Energy data management method and system

Abstract

The invention relates to the technical field of energy management, and discloses an energy data management method and system, and the method comprises the following steps: S1, data collection: collecting equipment energy consumption, environment and market data; s2, data processing: cleaning abnormal data, and extracting features; s3, optimization model construction: constructing a multi-objective optimization model; s4, optimization scheduling strategy generation: generating and executing a scheduling strategy; s5, dynamic feedback and optimization: adjusting the model and the strategy according to feedback; and S6, user interaction and visualization: providing visualization and target adjustment functions. By constructing a multi-target optimization model, the targets of economy, energy efficiency and environmental protection are comprehensively considered, and supply and demand balance, equipment limitation and environmental constraint are combined, so that global optimization of energy scheduling is realized. An optimization algorithm is adopted to generate a Pareto optimal solution set, a user is supported to dynamically balance multiple targets, an optimal strategy is selected, the comprehensive efficiency of an energy management system is remarkably improved, and diversified requirements are met.

Description

Technical Field

[0001] The present invention relates to the technical field of energy management, and in particular to an energy data management method and system. Background Art

[0002] Energy data management methods and systems play a vital role in the field of modern energy management. Its core function is to achieve efficient utilization and scientific allocation of energy resources by real-time collection, analysis, optimization and scheduling of data involved in the process of energy production, transmission, storage and consumption. With the increasing popularity of renewable energy and the gradual transformation of traditional energy, the complexity of energy data management systems has increased significantly. The coexistence of multiple energy forms (such as photovoltaics, wind power, traditional power grids, etc.) and the widespread distribution of large-scale energy-consuming equipment make energy data management a complex project with multiple dimensions and multiple objectives. At the same time, with the application of technologies such as the Internet of Things, cloud computing and artificial intelligence, energy data management systems are developing towards intelligence and digitalization. However, these advances also put forward higher requirements on the real-time, flexibility, economy and energy efficiency of the system.

[0003] Although existing technologies have achieved the collection and optimization of energy data to a certain extent, there are some common problems that need to be solved. First, traditional energy data management methods are mainly based on single-objective optimization, usually focusing on a single indicator of operating cost or energy efficiency, lacking comprehensive consideration of economy, energy efficiency and environmental protection, and difficult to adapt to complex and changing energy demand scenarios. Secondly, the real-time response capability and dynamic feedback mechanism of existing systems are relatively weak. When faced with fluctuations in energy supply (such as changes in photovoltaic or wind power generation) or changes in energy demand, the adjustment speed is often lagging, affecting the overall operating efficiency. In addition, in terms of user participation, most systems do not provide flexible interactive functions, making it impossible for users to adjust energy scheduling targets according to personalized needs, limiting the adaptability and application scope of the system. Finally, in terms of system architecture design, existing energy data management systems are mostly based on a single module, lacking modularity and synergy, resulting in insufficient scalability and difficulty in coping with increasingly complex energy management scenarios. These problems restrict the potential of energy management systems in efficiency improvement and sustainable development. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides an energy data management method and system, which solve the problems of insufficient multi-objective optimization, delayed real-time response, low user participation and poor system scalability in the existing energy data management methods.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: an energy data management method, comprising the following steps:

[0006] S1. Data collection: Collect energy consumption data, environmental data, and market dynamic data of energy equipment through a sensor network;

[0007] S2. Data processing: Clean, detect and repair anomalies, and extract features from the collected data;

[0008] S3. Optimization model construction: Construct a multi-objective optimization model based on energy operation cost, energy efficiency, and carbon emissions;

[0009] S4. Generation of optimized scheduling strategy: Use an optimization algorithm to generate an energy scheduling strategy to control the operation of energy equipment;

[0010] S5. Dynamic feedback and optimization: Dynamically adjust the optimization model according to the equipment operation status and feedback data, and update the energy scheduling strategy;

[0011] S6. User interaction and visualization: Provide data visualization through a user interface and allow users to adjust the target weights of the optimization model.

[0012] Preferably, in step S1, the sensor network realizes data transmission through the lightweight communication protocol MQTT, and uses the TLS protocol for data encryption to ensure data transmission security.

[0013] Preferably, in step S2, anomalies are detected through the Isolation Forest algorithm for cleaning, and its fitness function is:

[0014]

[0015] where x is the detected data point, k is the number of trees in the constructed Isolation Forest, T i is the i-th decision tree, and pathlength is the path length of the data point x in the tree T i ;

[0016] When S(x) > θ, x is considered an anomaly point, and the anomaly data is repaired using a time series prediction model, and its prediction formula is:

[0017]

[0018] where is the predicted value, and α is the smoothing coefficient.

[0019] Preferably, in step S2, feature extraction is performed on time series data through wavelet transform, and its transform formula is:

[0020]

[0021] where ψ is the mother wavelet function, and a and b are the scale and translation parameters respectively.

[0022] Preferably, in the step S3, the objective function of the multi-objective optimization model includes:

[0023] Minimize the operating cost C:

[0024]

[0025] where C is the total operating cost, is the energy consumption of the i-th device, and P i is the unit price of the i-th type of energy, and C switch,j is the start-stop cost of the j-th device;

[0026] Maximize the energy efficiency η:

[0027]

[0028] Minimize the carbon emission CO 2 :

[0029]

[0030] where CO 2 is the total carbon emission, and EF i is the carbon emission factor of the i-th type of energy.

[0031] Preferably, in the step S3, the multi-objective optimization model is set with the following constraint conditions:

[0032] Energy supply-demand balance constraint:

[0033]

[0034] where D t is the energy demand at time t;

[0035] Device operating state constraint:

[0036] P min,i ≤P i ≤P max,i

[0037] where P min,i and P max,i are the minimum and maximum powers of the i-th device respectively;

[0038] Environmental impact constraint:

[0039] E solar =k·I solar

[0040] where E solar is the electric energy generated by solar energy, I solar is the solar irradiance intensity, and k is the conversion efficiency.

[0041] Preferably, in step S4, the optimization algorithm adopts the non-dominated sorting genetic algorithm, and its optimization process includes the following steps:

[0042] a. Population initialization:

[0043] P 0 = {x 1 , x 2 , …, x N}

[0044] where P 0 is the initial population and N is the population size;

[0045] b. Calculate fitness:

[0046]

[0047] where F(x) is the fitness of solution x, w j is the weight of objective j, and f j (x) is the objective function value;

[0048] c. Generate the Pareto front solution set by non-dominated sorting.

[0049] Preferably, in step S5, the optimization model adopts a rolling optimization mechanism to dynamically predict the energy demand in future periods, and its prediction model is:

[0050]

[0051] where, is the predicted demand at time t, D t-k , D t-k+1 , …, D t-1 are the historical demands in the past k time periods.

[0052] Preferably, in step S6, the user interface provides real-time visualization of the optimization results, including:

[0053] Dynamic trend chart of energy consumption data;

[0054] Optimization results of each objective function;

[0055] Selection and recommendation of Pareto optimal solutions.

[0056] An energy data management system includes:

[0057] A data acquisition module for collecting device energy consumption data, environmental data, and market dynamic data through a sensor network;

[0058] A data processing module for cleaning, repairing, and feature extraction of the collected data;

[0059] An optimization decision-making module, which is used to construct a multi-objective optimization model and generate an energy scheduling strategy;

[0060] A strategy execution module, which is used to control the operation of energy equipment;

[0061] A user interaction module, which is used to provide visualization of the optimization results and allow users to adjust the weights of the optimization objectives.

[0062] The present invention provides an energy data management method and system. It has the following beneficial effects:

[0063] 1. By constructing a multi-objective optimization model, the present invention unifies and integrates the three core objectives of economy, energy efficiency, and environmental protection in energy management, and combines multiple constraints such as energy supply-demand balance, equipment operation limitations, and environmental conditions to achieve global optimization of energy scheduling. At the same time, an optimization algorithm is used to generate a Pareto optimal solution set, enabling users to dynamically balance between multiple objectives and select the optimal strategy, thereby improving the comprehensive efficiency of the energy management system and adapting to diverse application requirements.

[0064] 2. Through real-time data collection, cleaning, feature extraction, and feedback adjustment, the present invention realizes dynamic closed-loop control of energy scheduling. The system combines a rolling optimization mechanism and a time series prediction model, which can accurately predict future energy demands and dynamically update the optimization model parameters to adapt to complex scenarios such as energy price fluctuations, equipment failures, or environmental condition changes. Compared with traditional static optimization schemes, the closed-loop feedback mechanism of the present invention significantly improves the robustness and adaptive ability of the system, ensuring the real-time performance and continuous optimization effect of energy management.

[0065] 3. By displaying the optimization results and operating status of the energy system through a visualization interface and providing a target weight adjustment function, the present invention enables users to flexibly adjust the preferences of the optimization model according to actual needs. The adjusted weights by users can act on the optimization model in real time to dynamically generate a new scheduling plan, thereby enhancing the user-friendliness and application scope of the system. While meeting the personalized needs of users, it also improves the practical application value of energy scheduling. Brief Description of the Drawings

[0066] Figure 1 is a perspective view of the present invention;

[0067] Figure 2 is a schematic diagram of the present invention. Detailed Embodiments

[0068] Next, in combination with the accompanying drawings of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0069] Embodiment 1:

[0070] Please refer to the attached Figure 1 , the embodiment of the present invention provides an energy data management method, including the following steps:

[0071] S1. Data collection: Collect energy device energy consumption data, environmental data, and market dynamic data through a sensor network;

[0072] S2. Data processing: Clean, detect and repair anomalies in the collected data, and extract features;

[0073] S3. Optimization model construction: Build a multi-objective optimization model based on energy operation cost, energy efficiency, and carbon emissions;

[0074] S4. Optimization scheduling strategy generation: Use an optimization algorithm to generate an energy scheduling strategy to control the operation of energy devices;

[0075] S5. Dynamic feedback and optimization: Dynamically adjust the optimization model according to the device operation status and feedback data, and update the energy scheduling strategy;

[0076] S6. User interaction and visualization: Provide data visualization through a user interface and allow users to adjust the target weights of the optimization model;

[0077] In step S1, the sensor network realizes data transmission through the lightweight communication protocol MQTT, and uses the TLS protocol for data encryption to ensure data transmission security;

[0078] In step S2, anomalies are detected through the Isolation Forest algorithm for cleaning, and its fitness function is:

[0079]

[0080] where x is the detected data point, k is the number of trees in the constructed Isolation Forest, T i is the i-th decision tree, and pathlength is the path length of the data point x in the tree T i ;

[0081] When S(x)>θ, x is considered an anomaly point, and the anomaly data is repaired using a time series prediction model, and its prediction formula is:

[0082]

[0083] Among them, is the predicted value, and α is the smoothing coefficient;

[0084] In step S2, feature extraction is performed on the time series data through wavelet transform, and its transformation formula is:

[0085]

[0086] Among them, ψ is the mother wavelet function, and a and b are the scale and translation parameters respectively;

[0087] In step S3, the objective function of the multi-objective optimization model includes:

[0088] Minimize the operating cost C:

[0089]

[0090] C is the total operating cost, is the energy consumption of the i-th device, and P i is the unit price of the i-th type of energy, and C switch,j is the start-stop cost of the j-th device;

[0091] Maximize the energy efficiency η:

[0092]

[0093] Minimize the carbon emission CO 2 :

[0094]

[0095] Among them, CO 2 is the total carbon emission, and EF i is the carbon emission factor of the i-th type of energy;

[0096] In step S3, the multi-objective optimization model has the following constraint conditions:

[0097] Energy supply-demand balance constraint:

[0098]

[0099] Among them, D t is the energy demand at time t;

[0100] Device operating state constraint:

[0101] P min,i ≤P i ≤P max,i

[0102] Among them, Pmin,i and P max,i are the minimum and maximum powers of the i-th device, respectively;

[0103] Environmental impact constraint:

[0104] E solar = k·I solar

[0105] where E solar is the electric energy generated by solar energy, I solar is the solar irradiance intensity, and k is the conversion efficiency;

[0106] In step S4, the optimization algorithm uses the non-dominated sorting genetic algorithm, and its optimization process includes the following steps:

[0107] a. Population initialization:

[0108] P 0 = {x 1 , x 2 , …, x N}

[0109] where P 0 is the initial population and N is the population size;

[0110] b. Calculate fitness:

[0111]

[0112] where F(x) is the fitness of solution x, w j is the weight of objective j, and f j (x) is the objective function value;

[0113] c. Generate the Pareto front solution set by non-dominated sorting;

[0114] In step S5, the optimization model adopts a rolling optimization mechanism to dynamically predict the future energy demand, and its prediction model is:

[0115]

[0116] where, is the predicted demand at time t, D t-k , D t-k+1 , …, D t-1 are the historical demands in the past k time periods;

[0117] In step S6, the user interface provides real-time visualization of the optimization results, including:

[0118] Dynamic trend chart of energy consumption data;

[0119] Optimization results of each objective function;

[0120] Selection and recommendation of Pareto optimal solutions.

[0121] Specifically, in step S1, it collects energy-related data through a sensor network, including device energy consumption data (such as voltage, current, and power), environmental data (such as temperature, humidity, and solar radiation intensity), and market dynamic data (such as energy prices and carbon emission quotas). These data are uploaded to the data processing module through a lightweight communication protocol and encrypted through the TLS / SSL protocol to ensure data security. The sensor network realizes real-time monitoring and unified management of multiple energy systems, making the coverage of energy data collection wider and the response speed faster. At the same time, organizing data through time series facilitates subsequent analysis and processing, improving the timeliness and availability of data. This step realizes the automation and digitization of energy data collection, eliminating the problems of incomplete data and lagging updates in traditional energy management.

[0122] In step S2, during the data processing stage, it cleans, detects, and repairs anomalies in the collected raw data, and extracts key information through feature extraction techniques. The isolation forest algorithm is used in the cleaning stage to detect abnormal data points, ensuring the accuracy and reliability of energy consumption data. For abnormal data, it is repaired in combination with a time series prediction model, enabling the repaired data to retain historical trends and patterns to the greatest extent. In the feature extraction stage, wavelet transform is used to perform multi-scale decomposition on time series data to extract the trend and volatility characteristics of energy consumption data. After data cleaning and feature extraction, the data quality is significantly improved, effectively reducing the influence of noise and outliers, thus providing high-quality data input for subsequent optimization modeling.

[0123] In step S3, it constructs a multi-objective optimization model based on energy operating costs, energy efficiency, and carbon emissions. This model comprehensively considers the three core objectives of economy, energy efficiency, and environmental protection, and at the same time introduces conditions such as energy supply-demand balance, equipment operating status, and environmental constraints to ensure that the optimization results conform to the actual application scenario. In the optimization model, the objective of minimizing operating costs reduces energy procurement and equipment operating expenses; the objective of maximizing energy efficiency improves energy utilization rate and reduces energy waste; the objective of minimizing carbon emissions promotes low-carbon development and environmental protection. Through multi-objective modeling, the system can achieve dynamic balance among multiple optimization objectives, thus meeting customized requirements in different scenarios. This step provides a mathematical basis and decision-making basis for energy scheduling, enhancing the flexibility and intelligence of the system.

[0124] In step S4, the non-dominated sorting genetic algorithm is used to generate an energy scheduling strategy according to the optimization model. The optimization process starts from population initialization and finally obtains a set of Pareto optimal solutions through calculating the fitness function, non-dominated sorting, and Pareto front screening. These solutions provide multiple optimization schemes, and users can select strategies that emphasize economy, energy efficiency, or environmental protection according to actual needs. The scheduling strategy realizes the global optimization of the energy system by controlling the start and stop of equipment and energy distribution. This step combines a rolling optimization mechanism to dynamically predict future energy demands and adjust the scheduling strategy, enhancing the system's response ability to energy demand changes and solving the problems of insufficient rigidity and real-time performance in traditional energy management.

[0125] In step S5, after the energy scheduling strategy is executed, the system collects the operating status and execution results of the equipment through a feedback mechanism and compares them with the expected goals of the optimization model. If a deviation is found, the system will dynamically adjust the parameters of the optimization model to generate an updated energy scheduling strategy. The feedback mechanism ensures that the system can adapt to changes in the external environment and demands, such as energy price fluctuations, equipment failures, or environmental condition changes, enhancing the robustness and adaptability of the optimization model. Through the combination of rolling optimization and dynamic feedback, the system forms a closed-loop control mechanism to continuously optimize the energy management effect and effectively avoid the problem of efficiency decline during long-term operation.

[0126] In step S6, through the user interface, the system visually displays the optimization results in the forms of dynamic trend charts, optimization results of objective functions, and recommendations of Pareto optimal solutions. Users can adjust the objective weights of the optimization model through the interface and observe the real-time changes in the optimization results. The user interaction function provides flexibility and controllability for energy management, enabling users to participate in scheduling decisions and enhancing the practicality of the system. The visual display makes the complex optimization results more intuitive, facilitating user understanding and decision-making, thus enhancing the user experience and application value of the system.

[0127] Embodiment 2:

[0128] Please refer to the appendix Figure 2 , an energy data management system, including:

[0129] A data acquisition module, which is used to collect equipment energy consumption data, environmental data, and market dynamic data through a sensor network;

[0130] A data processing module, which is used to clean, repair, and extract features from the collected data;

[0131] An optimization decision module, which is used to construct a multi-objective optimization model and generate an energy scheduling strategy;

[0132] A strategy execution module, which is used to control the operation of energy equipment;

[0133] A user interaction module, which is used to provide visualization of optimization results and allow users to adjust the weights of optimization objectives.

[0134] Specifically, the data acquisition module is responsible for real-time acquisition of multi-source data of the energy system through a sensor network, including device energy consumption data (voltage, current, power, etc.), environmental data (temperature, humidity, solar irradiance intensity, etc.), and market dynamic data (energy price, carbon emission quota, etc.). This module uses lightweight communication protocols (such as MQTT, CoAP) to achieve efficient data transmission and uses the TLS / SSL protocol to encrypt the data to ensure the security of data transmission. The arrangement of sensor nodes and network design ensure a comprehensive perception of the energy system state, achieving high-timeliness and wide-coverage energy data acquisition. The collected data is organized by time serialization, facilitating the storage and analysis of the subsequent data processing module. The automated acquisition method of this module eliminates the delay and error problems of manual recording in traditional energy management, significantly improving the timeliness and accuracy of data.

[0135] The data processing module cleans, repairs, and extracts features from the collected raw data to ensure the integrity, accuracy, and availability of the data. In the cleaning stage, the Isolation Forest algorithm is used to detect abnormal data points, which can effectively identify data deviations caused by abnormal device operation or sensor failures. In the repair stage, a time series prediction model is used to complement missing or abnormal data, maximizing the retention of the historical patterns and trends of the data. In the feature extraction stage, wavelet transform is used to decompose the time series data to extract key features for capturing the periodicity and mutation characteristics of device operation. Through this module, the problems of noise, anomalies, and redundancy that may exist in the raw data are effectively solved, obtaining high-quality processing results, thus providing reliable data input for the subsequent optimization decision-making module.

[0136] The optimization decision-making module constructs a multi-objective optimization model and generates an energy scheduling strategy through an algorithm. The model comprehensively considers objectives such as minimizing operating costs, maximizing energy efficiency, and minimizing carbon emissions, and sets energy supply-demand balance, device operating status, and environmental constraint conditions to ensure the practical feasibility of the scheduling strategy. The NSGA-II (Non-dominated Sorting Genetic Algorithm) is used to optimize the model, and multiple alternative optimization schemes are generated through population initialization, fitness calculation, and Pareto optimal solution screening. These schemes cover different objective trade-offs, such as emphasizing cost savings, improving efficiency, or reducing carbon emissions, and users can select the best strategy according to specific needs. Through optimized scheduling, this module can achieve global coordination of the energy system, significantly improving the economy and environmental friendliness of energy management.

[0137] The policy execution module converts the energy scheduling policy generated by the optimization decision module into specific device operation instructions, and controls the start / stop of devices, load adjustment, and energy distribution through Internet of Things technology. The execution module supports distributed operation, processes local optimization tasks through edge computing devices, and operates in coordination with the cloud optimization results to achieve the dynamic combination of global and local policies. Key data during the execution process (such as device status and execution results) will be transmitted back to the system in real time for the feedback mechanism to analyze. The high-efficiency execution ability of this module ensures that the optimization policy can be accurately implemented into device operations, and further enhances the dynamic response ability of the scheduling effect through a closed-loop feedback mechanism.

[0138] The user interaction module displays the optimization results and the operating status of the energy system through a visual interface, and provides a target weight adjustment function, allowing users to customize the optimization model according to actual needs. The visual interface intuitively displays data analysis and optimization results in the form of dynamic trend charts, objective function results, and Pareto optimal solution recommendations, helping users quickly understand the dynamics of complex energy systems. By adjusting the weights of the optimization objectives, users can observe the changes in the optimization results under different weight configurations in real time, so as to select the scheduling scheme that best meets their needs. This module enhances the operability of the system, enabling users to flexibly participate in scheduling decisions, and combining professionalism and ease of use.

[0139] In this way, through the close cooperation of the above modules in the system, a closed-loop operation mechanism from data collection to execution feedback is formed. First, the data collection module obtains multi-source data in real time through the sensor network, providing comprehensive perception capabilities for the energy system. These data are transmitted to the data processing module for cleaning, repair, and feature extraction, providing high-quality input for the optimization decision module. The optimization decision module constructs a multi-objective optimization model, generates a set of Pareto optimal solutions through algorithms, and selects the optimal scheduling strategy according to system requirements. The policy execution module converts the selected scheduling scheme into device control instructions, implements them into specific energy distribution and device operations, and at the same time collects execution result data in real time and feeds it back to the optimization decision module. The user interaction module runs throughout the entire process of the system operation. Through visual analysis and target adjustment functions, it provides an interaction interface between users and the system, further improving the customization and dynamic adaptation capabilities of the scheduling strategy.

[0140] Embodiment 3: Traditional energy management method based on single-objective optimization

[0141] Specific steps:

[0142] Data collection: Collect device energy consumption data and environmental data, using periodic recording, and some data comes from manual statistics.

[0143] Data processing: Process the data through simple mean filtering, without feature extraction or repair mechanism.

[0144] Optimized Modeling: A single-objective optimization model is constructed with the sole goal of minimizing operating costs, without considering other objectives.

[0145] Optimized Scheduling: A static scheduling plan is generated using linear programming and directly executed after generation, without supporting real-time adjustment.

[0146] Execution and Analysis: The execution effect of the scheduling plan needs to be manually analyzed and cannot form dynamic feedback.

[0147] Summary:

[0148] Although the single goal of cost minimization is achieved, it lacks comprehensive considerations such as dynamic feedback, energy efficiency, and environmental protection, and it is difficult to meet the complex requirements of modern energy management.

[0149] Example 4: Energy Management Method Based on Static Scheduling

[0150] Specific Steps:

[0151] Data Collection: Sensors collect device and environmental data, but the data collection frequency is low and real-time monitoring cannot be achieved.

[0152] Data Processing: The data is simply screened by thresholds to detect anomalies, and no repair is performed when data is missing.

[0153] Optimized Modeling: A multi-objective optimization model is constructed, with objectives including operating costs, energy efficiency, and environmental protection, but dynamic adjustment is not supported.

[0154] Optimized Scheduling: A scheduling plan is generated based on the traditional genetic algorithm, and the scheduling plan is statically generated once.

[0155] Execution and Adjustment: There is no feedback mechanism, and the optimization model cannot be dynamically adjusted according to environmental or demand changes.

[0156] Summary:

[0157] Although a multi-objective optimization model is constructed, due to the lack of a dynamic feedback mechanism, it cannot adapt to environmental changes, and the real-time and sustainability of the optimization results are poor.

[0158] Example 5: Energy Management System Based on Standard Optimization Model

[0159] Specific Steps:

[0160] Data Collection: Device and environmental data are collected through a sensor network, but market data is not introduced and the data dimension is limited.

[0161] Data Processing: After cleaning the data, it is directly stored without feature extraction and anomaly repair steps.

[0162] Optimization modeling: Build an optimization model based on a single goal (maximizing energy efficiency), and the scheduling plan is generated statically.

[0163] Policy execution: Scheduling policies are applied directly to devices without user intervention and adjustment.

[0164] Result presentation: The optimization results are presented through fixed reports, and real-time visualization and dynamic interaction functions are not provided.

[0165] Summarize:

[0166] The system uses a simple optimization model and fixed display method, which fails to reflect flexibility and real-time performance and is unable to meet the needs of complex energy scenarios.

[0167] Summary:

[0168] Comparative analysis:

[0169] Multi-objective optimization capability: Embodiments 1 and 2 comprehensively consider economy, energy efficiency and environmental protection, and achieve a dynamic balance of multiple objectives; Embodiments 3 to 5 only focus on a single objective and cannot meet complex needs.

[0170] Real-time response capability: Embodiments 1 and 2 combine rolling optimization and feedback mechanisms, significantly improving the ability to adapt to changes in the external environment; Embodiments 4 and 5 lack a real-time feedback mechanism and are difficult to adjust dynamically.

[0171] User participation and flexibility: Embodiments 1 and 2 provide a user interaction module, support target weight adjustment and result visualization, and enhance the operability of the system; Embodiments 3 to 5 lack user participation functions.

[0172] Summarize:

[0173] The technical solution of the present invention (Example 1 and Example 2) is significantly superior to the traditional energy management solution (Examples 3 to 5), especially in terms of multi-objective optimization, real-time feedback, user interaction and system collaborative operation, and embodies advancement and innovation, and is suitable for modern complex energy management scenarios.

[0174] Test experiment:

[0175] Test objectives:

[0176] By testing the five embodiments in the same scenario, the performance of each embodiment is evaluated from four core dimensions: energy scheduling efficiency, multi-objective optimization effect, real-time response capability and user participation flexibility, and the technical advantages of embodiment one and embodiment two (the present invention) are verified.

[0177] Test experiment content:

[0178] Experimental scenario setup: A simulated scenario of an industrial park with 10 devices and 3 types of energy sources (photovoltaic, grid, and wind power) is adopted.

[0179] Test metrics:

[0180] Operating cost (unit: yuan);

[0181] Energy efficiency (percentage);

[0182] Carbon emissions (unit: kg);

[0183] Delay time of optimization and scheduling adjustment (unit: seconds);

[0184] Time to generate a new scheduling plan after user adjustment (unit: seconds).

[0185] Test process:

[0186] 1. Data collection:

[0187] Real-time energy consumption data (such as voltage, current, power) and environmental data (such as temperature, humidity, light intensity) of 10 devices are collected through sensor simulation;

[0188] Data on the dynamic changes in market energy prices are introduced, with the photovoltaic price set at 0 yuan / kWh, the grid price at 0.8 yuan / kWh, and the wind power price at 0.5 yuan / kWh;

[0189] The data is updated every minute, and the collection period is 1 hour (a total of 60 data points).

[0190] 2. Optimization modeling:

[0191] Example 1 and Example 2: A multi-objective optimization model is adopted, with the objectives of minimizing operating cost, maximizing energy efficiency, and minimizing carbon emissions, combined with dynamic feedback and user interaction;

[0192] Example 3: A single-objective optimization model with the sole objective of minimizing operating cost;

[0193] Example 4: A multi-objective optimization model without a dynamic feedback mechanism;

[0194] Example 5: A standard optimization model with the core of maximizing a single objective of energy efficiency and without user interaction.

[0195] 3. Scheduling execution:

[0196] Generate a scheduling plan according to the optimization model and control the start / stop of devices and energy distribution:

[0197] Photovoltaic is given priority, and the remainder is supplemented by the grid and wind power;

[0198] Set the environment to change dynamically. For example, at 30 minutes, the solar irradiance intensity drops sharply, resulting in a 50% reduction in photovoltaic power generation.

[0199] 4. Testing and data recording:

[0200] Record the operating cost, energy efficiency, carbon emissions, optimization adjustment delay time, and generation time after user adjustment every minute.

[0201] Test data and results:

[0202] Summary table of experimental results:

[0203]

[0204] Analysis of test results:

[0205] Example 1 and Example 2 (technical solution of the present invention):

[0206] Economy: The operating cost is 950 yuan. It preferentially uses low-cost energy (photovoltaic), and dynamically adjusts the usage ratio of wind power and the power grid when photovoltaic power is insufficient. The economy is close to that of Example 3.

[0207] Energy efficiency: The energy utilization rate reaches 92%, making full use of efficient clean energy, which is better than all other examples.

[0208] Environmental protection: The carbon emissions are 40 kg, significantly lower than those of Example 3 (60 kg), Example 4 (50 kg), and Example 5 (55 kg).

[0209] Real-time response ability: The adjustment delay time is 10 seconds, and it quickly responds to the sudden reduction in photovoltaic power generation through a dynamic feedback mechanism.

[0210] User flexibility: The time for the user to generate a new scheduling plan after adjusting the target weight is 15 seconds, which reflects strong flexibility.

[0211] Example 3:

[0212] Economy: The operating cost is 880 yuan. Single-objective optimization makes it perform best in cost control.

[0213] Energy efficiency and environmental protection: Due to the lack of optimization of energy efficiency and carbon emissions, the energy utilization rate is only 74%, and the carbon emissions reach 60 kg, which is significantly higher than other examples.

[0214] Real-time and flexibility: Lack of dynamic adjustment mechanism and user interaction function, unable to adapt to environmental changes.

[0215] Example 4:

[0216] Multi-objective optimization: There is a certain balance in operating costs, energy efficiency and carbon emissions, but due to the lack of feedback mechanism, the adjustment delay time is long (30 seconds) and it performs poorly in the scenario of sudden reduction of photovoltaic power.

[0217] User flexibility: No user interaction function is provided, making it difficult to meet diverse needs.

[0218] Embodiment five:

[0219] Standard optimization model: Under the optimization of maximizing the single objective energy efficiency, the energy efficiency reaches 89%, which is better than Examples 3 and 4. However, due to the unbalanced operation cost and carbon emissions, the cost reaches 1,050 yuan and the carbon emissions are 55 kg.

[0220] Real-time and flexibility: Similar to the fourth embodiment, it lacks dynamic adjustment and user interaction functions.

[0221] Summary:

[0222] Comparative analysis:

[0223] Multi-objective optimization capability: Embodiments 1 and 2 comprehensively consider economy, energy efficiency and environmental protection, and achieve a dynamic balance of multiple objectives; Embodiments 3 to 5 only focus on a single objective and cannot meet complex needs.

[0224] Real-time response capability: Embodiments 1 and 2 combine rolling optimization and feedback mechanisms, significantly improving the ability to adapt to changes in the external environment; Embodiments 4 and 5 lack a real-time feedback mechanism and are difficult to adjust dynamically.

[0225] User participation and flexibility: Embodiments 1 and 2 provide a user interaction module, support target weight adjustment and result visualization, and enhance the operability of the system; Embodiments 3 to 5 lack user participation functions.

[0226] Summarize:

[0227] The technical solution of the present invention (Example 1 and Example 2) is significantly superior to the traditional energy management solution (Examples 3 to 5), especially in terms of multi-objective optimization, real-time feedback, user interaction and system collaborative operation, and embodies advancement and innovation, and is suitable for modern complex energy management scenarios.

[0228] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An energy data management method, characterized in that: The following steps are involved: S1. Data collection: Collect energy consumption data of energy equipment, environmental data and market dynamic data through sensor networks; S2, data processing: cleaning, anomaly detection and repair of collected data, and feature extraction; S3. Optimization model construction: construct a multi-objective optimization model based on energy operation cost, energy efficiency and carbon emissions; S4, optimization scheduling strategy generation: using optimization algorithms to generate energy scheduling strategies and control the operation of energy equipment; S5. Dynamic feedback and optimization: Dynamically adjust the optimization model and update the energy scheduling strategy based on the equipment operation status and feedback data; S6. User interaction and visualization: Provide data visualization through the user interaction interface and allow users to adjust the target weights of the optimization model.

2. The energy data management method according to claim 1, characterized in that: In step S1, the sensor network implements data transmission through the lightweight communication protocol MQTT, and uses the TLS protocol to encrypt data to ensure data transmission security.

3. The energy data management method according to claim 1, characterized in that: In the step S2, the abnormal data points are detected by the isolation forest algorithm, and its fitness function is: Among them, x is the detection data point, k is the number of trees in the constructed isolation forest, and T i is the i-th decision tree, and the path length is the path length of the data point x in the tree T i The path length in ; When S(x)>θ, x is considered as an abnormal point, and the abnormal data is repaired using the time series prediction model. The prediction formula is: in, is the predicted value, and α is the smoothing coefficient.

4. The energy data management method according to claim 1, characterized in that: In step S2, feature extraction is performed on time series data by wavelet transformation, and the transformation formula is: Among them, ψ is the mother wavelet function, a and b are the scale and translation parameters respectively.

5. The energy data management method according to claim 1, characterized in that: In the step S3, the objective function of the multi-objective optimization model includes: Minimize operating costs C: C is the total operating cost, P is the energy consumption of the i-th device, i is the unit price of the i-th energy source, C switch,j is the start-up and shutdown cost of the j-th equipment; Maximizing energy efficiency η: Minimizing carbon emissions in, is the total carbon emissions, is the carbon emission factor of the i-th energy source.

6. The energy data management method according to claim 1, characterized in that: In the step S3, the multi-objective optimization model is set with the following constraints: Energy supply and demand balance constraints: Among them, D t is the energy demand at time t; Equipment operating status constraints: P min,i ≤P i ≤P max,i Among them, P min,i and P max,i are the minimum and maximum power of the i-th device respectively; Environmental impact constraints: E solar =k·I solar Among them, E solar The electricity generated by solar energy, I solar is the solar radiation intensity, and k is the conversion efficiency.

7. The energy data management method according to claim 1, characterized in that: In step S4, the optimization algorithm adopts a non-dominated sorting genetic algorithm, and the optimization process includes the following steps: a. Population initialization: P0={x1,x2,…,x N } Among them, P0 is the initial population, N is the population size; b. Calculate fitness: Among them, F(x) is the fitness of solution x, w j is the weight of target j, f j (x) is the objective function value; c. Non-dominated sorting generates the Pareto front solution set.

8. The energy data management method according to claim 1, characterized in that: In step S5, the optimization model adopts a rolling optimization mechanism to dynamically predict the energy demand in the future period, and its prediction model is: in, is the forecast demand at time t, D t-k ,D t-k+1 ,…,D t-1 is the historical demand in the past k time periods.

9. The energy data management method according to claim 1, characterized in that: In step S6, the user interaction interface provides real-time visualization of the optimization results, including: Dynamic trend chart of energy consumption data; Optimization results of each objective function; Selection and recommendation of Pareto optimal solution.

10. An energy data management system, according to an energy data management method according to any one of claims 1 to 9, characterized in that: include: A data collection module, which is used to collect equipment energy consumption data, environmental data and market dynamics data through a sensor network; Data processing module, which is used to clean, repair and extract features of collected data; Optimization decision module, which is used to build a multi-objective optimization model and generate energy scheduling strategies; A strategy execution module, which is used to control the operation of energy equipment; The user interaction module is used to provide visualization of the optimization results and allow users to adjust the weights of the optimization objectives.

Images