Energy conservation and carbon reduction multi-objective optimization methods, systems, storage media and computers

Abstract

The application provides an energy-saving and carbon-reducing multi-objective optimization method, system, storage medium and computer, which comprises the following steps: processing energy use data of a target region based on space-time characteristics to construct space-time data; splitting the space-time data based on energy use users to construct optimization tasks; building corresponding data constraints in the optimization tasks; using a feature processing model to process the space-time data to obtain an enhanced feature vector, and optimizing the optimization tasks after data constraints on the feature processing model, and adding an attention mechanism to the feature processing model to construct a multi-objective processing model; coupling the multi-objective processing model with a non-dominated sorting genetic algorithm to construct a multi-objective optimization model, and generating corresponding optimization decisions according to optimization targets of the target region and optimal solutions of the multi-objective optimization model. The application continuously predicts and optimizes the multi-objective processing model based on real-time states to improve the adaptability and robustness of the model.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a multi-objective optimization method, system, storage medium, and computer for energy saving and carbon reduction. Background Technology

[0002] Traditional energy-saving optimization methods typically employ two approaches: one relies on energy consumption models based on physical rules or simple empirical processing, while the other uses regression-based neural network models for energy consumption prediction. However, building an energy consumption model requires a large number of system parameters, making the modeling process difficult and costly in complex industrial environments. Furthermore, the constructed energy consumption model cannot adaptively optimize based on the aging of energy-consuming equipment and environmental conditions, leading to decreased model accuracy and unsustainable optimization effects. On the other hand, using neural network models for prediction usually involves single-task learning of energy consumption and carbon emissions, resulting in accuracy that fails to meet preset requirements and an inability to adjust based on the real-time status of energy consumption and carbon emission data. Summary of the Invention

[0003] Based on this, the purpose of the present invention is to provide a multi-objective optimization method, system, storage medium, and computer for energy saving and carbon reduction, so as to at least solve the shortcomings of the above-mentioned technologies.

[0004] This invention proposes a multi-objective optimization method for energy conservation and carbon reduction, comprising:

[0005] Energy consumption data of the target area is collected in real time through IoT terminals, and the energy consumption data is processed based on spatiotemporal characteristics to construct corresponding spatiotemporal data.

[0006] Based on the energy users in the target area, the spatiotemporal data is split to construct optimization tasks corresponding to each energy user;

[0007] The optimization tasks of each energy user are processed using a preset dynamic window algorithm to establish corresponding data constraints in the optimization tasks.

[0008] A feature processing model is constructed, and the feature processing model is used to perform feature processing on the spatiotemporal data to obtain the corresponding enhanced feature vector;

[0009] The feature processing model is optimized using the enhanced feature vector and the optimization task after data constraints, and an attention mechanism is added to the feature processing model to construct the corresponding multi-objective processing model;

[0010] The multi-objective processing model is coupled with a non-dominated sorting genetic algorithm to construct a corresponding multi-objective optimization model, and an optimization decision is generated based on the optimization objective of the target region and the optimal solution of the multi-objective optimization model.

[0011] Furthermore, the steps of collecting energy consumption data of the target area in real time through IoT terminals and processing the energy consumption data based on spatiotemporal characteristics to construct corresponding spatiotemporal data include:

[0012] Acquire energy consumption data collected by data sensors installed in the target area, wherein the energy consumption data includes equipment power, ambient temperature, and load status;

[0013] The energy consumption data is cleaned and format converted to obtain corresponding preprocessed data. The preprocessed data is then subjected to data interpolation and timestamp alignment to obtain preliminary data.

[0014] Based on the location information of the data sensors in the target area, the preliminary data is mapped to construct corresponding spatiotemporal features and obtain corresponding spatiotemporal data.

[0015] Furthermore, the step of splitting the spatiotemporal data based on the energy users in the target area to construct the optimization tasks corresponding to each energy user includes:

[0016] Based on the energy users in the target area, the spatiotemporal data is split to obtain independent subsets of energy consumption data for each energy user;

[0017] The energy consumption data subsets of each energy user are analyzed independently to extract the corresponding energy consumption characteristics, and corresponding optimization tasks are constructed based on the energy consumption needs of each energy user.

[0018] Furthermore, the step of processing the optimization tasks of each energy user using a preset dynamic window algorithm to establish corresponding data constraints in the optimization tasks includes:

[0019] A dynamic window algorithm is initialized for the optimization tasks of each energy user. At the start of each optimization, a dynamic time window is created, and data prediction is performed within the dynamic time window.

[0020] Based on the type of energy consumption data, corresponding data constraints are constructed, and the data constraints are added to the dynamic time window for iterative optimization to obtain the optimized task after data constraints.

[0021] Furthermore, the step of using the feature processing model to perform feature processing on the spatiotemporal data to obtain the corresponding enhanced feature vector includes:

[0022] The spatiotemporal data are processed using the aforementioned feature processing model to calculate the mean power and load fluctuation variance in the spatiotemporal data.

[0023] The frequency domain features of the spatiotemporal data are extracted, and the power mean, the load fluctuation variance, and the frequency domain features are fused to obtain the corresponding enhanced feature vector.

[0024] This invention also proposes a multi-objective optimization system for energy saving and carbon reduction, comprising:

[0025] The data acquisition module is used to collect energy consumption data of the target area in real time through IoT terminals, and process the energy consumption data based on spatiotemporal characteristics to construct corresponding spatiotemporal data.

[0026] The data splitting module is used to split the spatiotemporal data based on the energy users in the target area to construct optimization tasks corresponding to each energy user.

[0027] The data processing module is used to process the optimization tasks of each energy user using a preset dynamic window algorithm, so as to build corresponding data constraints in the optimization tasks.

[0028] The feature processing module is used to construct a feature processing model and use the feature processing model to perform feature processing on the spatiotemporal data to obtain the corresponding enhanced feature vector;

[0029] The model building module is used to optimize the feature processing model using the enhanced feature vector and the optimization task after data constraints, and to add an attention mechanism to the feature processing model to build a corresponding multi-objective processing model.

[0030] The objective optimization module is used to couple the multi-objective processing model with the non-dominated sorting genetic algorithm to construct a corresponding multi-objective optimization model, and to generate corresponding optimization decisions based on the optimization objective of the target region and the optimal solution of the multi-objective optimization model.

[0031] Furthermore, the data acquisition module is specifically used for:

[0032] Acquire energy consumption data collected by data sensors installed in the target area, wherein the energy consumption data includes equipment power, ambient temperature, and load status;

[0033] The energy consumption data is cleaned and format converted to obtain corresponding preprocessed data. The preprocessed data is then subjected to data interpolation and timestamp alignment to obtain preliminary data.

[0034] Based on the location information of the data sensors in the target area, the preliminary data is mapped to construct corresponding spatiotemporal features and obtain corresponding spatiotemporal data.

[0035] Furthermore, the data splitting module is specifically used for:

[0036] Based on the energy users in the target area, the spatiotemporal data is split to obtain independent subsets of energy consumption data for each energy user;

[0037] The energy consumption data subsets of each energy user are analyzed independently to extract the corresponding energy consumption characteristics, and corresponding optimization tasks are constructed based on the energy consumption needs of each energy user.

[0038] Furthermore, the data processing module is specifically used for:

[0039] A dynamic window algorithm is initialized for the optimization tasks of each energy user. At the start of each optimization, a dynamic time window is created, and data prediction is performed within the dynamic time window.

[0040] Based on the type of energy consumption data, corresponding data constraints are constructed, and the data constraints are added to the dynamic time window for iterative optimization to obtain the optimized task after data constraints.

[0041] Furthermore, the feature processing module is specifically used for:

[0042] The spatiotemporal data are processed using the aforementioned feature processing model to calculate the mean power and load fluctuation variance in the spatiotemporal data.

[0043] The frequency domain features of the spatiotemporal data are extracted, and the power mean, the load fluctuation variance, and the frequency domain features are fused to obtain the corresponding enhanced feature vector.

[0044] The present invention also proposes a storage medium on which a computer program is stored, which, when executed by a processor, implements the above-mentioned multi-objective optimization method for energy saving and carbon reduction.

[0045] The present invention also proposes a computer, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned energy-saving and carbon-reduction multi-objective optimization method.

[0046] The energy-saving and carbon-reduction multi-objective optimization method, system, storage medium, and computer of this invention collect energy consumption data of the target area in real time, add spatiotemporal features to the energy consumption data, and perform data segmentation based on the obtained spatiotemporal data and energy users to construct optimization tasks for each energy user. The search efficiency of task processing is improved through data constraints, thereby improving the processing accuracy and efficiency of the model. By constructing a multi-objective processing model based on an attention mechanism, energy consumption and carbon emissions are accurately predicted. The non-dominated sorting genetic algorithm is coupled with the multi-objective processing model to find the optimal solution that balances energy saving and carbon reduction objectives. The multi-objective processing model is continuously predicted and optimized based on real-time status to improve the model's adaptability and robustness. Attached Figure Description

[0047] Figure 1 This is a flowchart of the multi-objective optimization method for energy saving and carbon reduction in the first embodiment of the present invention;

[0048] Figure 2 This is a structural block diagram of the energy-saving and carbon-reduction multi-objective optimization system in the second embodiment of the present invention;

[0049] Figure 3 This is a structural block diagram of the computer in the third embodiment of the present invention.

[0050] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0051] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0053] Example 1

[0054] Please see Figure 1 The figure shows a multi-objective optimization method for energy saving and carbon reduction in the first embodiment of the present invention, the method specifically including steps S101 to S106:

[0055] S101, collect energy consumption data of the target area in real time through the Internet of Things terminal, and process the energy consumption data based on spatiotemporal characteristics to construct corresponding spatiotemporal data;

[0056] Furthermore, step S101 specifically includes steps S1011 to S1013:

[0057] S1011, acquire energy consumption data collected by data sensors installed in the target area, wherein the energy consumption data includes equipment power, ambient temperature and load status;

[0058] S1012, perform data cleaning and format conversion on the energy consumption data to obtain corresponding preprocessed data, and perform data interpolation and timestamp alignment on the preprocessed data to obtain preliminary data;

[0059] S1013, Based on the location information of the data sensor in the target area, the preliminary data is mapped to construct the corresponding spatiotemporal features to obtain the corresponding spatiotemporal data.

[0060] In practice, data is collected through data sensors pre-installed in the target area. These sensors are connected to smart IoT terminals in the target area, including smart meters, water meters, gas meters, and other sensors capable of collecting energy consumption data from different monitoring points in real time. During data collection, corresponding timestamps and spatial location identifiers are added to the data. The energy consumption data includes device power. Ambient temperature and load status Spatial location identifiers include terminal ID and coordinate information;

[0061] Specifically, the collected energy consumption data will be cleaned and format converted to unify the data and improve the accuracy of subsequent processing. To avoid data loss, the preprocessed data will be interpolated and the interpolated data will be timestamped to unify the time granularity and obtain the corresponding preliminary data.

[0062] Furthermore, based on the aforementioned spatial location identifiers, the data is mapped onto the aforementioned preliminary data. Relevant spatiotemporal information is extracted from the energy consumption data, and this spatiotemporal information is used to construct corresponding spatiotemporal features. The preliminary data, spatiotemporal features, and map data corresponding to the target area are then fused to obtain the corresponding spatiotemporal data. In the formula, Indicates the first Each sampling timestamp express 3D feature vectors , Indicates the first The energy consumption value of the target area at each sampling timestamp. Indicates the first The carbon emissions of the target area at each sampling timestamp.

[0063] S102, the spatiotemporal data is split based on the energy users in the target area to construct optimization tasks corresponding to each energy user;

[0064] Furthermore, step S102 specifically includes steps S1021 to S1022:

[0065] S1021, the spatiotemporal data is split based on the energy users in the target area to obtain independent energy consumption data subsets for each energy user;

[0066] S1022, independently analyze the energy consumption data subsets of each energy user to extract the corresponding energy consumption characteristics, and construct corresponding optimization tasks based on the energy consumption needs of each energy user.

[0067] In practice, spatiotemporal data is split based on energy users in the target area. Combined with a pre-built spatiotemporal database, an independent subset of energy consumption data for each energy user is obtained. After the splitting process, each energy user will obtain complete energy consumption data with timestamps and other relevant spatiotemporal information.

[0068] Specifically, the energy consumption data subset of each energy user is analyzed independently. The analysis includes information such as peak, trough, and seasonal changes in energy consumption data for each user, and the corresponding load characteristics are calculated, including maximum load demand, average load, and load factor. Based on the energy demand of the user, corresponding task target information is created, and the task target information, energy demand, and compliance characteristics are integrated to construct the corresponding optimization task.

[0069] S103, The optimization tasks of each energy user are processed using a preset dynamic window algorithm to build corresponding data constraints in the optimization tasks;

[0070] Furthermore, step S103 specifically includes steps S1031 to S1032:

[0071] S1031, initialize the dynamic window algorithm for the optimization task of each energy user, create a dynamic time window at the start of each optimization, and perform data prediction within the dynamic time window;

[0072] S1032, construct corresponding data constraints based on the type of energy consumption data, and add the data constraints into the dynamic time window for iterative optimization to obtain the optimized task after data constraints.

[0073] In practical implementation, a dynamic window algorithm is initialized for the optimization tasks of each energy user. The window length (in this embodiment, the window length is set to the next 4 hours) and the rolling step size (in this embodiment, the rolling step size is set to 15 minutes) are defined. It is understood that the window length determines the future time range covered by each optimization, while the rolling step size determines the frequency of the algorithm's forward progression. At the start of each optimization, the dynamic window algorithm creates a dynamic time window from the start time to the window length, predicts all data related to this window, and obtains the corresponding rules from the corresponding data constraint rules based on the type of energy consumption data (in this embodiment, the types of energy consumption data include power data type, environmental data type, and load data type). At each time step within the dynamic time window, the corresponding rules are iteratively optimized with the optimization task to solve the corresponding optimization problem, thereby obtaining the data-constrained optimization task.

[0074] S104, Construct a feature processing model, and use the feature processing model to perform feature processing on the spatiotemporal data to obtain the corresponding enhanced feature vector;

[0075] Furthermore, step S104 specifically includes steps S1041 to S1042:

[0076] S1041, The feature processing model is used to perform feature processing on the spatiotemporal data to calculate the power mean and load fluctuation variance in the spatiotemporal data;

[0077] S1042, extract the frequency domain features of the spatiotemporal data, and fuse the power mean, the load fluctuation variance and the frequency domain features to obtain the corresponding enhanced feature vector.

[0078] In practical implementation, a convolutional neural network model is defined, and a long short-term memory network and a gated recurrent algorithm are used to optimize the convolutional neural network model. Historical spatiotemporal data is used as input and imported into the optimized convolutional neural network model to obtain the feature processing model.

[0079] Furthermore, this feature processing model is used to perform feature processing on the spatiotemporal data to calculate the power mean of the spatiotemporal data. and load fluctuation variance Extracting frequency domain features from spatiotemporal data averaging power Load fluctuation variance and frequency domain characteristics The data is fused, and the fused data is then transformed into vectors to obtain the corresponding enhanced feature vectors. ,in, Represents the vector dimension.

[0080] S105, the feature processing model is optimized using the enhanced feature vector and the optimization task after data constraints, and an attention mechanism is added to the feature processing model to construct the corresponding multi-objective processing model;

[0081] In practical implementation, the enhanced feature vectors obtained above, the data-constrained optimization task, and the corresponding attention mechanism are used to construct the corresponding energy consumption prediction algorithm and carbon emission prediction algorithm. The calculation formula for the energy consumption prediction algorithm is as follows:

[0082] ;

[0083] In the formula, Indicates the first Predicted energy consumption values ​​for individual energy users This represents the activation function of the first fully connected layer in the feature processing model. Represents the energy consumption weight matrix. This represents the Long Short-Term Memory (LSTM) network algorithm. Indicates the first The number of energy-related task data in the optimization task after data constraints for each energy user. express Data weights, Represents the energy consumption bias vector;

[0084] The calculation formula for the carbon emission prediction algorithm is as follows:

[0085] ;

[0086] In the formula, Indicates the first Carbon emission forecasts for individual energy users This represents the activation function of the second fully connected layer in the feature processing model. Represents the carbon emission weight matrix. This represents the attention mechanism function. Indicates the first The amount of carbon emission-related task data in the optimization task after data constraints for each energy user. express Data weights, Represents the carbon emission bias vector;

[0087] Based on the energy consumption prediction algorithm and carbon emission prediction algorithm described above, a loss function for the multi-objective processing model is defined, and the corresponding multi-objective processing model is constructed according to the obtained loss function:

[0088] ;

[0089] In the formula, , The weight representing energy loss, The weights representing carbon emission losses Indicates the first Energy consumption value of an individual energy user Indicates the first Carbon emissions per energy user express Regularization coefficient, The Frobenius norm of the weight matrix 。

[0090] S106, the multi-objective processing model is coupled with the non-dominated sorting genetic algorithm to construct a corresponding multi-objective optimization model, and a corresponding optimization decision is generated based on the optimization objective of the target region and the optimal solution of the multi-objective optimization model.

[0091] In practical implementation, a non-dominated sorting genetic algorithm and a multi-objective processing model are coupled to construct the corresponding objective optimization model. Specifically, the population is initialized and corresponding reference points are generated. The neighborhood search behavior of single-point binary crossover is simulated to obtain the corresponding offspring information based on the initialized population. Small random perturbations are introduced to maintain diversity and explore the region near the parent generation. Polynomial mutation is performed on each offspring information to obtain the corresponding offspring population. The initialized population and the offspring population are merged and the merged population is compared pairwise. The population is stratified according to the Pareto dominance relationship to obtain several non-dominated layers.

[0092] Specifically, a preliminary optimization model is constructed by associating a reference point with several non-dominated layers and using the Nicholson operation and iterative processing. A rolling optimization strategy is then incorporated into this preliminary model, using the current state as the initial value, to solve the finite-time optimization problem, thereby constructing the corresponding multi-objective optimization model. The calculation formula for the rolling optimization strategy is as follows:

[0093] ;

[0094] In the formula, Indicates the discount factor. Represents the prediction time domain, where is the total number of time steps. Indicates the rolling optimization time interval. Indicates time The predicted energy consumption value, Indicates time The predicted carbon emissions , These represent the predicted energy consumption values. and carbon emission forecasts Optimization weights, A set of vectors representing spatiotemporal data;

[0095] Based on the optimization objectives of the target area and the optimal solution of the multi-objective optimization model, corresponding optimization decisions are generated. These optimization decisions are then used to optimize the energy consumption data of each energy user in the target area to obtain the optimal combination.

[0096] In summary, the energy-saving and carbon-reduction multi-objective optimization method in the above embodiments of the present invention collects energy consumption data of the target area in real time, adds spatiotemporal features to the energy consumption data, and performs data splitting based on the obtained spatiotemporal data and energy users to construct optimization tasks for each energy user. The search efficiency of task processing is improved through data constraints, thereby improving the processing accuracy and efficiency of the model. By constructing a multi-objective processing model based on an attention mechanism, energy consumption and carbon emissions are accurately predicted. The non-dominated sorting genetic algorithm is coupled with the multi-objective processing model to find the optimal solution that balances energy saving and carbon reduction objectives. The multi-objective processing model is continuously predicted and optimized based on real-time status to improve the adaptability and robustness of the model.

[0097] Example 2

[0098] In another aspect, this invention proposes an energy-saving and carbon-reducing multi-objective optimization system, please refer to [link / reference needed]. Figure 2 The figure shows a multi-objective optimization system for energy saving and carbon reduction according to a second embodiment of the present invention. The system includes:

[0099] The data acquisition module 11 is used to collect energy consumption data of the target area in real time through the Internet of Things terminal, and process the energy consumption data based on spatiotemporal characteristics to construct corresponding spatiotemporal data.

[0100] Furthermore, the data acquisition module 11 is specifically used for:

[0101] Acquire energy consumption data collected by data sensors installed in the target area, wherein the energy consumption data includes equipment power, ambient temperature, and load status;

[0102] The energy consumption data is cleaned and format converted to obtain corresponding preprocessed data. The preprocessed data is then subjected to data interpolation and timestamp alignment to obtain preliminary data.

[0103] Based on the location information of the data sensors in the target area, the preliminary data is mapped to construct corresponding spatiotemporal features and obtain corresponding spatiotemporal data.

[0104] The data splitting module 12 is used to split the spatiotemporal data based on the energy users in the target area to construct optimization tasks corresponding to each energy user.

[0105] Furthermore, the data splitting module 12 is specifically used for:

[0106] Based on the energy users in the target area, the spatiotemporal data is split to obtain independent subsets of energy consumption data for each energy user;

[0107] The energy consumption data subsets of each energy user are analyzed independently to extract the corresponding energy consumption characteristics, and corresponding optimization tasks are constructed based on the energy consumption needs of each energy user.

[0108] The data processing module 13 is used to process the optimization tasks of each energy user using a preset dynamic window algorithm, so as to build corresponding data constraints in the optimization tasks.

[0109] Furthermore, the data processing module 13 is specifically used for:

[0110] A dynamic window algorithm is initialized for the optimization tasks of each energy user. At the start of each optimization, a dynamic time window is created, and data prediction is performed within the dynamic time window.

[0111] Based on the type of energy consumption data, corresponding data constraints are constructed, and the data constraints are added to the dynamic time window for iterative optimization to obtain the optimized task after data constraints.

[0112] The feature processing module 14 is used to construct a feature processing model and use the feature processing model to perform feature processing on the spatiotemporal data to obtain the corresponding enhanced feature vector.

[0113] Furthermore, the feature processing module 14 is specifically used for:

[0114] The spatiotemporal data are processed using the aforementioned feature processing model to calculate the mean power and load fluctuation variance in the spatiotemporal data.

[0115] The frequency domain features of the spatiotemporal data are extracted, and the power mean, the load fluctuation variance, and the frequency domain features are fused to obtain the corresponding enhanced feature vector.

[0116] The model building module 15 is used to optimize the feature processing model using the enhanced feature vector and the optimization task after data constraints, and to add an attention mechanism to the feature processing model to build a corresponding multi-objective processing model.

[0117] The objective optimization module 16 is used to couple the multi-objective processing model with the non-dominated sorting genetic algorithm to construct a corresponding multi-objective optimization model, and generate corresponding optimization decisions based on the optimization objective of the target region and the optimal solution of the multi-objective optimization model.

[0118] The functions or operation steps implemented by the above modules and units are largely the same as those in the above method embodiments, and will not be repeated here.

[0119] The energy-saving and carbon-reduction multi-objective optimization system provided in this embodiment of the invention has the same implementation principle and technical effects as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the system embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0120] Example 3

[0121] This invention also proposes a computer, please refer to [link / reference]. Figure 3 The computer shown in the third embodiment of the present invention includes a memory 10, a processor 20, and a computer program 30 stored in the memory 10 and executable on the processor 20. When the processor 20 executes the computer program 30, it implements the above-mentioned energy-saving and carbon-reduction multi-objective optimization method.

[0122] The memory 10 includes at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 10 can be an internal storage unit of a computer, such as the computer's hard disk. In other embodiments, the memory 10 can be an external storage device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Furthermore, the memory 10 can include both internal and external storage units of the computer. The memory 10 can be used not only to store application software and various types of data installed on the computer, but also to temporarily store data that has been output or will be output.

[0123] In some embodiments, the processor 20 may be an electronic control unit (ECU), a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip, used to run program code stored in the memory 10 or process data, such as executing access restriction programs.

[0124] It should be pointed out that, Figure 3 The structure shown does not constitute a limitation on the computer. In other embodiments, the computer may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0125] This invention also proposes a storage medium storing a computer program that, when executed by a processor, implements the energy-saving and carbon-reduction multi-objective optimization method described above.

[0126] Those skilled in the art will understand that the logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0127] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0128] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0129] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0130] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A multi-objective optimization method for energy saving and carbon reduction, characterized in that, The application relates to a method for constructing a multi-objective optimization model based on a target region. Real-time energy consumption data of a target region is collected through an Internet of Things terminal, and the energy consumption data is processed based on space-time characteristics to construct corresponding space-time data. The space-time data is split based on energy consumption users of the target region to construct optimization tasks corresponding to the energy consumption users. A preset dynamic window algorithm is used to process the optimization tasks of the energy consumption users to build corresponding data constraints in the optimization tasks. A feature processing model is constructed, and the space-time data is processed by using the feature processing model to obtain a corresponding enhanced feature vector. The enhanced feature vector, the optimization tasks after data constraint and a corresponding attention mechanism are used to construct a corresponding energy consumption prediction algorithm and a carbon emission prediction algorithm, wherein a calculation formula of the energy consumption prediction algorithm is: ； In the formula, represents the energy consumption prediction value of the first energy consumer, represents the activation function of the first full connection layer in the feature processing model, represents the energy consumption weight matrix, represents the long short-term memory network algorithm, represents the enhanced feature vector, represents the number of energy consumption related task data in the optimized task of the first energy consumer after data constraint, represents the data weight of , and represents the energy consumption bias vector. A calculation formula of the carbon emission prediction algorithm is: ； In the formula, represent the carbon emission prediction of the first energy user, represent the second full connection layer activation function in the feature processing model, represent the carbon emission weight matrix, represent the attention mechanism function, represent the number of task data related to carbon emissions in the optimized task after data constraint of the first energy user, represent the data weight of , and represent the carbon emission bias vector. A loss function of a multi-objective processing model is defined based on the energy consumption prediction algorithm and the carbon emission prediction algorithm, and a corresponding multi-objective processing model is constructed according to the obtained loss function: ； In the formula, , The weight representing energy loss, The weights representing carbon emission losses Indicates the first Energy consumption value of an individual energy user Indicates the first Carbon emissions per energy user express Regularization coefficient, The Frobenius norm of the weight matrix; The multi-objective processing model is coupled with a non-dominated sorting genetic algorithm to construct a corresponding multi-objective optimization model, and an optimization decision is generated according to an optimization target of the target region and an optimal solution of the multi-objective optimization model. The step of processing the space-time data by using the feature processing model to obtain a corresponding enhanced feature vector comprises the following steps: The space-time data is processed by using the feature processing model to calculate a power mean value and a load fluctuation variance in the space-time data. Frequency domain features of the space-time data are extracted, and the power mean value, the load fluctuation variance and the frequency domain features are fused to obtain a corresponding enhanced feature vector.

2. The energy-saving and carbon-reducing multi-objective optimization method according to claim 1, characterized in that, The step of collecting real-time energy consumption data of a target region through an Internet of Things terminal and processing the energy consumption data based on space-time characteristics to construct corresponding space-time data comprises the following steps: Energy consumption data collected by a data sensor installed in the target region is obtained, wherein the energy consumption data comprises device power, environmental temperature and load state. The energy consumption data is subjected to data cleaning and format conversion to obtain corresponding pretreatment data, and the pretreatment data is subjected to data interpolation processing and timestamp alignment to obtain initial data. The initial data is subjected to data mapping based on position information of the data sensor of the target region to construct corresponding space-time features to obtain corresponding space-time data.

3. The energy-saving and carbon-reducing multi-objective optimization method according to claim 1, characterized in that, The step of splitting the space-time data based on energy consumption users of the target region to construct optimization tasks corresponding to the energy consumption users comprises the following steps: The space-time data is split based on energy consumption users of the target region to obtain independent energy consumption data subsets of the energy consumption users. The energy consumption data subsets of the energy consumption users are independently analyzed to extract corresponding energy consumption features, and optimization tasks are constructed based on energy consumption demands of the energy consumption users.

4. The energy-saving and carbon-reducing multi-objective optimization method according to claim 1, characterized in that, The step of processing the optimization task of each energy-using user by using a preset dynamic window algorithm to build corresponding data constraints in the optimization task comprises: initializing the dynamic window algorithm for the optimization task of each energy-using user, creating a dynamic time window at the start of each optimization, and performing data prediction within the dynamic time window; based on the type of energy-using data, build corresponding data constraints, and add the data constraints to the dynamic time window for iterative optimization to obtain the optimization task after data constraints.

5. An energy saving and carbon reduction multi-objective optimization system, characterized by, Comprise: a data acquisition module for acquiring energy-using data of a target area in real time through an Internet of Things terminal, and processing the energy-using data based on space-time characteristics to build corresponding space-time data; a data splitting module for splitting the space-time data based on energy-using users of the target area to build optimization tasks corresponding to each energy-using user; a data processing module for processing the optimization task of each energy-using user by using a preset dynamic window algorithm to build corresponding data constraints in the optimization task; a feature processing module for building a feature processing model and processing the space-time data using the feature processing model to obtain a corresponding enhanced feature vector; a model building module for building a corresponding energy consumption prediction algorithm and carbon emission prediction algorithm using the enhanced feature vector, the optimization task after data constraints, and the corresponding attention mechanism, wherein the calculation formula of the energy consumption prediction algorithm is: ； In the formula, represents the energy consumption prediction value of the first energy consumer, represents the activation function of the first full connection layer in the feature processing model, represents the energy consumption weight matrix, represents the long short-term memory network algorithm, represents the enhanced feature vector, represents the number of energy-related task data in the optimized task of the first energy consumer after data constraint, represents the data weight of , and represents the energy consumption bias vector. the calculation formula of the carbon emission prediction algorithm is: ； In the formula, represents the carbon emission prediction of the first energy user, represents the second fully connected layer activation function in the feature processing model, represents the carbon emission weight matrix, represents the attention mechanism function, represents the number of task data related to carbon emission in the optimized task of the first energy user after data constraint, represents the data weight of represents the carbon emission bias vector;​​​ define the loss function of the multi-objective processing model based on the energy consumption prediction algorithm and the carbon emission prediction algorithm, and build the corresponding multi-objective processing model according to the obtained loss function: ； In the formula, , The weight representing energy loss, The weights representing carbon emission losses Indicates the first Energy consumption value of an individual energy user Indicates the first Carbon emissions per energy user express Regularization coefficient, The Frobenius norm of the weight matrix; a target optimization module for coupling the multi-objective processing model with a non-dominated sorting genetic algorithm to build a corresponding multi-objective optimization model, and generating a corresponding optimization decision according to the optimization target of the target area and the optimal solution of the multi-objective optimization model; wherein the feature processing module is specifically configured to: use the feature processing model to process the space-time data to calculate the power mean and load fluctuation variance in the space-time data; extract the frequency domain features of the space-time data, and fuse the power mean, the load fluctuation variance and the frequency domain features to obtain a corresponding enhanced feature vector.

6. The energy saving and carbon reduction multi-objective optimization system according to claim 5, wherein, The data acquisition module is specifically configured to: acquire energy-using data collected by data sensors installed in the target area, wherein the energy-using data includes device power, environmental temperature and load state; perform data cleaning and format conversion on the energy-using data to obtain corresponding preprocessed data, and perform data interpolation processing and timestamp alignment on the preprocessed data to obtain preliminary data; perform data mapping on the preliminary data based on the position information of the data sensors in the target area to build corresponding space-time characteristics to obtain corresponding space-time data.

7. The energy saving and carbon reduction multi-objective optimization system according to claim 5, wherein, The data splitting module is specifically configured to: performing data splitting on the spatio-temporal data based on energy-using users of the target region to obtain energy-using data subsets of the energy-using users independently; performing independent analysis on the energy-using data subsets of the energy-using users to extract corresponding energy-using features, and constructing corresponding optimization tasks based on energy-using demands of the energy-using users.

8. A storage medium having stored thereon a computer program, characterized in that The program is executed by the processor to implement the energy-saving and carbon-reducing multi-objective optimization method of any one of claims 1 to 4.

9. A computer comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the energy-saving and carbon-reducing multi-objective optimization method of any one of claims 1 to 4.

Images