Integrated method and system for dynamic monitoring and analysis of energy consumption in buildings

Abstract

This invention proposes a method and system for dynamic monitoring and analysis of energy consumption in integrated buildings, belonging to the field of dynamic energy consumption monitoring technology. It involves layered data collection from the integrated building, analysis and control of the collection frequency and accuracy based on the layered data, edge preprocessing of the layered data based on the collected data, secondary optimization of the edge preprocessed data in the cloud, classification and storage of the optimized cloud data, energy consumption analysis based on the classified and stored data, trend, distribution, anomaly, and optimization analysis based on the energy consumption analysis data, proactive prediction, collaborative linkage, mobile status, and emergency regulation based on the energy consumption prediction analysis data, obtaining multi-dimensional regulation data, and iterative optimization of the regulation effect based on the multi-dimensional regulation data, obtaining iterative optimization data. This invention can achieve multi-dimensional and comprehensive identification of insufficient energy consumption regulation in integrated buildings, covering key aspects such as parameters, logic, scenario adaptation, and models.

Description

Technical Field

[0001] This invention proposes an integrated method and system for dynamic monitoring and analysis of energy consumption within buildings, relating to the field of dynamic monitoring and analysis technology, specifically to the field of integrated dynamic monitoring and analysis of energy consumption within buildings. Background Technology

[0002] Traditional methods for analyzing energy consumption control deficiencies in integrated housing have significant limitations. Their analytical approaches are simplistic and one-sided, often only able to identify surface-level issues during the control process, failing to comprehensively identify deficiencies from core dimensions. Existing technologies lack a multi-dimensional identification system, making it difficult to cover key aspects such as parameter settings, control logic, scenario adaptation, and model performance. Furthermore, the lack of in-depth analysis of the root causes of deficiencies leads to a lack of clear direction for control optimization, resulting in blind optimization. Simultaneously, due to the mobile and dynamic nature of integrated housing, traditional analysis methods cannot adapt to its dynamic operational needs, further reducing the targetedness and effectiveness of control optimization. This makes it difficult to continuously improve energy consumption control performance and fails to meet the needs of efficient and intelligent energy management in integrated housing. Summary of the Invention

[0003] This invention provides a method and system for integrated dynamic monitoring and analysis of indoor energy consumption, in order to solve the above-mentioned problems: The present invention proposes an integrated method and system for dynamic monitoring and analysis of indoor energy consumption, wherein the method includes: S1. Collect data from the prefabricated houses in layers to obtain layered collection data. Analyze and control the collection frequency and accuracy based on the layered collection data to obtain collection analysis and control data. S2. Based on the collected and analyzed control data, perform edge preprocessing on the layered collected data to obtain edge preprocessed data. Perform secondary optimization on the edge preprocessed data in the cloud to obtain cloud optimized data. Classify and store the cloud optimized data to obtain classified storage data. S3. Perform energy consumption analysis based on the classified storage data to obtain energy consumption analysis data. Based on the energy consumption analysis data, perform trend, distribution, anomaly and optimization analysis to obtain dynamic energy consumption analysis data. S4. Based on energy consumption prediction and analysis data, perform proactive prediction, coordinated linkage, mobile status and emergency control to obtain multi-dimensional control data. Based on the multi-dimensional control data, iteratively optimize the control effect to obtain iterative optimization data.

[0004] Furthermore, the system includes: The data acquisition and control module is used to collect data from the integrated house in layers, obtain layered data, and analyze and control the acquisition frequency and accuracy based on the layered data to obtain data acquisition analysis and control data. The classification and storage module is used to perform edge preprocessing on the hierarchical collected data based on the collected analysis and control data to obtain edge preprocessed data, perform secondary optimization on the edge preprocessed data in the cloud to obtain cloud optimized data, and classify and store the cloud optimized data to obtain classified storage data. The energy consumption analysis module is used to perform energy consumption analysis based on the classified stored data, obtain energy consumption analysis data, and perform trend, distribution, anomaly and optimization analysis based on the energy consumption analysis data to obtain dynamic energy consumption analysis data. The effect iteration module is used to perform proactive prediction, coordinated linkage, mobile status and emergency control based on energy consumption prediction and analysis data, obtain multi-dimensional control data, and perform iterative optimization of control effect based on multi-dimensional control data to obtain effect iteration optimization data.

[0005] The beneficial effects of this invention are as follows: This method solves the technical problems of traditional energy consumption monitoring methods, such as poor adaptability, disconnected processes, inability to meet the dynamic needs of integrated housing such as relocation and dismantling, low data quality, incomplete analysis, lagging regulation, and lack of iterative optimization mechanisms. It realizes full-process dynamic monitoring and accurate analysis of energy consumption in integrated housing, improves the reliability of energy consumption data and the comprehensiveness of analysis results, and ensures the initiative, coordination, and emergency response capabilities of energy consumption regulation. It effectively improves the intelligence level of energy consumption management in integrated housing, reduces energy waste, reduces the cost of manual intervention, and improves the adaptability and effectiveness of energy consumption regulation, realizing continuous optimization of energy consumption management and adapting to the multi-scenario and dynamic operation needs of integrated housing. Attached Figure Description

[0006] Figure 1 A schematic diagram of a method for dynamic monitoring and analysis of energy consumption within a building. Detailed Implementation

[0007] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0008] Example 1 In one embodiment of the present invention, the method and system for dynamic monitoring and analysis of indoor energy consumption proposed by the present invention include: S1. Collect data from the prefabricated houses in layers to obtain layered collection data. Analyze and control the collection frequency and accuracy based on the layered collection data to obtain collection analysis and control data. S2. Based on the collected and analyzed control data, perform edge preprocessing on the layered collected data to obtain edge preprocessed data. Perform secondary optimization on the edge preprocessed data in the cloud to obtain cloud optimized data. Classify and store the cloud optimized data to obtain classified storage data. S3. Perform energy consumption analysis based on the classified storage data to obtain energy consumption analysis data. Based on the energy consumption analysis data, perform trend, distribution, anomaly and optimization analysis to obtain dynamic energy consumption analysis data. S4. Based on energy consumption prediction and analysis data, perform proactive forecasting, coordinated action, mobile status monitoring, and emergency control to obtain multi-dimensional control data. Then, based on this multi-dimensional control data, iteratively optimize the control effect to obtain iterative optimization data, such as… Figure 1 As shown.

[0009] The working principle and technical effects of the above-mentioned technical solution are as follows: This method acquires energy consumption and related data of each level of the integrated house through layered collection, and simultaneously optimizes the collection frequency and accuracy to ensure the relevance and reliability of the collected data; then, the collected data undergoes edge preprocessing and cloud-based secondary optimization to remove invalid interference, improve data quality, and classify and store the data; based on the classified and stored data, a comprehensive energy consumption analysis is carried out, covering trends, distribution, anomalies, and optimization directions, forming comprehensive dynamic energy consumption analysis data; finally, based on the analysis data, proactive prediction and multi-dimensional regulation are carried out, and the regulation effect is evaluated, shortcomings are identified, and iterative optimization is performed simultaneously to achieve dynamic cyclical upgrading of energy consumption monitoring and regulation, ensuring that the entire method is adapted to the dynamic operation scenario of the integrated house and forms a closed loop of energy consumption management throughout the entire process and in all dimensions.

[0010] This method addresses the technical problems of traditional energy consumption monitoring methods, such as poor adaptability, disjointed processes, inability to meet the dynamic needs of integrated housing such as relocation and dismantling, low data quality, incomplete analysis, lagging control, and lack of iterative optimization mechanisms. It achieves full-process dynamic monitoring and precise analysis of integrated housing energy consumption, improves the reliability of energy consumption data and the comprehensiveness of analysis results, and ensures the initiative, coordination, and emergency response capabilities of energy consumption control. It effectively improves the intelligence level of integrated housing energy consumption management, reduces energy waste, reduces manual intervention costs, and enhances the adaptability and effectiveness of energy consumption control, enabling continuous optimization of energy consumption management and adapting to the multi-scenario and dynamic operation needs of integrated housing.

[0011] In one embodiment of the present invention, S1 includes: Data is collected from the perception layer, adaptation layer, transmission layer and edge processing layer of the integrated house to obtain layered data. Based on the stratified data collection, frequency adjustment needs analysis is performed to obtain adjustment needs analysis data; Obtain basic frequency standard information, and combine the basic frequency standard information with adjustment demand analysis data to conduct dynamic adjustment rule analysis and obtain dynamic adjustment analysis data; Frequency control verification is performed based on dynamic adjustment analysis data to obtain frequency control verification information. Based on the frequency control verification information, the acquisition accuracy is analyzed to obtain acquisition accuracy analysis data; Based on the collected accuracy analysis data, a dual calibration and deviation handling process is performed to obtain calibration handling information; Based on the calibration and processing information, accuracy optimization iterations and outputs are performed to obtain acquired, analyzed, and controlled data.

[0012] The working principle and technical effects of the above technical solution are as follows: Layered data collection covers all levels of energy consumption related to integrated housing, ensuring no omissions in the collection scope. The sensing layer collects basic energy consumption data, the adaptation layer collects data related to device adaptation, the transmission layer collects data transmission status data, and the edge processing layer collects intermediate data after preliminary processing. This data is then integrated to form layered collection data. Based on this layered collection data, the matching degree between the current collection frequency and the actual scenario is analyzed to clarify the frequency adjustment needs. Combined with a preset basic frequency standard, dynamic frequency adjustment rules conforming to the dynamic scenario of integrated housing are formulated, forming dynamic adjustment analysis data. Through a frequency control verification process, the rationality and effectiveness of the adjustment rules are verified, generating verification information. Based on this verification information, it is analyzed whether the collection accuracy meets the subsequent analysis requirements. Dual calibration and deviation handling are performed for accuracy deviations to eliminate potential accuracy issues. Through continuous iterative optimization, the final output is collection, analysis, and control data that combines reasonable frequency and precise accuracy.

[0013] This method addresses the technical problem of traditional data acquisition methods using fixed frequencies and precision, which cannot adapt to dynamic scenarios of integrated housing (such as movement, dismantling, and environmental changes), leading to redundant or missing data, insufficient precision, and affecting analysis and control effects. It achieves dynamic adjustment of the acquisition frequency and precise control of acquisition precision, improving the relevance and reliability of the acquired data, avoiding resource waste caused by ineffective acquisition, and reducing analytical biases due to insufficient acquisition precision. It also reduces energy consumption and resource usage in the acquisition process, improves acquisition efficiency, and ensures the stability and accuracy of the entire energy consumption monitoring system.

[0014] In one embodiment of the present invention, the step of performing frequency control verification based on dynamic adjustment analysis data to obtain frequency control verification information includes: Based on the dynamic adjustment analysis data, verification data collection and organization are carried out to obtain verification collection and organization data; Based on the collected and organized verification data, multi-dimensional verification analysis was conducted to obtain multi-dimensional verification analysis data. Based on the multi-dimensional verification analysis data, multi-dimensional verification results are determined and feedback is provided to obtain determination feedback data. Based on the judgment feedback data, verification information is output to obtain frequency control verification information.

[0015] Specifically, based on the judgment feedback data, verification information is output to obtain frequency control verification information, including: Integrate and verify collected data, multi-dimensional verification analysis data, and judgment feedback data to form complete frequency control verification information, covering data and results throughout the entire verification process.

[0016] The output specifications include the validation dataset, 3D validation model parameters, validation result judgment report (including grade, deviation details, and non-compliance items), adjustment rule optimization suggestions, and feedback execution records. All information is timestamped and batch-identified to ensure traceability.

[0017] The frequency control verification information is synchronized to the data acquisition accuracy analysis stage; it is also synchronized to the AI ​​frequency adjustment model for model parameter iteration and continuous optimization of the frequency adjustment logic.

[0018] The working principle and technical effects of the above technical solution are as follows: Based on dynamic adjustment analysis data, frequency adjustment-related records, energy consumption fluctuation data, and scenario change data are collected synchronously, categorized, and invalid information is removed to form standardized verification collection and processing data. Subsequently, a multi-dimensional verification model is constructed to comprehensively analyze the verification collection and processing data from three dimensions: the adaptability of frequency adjustment to the scenario, the effectiveness of adjustment actions, and the rationality of adjustment logic. This clarifies the compliance status of adjustment rules and forms multi-dimensional verification analysis data. Based on the analysis data, a graded judgment standard is formulated to grade the verification results, marking non-compliant items and their reasons. The judgment results are then fed back to the frequency adjustment module, forming judgment feedback data. Finally, all types of data and results from the entire verification process are integrated, output content is organized according to specifications, timestamps and batch identifiers are added to ensure traceability, and verification information is synchronized to the data acquisition accuracy analysis stage and the AI ​​frequency adjustment model to achieve closed-loop optimization of frequency adjustment.

[0019] This method addresses the technical problems of traditional frequency adjustment, such as the lack of an effective verification mechanism, the inability to guarantee the rationality of adjustment rules, and the lack of traceability and support for optimization of verification results, leading to poor adaptability and frequent invalid adjustments. It achieves multi-dimensional and accurate verification of frequency adjustment rules, ensuring that the adjustment rules meet the dynamic needs of integrated housing scenarios, thereby improving the effectiveness and rationality of frequency adjustment. Through standardized verification information output and traceability mechanisms, it reduces the difficulty of troubleshooting, continuously improves the intelligence level of frequency adjustment, reduces invalid data collection and collection deviation, further improves the quality of collected data, and reduces resource waste in the data collection process.

[0020] In one embodiment of the present invention, S2 includes: The collected, analyzed, and controlled data are cleaned and standardized to obtain preliminary processed data. Anomaly identification, labeling, and output are performed on the initially processed data to obtain edge preprocessing data; Data noise reduction and correlation optimization are performed on the edge preprocessed data to obtain correlation-optimized data; Based on the correlation optimization data, perform secondary anomaly identification and classification to obtain anomaly identification and classification data; The anomaly identification and classification data are further optimized and output to obtain cloud-optimized data. Classify and organize the optimized data in the cloud to obtain the categorized and organized data. The categorized and organized data is stored and managed in two ways to obtain storage management data; Output the storage management data to obtain categorized storage data.

[0021] The working principle and technical effects of the above technical solution are as follows: The collected raw data undergoes layer-by-layer optimization and standardized management. Based on the collected analysis and control data, data cleaning is performed to remove redundant, erroneous, and missing data, while standardization is carried out to unify data formats and definitions, obtaining preliminary processed data. Anomaly identification is performed on the preliminary processed data, marking and outputting abnormal data to form edge preprocessing data, achieving initial screening of abnormal data. Subsequently, noise reduction processing is performed on the edge preprocessing data to remove various interference noises, while data association logic is constructed to achieve data association optimization, improving data integrity and relevance. Secondary anomaly identification is performed based on the association-optimized data, classifying abnormal data, clarifying anomaly types and characteristics, forming anomaly identification classification data. Secondary optimization is performed on the classified abnormal data, eliminating invalid anomalies and correcting optimizable anomalies, outputting cloud-optimized data. Finally, the cloud-optimized data is classified and organized, divided according to data type and purpose, employing a dual storage mechanism to ensure data security, and undergoing standardized management, ultimately outputting classified storage data to ensure data can be quickly retrieved and is traceable.

[0022] This method addresses the technical problems of redundant, erroneous, and noisy data in the collected data, inconsistent data formats, poor correlation, incomplete identification of abnormal data, and chaotic and insecure data storage, which lead to large deviations and low efficiency in energy consumption analysis results. It achieves layer-by-layer optimization of the collected data, improving data quality and correlation, and ensuring data accuracy, completeness, and standardization. It comprehensively identifies and classifies abnormal data, reducing the difficulty of anomaly analysis. Through classified storage and dual management, it improves the security and standardization of data storage, reduces the risk of data loss and corruption, and improves data retrieval efficiency. This provides high-quality data support for subsequent energy consumption analysis and trend prediction, ensuring the efficient operation of the entire energy consumption monitoring system.

[0023] In one embodiment of the present invention, the step of performing data denoising and correlation optimization on the edge preprocessing data to obtain correlation-optimized data includes: Noise identification and classification are performed on the edge preprocessing data to obtain noise identification and classification data; Adaptive noise reduction is performed based on noise recognition and classification data to obtain noise reduction execution data. The noise reduction effect is verified based on the noise reduction execution data to obtain noise reduction effect verification data. After the noise reduction effect verification data passes, the model with associated parameters is imported to obtain the model import information; Based on the imported model information, construct the association logic to obtain the association logic construction information; Based on the association logic, information is constructed to complete and optimize the data, thereby obtaining data completion and optimization information; Based on the data completion and optimization information, content verification and output are performed to obtain related optimized data.

[0024] This includes content verification and output based on data completion and optimization information to obtain related optimized data, including: The completed data is fully validated, including data accuracy (deviation ≤ 0.3%), completeness (missing rate ≤ 0.05%), and consistency of associations (logical matching degree of association ≥ 97%).

[0025] Any correlation anomalies discovered during the verification process (such as mismatched parameter correlation logic) are marked and recorded, and then synchronously fed back to the correlation model for optimization.

[0026] The noise-reduced data, supplementary data, associated parameters, verification reports, and anomaly markers are integrated to form associated optimized data, which is then synchronously transmitted to the anomaly secondary identification stage.

[0027] The working principle and technical effects of the above technical solution are as follows: Comprehensive noise identification is performed on the edge preprocessing data to distinguish different types of noise, clarify noise characteristics and sources, and form noise identification classification data; based on noise type and intensity, differentiated noise reduction strategies are automatically matched, adaptive noise reduction processing is carried out, and the noise reduction process and results are recorded to form noise reduction execution data; the noise reduction effect is comprehensively verified to ensure that there is no residual noise or loss of effective information after noise reduction, forming noise reduction effect verification data. After the noise reduction effect meets the standard, related parameters such as building characteristics, equipment compatibility, and environment are imported to construct a multi-dimensional data association model, explore the inherent relationships between various data, and form association logic construction information; based on the association logic, missing items in the data are identified and completed to optimize data integrity, forming data completion optimization information. Finally, the completed data is comprehensively verified to check data accuracy, integrity, and association consistency, mark association anomalies and feed them back to the association model optimization, integrate various related data and reports to form association optimization data and transmit it to the anomaly secondary identification stage to complete the deep optimization of the data.

[0028] This method addresses the technical challenges of data acquisition in scenarios such as mobile integrated housing and extreme environments, where noise, poor data correlation, and missing data lead to low data quality and affect the accuracy of anomaly identification and energy consumption analysis. It achieves accurate noise identification and adaptive noise reduction, avoiding the poor adaptability of traditional fixed noise reduction strategies, improving noise reduction effectiveness, and preserving valuable information in the data. Through the construction of correlation logic and data completion, it enhances data correlation and completeness, resolving data fragmentation issues. Furthermore, through comprehensive verification and anomaly feedback optimization, it further improves data quality, reduces bias in subsequent analysis, and enhances the reliability and efficiency of the entire data processing process.

[0029] In one embodiment of the present invention, S3 includes: Based on the categorized and stored data, basic energy consumption statistics are obtained; Compliance verification is conducted based on basic energy consumption statistics to obtain compliance verification data; Data is filtered and extracted based on compliance verification data to obtain filtered and extracted data; Output energy consumption analysis data based on the filtered and extracted data; Establish an AI energy consumption prediction model, and use the AI ​​energy consumption prediction model in conjunction with energy consumption analysis data to predict energy consumption trends and obtain energy consumption trend prediction information. The impact of energy consumption trend forecast information is quantified and the results are output to obtain trend analysis data; Perform distribution analysis on the trend analysis data to obtain distribution analysis data; Perform anomaly analysis on the distribution analysis data to obtain anomaly analysis data; Perform optimization analysis on the anomaly analysis data to obtain optimized analysis data; By integrating energy consumption analysis data, trend analysis data, distribution analysis data, anomaly analysis data, and optimization analysis data, dynamic energy consumption analysis data is obtained.

[0030] The working principle and technical effects of the above technical solution are as follows: Based on high-quality data stored in a categorized manner, comprehensive and in-depth energy consumption analysis is conducted to form comprehensive dynamic energy consumption analysis data. Based on the categorized data, basic energy consumption statistics are conducted, core energy consumption-related information is sorted out, and basic energy consumption statistics are formed. The statistical data undergoes compliance verification, identifying data that does not comply with energy consumption management regulations, forming compliance verification data. Based on the verification data, valid data is filtered and extracted, outputting energy consumption analysis data to clarify the current basic energy consumption situation. Subsequently, an AI energy consumption prediction model is constructed, combining the energy consumption analysis data to conduct energy consumption trend prediction, obtaining trend prediction information. The impact of the prediction information is quantified, analyzing the impact of energy consumption trends on building operation, outputting trend analysis data. Distribution analysis is performed on the trend analysis data to clarify the distribution characteristics of energy consumption in different time periods, regions, and devices, obtaining distribution analysis data. Based on the distribution data, energy consumption anomalies are identified, the causes and impacts of the anomalies are analyzed, obtaining anomaly analysis data. For anomalies and energy consumption optimization space, optimization analysis is conducted, proposing optimization directions and ideas, obtaining optimization analysis data. Finally, by integrating various analytical data, a comprehensive and systematic dynamic energy consumption analysis is formed, which fully presents the current status, trends, anomalies, and optimization directions of energy consumption.

[0031] This method addresses the shortcomings of traditional energy consumption analysis methods, which are often limited in scope and cannot predict trends or identify potential anomalies. They also lack targeted optimization analysis, leading to passive energy management and unclear optimization directions. This new method achieves comprehensive and in-depth energy consumption analysis, enabling a grasp of current energy consumption levels, accurate prediction of trends, identification of anomalies, and clarification of optimization directions. It enhances the comprehensiveness and foresight of energy consumption analysis, reduces the workload and biases of manual analysis, improves efficiency and accuracy, and promotes a shift in energy management from passive response to proactive prediction and precise optimization, further enhancing the intelligence level of integrated housing energy management.

[0032] In one embodiment of the present invention, the step of establishing an AI energy consumption prediction model and predicting energy consumption trends based on the AI ​​energy consumption prediction model and energy consumption analysis data to obtain energy consumption trend prediction information includes: An AI energy consumption prediction model is constructed based on LSTM neural network and integrated housing features; Obtain housing usage demand information, combine housing usage demand information with AI energy consumption prediction model, perform multi-timescale energy consumption trend prediction, and obtain multi-time prediction data; Model validation is performed on multi-time forecast data to obtain multi-time model validation data. Energy consumption trend prediction information is output based on the multi-time model validation data.

[0033] This includes acquiring housing usage demand information, combining this information with an AI energy consumption prediction model to predict energy consumption trends across multiple time scales, and obtaining multi-time-scale prediction data, including: Collect information on the usage needs of integrated housing, including core elements such as usage scenarios, usage plans, and energy consumption control targets, and then standardize and organize this information.

[0034] Energy consumption analysis data is extracted, sorted, and then imported into the established AI energy consumption prediction model along with standardized housing usage demand information.

[0035] By combining the two types of data to start the model, we can carry out energy consumption trend prediction at multiple time scales in the short and medium term and analyze the energy consumption fluctuation patterns in different periods.

[0036] By combining housing usage plans and scenario changes, we can predict energy consumption trends in different time periods, identify potential energy consumption anomalies, and generate multi-time forecast data.

[0037] The working principle and technical effects of the above solution are as follows: An AI energy consumption prediction model adapted to the characteristics of integrated housing is constructed to achieve accurate energy consumption trend prediction across multiple time scales. Using an LSTM neural network as the core, and considering the characteristics of integrated housing—its mobility, varied scenarios, and complex energy consumption fluctuations—the model structure is optimized, and scenario adaptation and data migration modules are added to construct an AI energy consumption prediction model tailored to the needs of integrated housing, distinct from conventional general-purpose prediction models. Subsequently, housing usage demand information is collected, including core content such as usage scenarios, usage plans, and energy consumption control targets, and standardized to ensure matching with model input requirements. Energy consumption analysis data generated in the S3 stage is extracted, sorted, and imported into the AI ​​energy consumption prediction model along with the standardized usage demand information, making the prediction results more closely aligned with actual usage needs. The model is launched and runs, combining two types of data to conduct short-term and medium-term multi-timescale energy consumption trend predictions, analyze the energy consumption fluctuation patterns in different periods, and predict energy consumption trends in each period by combining housing usage plans and scenario changes, identify potential energy consumption anomaly risks, and generate multi-time forecast data; the forecast data is then used to validate the model, correct prediction biases, ensure the accuracy of the forecast results, and finally output complete energy consumption trend forecast information.

[0038] This method addresses the technical problems of traditional energy consumption prediction models, which suffer from poor adaptability, inability to meet the dynamic needs of integrated housing scenarios, single prediction dimensions, low accuracy, and failure to consider housing usage requirements, leading to inconsistencies between prediction results and reality, and thus failing to provide effective support for energy consumption regulation. It achieves the construction of an AI energy consumption prediction model adapted to the characteristics of integrated housing, improving the adaptability and generalization ability of the prediction model; it conducts multi-timescale predictions based on housing usage requirements, improving the relevance and accuracy of the prediction results, enabling precise prediction of energy consumption trends and identification of potential anomalies; it reduces the bias in energy consumption prediction, facilitating the early formulation of control strategies, reducing energy waste, enhancing the initiative and scientific nature of energy consumption management, and further optimizing the energy consumption management level of integrated housing.

[0039] In one embodiment of the present invention, S4 includes: Energy consumption prediction information is extracted from dynamic energy consumption analysis data to obtain energy consumption prediction extraction information. Based on energy consumption prediction, information is extracted for multi-dimensional energy consumption regulation to obtain multi-dimensional regulation data. The effect of regulation is quantitatively evaluated based on multidimensional regulation data to obtain quantitative evaluation data of the effect. Based on the quantitative evaluation data of the effects, an analysis of insufficient regulation was conducted to obtain data on the analysis of insufficient regulation. Based on the analysis of insufficient regulation data, the regulation strategy is iteratively optimized to obtain data on the effect of iterative optimization.

[0040] The working principle and technical effects of the above-mentioned technical solution are as follows: Energy consumption prediction information is extracted from dynamic energy consumption analysis data to clarify core content such as energy consumption trends and abnormal risks, forming energy consumption prediction extraction information. Based on this extraction information, multi-dimensional energy consumption control strategies are formulated and implemented from four dimensions: proactive prediction and control, coordinated linkage of various links, mobile status adaptation, and emergency anomaly handling, forming multi-dimensional control data to ensure control coverage of all scenarios and all stages. Subsequently, the control effect is quantitatively evaluated, analyzing the implementation effect of the control strategy to form quantitative effect evaluation data. Based on the evaluation data, deficiencies in the control process are comprehensively investigated, and the root causes of these deficiencies are analyzed, forming control deficiency analysis data. For the identified deficiencies, the control strategy is optimized, and control parameters and logic are updated, forming effect iterative optimization data to achieve continuous upgrading of the control strategy, ensuring continuous improvement in control effect and adapting to changes in the dynamic operation scenario of integrated housing.

[0041] This method addresses the technical problems of traditional energy consumption control methods, which are often limited in scope and lack specificity, failing to adapt to the dynamic scenarios of integrated housing, lacking coordination and emergency response capabilities, and lacking effectiveness evaluation and iterative optimization mechanisms, resulting in poor control effects and an inability to continuously improve. It achieves multi-dimensional and precise energy consumption control, improving the targeting, coordination, and emergency response capabilities of control, effectively addressing energy consumption changes in dynamic scenarios such as the movement and anomalies of integrated housing. Through control effect evaluation and deficiency analysis, it identifies control shortcomings, enabling continuous iterative optimization of control strategies, continuously improving control effectiveness, reducing energy waste, and lowering control costs. Simultaneously, it enhances the intelligence and adaptability of energy consumption control, ensuring that the entire energy consumption monitoring and analysis system can continuously adapt to the operational needs of integrated housing, achieving continuous optimization of energy consumption management.

[0042] In one embodiment of the present invention, the step of performing a regulation insufficiency analysis based on the effect quantification evaluation data to obtain regulation insufficiency analysis data includes: The data from the quantitative evaluation of the effects are structured and classified to obtain classification information; The system identifies deficiencies in the sorted and categorized information by adjusting parameters, logic, scenario adaptation, and model dimensions, thus obtaining multi-dimensional deficiency identification data. The system performs root cause analysis and outputs data on multi-dimensional deficiency identification data to obtain data for regulating deficiency analysis.

[0043] This includes identifying deficiencies in the sorting and categorizing of information by adjusting parameters, logic, scenario adaptation, and model dimensions, resulting in multi-dimensional deficiency identification data, including: Investigate and identify problems with unreasonable control parameter settings, such as deviations in control thresholds, lead times, and priority settings, and clarify the deficiencies at the parameter level.

[0044] Investigate and address any shortcomings in the regulatory logic, such as data interaction delays, unclear regulatory grading standards, and disconnects in coordination among different stages, and identify and address any logical flaws.

[0045] The investigation focused on issues such as insufficient adaptation of control strategies to the dynamic scenarios of integrated housing, including the lack of control measures for relocation and dismantling scenarios, and untimely control measures for extreme environments.

[0046] Investigate issues such as aging AI control models, untimely parameter iteration, and declining prediction accuracy to identify model-level deficiencies and generate multi-dimensional deficiency identification data.

[0047] The working principle and technical effects of the above-mentioned technical solution are as follows: The quantitative evaluation data of the control effect is structured and classified according to evaluation dimensions and control links to form a comprehensive classification information, ensuring the comprehensiveness of deficiency identification. Deficiency identification is then carried out from four core dimensions: parameter dimension, investigating unreasonable control parameter settings and clarifying parameter-level deviations; logic dimension, investigating incomplete control logic and poor coordination between links, and identifying logical loopholes; scenario adaptation dimension, investigating insufficient adaptation of control strategies to dynamic scenarios such as integrated housing movement, dismantling, and extreme environments, and identifying scenario adaptation shortcomings; and model dimension, investigating issues such as aging AI control models and untimely parameter iterations, and identifying model-level deficiencies, integrating these into multi-dimensional deficiency identification data. Finally, in-depth root cause analysis is conducted on the identified deficiencies to clarify the core causes of each deficiency, avoiding superficial analysis, and compiling the deficiency identification results, root cause analysis, and improvement directions to form control deficiency analysis data.

[0048] This method addresses the shortcomings of traditional control deficiency analysis methods, which are often simplistic and one-sided, only able to identify surface problems and unable to identify deficiencies from multiple dimensions or delve into their root causes. This results in poor targeting of control optimization and a lack of sustained improvement in effectiveness. This method achieves multi-dimensional and comprehensive identification of control deficiencies, covering four core levels: parameters, logic, scenario adaptation, and models, ensuring no omissions or blind spots. By deeply exploring the root causes of deficiencies, it clarifies the core direction of control optimization, avoids blind optimization, and improves the targeting and effectiveness of iterative optimization of control strategies. It reduces the cost and workload of control optimization, continuously improves the adaptability and effectiveness of energy consumption control, ensures that energy consumption control can continuously adapt to changes in the dynamic scenarios of integrated housing, and further enhances the stability and intelligence level of the entire energy management system.

[0049] According to one embodiment of the present invention, the system includes: The data acquisition and control module is used to collect data from the integrated house in layers, obtain layered data, and analyze and control the acquisition frequency and accuracy based on the layered data to obtain data acquisition analysis and control data. The classification and storage module is used to perform edge preprocessing on the hierarchical collected data based on the collected analysis and control data to obtain edge preprocessed data, perform secondary optimization on the edge preprocessed data in the cloud to obtain cloud optimized data, and classify and store the cloud optimized data to obtain classified storage data. The energy consumption analysis module is used to perform energy consumption analysis based on the classified stored data, obtain energy consumption analysis data, and perform trend, distribution, anomaly and optimization analysis based on the energy consumption analysis data to obtain dynamic energy consumption analysis data. The effect iteration module is used to perform proactive prediction, coordinated linkage, mobile status and emergency control based on energy consumption prediction and analysis data, obtain multi-dimensional control data, and perform iterative optimization of control effect based on multi-dimensional control data to obtain effect iteration optimization data.

[0050] The working principle and technical effects of the above-mentioned technical solution are as follows: This system acquires energy consumption and related data of each level of the integrated house through layered collection, and simultaneously optimizes the collection frequency and accuracy to ensure the relevance and reliability of the collected data; then, the collected data undergoes edge preprocessing and cloud-based secondary optimization to remove invalid interference, improve data quality, and classify and store the data; based on the classified and stored data, a comprehensive energy consumption analysis is carried out, covering trends, distribution, anomalies, and optimization directions, forming comprehensive dynamic energy consumption analysis data; finally, based on the analysis data, proactive prediction and multi-dimensional regulation are carried out, and the regulation effect is evaluated, shortcomings are identified, and iterative optimization is performed simultaneously to achieve dynamic cyclical upgrading of energy consumption monitoring and regulation, ensuring that the entire method is adapted to the dynamic operation scenario of the integrated house and forms a closed loop of energy consumption management throughout the entire process and in all dimensions.

[0051] This system addresses the technical problems of traditional energy consumption monitoring methods, such as poor adaptability, disjointed processes, inability to meet the dynamic needs of integrated housing such as relocation and dismantling, low data quality, incomplete analysis, lagging control, and lack of iterative optimization mechanisms. It achieves full-process dynamic monitoring and precise analysis of integrated housing energy consumption, improving the reliability of energy consumption data and the comprehensiveness of analysis results, ensuring proactive, collaborative, and emergency response capabilities in energy consumption control. It effectively enhances the intelligence level of integrated housing energy consumption management, reduces energy waste, minimizes manual intervention costs, and improves the adaptability and effectiveness of energy consumption control, enabling continuous optimization of energy consumption management and adapting to the multi-scenario and dynamic operational needs of integrated housing.

[0052] Example 2 Another embodiment of the present invention further identifies the model inaccuracy state in the energy consumption trend prediction process in advance, and dynamically corrects the input weights, scenario adaptation parameters and prediction confidence of the artificial intelligence energy consumption prediction model before the prediction error expands into the control failure.

[0053] By adding a model inaccuracy judgment and adaptive correction process between the original energy consumption dynamic analysis and multi-dimensional control process, the artificial intelligence energy consumption prediction model can be adjusted in advance according to the actual scenario migration status of the integrated house.

[0054] Specifically, the integrated house energy consumption dynamic monitoring and analysis method of this embodiment includes the following steps S1-S7.

[0055] S1. Collect data in layers for the prefabricated houses to obtain layered data. Analyze and control the collection frequency and accuracy based on the layered data to obtain collection analysis and control data.

[0056] S2. Based on the collected and analyzed control data, perform edge preprocessing on the layered collected data to obtain edge preprocessed data. Perform secondary optimization on the edge preprocessed data in the cloud to obtain cloud optimized data. Classify and store the cloud optimized data to obtain classified stored data.

[0057] S3. Perform energy consumption analysis based on the classified storage data to obtain energy consumption analysis data. Based on the energy consumption analysis data, perform trend, distribution, anomaly and optimization analysis to obtain dynamic energy consumption analysis data.

[0058] S4. Based on the energy consumption prediction information in the dynamic energy consumption analysis data, perform proactive prediction, coordinated linkage, mobile status and emergency regulation to obtain multi-dimensional regulation data, and perform iterative optimization of regulation effect based on the multi-dimensional regulation data to obtain effect iterative optimization data.

[0059] S1-S4 are the same as in Example 1.

[0060] S5. Based on the degree of difference in scene migration of integrated houses, the degree of prediction error offset within the continuous prediction time window, and the degree of change in equipment combination, the model inaccuracy is determined, the model inaccuracy determination quantity is obtained, and the original artificial intelligence energy consumption prediction model is judged to have inaccuracy risk based on the model inaccuracy determination quantity.

[0061] Specifically, the degree of difference in scene migration is used to characterize the difference between the current usage scenario of the integrated house and the original model training scenario. It is determined based on at least two of the following: changes in house use, changes in installation environment, changes in deployment location, changes in building envelope status, changes in outdoor temperature and humidity range, and changes in user population density. The degree of prediction error offset within the continuous prediction time window is used to characterize the cumulative deviation between the predicted energy consumption and the actual energy consumption of the artificial intelligence energy consumption prediction model within the continuous prediction time window. The degree of change in equipment combination is used to characterize the impact of the access, withdrawal, replacement, changes in rated power, and changes in operating priority of energy-consuming equipment in the integrated house on the original model input structure.

[0062] In practice, the degree of difference in scene migration, the degree of prediction error shift within the continuous prediction time window, and the degree of change in equipment combination are first normalized to ensure that data from different sources can be compared under the same evaluation criteria. Then, the three types of data are weighted and fused according to preset inaccuracy judgment weights to form a model inaccuracy judgment metric. The inaccuracy judgment weights can be calibrated using historical scene migration samples. During calibration, scene migration records, equipment combination change records, and actual energy consumption deviation records before the model prediction error amplifies are used as samples, and the weight combination that can reduce the inaccuracy missed judgment rate and false judgment rate is selected as the current weight.

[0063] When formulating the model inaccuracy criterion, the impact of not only the degree of scene migration difference, the degree of prediction error shift within the continuous prediction time window, and the degree of equipment combination change on model inaccuracy is considered, but also the combined effect between the degree of scene migration difference and the degree of prediction error shift is further considered. When the integrated house experiences only slight scene changes but the prediction error does not continue to increase, the model inaccuracy criterion remains at a low level; when the integrated house experiences significant scene migration and the prediction error increases synchronously within the continuous prediction time window, the model inaccuracy criterion increases accordingly, thus reflecting in advance the risk of inaccuracy of the original artificial intelligence energy consumption prediction model in the current scenario.

[0064] If the model's inaccuracy determination is less than the first inaccuracy threshold, the input weights and scenario adaptation parameters of the original AI energy consumption prediction model are maintained, and only the current scenario migration information is recorded. If the model's inaccuracy determination is greater than or equal to the first inaccuracy threshold and less than the second inaccuracy threshold, the scenario adaptation correction process is initiated. If the model's inaccuracy determination is greater than or equal to the second inaccuracy threshold, the original AI energy consumption prediction model is marked as having a high inaccuracy risk, and the proportion of its prediction results being directly used in subsequent multidimensional regulation data is reduced.

[0065] S6. Based on the degree of difference in scene migration, the degree of prediction error offset within the continuous prediction time window, and the degree of usage plan offset, perform scene adaptation correction analysis to obtain the scene adaptation correction amount, and dynamically correct the input weights and scene adaptation parameters of the artificial intelligence energy consumption prediction model based on the scene adaptation correction amount.

[0066] Specifically, the degree of usage plan offset is used to characterize the change in the current usage plan of the integrated housing relative to the corresponding usage plan in the original model. The usage plan includes at least one of the following: usage time period, number of people, equipment operation plan, air conditioning setting target, fresh air operation plan, lighting usage plan, emergency support needs, and energy consumption control target. When the integrated housing is changed from office use to accommodation use, from ordinary temporary housing to emergency support housing, or from high-frequency daytime use to high-frequency nighttime use, the degree of usage plan offset increases accordingly.

[0067] In practice, the degree of deviation from the usage plan is first normalized and then input into the scene adaptation correction analysis process along with the degree of difference in scene migration and the degree of prediction error deviation within the continuous prediction time window. The system comprehensively evaluates the degree of difference in scene migration, the degree of prediction error deviation, and the degree of deviation from the usage plan according to preset scene adaptation correction weights, forming the scene adaptation correction amount. The scene adaptation correction weights can be determined based on iterative optimization data of historical control effects, or they can be grouped and calibrated based on actual operating samples of different types of integrated housing.

[0068] When formulating the scenario adaptation correction, the degree of difference in scenario migration and the degree of prediction error shift within the continuous prediction time window continue to be used as common basic parameters in the analysis, ensuring continuity between the scenario adaptation correction analysis in S6 and the model inaccuracy determination in S5. Simultaneously, S6 further introduces the degree of usage plan shift to differentiate the impact of different usage plan changes on the model correction direction when both scenario migration and prediction error shift occur. For example, when the installation environment of the integrated housing changes significantly but the usage plan remains largely unchanged, the scenario adaptation correction primarily favors environmental impact parameters; when the usage plan changes significantly and the prediction error expands synchronously, the scenario adaptation correction primarily favors the input weights corresponding to usage time periods, personnel density, and equipment operation plans.

[0069] Based on the scenario adaptation correction amount, the AI ​​energy consumption prediction model is tiered for correction. When the scenario adaptation correction amount is in the low correction range, only the input weights corresponding to the housing usage demand information are fine-tuned, while maintaining the original prediction confidence. When the scenario adaptation correction amount is in the medium correction range, the input weights of environmental features, equipment combination features, and usage plan features are adjusted simultaneously, and the scenario adaptation parameters are updated. When the scenario adaptation correction amount is in the high correction range, in addition to adjusting the input weights and scenario adaptation parameters, the confidence of the original prediction results is reduced, and the participation ratio of recent measured energy consumption data in subsequent model iterations is increased.

[0070] S7. Based on the model inaccuracy judgment quantity and the scenario adaptation correction quantity, form the model adaptive optimization quantity, and generate the model adaptive optimization result based on the model adaptive optimization quantity. Then, synchronize the model adaptive optimization result to the energy consumption trend prediction and multi-dimensional control process.

[0071] Specifically, the model adaptive optimization quantity is used to comprehensively reflect the inaccuracy risk and correction needs of the original AI energy consumption prediction model under the current scenario migration conditions. The system comprehensively evaluates the model inaccuracy judgment quantity and the scenario adaptation correction quantity to form the model adaptive optimization quantity. When forming the model adaptive optimization quantity, not only are the magnitudes of the model inaccuracy judgment quantity and the scenario adaptation correction quantity considered separately, but also the superimposed effect when both are high. When the model inaccuracy risk is high and the scenario adaptation correction need is large, the model adaptive optimization quantity is increased accordingly, enabling the system to respond prominently to states where the model has already shown inaccuracy risk and requires significant correction.

[0072] The adaptive optimization results are generated based on the adaptive optimization amount. These results include model input weight correction information, scene adaptation parameter correction information, prediction confidence correction information, and model iteration labeling information. Specifically, when the adaptive optimization amount is less than the first optimization threshold, the main parameters of the original AI energy consumption prediction model remain unchanged, and only the current scene migration data is written into the classification storage data. When the adaptive optimization amount is greater than or equal to the first optimization threshold and less than the second optimization threshold, the input weights of environmental features, equipment combination features, and usage plan features are dynamically corrected, and the scene adaptation parameters are updated. When the adaptive optimization amount is greater than or equal to the second optimization threshold, the prediction confidence of the original AI energy consumption prediction model's output is reduced, and recently measured energy consumption data, the degree of scene migration difference, the degree of prediction error shift within the continuous prediction time window, the degree of equipment combination change, and the degree of usage plan shift are used as new iteration samples and written into the effect iteration optimization data.

[0073] Furthermore, after generating the adaptive optimization results of the model, the system synchronizes the corrected input weights, scenario adaptation parameters, and prediction confidence to the energy consumption trend prediction process. This ensures that subsequent energy consumption trend prediction information no longer solely relies on the historical scenario distribution corresponding to the original model, but rather is output in conjunction with the actual migration status of the current integrated housing. Simultaneously, the system synchronizes the prediction confidence to the multi-dimensional control data generation process. When the prediction confidence decreases, it proactively reduces the execution ratio of aggressive control strategies and increases the participation of conservative control strategies, phased control strategies, or manual review strategies. This prevents erroneous control of air conditioning, fresh air systems, lighting, energy storage, or temporary power supply equipment due to model inaccuracies.

[0074] The working principle and technical effects of the above technical solution are as follows: This embodiment does not simply add an error check after the original artificial intelligence energy consumption prediction model. Instead, when the integrated house experiences scene migration, changes in installation environment, changes in equipment combination, or changes in usage plan, the degree of difference in scene migration, the degree of prediction error deviation within the continuous prediction time window, the degree of change in equipment combination, and the degree of deviation in usage plan are introduced into the same model inaccuracy judgment and correction link. S5 identifies in advance whether there is a risk of inaccuracy in the original artificial intelligence energy consumption prediction model through the model inaccuracy judgment quantity; S6 determines how to correct the model input weights and scene adaptation parameters through the scene adaptation correction quantity; S7 uniformly determines the degree of correction of the model input weights, scene adaptation parameters, and prediction confidence through the model adaptive optimization quantity, and synchronizes the correction results to the energy consumption trend prediction and multi-dimensional control process.

[0075] Through the above methods, when integrated housing is moved from one usage scenario to another, or when its purpose, installation environment, equipment combination, or usage plan changes, the system can identify the inaccuracy of the original AI energy consumption prediction model in advance, before the prediction error expands to the point of control failure, and dynamically correct the model input weights, scenario adaptation parameters, and prediction confidence. This avoids the original model from continuing to predict energy consumption trends based on old scenario patterns in the new scenario, and also avoids directly implementing control strategies based on low-reliability prediction results from multi-dimensional control data, thereby improving the reliability of energy consumption trend prediction information and the adaptability of subsequent control strategies.

[0076] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for integrated dynamic monitoring and analysis of energy consumption in a building, characterized in that, The method includes: S1. Collect data from the prefabricated houses in layers to obtain layered collection data. Analyze and control the collection frequency and accuracy based on the layered collection data to obtain collection analysis and control data. S2. Based on the collected and analyzed control data, perform edge preprocessing on the layered collected data to obtain edge preprocessed data. Perform secondary optimization on the edge preprocessed data in the cloud to obtain cloud optimized data. Classify and store the cloud optimized data to obtain classified storage data. S3. Perform energy consumption analysis based on the classified storage data to obtain energy consumption analysis data. Based on the energy consumption analysis data, perform trend, distribution, anomaly and optimization analysis to obtain dynamic energy consumption analysis data. S4. Based on energy consumption prediction and analysis data, perform proactive prediction, coordinated linkage, mobile status and emergency control to obtain multi-dimensional control data. Based on the multi-dimensional control data, iteratively optimize the control effect to obtain iterative optimization data.

2. The method of claim 1, wherein the method further comprises: S1 includes: Data is collected from the perception layer, adaptation layer, transmission layer and edge processing layer of the integrated house to obtain layered data. Based on the stratified data collection, frequency adjustment needs analysis is performed to obtain adjustment needs analysis data; Obtain basic frequency standard information, and combine the basic frequency standard information with adjustment demand analysis data to conduct dynamic adjustment rule analysis and obtain dynamic adjustment analysis data; Frequency control verification is performed based on dynamic adjustment analysis data to obtain frequency control verification information. Based on the frequency control verification information, the acquisition accuracy is analyzed to obtain acquisition accuracy analysis data; Based on the collected accuracy analysis data, a dual calibration and deviation handling process is performed to obtain calibration handling information; Based on the calibration and processing information, accuracy optimization iterations and outputs are performed to obtain acquired, analyzed, and controlled data.

3. The method of claim 2, wherein the method further comprises: The step of verifying frequency control based on dynamically adjusted analysis data to obtain frequency control verification information includes: Based on the dynamic adjustment analysis data, verification data collection and organization are carried out to obtain verification collection and organization data; Based on the collected and organized verification data, multi-dimensional verification analysis was conducted to obtain multi-dimensional verification analysis data. Based on the multi-dimensional verification analysis data, multi-dimensional verification results are determined and feedback is provided to obtain determination feedback data. Based on the judgment feedback data, verification information is output to obtain frequency control verification information.

4. The method of claim 1, wherein the method further comprises: S2 includes: The collected, analyzed, and controlled data are cleaned and standardized to obtain preliminary processed data. Anomaly identification, labeling, and output are performed on the initially processed data to obtain edge preprocessing data; Data noise reduction and correlation optimization are performed on the edge preprocessed data to obtain correlation-optimized data; Based on the correlation optimization data, perform secondary anomaly identification and classification to obtain anomaly identification and classification data; The anomaly identification and classification data are further optimized and output to obtain cloud-optimized data. Classify and organize the optimized data in the cloud to obtain the categorized and organized data. The categorized and organized data is stored and managed in two ways to obtain storage management data; Output the storage management data to obtain categorized storage data.

5. The method of claim 4, wherein the method further comprises: The step of performing data denoising and correlation optimization on the edge preprocessed data to obtain correlation-optimized data includes: Noise identification and classification are performed on the edge preprocessing data to obtain noise identification and classification data; Adaptive noise reduction is performed based on noise recognition and classification data to obtain noise reduction execution data. The noise reduction effect is verified based on the noise reduction execution data to obtain noise reduction effect verification data. After the noise reduction effect verification data passes, the model with associated parameters is imported to obtain the model import information; Based on the imported model information, construct the association logic to obtain the association logic construction information; Based on the association logic, information is constructed to complete and optimize the data, thereby obtaining data completion and optimization information; Based on the data completion and optimization information, content verification and output are performed to obtain related optimized data.

6. The method of claim 1, wherein the method further comprises: S3 includes: Based on the categorized and stored data, basic energy consumption statistics are obtained; Compliance verification is conducted based on basic energy consumption statistics to obtain compliance verification data; Data is filtered and extracted based on compliance verification data to obtain filtered and extracted data; Output energy consumption analysis data based on the filtered and extracted data; Establish an AI energy consumption prediction model, and use the AI ​​energy consumption prediction model in conjunction with energy consumption analysis data to predict energy consumption trends and obtain energy consumption trend prediction information. The impact of energy consumption trend forecast information is quantified and the results are output to obtain trend analysis data; Perform distribution analysis on the trend analysis data to obtain distribution analysis data; Perform anomaly analysis on the distribution analysis data to obtain anomaly analysis data; Perform optimization analysis on the anomaly analysis data to obtain optimized analysis data; By integrating energy consumption analysis data, trend analysis data, distribution analysis data, anomaly analysis data, and optimization analysis data, dynamic energy consumption analysis data is obtained.

7. The method of claim 6, wherein the method further comprises: The establishment of an AI energy consumption prediction model, combined with energy consumption analysis data, is used to predict energy consumption trends and obtain energy consumption trend prediction information, including: An AI energy consumption prediction model is constructed based on LSTM neural network and integrated housing features; Obtain housing usage demand information, combine housing usage demand information with AI energy consumption prediction model, perform multi-timescale energy consumption trend prediction, and obtain multi-time prediction data; Model validation is performed on multi-time forecast data to obtain multi-time model validation data. Energy consumption trend prediction information is output based on the multi-time model validation data.

8. The method of claim 1, wherein the method further comprises: S4 includes: Energy consumption prediction information is extracted from dynamic energy consumption analysis data to obtain energy consumption prediction extraction information. Based on energy consumption prediction, information is extracted for multi-dimensional energy consumption regulation to obtain multi-dimensional regulation data. The effect of regulation is quantitatively evaluated based on multidimensional regulation data to obtain quantitative evaluation data of the effect. Based on the quantitative evaluation data of the effects, an analysis of insufficient regulation was conducted to obtain data on the analysis of insufficient regulation. Based on the analysis of insufficient regulation data, the regulation strategy is iteratively optimized to obtain data on the effect of iterative optimization.

9. The method of claim 8, wherein the method further comprises: The step of performing a regulatory insufficiency analysis based on the quantitative evaluation data of the effect to obtain regulatory insufficiency analysis data includes: The data from the quantitative evaluation of the effects are structured and classified to obtain classification information; The system identifies deficiencies in the sorted and categorized information by adjusting parameters, logic, scenario adaptation, and model dimensions, thus obtaining multi-dimensional deficiency identification data. The system performs root cause analysis and outputs data on multi-dimensional deficiency identification data to obtain data for regulating deficiency analysis.

10. The integrated dynamic monitoring and analysis system for energy consumption in a building, characterized in that, The system includes: The data acquisition and control module is used to collect data from the integrated house in layers, obtain layered data, and analyze and control the acquisition frequency and accuracy based on the layered data to obtain data acquisition analysis and control data. The classification and storage module is used to perform edge preprocessing on the hierarchical collected data based on the collected analysis and control data to obtain edge preprocessed data, perform secondary optimization on the edge preprocessed data in the cloud to obtain cloud optimized data, and classify and store the cloud optimized data to obtain classified storage data. The energy consumption analysis module is used to perform energy consumption analysis based on the classified stored data, obtain energy consumption analysis data, and perform trend, distribution, anomaly and optimization analysis based on the energy consumption analysis data to obtain dynamic energy consumption analysis data. The effect iteration module is used to perform proactive prediction, coordinated linkage, mobile status and emergency control based on energy consumption prediction and analysis data, obtain multi-dimensional control data, and perform iterative optimization of control effect based on multi-dimensional control data to obtain effect iteration optimization data.

Images