Intelligent energy management system based on dynamic fusion of multi-source heterogeneous data

Abstract

The present application relates to the technical field of energy management, and especially relates to a smart energy management system based on dynamic fusion of multi-source heterogeneous data, which comprises a data acquisition unit, a data preprocessing unit, a time period prediction unit, an energy regulation unit, a dynamic evaluation unit, a verification feedback unit and a back-end visual unit; the present application effectively solves the problem of non-uniform data format through dynamic fusion of multi-source heterogeneous data, and reduces energy cost based on energy consumption prediction of the processed multi-source heterogeneous data; each energy regulation scheme is selected and screened based on the information progression mode to obtain a preferred management scheme, that is, to ensure that the energy regulation scheme meets the preset target and realizes fine management of energy, and to analyze the stability and optimization of energy from two points, and then to manage the problems of "stability" and "highness" of energy management according to the information feedback, so as to realize continuous optimization under the premise of stable operation.

Description

Technical Field

[0001] This invention relates to the field of energy management technology, and in particular to a smart energy management system based on the dynamic fusion of multi-source heterogeneous data. Background Technology

[0002] With the rapid development of the social economy, energy consumption is increasing day by day, and the importance of energy management is becoming more and more prominent. At the same time, with the continuous growth of energy demand and the advancement of the "dual carbon" goal, the energy system is transforming from traditional extensive management to refined and intelligent management. As a core means to improve energy utilization efficiency and reduce carbon emissions, smart energy management faces the key challenge of multi-source heterogeneous data processing in its technological development and application.

[0003] Smart energy management systems achieve efficient energy utilization and optimized management through the collection, analysis, and processing of energy data. However, in actual energy management, energy data comes from a wide range of sources, including electricity, gas, water, and renewable energy, and the diverse data types result in multi-source heterogeneous data. This makes it difficult to solve the problems of inconsistent energy data formats and varying quality. At the same time, it is difficult to accurately predict and manage energy consumption and select the best management solutions, leading to reduced energy management efficiency. Furthermore, it is difficult to address the problems of unstable and unsatisfactory energy management, thus hindering the achievement of stable energy operation and continuous optimization.

[0004] To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a smart energy management system based on the dynamic fusion of multi-source heterogeneous data to solve the aforementioned technical defects. This invention effectively solves the problems of inconsistent energy data formats and varying quality by dynamically fusing multi-source heterogeneous data. At the same time, it selects the best energy control scheme based on a progressive information approach to obtain the preferred management scheme. Furthermore, it analyzes the stability and optimization of energy to address the problems of "unstable" and "unimproved" energy management, ultimately achieving continuous optimization under the premise of stable operation.

[0006] The objective of this invention can be achieved through the following technical solution: a smart energy management system based on dynamic fusion of multi-source heterogeneous data, including a data acquisition unit, a data preprocessing unit, a time period prediction unit, an energy regulation unit, a dynamic evaluation unit, a verification feedback unit, and a back-end visualization unit;

[0007] The data acquisition unit is used to collect heterogeneous energy data from different energy monitoring points;

[0008] The data preprocessing unit cleans, transforms, and standardizes the collected heterogeneous energy data;

[0009] The time period prediction unit is used to perform energy prediction and time period segmentation management analysis on the preprocessed heterogeneous energy data to obtain peak and trough time periods.

[0010] The energy control unit is used to perform optimization screening analysis on the basic parameter information of each energy control scheme to obtain the preferred management scheme. The dynamic evaluation unit is used to perform joint evaluation and analysis of the stability and optimization of heterogeneous energy data to determine whether the stability and optimization of energy are normal and obtain risk signals or stable signals. When a risk signal is generated, the collected energy efficiency improvement rate is further processed to obtain compliance signals or non-compliance signals.

[0011] When a risk signal is generated, the verification feedback unit is used to conduct interference source tracing and degree assessment analysis on the energy corresponding to the risk signal, to obtain a single impact item or a synergistic impact item, and further to discriminate the collected optimization and adaptation score to obtain a bottleneck signal or a defect signal.

[0012] Preferably, the analysis process of the time period prediction unit is as follows:

[0013] Based on the preprocessed heterogeneous energy data, structured time-series data and unstructured time-series data of each energy source are obtained. The structured time-series data means that the structured data in the heterogeneous energy data is sorted according to the timestamp and converted into a numerical sequence of a preset fixed length. The unstructured time-series data means that the unstructured data in the heterogeneous energy data is converted into a semantic vector sequence through a pre-set language model or into a feature sequence through a feature extractor.

[0014] Structured and unstructured time-series data are uniformly converted into a set of multimodal input sequences. The set of multimodal input sequences is then processed, and the fusion feature vectors of each energy source are obtained based on the processed set of multimodal input sequences.

[0015] The obtained fused feature vector is input into a pre-set energy consumption prediction model, and the predicted energy consumption report output by the pre-set energy consumption prediction model is obtained.

[0016] Based on the comparison between the predicted energy consumption characteristic curve in the predicted energy consumption report and the historical energy consumption characteristic curve, the peak and trough periods in the predicted energy consumption characteristic curve are obtained.

[0017] Preferably, the analysis process of the energy control unit is as follows:

[0018] Energy control plans are generated based on predicted energy consumption reports. Basic parameter information for each energy control plan is obtained, including target achievement rate and feasibility score. The product of each parameter in the basic parameter information and its corresponding pre-set weight coefficient is set as the plan priority coefficient. The plan priority coefficients are sorted in descending order, and the difference between the highest and second highest scores in the plan priority coefficients is obtained. The difference between the highest and second highest scores in the plan priority coefficients is set as the dynamic priority coefficient. The dynamic priority coefficient is checked to see if it is less than a preset dynamic priority coefficient threshold. If not, the energy control plan corresponding to the highest score in the plan priority coefficients is determined as the preferred management plan. If so, the influence coefficients of the energy control plans corresponding to the highest and second highest scores in the plan priority coefficients are obtained, and the energy control plan corresponding to the maximum value of the influence coefficient is set as the preferred management plan.

[0019] Preferably, the execution feasibility score represents the difference between the product of the device response latency rate and the instruction transmission failure rate multiplied by the corresponding pre-set weighting coefficients and the full score; the influence coefficient represents the value obtained by subtracting the equipment loss cost from the energy saving cost under the simulation scheme.

[0020] Preferably, the analysis process of the dynamic evaluation unit is as follows:

[0021] Based on the preprocessed heterogeneous energy data, the dynamic energy efficiency coefficients of each energy source during its operating period are obtained. The dynamic energy efficiency coefficient represents the ratio between the standard deviation and the average value of the energy utilization rate or renewable energy substitution rate. The dynamic energy efficiency coefficient is judged to see if it exceeds the preset dynamic energy efficiency coefficient threshold. If not, a stable signal is generated; if so, a risk signal is generated.

[0022] Preferably, when a risk signal is generated, the energy efficiency improvement rate during the energy operation period corresponding to the risk signal is obtained based on the preprocessed heterogeneous energy data. The energy efficiency improvement rate represents the value obtained by dividing the difference between the energy efficiency before and after the energy efficiency scheme is improved by the energy efficiency before the improvement. The system then judges whether the energy efficiency improvement rate exceeds the preset energy efficiency improvement rate threshold. If it does, a compliance signal is generated; otherwise, a non-compliance signal is generated.

[0023] Preferably, the analysis process of the verification feedback unit is as follows:

[0024] Based on the preprocessed heterogeneous energy data, multiple verification results of the energy corresponding to the risk signal were obtained. The multiple verifications include data layer verification, equipment layer verification and system layer verification. The multiple verification results include normal and abnormal.

[0025] Based on the implementation order of multiple verifications, verifications with abnormal results are set as interference items, and the number of interference items is judged. If the number of interference items is equal to 1, it is judged as a single influencing item; if the number of interference items is not equal to 1, it is judged as a synergistic influencing item.

[0026] Preferably, the optimized adaptation score of the energy corresponding to the non-compliant signal is obtained. The optimized adaptation score is the product of the optimized execution score and the execution adaptability score and the corresponding preset weight coefficient. The optimized execution score represents the degree of fit between the actual energy efficiency scheme execution and the preset energy efficiency scheme execution. The execution adaptability score represents the degree of matching between the energy efficiency scheme and the current energy management system.

[0027] The system then determines whether the optimization and adaptation score exceeds the preset optimization and adaptation score threshold. If it does, a bottleneck signal is generated; otherwise, a defect signal is generated.

[0028] The beneficial effects of this invention are as follows:

[0029] This invention effectively solves the problems of inconsistent energy data formats and varying quality by dynamically fusing multi-source heterogeneous data, thereby improving the usability and analytical value of the data. Based on the processed multi-source heterogeneous data, energy consumption is predicted, and energy allocation is adjusted in a timely manner during the prediction period to achieve optimal energy allocation, effectively improve energy efficiency, reduce energy costs, and help achieve the goal of energy conservation and emission reduction.

[0030] This invention selects the best energy control scheme based on a progressive information approach, analyzes it through multi-dimensional data fusion and evaluation, and analyzes the impact of scheme implementation to obtain the preferred management scheme. This ensures that the energy control scheme not only meets the preset goals but also achieves intelligent and refined energy management, thereby improving the management effect of each energy source.

[0031] This invention analyzes energy from the perspectives of stability and optimization, enabling timely early warning and targeted management of existing defects. It uses an in-depth approach to identify energy disturbances, determining whether the energy anomalies corresponding to risk signals are caused by a single or synergistic effect. Based on the feedback, it then addresses the issues of "unstable" and "unimproved" energy management, ultimately achieving continuous optimization under the premise of stable operation. Attached Figure Description

[0032] The invention will now be further described with reference to the accompanying drawings;

[0033] Figure 1 This is a flowchart of the system of the present invention;

[0034] Figure 2 This is a partial analysis reference diagram of the present invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments;

[0037] Example 1:

[0038] Please see Figures 1 to 2 As shown, the present invention is a smart energy management system based on dynamic fusion of multi-source heterogeneous data, including a data acquisition unit, a data preprocessing unit, a time period prediction unit, an energy regulation unit, a dynamic evaluation unit, a verification feedback unit, and a back-end visualization unit. The data acquisition unit and the data preprocessing unit have a one-way communication connection. The data preprocessing unit has a one-way communication connection with both the time period prediction unit and the dynamic evaluation unit. The time period prediction unit has a one-way communication connection with both the energy regulation unit and the back-end visualization unit. The energy regulation unit has a one-way communication connection with the back-end visualization unit. The dynamic evaluation unit has a one-way communication connection with both the verification feedback unit and the back-end visualization unit. The verification feedback unit has a one-way communication connection with the back-end visualization unit.

[0039] The data acquisition unit is used to collect heterogeneous energy data from different energy monitoring points. Heterogeneous energy data refers to the collective information such as environmental data, equipment operating parameters, and consumption data of each energy source.

[0040] For example, by setting up various sensors such as temperature sensors and humidity sensors at various energy monitoring points, real-time data on the energy usage environment (temperature, humidity), equipment operating parameters (operating temperature, operating power), and consumption data (energy consumption per unit time, electricity consumption) can be collected; at the same time, energy data such as voltage, current, gas flow, water consumption, and solar power generation can be received from energy supply ends such as the power grid, gas company, and water company through data interfaces.

[0041] The data preprocessing unit performs preprocessing such as cleaning, conversion, and standardization on the collected heterogeneous energy data;

[0042] The time-period prediction unit is used to perform energy forecasting and time-period segmentation management analysis on the preprocessed heterogeneous energy data to obtain peak and trough periods. The specific energy forecasting and time-period segmentation management analysis process is as follows:

[0043] Based on the preprocessed heterogeneous energy data, structured time-series data and unstructured time-series data of each energy source are obtained. The structured time-series data means that the structured data (such as electricity, temperature, etc.) in the heterogeneous energy data is sorted according to the timestamp and converted into a numerical sequence of a preset fixed length. The unstructured time-series data means that the unstructured data (equipment fault log text, video monitoring, etc.) in the heterogeneous energy data is converted into a semantic vector sequence through a pre-set language model (such as BERT) or into a feature sequence through a feature extractor (such as CNN to extract video frame features).

[0044] Structured and unstructured time-series data are uniformly converted into a set of multimodal input sequences. The set of multimodal input sequences is then processed, including data format standardization and spatiotemporal dimension alignment. Based on the processed set of multimodal input sequences, the fusion feature vectors of each energy source are obtained.

[0045] The obtained fused feature vector is input into a pre-set energy consumption prediction model, and the predicted energy consumption report output by the pre-set energy consumption prediction model is obtained.

[0046] Based on the comparison between the predicted energy consumption characteristic curve in the predicted energy consumption report and the historical energy consumption characteristic curve, the peak and trough periods in the predicted energy consumption characteristic curve are obtained. The back-end visual unit immediately responds to the peak and trough periods and marks them in red. Then, based on the peak and trough periods, energy is precisely adjusted, which helps to provide a basis for energy dispatch and thus helps to improve the efficiency of energy management.

[0047] Example 2:

[0048] The energy control unit is used to perform a selection and screening analysis on the basic parameter information of various energy control schemes to obtain the preferred management scheme. The specific selection and screening analysis process is as follows:

[0049] Energy control schemes are generated based on predicted energy consumption reports. Basic parameter information for each energy control scheme is obtained, including target achievement and feasibility score. The product of each parameter in the basic parameter information and its corresponding pre-set weight coefficient is set as the scheme priority coefficient. The scheme priority coefficients are sorted in descending order, and the difference between the highest and second highest scores in the scheme priority coefficients is obtained. This difference is set as the dynamic priority coefficient, and the dynamic priority coefficient is checked to see if it is less than a preset dynamic priority coefficient threshold. If not, the energy control scheme corresponding to the highest score in the scheme priority coefficient is determined as the preferred management scheme. If so, the influence coefficients of the energy control schemes corresponding to the highest and second highest scores in the scheme priority coefficients are obtained. The influence coefficient represents the value obtained by subtracting equipment loss cost from the energy saving cost under the simulated scheme. The energy control scheme corresponding to the maximum value of the influence coefficient is set as the preferred management scheme. The back-end visual unit immediately responds to the preferred management scheme and executes it, thus ensuring that the energy control scheme meets the preset target and realizes intelligent and refined energy management to improve the management effect of each energy source.

[0050] Among them, the execution feasibility score (0-100) represents the difference between the product of the device response latency rate and the command transmission failure rate and the corresponding pre-set weight coefficient and the full score.

[0051] Example 3:

[0052] The dynamic evaluation unit is used to perform joint stability and optimization evaluation analysis on heterogeneous energy data to determine whether the stability and optimization of the energy are normal. The specific joint stability and optimization evaluation analysis process is as follows:

[0053] Based on the preprocessed heterogeneous energy data, the dynamic energy efficiency coefficients of each energy source during its operating period are obtained. The dynamic energy efficiency coefficient represents the ratio between the standard deviation and the average value of energy utilization rate or renewable energy substitution rate. It should be noted that the dynamic energy efficiency coefficient is a key parameter for measuring the stability of energy efficiency indicators.

[0054] It also determines whether the dynamic energy efficiency coefficient exceeds the preset dynamic energy efficiency coefficient threshold. If not, a stable signal is generated; if so, a risk signal is generated.

[0055] When a risk signal is generated, the energy efficiency improvement rate within the corresponding energy operating period is obtained based on the preprocessed heterogeneous energy data. The energy efficiency improvement rate represents the value obtained by dividing the difference between the energy efficiency before and after the energy efficiency scheme is improved by the energy efficiency before the improvement. The system also judges whether the energy efficiency improvement rate exceeds the preset energy efficiency improvement rate threshold. If it does, a compliance signal is generated; otherwise, a non-compliance signal is generated.

[0056] The back-end visual unit immediately responds to the compliance signal or non-compliance signal and immediately displays the preset warning text corresponding to the compliance signal or non-compliance signal, so as to intuitively understand whether the stability and optimization of each energy source meet the standards based on the information feedback, so as to provide timely warning feedback and targeted management for any defects.

[0057] When a risk signal is generated, the verification feedback unit is used to perform interference source tracing and severity assessment analysis on the energy source corresponding to the risk signal, and to determine whether the energy anomaly corresponding to the risk signal is a single impact or a synergistic impact. The specific interference source tracing and severity assessment analysis process is as follows:

[0058] Based on the preprocessed heterogeneous energy data, multiple verification results of the energy corresponding to the risk signal were obtained. The multiple verifications include data layer verification, equipment layer verification and system layer verification. The multiple verification results include normal and abnormal.

[0059] For example; data layer verification: whether the data collected by the data acquisition device is valid; device layer verification: whether the operating status of the device is normal; system layer verification: whether the energy input is stable and whether the environment is abnormal, etc.

[0060] Based on the implementation order of multiple verifications, the verification corresponding to the abnormal results of multiple verifications is set as interference items, and the number of interference items is judged. If the number of interference items is equal to 1, it is judged as a single impact item. If the number of interference items is not equal to 1, it is judged as a collaborative impact item. The back-end visual unit immediately displays the interference items corresponding to the single impact item or the collaborative impact item. On the one hand, this helps to accurately verify the interference items corresponding to the abnormal energy stability, and on the other hand, it helps to understand the degree of interference to the abnormal energy stability.

[0061] At the same time, the optimization and adaptation score of the energy corresponding to the non-compliant signal is obtained. The optimization and adaptation score is the product of the optimization execution score and the execution adaptability score and the corresponding preset weight coefficient. The optimization execution score represents the degree of fit between the actual energy efficiency scheme execution and the preset energy efficiency scheme execution. The execution adaptability score represents the degree of matching between the energy efficiency scheme and the current energy management system.

[0062] The system then determines whether the optimization and adaptation score exceeds the preset optimization and adaptation score threshold. If it does, a bottleneck signal is generated; otherwise, a defect signal is generated. The backend visual unit immediately displays the preset warning text corresponding to the bottleneck signal or defect signal. Specifically, the preset warning text corresponding to the bottleneck signal is: "Energy management has reached its limit," and the preset warning text corresponding to the defect signal is: "Optimization deviation." Based on the information feedback, the system then manages the issues of "unstable" and "unimproved" energy management systems in a targeted manner, ultimately achieving continuous optimization under the premise of stable operation.

[0063] In summary, this invention effectively solves the problems of inconsistent energy data formats and varying quality by dynamically fusing multi-source heterogeneous data, improving data usability and analytical value. Furthermore, energy consumption prediction based on the processed multi-source heterogeneous data provides a scientific basis for energy management strategy formulation, enhancing the intelligence level of energy management. Timely adjustments to energy allocation during the prediction period optimize energy configuration, effectively improving energy efficiency, reducing energy costs, and contributing to energy conservation and emission reduction goals. Finally, the invention employs a progressive information approach to select the best energy control schemes and analyzes them through multi-dimensional data fusion and evaluation. Along with the implementation of the plan, an impact analysis is conducted to determine the preferred management plan. This plan ensures that the energy regulation plan not only meets the preset objectives but also achieves intelligent and refined energy management to improve the management effectiveness of each energy source. The analysis also considers the stability and optimization of energy to provide timely early warning feedback and targeted management for any deficiencies. Through in-depth analysis, energy interference is identified to determine whether the energy anomalies corresponding to risk signals are caused by a single or synergistic effect. Based on the feedback information, targeted management is implemented to address the issues of "unstable" and "unimproved" energy management, ultimately achieving continuous optimization under the premise of stable operation.

[0064] The threshold is set for comparative analysis of results to determine whether they are good or bad. The value of the threshold is determined by a combination of large-scale model analysis of sample data and human experience. It can also be adjusted appropriately based on seasonal or common-sense influencing factors.

[0065] The size of the coefficient is a specific value obtained by quantifying each parameter to facilitate subsequent comparison. The size of the coefficient depends on the amount of sample data and the corresponding operating coefficient initially set by those skilled in the art for each set of sample data; as long as it does not affect the proportional relationship between the parameter and the quantified value.

[0066] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A smart energy management system based on dynamic fusion of multi-source heterogeneous data, characterized in that, It includes a data acquisition unit, a data preprocessing unit, a time period prediction unit, an energy control unit, a dynamic evaluation unit, a verification and feedback unit, and a back-end visualization unit; The data acquisition unit is used to collect heterogeneous energy data from different energy monitoring points; The data preprocessing unit cleans, transforms, and standardizes the collected heterogeneous energy data; The time period prediction unit is used to perform energy prediction and time period segmentation management analysis on the preprocessed heterogeneous energy data to obtain peak and trough time periods. The energy control unit is used to perform optimization screening analysis on the basic parameter information of each energy control scheme to obtain the preferred management scheme. The dynamic evaluation unit is used to perform joint evaluation and analysis of the stability and optimization of heterogeneous energy data to determine whether the stability and optimization of energy are normal and obtain risk signals or stable signals. When a risk signal is generated, the collected energy efficiency improvement rate is further processed to obtain compliance signals or non-compliance signals. When a risk signal is generated, the verification feedback unit is used to conduct interference source tracing and degree assessment analysis on the energy corresponding to the risk signal, to obtain a single impact item or a synergistic impact item, and further to discriminate the collected optimization and adaptation score to obtain a bottleneck signal or a defect signal. The analysis process of the energy control unit is as follows: Energy control schemes are generated based on predicted energy consumption reports. Basic parameter information for each energy control scheme is obtained, including target achievement and feasibility score. The product of each parameter in the basic parameter information and its corresponding pre-set weight coefficient is set as the scheme priority coefficient. The scheme priority coefficients are sorted in descending order. The difference between the highest and second highest scores in the scheme priority coefficients is obtained and set as the dynamic priority coefficient. The dynamic priority coefficient is then judged to see if it is less than the preset dynamic priority coefficient threshold. If not, the energy control scheme corresponding to the highest score in the scheme priority coefficient is determined as the preferred management scheme. If so, the influence coefficients of the energy control schemes corresponding to the highest and second highest scores in the scheme priority coefficients are obtained, and the energy control scheme corresponding to the maximum value of the influence coefficient is set as the preferred management scheme. The analysis process of the verification feedback unit is as follows: Based on the preprocessed heterogeneous energy data, multiple verification results of the energy corresponding to the risk signal were obtained. The multiple verifications include data layer verification, equipment layer verification and system layer verification. The multiple verification results include normal and abnormal. Based on the implementation order of multiple verifications, the verifications corresponding to abnormal results of multiple verifications are set as interference items, and the number of interference items is judged. If the number of interference items is equal to 1, it is judged as a single influencing item; if the number of interference items is not equal to 1, it is judged as a synergistic influencing item. The optimization and adaptation score of the energy corresponding to the non-compliance signal is obtained. The optimization and adaptation score is the product of the optimization execution score and the execution adaptability score and the corresponding preset weight coefficient. The optimization execution score represents the degree of fit between the actual energy efficiency scheme execution and the preset energy efficiency scheme execution. The execution adaptability score represents the degree of matching between the energy efficiency scheme and the current energy management system. The system then determines whether the optimization and adaptation score exceeds the preset optimization and adaptation score threshold. If it does, a bottleneck signal is generated; otherwise, a defect signal is generated.

2. The smart energy management system based on dynamic fusion of multi-source heterogeneous data according to claim 1, characterized in that, The analysis process of the time period prediction unit is as follows: Based on the preprocessed heterogeneous energy data, structured time-series data and unstructured time-series data of each energy source are obtained. The structured time-series data means that the structured data in the heterogeneous energy data is sorted according to the timestamp and converted into a numerical sequence of a preset fixed length. The unstructured time-series data means that the unstructured data in the heterogeneous energy data is converted into a semantic vector sequence through a pre-set language model or into a feature sequence through a feature extractor. Structured and unstructured time-series data are uniformly converted into a set of multimodal input sequences. The set of multimodal input sequences is then processed, and the fusion feature vectors of each energy source are obtained based on the processed set of multimodal input sequences. The obtained fused feature vector is input into a pre-set energy consumption prediction model, and the predicted energy consumption report output by the pre-set energy consumption prediction model is obtained. Based on the comparison between the predicted energy consumption characteristic curve in the predicted energy consumption report and the historical energy consumption characteristic curve, the peak and trough periods in the predicted energy consumption characteristic curve are obtained.

3. The smart energy management system based on dynamic fusion of multi-source heterogeneous data according to claim 1, characterized in that, The execution feasibility score represents the difference between the product of the device response latency rate and the command transmission failure rate multiplied by the corresponding pre-set weighting coefficients and the full score; the impact coefficient represents the value obtained by subtracting the equipment loss cost from the energy saving cost under the simulation scheme.

4. The smart energy management system based on dynamic fusion of multi-source heterogeneous data according to claim 1, characterized in that, The analysis process of the dynamic evaluation unit is as follows: Based on the preprocessed heterogeneous energy data, the dynamic energy efficiency coefficients of each energy source during its operating period are obtained. The dynamic energy efficiency coefficient represents the ratio between the standard deviation and the average value of the energy utilization rate or renewable energy substitution rate. The dynamic energy efficiency coefficient is judged to see if it exceeds the preset dynamic energy efficiency coefficient threshold. If not, a stable signal is generated; if so, a risk signal is generated.

5. The smart energy management system based on dynamic fusion of multi-source heterogeneous data according to claim 4, characterized in that, When a risk signal is generated, the energy efficiency improvement rate during the corresponding energy operation period is obtained based on the preprocessed heterogeneous energy data. The energy efficiency improvement rate is the value obtained by dividing the difference between the energy efficiency before and after the energy efficiency scheme is improved by the energy efficiency before the improvement. The system then judges whether the energy efficiency improvement rate exceeds the preset energy efficiency improvement rate threshold. If it does, a compliance signal is generated; otherwise, a non-compliance signal is generated.

Images