, such as the current low energy consumption mode; and energy consumption anomaly indicators, such as classification based on the degree of deviation (e.g., slight anomaly, severe anomaly), or determination of energy consumption anomaly based on the high energy consumption mode identification results of empty conference rooms. This application does not impose any restrictions on these.

[0043] S40: Generate an intelligent energy consumption control strategy for the building to be detected based on the target energy consumption status data.

[0044] Furthermore, based on the energy consumption status analysis results (i.e., target energy consumption status data), an optimized energy use operation plan for the building under test can be generated, which is the intelligent energy consumption control strategy. This strategy aims to better and more accurately control the energy consumption of the building under test, thereby reducing energy consumption, improving energy utilization efficiency, and eliminating abnormal energy consumption.

[0045] Intelligent energy consumption control strategies may include, but are not limited to, adjustment strategies for equipment parameters, such as air conditioning system adjustment strategies, such as dynamically adjusting the temperature setpoint based on personnel density (e.g., adjusting the temperature from 24℃ to 26℃ when personnel decrease), lighting system adjustment strategies, such as automatically adjusting brightness through light sensors, or zoned control based on personnel distribution (e.g., turning off 80% of the lighting in unoccupied areas); energy dispatch optimization strategies, such as peak-shifting power consumption strategies, such as operating high-power equipment (e.g., water heaters, charging piles) during off-peak electricity price periods (e.g., at night), multi-energy coordination strategies, such as combining solar energy, energy storage batteries, etc., and prioritizing the use of renewable energy (e.g., reducing grid power purchases when photovoltaic power meets 30% of the load); and abnormal response mechanisms, such as automatically alarming when energy consumption is abnormal and triggering a fault investigation process (e.g., dispatching a work order to maintenance personnel), such as automatically cutting off non-critical loads in emergencies (e.g., only keeping power to fire-fighting equipment supplied during a fire). This application does not impose any restrictions on these.

[0046] The method provided in this application can realize a closed loop of "data-driven modeling → real-time status analysis → intelligent strategy execution", which can significantly improve the intelligence level of building energy consumption management; reduce unnecessary energy waste (such as standby power loss of equipment in unmanned areas) through dynamic regulation, which can further save energy and reduce costs; automated analysis and decision-making can replace manual inspection, which is conducive to shortening the response time of anomalies and improving management efficiency; and provide data support for the buildings to be inspected to obtain green certification (such as pioneer certification in energy and environmental design, and the Building Research Establishment Environmental Assessment Method (BREEAM)).

[0047] In this embodiment, by acquiring historical energy consumption data, historical personnel distribution data, historical environmental data, current energy consumption data, current environmental data, and current personnel distribution data corresponding to the building to be tested, a target energy consumption status analysis model can be constructed based on the historical energy consumption data, historical personnel distribution data, and historical environmental data. Then, based on the target energy consumption status analysis model, energy consumption status analysis can be performed according to the current energy consumption data, current environmental data, and current personnel distribution data to obtain target energy consumption status data. Furthermore, based on the target energy consumption status data, an intelligent energy consumption control strategy corresponding to the building to be tested can be generated, which is beneficial for obtaining a more accurate and comprehensive intelligent energy consumption control strategy and can improve the accuracy of the intelligent energy consumption control process for the building to be tested.

[0048] In one possible implementation, when constructing the target energy consumption state analysis model, a target energy consumption prediction module for energy consumption prediction processing, a target energy consumption pattern classification module for energy consumption pattern classification processing, and a target energy consumption anomaly detection module for energy consumption anomaly detection processing can be constructed separately, thereby further obtaining the target energy consumption state analysis model. Specifically, a method for constructing a target energy consumption state analysis model based on the historical energy consumption related data, the historical personnel distribution data, and the historical environmental related data includes:

[0049] A1. Preprocess the historical energy consumption data, the historical personnel distribution data, and the historical environment data to obtain a set of historical training sample data.

[0050] A2. Perform energy consumption prediction processing based on the historical training sample data set to obtain the target energy consumption prediction module;

[0051] A3. Perform energy consumption pattern classification processing based on the historical training sample data set to obtain the target energy consumption pattern classification module;

[0052] A4. Perform energy consumption anomaly detection processing based on the historical training sample data set to obtain the target energy consumption anomaly detection module;

[0053] A5. Construct a target energy consumption status analysis model based on the target energy consumption prediction module, the target energy consumption pattern classification module, and the target energy consumption anomaly detection module.

[0054] Data preprocessing can be understood as cleaning, transforming, and extracting features from raw historical energy consumption, personnel distribution, and environmental data to make them suitable for model training. Specifically, missing value handling can be employed, such as using interpolation methods (e.g., linear interpolation, Kalman filtering) to fill missing values ​​in sensor data; outlier filtering can be used, such as detecting and correcting abnormal energy consumption data points using the three-standard-deviation (3sigma, 3σ) rule or isolation forest; data standardization can be used, such as normalizing features of different dimensions (e.g., temperature, number of people, energy consumption) to the same scale to avoid the model being biased towards high-value features; and time series alignment can be used, such as unifying data with different sampling frequencies (e.g., 15-minute energy consumption data and hourly personnel data) to the same time granularity. This application does not impose any restrictions on this.

[0055] The historical training sample dataset can be understood as a multidimensional feature dataset used for training the model after preprocessing. This historical training sample dataset may contain, but is not limited to, various feature vectors, such as energy consumption features: total energy consumption, energy consumption by component (air conditioning / lighting / elevator), energy consumption change rate; personnel features: population density in each area, peak hour distribution, and dwell time; environmental features: temperature, humidity, light intensity, and weather type (sunny / rainy / snowy); or it may also include label data, such as energy consumption mode classification labels (e.g., "office peak mode", "nighttime low energy consumption mode"), or anomaly detection labels (1 = abnormal, 0 = normal), etc. This application does not impose any restrictions on this.

[0056] Energy consumption prediction processing can be understood as the process of predicting energy consumption values ​​at a specific future point in time or over a period of time based on historical data (i.e., a set of historical training sample data). By performing energy consumption prediction processing based on the historical training sample data set, a target energy consumption prediction module can be obtained. Specifically, it can be based on time-series models, such as Autoregressive Integrated Moving Average (ARIMA) or Seasonal Autoregressive Integrated Moving Average (SARIMA), to capture the seasonality and trends of energy consumption data; or on Long Short-Term Memory (LSTM) networks or Gated Recurrent Units (GRUs), to handle long-term and short-term dependencies; or through ensemble learning, such as extreme gradient boosting (XGBoost) or Light Gradient Boosting Machine (LightGBM), to fuse multiple features (such as temperature, time, and personnel) for prediction. This application does not impose any limitations on this approach.

[0057] The target energy consumption prediction module can be understood as the energy consumption prediction model obtained after training. This target energy consumption prediction module can perform point predictions, such as the energy consumption value at a specific time (e.g., predicting that the air conditioner energy consumption at 10:00 AM tomorrow will be 85kW), interval predictions, such as the confidence interval of the predicted energy consumption (e.g., a 95% confidence interval [80kW, 90kW]), or uncertainty quantification, such as the prediction error distribution, to evaluate the reliability of the prediction. This application does not impose any limitations on these aspects.

[0058] Energy consumption pattern classification can be understood as the process of dividing energy consumption patterns into categories based on historical training sample data, in order to identify different pattern categories and determine typical energy consumption behaviors. Specifically, unsupervised learning can be used, such as the K-means clustering algorithm (K-means), to cluster based on the shape of the energy consumption curve (e.g., "stable pattern" or "fluctuating pattern"), or the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm, to discover densely connected pattern regions and handle irregular distributions. Feature extraction methods can be used, such as extracting time-domain features, which can be achieved by calculating the mean, standard deviation, kurtosis, etc.; extracting frequency-domain features, which can be achieved by using Fourier transform to extract periodic (e.g., weekday / weekend pattern) features. This application does not impose any restrictions on these methods.

[0059] The target energy consumption pattern classification module can be understood as a trained energy consumption pattern classification model. This module can perform pattern recognition, such as determining which predefined pattern the current energy consumption data belongs to (e.g., "lunch break mode," "equipment maintenance mode," etc.); it can perform pattern evolution analysis, such as tracking pattern changes over time (e.g., differences between summer and winter patterns); and it can perform pattern rule generation, such as outputting classification rules (e.g., "temperature > 30℃ and personnel density > 0.8 people / m²"). 2 →Air conditioning peak mode), this application does not restrict this.

[0060] Energy consumption anomaly detection and processing can be understood as the process of identifying energy consumption data points or sequences that deviate from the normal state. Specifically, statistical methods, such as the 3σ rule, can be used, where data exceeding the mean ± 3 times the standard deviation are considered anomalies; machine learning methods, such as isolated forests, can quickly identify outliers through tree structures; and autoencoders can reconstruct data with errors exceeding a threshold, which can be considered anomalies. This application does not impose any restrictions on these methods.

[0061] The target energy consumption anomaly detection module can be understood as a trained anomaly detection system. This target energy consumption anomaly detection module can perform real-time monitoring, such as judging anomalies in real time through real-time energy consumption data; it can locate anomalies, such as pointing out the area where the anomaly occurred (e.g., the air conditioning system on a certain floor) and the time point; it can classify anomalies, such as classifying them according to severity (e.g., mild anomaly → yellow warning, severe anomaly → red alarm), etc., and this application does not impose any limitations on these aspects.

[0062] A multi-module fusion mechanism can be used to further integrate the three sub-modules of energy consumption prediction, energy consumption pattern classification, and energy consumption anomaly detection to form a target energy consumption state analysis model with comprehensive analytical capabilities. Specifically, the different outputs of the target energy consumption prediction module, the target energy consumption pattern classification module, and the target energy consumption anomaly detection module can be output in parallel to obtain the output of the target energy consumption state analysis model. Optionally, a decision tree fusion approach can be used to formulate priority rules based on the outputs of the target energy consumption prediction module, the target energy consumption pattern classification module, and the target energy consumption anomaly detection module, and output the results according to the priority rules to obtain the output of the target energy consumption state analysis model.

[0063] Through the collaboration of the above multiple modules, the resulting target energy consumption status analysis model can achieve comprehensive perception of energy consumption of the building under test, energy consumption prediction, energy consumption pattern classification and anomaly identification, and can provide data-driven decision support for subsequent intelligent regulation.

[0064] In one possible implementation, when constructing the target energy consumption prediction module, a supervised learning modeling framework can be used to train and obtain a target energy consumption prediction module that can accurately predict energy consumption using historical data (i.e., a historical training sample dataset). The target energy consumption prediction module can also be understood as a target energy consumption prediction model, which is not limited in this application. Optionally, its construction steps can follow a machine learning process of "data preprocessing → feature engineering → model definition → model training," which is not limited in this application. Specifically, a method for performing energy consumption prediction processing based on the historical training sample dataset to obtain a target energy consumption prediction module includes:

[0065] B1. Extract energy consumption prediction-related features from each historical training sample data in the historical training sample data set to obtain an energy consumption prediction feature set;

[0066] B2. Evaluate the importance of each energy consumption prediction feature in the energy consumption prediction feature set to obtain an important energy consumption prediction feature set;

[0067] B3. Perform key feature screening on each important performance consumption prediction feature in the set of important performance consumption prediction features to obtain a set of key energy consumption prediction features.

[0068] B4. Based on the initial energy consumption prediction module and the energy consumption prediction task, construct the energy consumption prediction loss function;

[0069] B5. Train the initial energy consumption prediction module based on the key energy consumption prediction feature set and the energy consumption prediction loss function to obtain the target energy consumption prediction module.

[0070] Specifically, by extracting features relevant to the energy consumption prediction task (such as time, temperature, equipment power, and number of personnel) from historical training sample data, an energy consumption prediction feature set can be formed. By extracting and processing these energy consumption prediction-related features, the original data (historical training sample data) can be transformed into input dimensions that the energy consumption prediction module can understand.

[0071] Statistical methods or evaluation algorithms (such as correlation coefficients, random forest feature importance, Shapley Additive Explanations (SHAP) values, etc.) can be used to assess the contribution of each feature to energy consumption prediction, thereby obtaining a set of important energy consumption prediction features. This allows for the selection of highly important features and the removal of redundant or irrelevant features in subsequent steps. By evaluating the importance of each energy consumption prediction feature, not only can the feature dimensionality be reduced and computational complexity decreased, but the generalization ability of the constructed target energy consumption prediction module can also be improved. Furthermore, importance evaluation can also identify key factors affecting energy consumption; for example, it may be determined that "outdoor temperature" has a greater impact than "date," but this application does not impose any limitations on this.

[0072] After obtaining the set of important performance consumption prediction features through importance assessment, the most representative core features can be further selected to obtain the key energy consumption prediction feature set. Specifically, this can be achieved through feature filtering based on business knowledge, such as eliminating redundant features: for example, if both "temperature ℃" and "temperature ℉" exist, only one needs to be retained; or through feature simplification based on statistical analysis, such as using a correlation coefficient matrix to calculate the correlation coefficient between features and then eliminating highly correlated features; or through cross-validation, such as k-fold cross-validation (e.g., k=5), to compare the generalization ability of the module under different feature subsets. Through key feature selection, the energy consumption prediction feature set can be further optimized to balance module efficiency and performance.

[0073] The initial energy consumption prediction module can employ linear regression, neural network, XGBoost, or other similar methods; this application does not impose any restrictions on this. The energy consumption prediction task can be a regression-type task. Based on the type of the initial energy consumption prediction module and the type of the energy consumption prediction task, a function can be constructed to measure the difference between the predicted and actual values ​​of the energy prediction module; this function is the energy consumption prediction loss function.

[0074] Optionally, taking the initial energy consumption prediction module as a linear regression module and the energy consumption prediction task as a regression-type task as an example, the process of constructing the energy consumption prediction loss function based on the initial energy consumption prediction module and the energy consumption prediction task can be seen in the following formula:

[0075]

[0076] Among them, L f Energy consumption prediction loss function; n is the number of key energy consumption prediction features in the key energy consumption prediction feature set; i is the index of the key energy consumption prediction feature; y i The actual energy consumption value of the i-th key energy consumption prediction feature; The predicted energy consumption value for the i-th key energy consumption prediction feature; E is the squared error of the i-th key energy consumption prediction feature; x This application does not impose any restrictions on the pre-set energy consumption threshold, such as 1.5 times the average energy consumption value.

[0077] It should be noted that in the formula for the energy consumption prediction loss function above, when the actual energy consumption y i When the energy consumption threshold is exceeded, the weight is 2, and the error can be amplified; when y i When the energy consumption threshold is not exceeded, the weight is 1, and the error can remain unchanged. In energy consumption prediction, prediction errors during high energy consumption periods (such as summer peaks or full-load operation of industrial equipment) may lead to more serious resource waste or equipment failure. Therefore, this application pays more attention to high energy consumption scenarios to achieve more accurate predictions.

[0078] Using the selected set of key energy consumption prediction features as input, the loss function is minimized through an optimization algorithm (such as gradient descent), and the parameters (such as weights and biases) of the initial energy consumption prediction module are iteratively updated until the module converges or reaches a preset performance index, thus obtaining the target energy consumption prediction module. It can be understood that this target energy consumption prediction module can learn the mapping relationship between features and energy consumption from historical training sample data sets, and is an energy consumption prediction module with energy prediction capabilities.

[0079] In one possible implementation, when constructing the target energy consumption pattern classification module, energy consumption pattern-related features can first be extracted from historical training samples to form a feature set. Then, a clustering algorithm is used to perform unsupervised partitioning of the feature set to obtain energy consumption pattern category data. The category data and feature set can then be used as labels and inputs to train an initial classifier. The model performance can then be evaluated using a validation set, and parameters or features can be optimized or adjusted to obtain a target energy consumption pattern classification module that can accurately identify different energy consumption patterns (such as peak office hours and low nighttime consumption). A method for performing energy consumption pattern classification processing based on the historical training sample data set to obtain a target energy consumption pattern classification module includes:

[0080] C1. Extract basic features from each historical training sample in the historical training sample data set to obtain a basic energy consumption pattern feature set.

[0081] C2. Perform feature transformation processing on each historical training sample data in the historical training sample data set to obtain a derived energy consumption mode feature set.

[0082] C3. Perform energy consumption pattern clustering processing based on the basic energy consumption pattern feature set and the derived energy consumption pattern feature set to obtain energy consumption pattern category data;

[0083] C4. Based on the energy consumption mode category data, the basic energy consumption mode feature set, and the derived energy consumption mode feature set, train the initial classifier to obtain the initial energy consumption mode classification module.

[0084] C5. Verify and optimize the initial energy consumption classification module to obtain the target energy consumption mode classification module.

[0085] The basic energy consumption pattern feature set can include one or more basic energy consumption pattern features. These features can be understood as features with clear physical meaning related to energy consumption patterns, directly extracted from historical training sample data sets. Specifically, they can include statistical features such as mean, standard deviation, maximum value, minimum value, kurtosis, and skewness; time-domain features such as slope (energy consumption change rate), inflection point (energy consumption mutation point), and periodic indicators (such as Fourier transform frequency); and environmental correlation features such as the correlation coefficients between environmental parameters such as temperature, humidity, and light intensity and energy consumption.

[0086] For example, “daily average energy consumption” and “hourly fluctuation coefficient” can be extracted from air conditioning energy consumption data, and “peak hour population density” and “regional activity” can be extracted from personnel distribution data, thereby obtaining the above-mentioned basic energy consumption pattern feature set.

[0087] The derived energy consumption pattern feature set may include one or more derived energy consumption pattern features. These derived energy consumption pattern features can be understood as features related to energy consumption patterns that are generated through mathematical transformations or combinations, abstracting or enhancing the original features (such as basic energy consumption pattern features). Specifically, transformation methods may include, but are not limited to, dimensionality reduction transformations, such as Principal Component Analysis (PCA), nonlinear dimensionality reduction methods (t-distributed stochastic neighbor embedding (t-SNE)); temporal transformations, such as differencing (energy consumption difference between adjacent times), moving average (smoothing energy consumption curves); feature cross-referencing, such as multiplying "temperature" and "air conditioning energy consumption" to generate a "temperature energy consumption index"; and deep learning features, such as low-dimensional representations extracted by autoencoders and temporal patterns learned by LSTM. This application does not impose any limitations on these.

[0088] For example, wavelet transform can be used to decompose historical training sample data into sub-bands of different frequencies to extract "high-frequency fluctuation features"; or the "energy consumption ratio when the temperature is >28℃" can be calculated to capture the energy consumption characteristics under high temperature conditions.

[0089] Energy consumption pattern category data can be understood as unsupervised classification labels formed by dividing energy consumption patterns into multiple different categories based on feature sets (i.e., the basic energy consumption pattern feature set and the derived energy consumption pattern feature set). Specifically, distance-based clustering, such as K-means and DBSCAN, can be used; hierarchical clustering, such as agglomerative clustering (gradually merging from single points) and divisive clustering (gradually splitting from the whole), can be used; and probability-based clustering, such as Gaussian Mixture Model (GMM), can output the probability of each sample belonging to each category. This application does not impose any restrictions on this.

[0090] The energy consumption pattern category data may include cluster centers, such as feature vectors of typical energy consumption patterns, and category labels, such as “office peak mode”, “nighttime maintenance mode”, or “weekend low energy consumption mode”, etc. This application does not impose any restrictions on this.

[0091] The initial energy consumption pattern classification module can be understood as a model used for pattern classification by performing supervised energy consumption pattern classification processing on the basic energy consumption pattern feature set and the derived energy consumption pattern feature set based on the clustering results (i.e., energy consumption pattern category data). Specifically, the feature set (basic energy consumption pattern feature set + derived energy consumption pattern feature set) can be used as input x, and the cluster label as y; and the initial classifier can be trained by dividing it into a training set (70%), a validation set (15%), and a test set (15%) to obtain the target energy consumption pattern classification module. Specifically, the initial classifier can be a classifier built based on decision trees, random forests, or support vector machines (SVM), and this application does not impose any restrictions on this.

[0092] Various validation metrics can be used to validate the initial energy consumption pattern classification module, thereby obtaining a more accurate target energy consumption pattern classification module. For example, accuracy (the proportion of correctly classified samples), F1 score (which balances precision and recall to address class imbalance), and Kappa coefficient (which considers the impact of random guessing and evaluates the actual performance of the classifier) ​​can be used. This application does not impose any restrictions on these metrics.

[0093] In one possible implementation, when constructing the target energy consumption anomaly detection module, it is necessary to first clarify the definition of energy consumption anomalies and business objectives, collect and preprocess energy consumption and related data, further select a suitable algorithm for modeling, and then detect data in real time after training and optimization to trigger hierarchical alarms according to rules. Finally, the model is iteratively optimized based on feedback to achieve closed-loop management from data to anomaly response. Specifically, a method for performing energy consumption anomaly detection processing based on the historical training sample data set to obtain the target energy consumption anomaly detection module includes:

[0094] D1. Perform anomaly feature identification processing on each historical training sample data in the historical training sample data set to obtain an energy consumption anomaly feature data set.

[0095] D2. Define anomaly labels for each energy consumption anomaly feature in the energy consumption anomaly feature data set to obtain energy consumption anomaly label data;

[0096] D3. Based on the energy consumption anomaly feature data set and the energy consumption anomaly label data, construct and obtain the energy consumption anomaly loss function;

[0097] D4. Based on the energy consumption anomaly loss function, train the initial energy consumption anomaly detection module to obtain the target energy consumption anomaly detection module.

[0098] Anomaly feature identification can be understood as extracting features that reflect energy consumption anomalies from the original training samples. The energy consumption anomaly feature dataset may include one or more energy consumption anomaly feature data. This energy consumption anomaly feature data can be understood as data containing anomalies after feature identification, which can usually be expressed as statistical characteristics or patterns that deviate from the normal pattern. Specifically, anomalies in statistical characteristics can be identified, such as using the Z-score / Standard Score method to calculate the standard deviation multiple of each data point from the mean (e.g., |Z|>3 is considered an anomaly), or using moving window statistics, such as calculating the maximum, minimum, and variance within the window (e.g., a sudden increase in energy consumption variance within 1 hour); or anomalies in time-series characteristics, such as identifying the rate of change of trends, such as energy consumption growth exceeding a threshold (e.g., a 20% increase compared to the previous week), seasonal deviations, such as differences compared to historical data for the same period (e.g., air conditioning energy consumption on summer weekdays is 30% lower than the historical average); or anomalies in correlation characteristics, such as environmental mismatch, such as excessively high air conditioning energy consumption at an outdoor temperature of 25°C (normally it should increase with rising temperature), or abnormal equipment status, such as energy consumption occurring even when an elevator is not running, etc. This application does not impose any restrictions on these aspects.

[0099] Anomaly label definition can be achieved by assigning a label (usually a binary label: 1 = anomaly, 0 = normal) to each energy consumption anomaly feature data. Specifically, it can be done using a threshold-based approach or an experience-based approach, such as manually labeled typical anomaly cases (e.g., energy consumption data during equipment failure periods), or by using established business rules (e.g., "energy consumption during non-working hours > baseline value 50% → anomaly"), to generate labels and thus obtain energy consumption anomaly label data. This application does not impose any restrictions on this.

[0100] The energy consumption anomaly tag dataset may include one or more energy consumption anomaly tag datasets, which can be data containing anomaly tags. It is understood that this energy consumption anomaly tag dataset typically corresponds to an energy consumption anomaly feature dataset.

[0101] The energy consumption anomaly loss function can be used to quantify the difference between the model's predicted anomaly labels and the true labels, guiding the model to learn to distinguish between normal and abnormal patterns. Specifically, it can be constructed using classification losses used in supervised learning, such as binary cross-entropy loss; or reconstruction impairments used in unsupervised or self-supervised learning, such as mean squared error loss, etc. This application does not impose any restrictions on this.

[0102] The initial energy consumption anomaly detection module can be a supervised learning model, such as logistic regression, random forest, or gradient boosting tree, or it can be an unsupervised or self-supervised model, such as an autoencoder, variational autoencoder (VAE), or isolated forest. This application does not impose any restrictions on this.

[0103] The trained target energy consumption anomaly detection module can provide early warnings of abnormal equipment, such as detecting decreased efficiency of air conditioning compressors or electrical leakage; it can detect energy waste, such as identifying idling of equipment in unoccupied areas or excessive lighting; and it can identify energy fraud, such as detecting illegal electricity use or abnormal metering equipment. By promptly detecting and resolving anomalies, unnecessary energy consumption is reduced, achieving the goal of energy conservation and cost reduction; predictive maintenance reduces unexpected failures, extends equipment life, and achieves better equipment maintenance optimization; and it ensures that energy use complies with regulatory requirements, avoiding safety hazards.

[0104] For examples consistent with the above embodiments, please refer to... Figure 2 , Figure 2 This is a schematic diagram of the structure of a terminal provided in an embodiment of this application, such as... Figure 2 As shown, it includes a processor, an input device, an output device, and a memory, which are interconnected. The memory is used to store a computer program, which includes program instructions. The processor is configured to call the program instructions. The program includes instructions for performing the following steps.

[0105] Obtain the electricity demand information of the target microgrid on the consumer side, and obtain the network parameter information of the target microgrid;

[0106] The carbon emission cost information and energy consumption cost information of the target microgrid are determined based on the network parameter information.

[0107] The target cost function of the target microgrid is determined based on the electricity demand information on the electricity consumption side, the carbon emission cost information, and the energy consumption cost information.

[0108] The optimized particle swarm optimization algorithm is used to solve the low-carbon scheduling model using the objective cost function to obtain the low-carbon scheduling parameters.

[0109] The target microgrid is scheduled using the aforementioned low-carbon scheduling parameters.

[0110] The above mainly describes the solutions of the embodiments of this application from the perspective of the method execution process. It is understood that, in order to achieve the above functions, the terminal includes the corresponding hardware structure and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the units and algorithm steps of the various examples described in the embodiments provided herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0111] This application embodiment can divide the terminal into functional units according to the above method example. For example, each function can be divided into a separate functional unit, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or as a software functional unit. It should be noted that the unit division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0112] For those consistent with the above, please refer to Figure 3 , Figure 3 This application provides a schematic diagram of the structure of an artificial intelligence-based intelligent building energy consumption control device. For example... Figure 3 As shown, the device includes:

[0113] The acquisition unit 101 is used to acquire historical energy consumption data, historical personnel distribution data, historical environmental data, current energy consumption data, current environmental data, and current personnel distribution data corresponding to the building to be detected.

[0114] The first processing unit 102 is used to construct a target energy consumption status analysis model based on the historical energy consumption data, the historical personnel distribution data, and the historical environmental data.

[0115] The second processing unit 103 is used to perform energy consumption status analysis based on the target energy consumption status analysis model, according to the current energy consumption related data, the current environment related data and the current personnel distribution data, to obtain target energy consumption status data.

[0116] The third processing unit 104 is used to generate an intelligent energy consumption control strategy for the building to be detected based on the target energy consumption status data.

[0117] In one possible implementation, the first processing unit 102 is configured to construct a target energy consumption status analysis model based on the historical energy consumption data, the historical personnel distribution data, and the historical environmental data, specifically for:

[0118] The historical energy consumption data, the historical personnel distribution data, and the historical environment data are preprocessed to obtain a set of historical training sample data.

[0119] Based on the historical training sample data set, energy consumption prediction processing is performed to obtain the target energy consumption prediction module;

[0120] Based on the historical training sample data set, energy consumption pattern classification processing is performed to obtain the target energy consumption pattern classification module;

[0121] Based on the historical training sample data set, energy consumption anomaly detection processing is performed to obtain the target energy consumption anomaly detection module;

[0122] A target energy consumption status analysis model is constructed based on the target energy consumption prediction module, the target energy consumption pattern classification module, and the target energy consumption anomaly detection module.

[0123] In one possible implementation, the first processing unit 102 is configured to perform energy consumption prediction processing based on the historical training sample data set to obtain a target energy consumption prediction module, specifically for:

[0124] Energy consumption prediction-related features are extracted from each historical training sample in the historical training sample data set to obtain an energy consumption prediction feature set.

[0125] The importance of each energy consumption prediction feature in the energy consumption prediction feature set is evaluated to obtain the important energy consumption prediction feature set;

[0126] For each important performance consumption prediction feature in the set of important performance consumption prediction features, key features are filtered to obtain a set of key energy consumption prediction features.

[0127] Based on the initial energy consumption prediction module and the energy consumption prediction task, an energy consumption prediction loss function is constructed.

[0128] The initial energy consumption prediction module is trained based on the key energy consumption prediction feature set and the energy consumption prediction loss function to obtain the target energy consumption prediction module.

[0129] In one possible implementation, the first processing unit 102 is configured to perform energy consumption pattern classification processing based on the historical training sample data set to obtain a target energy consumption pattern classification module, specifically for:

[0130] Basic features are extracted from each historical training sample in the historical training sample data set to obtain a basic energy consumption pattern feature set.

[0131] Perform feature transformation processing on each historical training sample data in the historical training sample data set to obtain a derived energy consumption pattern feature set.

[0132] Energy consumption pattern clustering is performed based on the basic energy consumption pattern feature set and the derived energy consumption pattern feature set to obtain energy consumption pattern category data.

[0133] The initial classifier is trained based on the energy consumption mode category data, the basic energy consumption mode feature set, and the derived energy consumption mode feature set to obtain the initial energy consumption mode classification module.

[0134] The initial energy consumption classification module is verified and optimized to obtain the target energy consumption mode classification module.

[0135] In one possible implementation, the first processing unit 102 is configured to perform energy consumption anomaly detection processing based on the historical training sample data set to obtain a target energy consumption anomaly detection module, specifically for:

[0136] Anomaly feature identification processing is performed on each historical training sample data in the historical training sample data set to obtain an energy consumption anomaly feature data set.

[0137] An anomaly label is defined for each energy consumption anomaly feature data in the energy consumption anomaly feature data set to obtain an energy consumption anomaly label data set.

[0138] Based on the energy consumption anomaly feature data set and the energy consumption anomaly label data set, an energy consumption anomaly loss function is constructed and obtained;

[0139] The initial energy consumption anomaly detection module is trained based on the energy consumption anomaly loss function to obtain the target energy consumption anomaly detection module.

[0140] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the artificial intelligence-based integrated intelligent control methods for building energy consumption described in the above method embodiments.

[0141] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program that causes a computer to perform some or all of the steps of any of the artificial intelligence-based integrated intelligent control methods for building energy consumption described in the above method embodiments.

[0142] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0143] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0144] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0145] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0146] Furthermore, the functional units in the various embodiments of the application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software program module.

[0147] If the integrated unit is implemented as a software program module and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0148] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage device, which may include: a flash drive, a read-only memory, a random access memory, a magnetic disk, or an optical disk, etc.

[0149] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A comprehensive intelligent control method for building energy consumption based on artificial intelligence, characterized in that, The method includes: Obtain historical energy consumption data, historical personnel distribution data, historical environmental data, current energy consumption data, current environmental data, and current personnel distribution data for the building to be tested; Based on the historical energy consumption data, the historical personnel distribution data, and the historical environmental data, a target energy consumption status analysis model is constructed. Based on the target energy consumption status analysis model, energy consumption status analysis is performed according to the current energy consumption related data, the current environment related data, and the current personnel distribution data to obtain target energy consumption status data; Based on the target energy consumption status data, an intelligent energy consumption control strategy is generated for the building to be tested. The step of constructing a target energy consumption status analysis model based on the historical energy consumption data, the historical personnel distribution data, and the historical environmental data includes: The historical energy consumption data, the historical personnel distribution data, and the historical environment data are preprocessed to obtain a set of historical training sample data. Energy consumption prediction is performed based on the historical training sample data set to obtain the target energy consumption prediction module. Based on the historical training sample data set, energy consumption pattern classification processing is performed to obtain the target energy consumption pattern classification module; Based on the historical training sample data set, energy consumption anomaly detection processing is performed to obtain the target energy consumption anomaly detection module; Based on the target energy consumption prediction module, the target energy consumption pattern classification module, and the target energy consumption anomaly detection module, a target energy consumption status analysis model is constructed. The step of performing energy consumption anomaly detection processing based on the historical training sample data set to obtain the target energy consumption anomaly detection module includes: Anomaly feature identification processing is performed on each historical training sample data in the historical training sample data set to obtain an energy consumption anomaly feature data set. An anomaly label is defined for each energy consumption anomaly feature data in the energy consumption anomaly feature data set to obtain an energy consumption anomaly label data set. Based on the energy consumption anomaly feature data set and the energy consumption anomaly label data set, an energy consumption anomaly loss function is constructed and obtained; The initial energy consumption anomaly detection module is trained based on the energy consumption anomaly loss function to obtain the target energy consumption anomaly detection module.

2. The intelligent building energy consumption control method based on artificial intelligence according to claim 1, characterized in that, The energy consumption prediction module, which performs energy consumption prediction processing based on the historical training sample data set to obtain the target energy consumption prediction module, includes: Energy consumption prediction-related features are extracted from each historical training sample in the historical training sample data set to obtain an energy consumption prediction feature set. The importance of each energy consumption prediction feature in the energy consumption prediction feature set is evaluated to obtain the important energy consumption prediction feature set; For each important performance consumption prediction feature in the set of important performance consumption prediction features, key features are filtered to obtain a set of key energy consumption prediction features. Based on the initial energy consumption prediction module and the energy consumption prediction task, an energy consumption prediction loss function is constructed. The initial energy consumption prediction module is trained based on the key energy consumption prediction feature set and the energy consumption prediction loss function to obtain the target energy consumption prediction module.

3. The building energy consumption integrated intelligent control method based on artificial intelligence according to claim 2, characterized in that, The energy consumption pattern classification module, which performs energy consumption pattern classification processing based on the historical training sample data set to obtain the target energy consumption pattern classification module, includes: Basic features are extracted from each historical training sample in the historical training sample data set to obtain a basic energy consumption pattern feature set. Perform feature transformation processing on each historical training sample data in the historical training sample data set to obtain a derived energy consumption pattern feature set. Energy consumption pattern clustering is performed based on the basic energy consumption pattern feature set and the derived energy consumption pattern feature set to obtain energy consumption pattern category data. The initial classifier is trained based on the energy consumption mode category data, the basic energy consumption mode feature set, and the derived energy consumption mode feature set to obtain the initial energy consumption mode classification module. The initial energy consumption mode classification module is verified and optimized to obtain the target energy consumption mode classification module.

4. A building energy consumption integrated intelligent control device based on artificial intelligence, characterized in that, The device includes: The acquisition unit is used to acquire historical energy consumption data, historical personnel distribution data, historical environmental data, current energy consumption data, current environmental data, and current personnel distribution data for the building to be tested. The first processing unit is used to construct a target energy consumption status analysis model based on the historical energy consumption data, the historical personnel distribution data, and the historical environmental data. The second processing unit is used to perform energy consumption status analysis based on the target energy consumption status analysis model, according to the current energy consumption related data, the current environment related data, and the current personnel distribution data, to obtain target energy consumption status data; The third processing unit is used to generate an intelligent energy consumption control strategy for the building to be detected based on the target energy consumption status data. The first processing unit is configured to construct a target energy consumption status analysis model based on the historical energy consumption data, the historical personnel distribution data, and the historical environmental data, specifically for: The historical energy consumption data, the historical personnel distribution data, and the historical environment data are preprocessed to obtain a set of historical training sample data. Energy consumption prediction is performed based on the historical training sample data set to obtain the target energy consumption prediction module. Based on the historical training sample data set, energy consumption pattern classification processing is performed to obtain the target energy consumption pattern classification module; Based on the historical training sample data set, energy consumption anomaly detection processing is performed to obtain the target energy consumption anomaly detection module; Based on the target energy consumption prediction module, the target energy consumption pattern classification module, and the target energy consumption anomaly detection module, a target energy consumption status analysis model is constructed. The first processing unit is used to perform energy consumption anomaly detection processing based on the historical training sample data set to obtain the target energy consumption anomaly detection module, specifically for: Anomaly feature identification processing is performed on each historical training sample data in the historical training sample data set to obtain an energy consumption anomaly feature data set. An anomaly label is defined for each energy consumption anomaly feature data in the energy consumption anomaly feature data set to obtain an energy consumption anomaly label data set. Based on the energy consumption anomaly feature data set and the energy consumption anomaly label data set, an energy consumption anomaly loss function is constructed and obtained; The initial energy consumption anomaly detection module is trained based on the energy consumption anomaly loss function to obtain the target energy consumption anomaly detection module.

5. The building energy consumption integrated intelligent control device based on artificial intelligence according to claim 4, characterized in that, The first processing unit is used to perform energy consumption prediction processing based on the historical training sample data set to obtain the target energy consumption prediction module, specifically for: Energy consumption prediction-related features are extracted from each historical training sample in the historical training sample data set to obtain an energy consumption prediction feature set. The importance of each energy consumption prediction feature in the energy consumption prediction feature set is evaluated to obtain the important energy consumption prediction feature set; For each important performance consumption prediction feature in the set of important performance consumption prediction features, key features are filtered to obtain a set of key energy consumption prediction features. Based on the initial energy consumption prediction module and the energy consumption prediction task, an energy consumption prediction loss function is constructed. The initial energy consumption prediction module is trained based on the key energy consumption prediction feature set and the energy consumption prediction loss function to obtain the target energy consumption prediction module.

6. A terminal, characterized in that, The system includes a processor, an input device, an output device, and a memory, which are interconnected. The memory stores a computer program, which includes program instructions. The processor is configured to invoke the program instructions to execute the intelligent building energy consumption control method based on artificial intelligence as described in any one of claims 1-3.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, the computer program including program instructions, which, when executed by a processor, cause the processor to perform the intelligent building energy consumption control method based on artificial intelligence as described in any one of claims 1-3.

Images