Distributed energy optimization scheduling and management method and platform based on AI

Abstract

The invention provides an AI-based distributed energy optimization scheduling and management method and platform, and the method comprises the steps: collecting data from a plurality of heterogeneous data sources, and carrying out the preprocessing, and obtaining a data set; constructing a multi-target optimization scheduling model according to the data set, and determining an optimization scheduling scheme; obtaining an optimization scheduling scheme, designing a cooperation measurement function, and designing a sharing optimization objective function to optimize energy flow; and collecting real-time feedback of the updated scheduling scheme, and dynamically adjusting the scheduling scheme by measuring the difference between the power demand and the actual scheduling power. According to the invention, through comprehensive application of multi-objective optimization, an intelligent cooperation mechanism and a big data analysis technology, scheduling and management of distributed energy are optimized, and the intelligent level and flexibility of the system are improved.

Description

Technical Field

[0001] The present invention belongs to the field of distributed energy optimization, and particularly relates to an AI-based distributed energy optimization scheduling and management method and platform. Background Art

[0002] With the gradual transformation of the global energy structure, the application of renewable energy has become an important direction to promote the energy revolution worldwide. As a core component of the smart grid, the distributed energy system has broad prospects and potential. The distributed energy system usually includes various forms of energy production and storage units such as solar energy, wind energy, energy storage devices, and micro combined heat and power. Its greatest feature is wide distribution, strong flexibility, and the ability to interact bidirectionally with the power grid. With the rapid development of information technology and the smart grid, the production and consumption management of distributed energy has begun to become more refined and intelligent. However, there are still many bottlenecks in the existing technology for the scheduling and management of distributed energy, which hinder its wide application and optimization.

[0003] First of all, the existing distributed energy scheduling methods mainly focus on the traditional centralized scheduling framework, and usually uniformly schedule each energy node through a central control system. Although this method can achieve the optimization of energy scheduling to a certain extent, due to the lack of sufficient flexibility and the ability to respond to changes, it cannot well adapt to the diverse and complex demands and changes in the distributed energy system. For example, the traditional centralized scheduling system often relies on relatively fixed parameters and rules, and cannot reflect the actual demands and operating state changes of each node (such as users, production equipment, energy storage systems, etc.) in the system in real time, resulting in the inability to achieve global optimal scheduling. In addition, this method usually ignores the synergistic effect between distributed nodes, making it difficult to give full play to the advantages of each node, resulting in low energy sharing efficiency and insufficient resource utilization.

[0004] Secondly, the existing distributed energy management platforms generally have problems such as insufficient data processing capabilities and serious information islands. Traditional energy scheduling platforms often do not have the ability to quickly process a large amount of heterogeneous data from different nodes in a large-scale distributed system. Especially in the face of highly dynamic factors such as user demands, weather conditions, and power market price fluctuations, the existing systems often lack flexible scheduling strategies and response mechanisms. In addition, the information exchange and coordination mechanism between distributed energy nodes is relatively weak, resulting in the inability of each node to achieve intelligent collaborative scheduling. The sharing and optimal scheduling of energy in the distributed system have not been fully utilized, resulting in energy waste and system instability.

[0005] Finally, most of the current research on the intelligent scheduling of distributed energy systems focuses on single-objective optimization, such as cost, efficiency, load balancing, etc. However, there are few methods that can simultaneously take into account multiple optimization objectives. For example, how to ensure the stability of the power grid, reduce carbon emissions, and ensure that the energy needs of each user are met while reducing energy costs. Existing technologies often optimize with a single objective as the main focus, ignoring the overall coordination and stability of the system, and the solution of multi-objective optimization problems often does not take into account the dynamics and real-time nature of the system. Under multi-objective scheduling, traditional methods fail to effectively handle the contradiction between balancing short-term economic benefits and long-term environmental protection and system stability, resulting in the actual application effect of distributed energy not meeting expectations. Summary of the Invention

[0006] The object of the present invention is to propose an AI-based distributed energy optimization scheduling and management method and platform. By comprehensively applying multi-objective optimization, intelligent cooperation mechanisms, and big data analysis technologies, it not only optimizes the scheduling and management of distributed energy but also improves the intelligent level and flexibility of the system.

[0007] To achieve the above object, in the first aspect of the present invention, an AI-based distributed energy optimization scheduling and management method is provided. The method includes the following steps:

[0008] Collect data from multiple heterogeneous data sources and perform preprocessing to obtain a data set;

[0009] Construct a multi-objective optimization scheduling model based on the data set to determine an optimized scheduling plan; wherein, the objectives of the multi-objective optimization scheduling model include energy cost, energy efficiency, load balance, and the proportion of green energy use; the optimized scheduling plan includes node power production, consumption, and load information;

[0010] Obtain the optimized scheduling plan, design a cooperation metric function, and based on the cooperation metric function, dynamically adjust the cooperation weights between nodes, optimize the energy transmission and sharing strategy between multiple nodes, and obtain an updated scheduling plan; after obtaining the updated scheduling plan, in view of the possible energy redundancy or shortage between nodes, design a shared optimization objective function to optimize the energy flow;

[0011] Collect the real-time feedback of the updated scheduling plan, dynamically adjust the scheduling plan by measuring the difference between the power demand and the actual scheduling power, and design an adjustment factor according to the difference between the power demand and the actual scheduling power to adjust the node power output. Combine the real-time feedback and the adjustment factor to further update the updated scheduling plan, and introduce multi-objective optimization to further optimize the scheduling plan to obtain the final scheduling plan, and ensure that the final scheduling plan meets all constraint conditions, including power upper limit, load balance, and transmission loss.

[0012] Further, collecting data from multiple heterogeneous data sources, including weather, load, battery status, and electricity market prices;

[0013] The preprocessing includes:

[0014] Determining the weight of the corresponding data source based on the quality, timeliness, and historical performance of the data source;

[0015] Weighted fusion of different data according to the weight of the corresponding data source to obtain the fused data;

[0016] Obtaining the fused data and using the method based on adaptive weighted Kalman filter to remove noise to obtain the filtered data.

[0017] Further, after obtaining the preprocessed data, based on the feature selection method of maximizing mutual information and regularization, screening out the most effective feature set as the data set; the feature selection method based on maximizing mutual information and regularization is expressed as:

[0018]

[0019] where I(X i ; Y) represents the mutual information between feature X i and the target variable Y, P(x i , y) is the joint probability distribution of feature X i and the target variable Y, and P(x i ) and P(y) are their marginal probability distributions respectively;

[0020] At the same time, introducing the regularization L reg term to reduce the influence of redundant features on the model:

[0021]

[0022] where λ is the regularization coefficient, and Redundancy(X i ) represents the redundancy of feature X i with other features.

[0023] Further, the comprehensive objective function of the multi-objective optimization scheduling model is the weighted sum of the objectives of the multi-objective optimization scheduling model, and the weights are adjusted according to the real-time data and status:

[0024]

[0025] where w1(t), w2(t), w3(t), w4(t) are the dynamic weights of the objectives, and the specific values are adjusted according to the real-time data and system status.

[0026] Furthermore, constructing a multi-objective optimal scheduling model based on the data set and determining the optimal scheduling scheme specifically includes:

[0027] Design a feedback mechanism for the dynamic weight of the objective, and the feedback mechanism automatically adjusts the weight of the objective according to real-time data; wherein, the real-time data includes market price and equipment health status;

[0028] During the optimization process, it is necessary to ensure that the constraints are met, including: power balance constraint, energy storage device constraint, and green energy ratio;

[0029] Among them,

[0030] The power balance constraint is used to ensure that the total generated power is equal to the demanded power;

[0031] The energy storage device constraint is used to limit the charging and discharging capabilities of the energy storage device to avoid overcharging or over-discharging;

[0032] The green energy ratio is used to ensure that the usage amount of green energy reaches a predetermined ratio;

[0033] For the comprehensive objective function, an adaptive particle swarm optimization algorithm is adopted, which combines a feedback mechanism to dynamically adjust the search strategy of the particles; wherein, the update rule of the adaptive particle swarm optimization algorithm is as follows:

[0034] v i (t + 1) = w(t)·v i (t) + c1·r1·(p i -x i (t)) + c2·r2·(g - x i (t))

[0035] Among them, v i (t + 1) is the updated particle velocity, v i (t) is the particle velocity, x i (t) is the particle position, p i and g are respectively the historical best and global best positions of the particle, w(t) is the inertia weight, and r1 and r2 are random numbers.

[0036] Furthermore, the collaboration metric function is expressed as:

[0037]

[0038] Among them, w ij (t) is the collaboration weight between node i and node j, indicating the tightness of collaboration; γ ij (t) is the transmission loss factor, reflecting the energy transmission loss between node i and node j; α ij is the adjustment factor, controlling the impact of transmission loss on collaboration; Pi (t) and P j (t) are the power production or consumption of node i and node j respectively;

[0039] Meanwhile, the collaboration weights are dynamically updated; when the collaboration requirement of node i increases, its weight w ij (t) will increase, thus enhancing the collaboration with other nodes;

[0040] Combining the collaboration weights and the collaboration metric function, the scheduling scheme of the node is updated, expressed as:

[0041]

[0042] where P optimized (t) is the optimized scheduling scheme, and α is an adjustment factor used to balance the relationship between the optimization scheme and the collaboration requirement.

[0043] Furthermore, the shared optimization objective function is expressed as:

[0044] Combining the transmission loss and the sharing efficiency,

[0045]

[0046] where η ij (t) is the sharing efficiency coefficient, representing the efficiency of node i sharing energy with node j, considering the loss in transmission.

[0047] Furthermore, the multi-objective optimization is expressed as:

[0048]

[0049] where is the power loss of node i at time t; is the load fluctuation metric of node i at time t; α1 and α2 are weighting coefficients to control the importance of power loss and load fluctuation.

[0050] Furthermore, when performing the multi-objective optimization adjustment, the constraint needs to be satisfied:

[0051]

[0052] where P max (t) is the maximum power output capacity of node i; is the power constraint condition of node i, and P new (t) is the final scheduling scheme.

[0053] In the second aspect of the present invention, an AI-based distributed energy optimization scheduling and management platform is provided. The platform includes:

[0054] A heterogeneous data source acquisition module for collecting data from multiple heterogeneous data sources and performing preprocessing to obtain a data set;

[0055] A multi-objective optimization module for constructing a multi-objective optimization scheduling model based on the data set to determine an optimized scheduling plan. Among them, the objectives of the multi-objective optimization scheduling model include energy cost, energy efficiency, load balance, and the proportion of green energy use; the optimized scheduling plan includes node power production, consumption, and load information;

[0056] A collaborative scheduling module for obtaining the optimized scheduling plan, designing a collaborative metric function, and based on the collaborative metric function, dynamically adjusting the collaborative weights between nodes to optimize the energy transmission and sharing strategy among multiple nodes to obtain an updated scheduling plan; after obtaining the updated scheduling plan, in view of the possible energy redundancy or shortage between nodes, a shared optimization objective function is designed to optimize the energy flow;

[0057] A scheduling optimization module for collecting real-time feedback of the updated scheduling plan, dynamically adjusting the scheduling plan by measuring the difference between the power demand and the actual scheduling power, and designing an adjustment factor according to the difference between the power demand and the actual scheduling power to adjust the node power output. Combining the real-time feedback and the adjustment factor to further update the updated scheduling plan, and introducing multi-objective optimization to further optimize the scheduling plan to obtain the final scheduling plan, and ensuring that the final scheduling plan meets all constraints, and the constraints include power upper limit, load balance, and transmission loss.

[0058] The beneficial technical effects of the present invention are at least as follows:

[0059] First of all, by introducing a multi-objective optimization algorithm, the present invention solves the problem that traditional methods cannot find a balance among multiple optimization objectives. The platform not only optimizes the energy cost, but also comprehensively considers various objectives such as the load balance, stability, grid security, and environmental protection benefits of the system. Through advanced methods such as Pareto front optimization and particle swarm optimization algorithm (PSO), it ensures that the energy scheduling cost is minimized and carbon emissions are reduced to the greatest extent on the premise of ensuring grid stability, and can effectively cope with the fluctuations of energy demand. In addition, the system can perform dynamic scheduling according to various factors such as real-time electricity demand, renewable energy generation, and weather changes, improving the real-time performance and flexibility of scheduling.

[0060] Secondly, through the intelligent collaboration mechanism and distributed ledger technology (such as blockchain), the present invention breaks through the data island problem in the existing distributed energy management platform. Energy sharing is achieved between distributed nodes through smart contracts or collaboration protocols, ensuring transparent, fair, and secure energy exchange for each node. In the case of energy surplus, nodes can automatically allocate the excess energy to other needy nodes according to the scheduling suggestions of the platform, thus realizing the optimal scheduling of energy for the entire system. This mechanism not only improves the energy utilization efficiency but also greatly enhances the overall flexibility and intelligent collaboration ability of the distributed system. By automatically executing energy trading and scheduling instructions through smart contracts, the interference and errors caused by human operations are effectively reduced.

[0061] Finally, through big data analysis and collaborative filtering algorithms, the platform realizes intelligent decision-making in demand forecasting and system optimization. The system can analyze and predict the energy demand of each node in real time based on multi-dimensional data such as historical data, weather information, and electricity market prices, further improving the accuracy and predictability of scheduling. The collaborative filtering algorithm helps the platform intelligently infer the future needs of users according to the needs and electricity consumption patterns of different nodes, and accurately allocate the energy resources of each node, thus avoiding over-scheduling or energy waste and ensuring the optimal operation of the entire distributed energy system. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The present invention will be further described with reference to the accompanying drawings. However, the embodiments in the drawings do not constitute any limitation to the present invention. For those of ordinary skill in the art, other drawings can be obtained according to the following drawings without creative efforts.

[0063] Figure 1 Flowchart of the AI-based distributed energy optimal scheduling and management method disclosed in the embodiments of the present invention

[0064] Figure 2 Framework diagram of the AI-based distributed energy optimal scheduling and management platform disclosed in the embodiments of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0065] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation to the present invention.

[0066] Embodiment 1

[0067] As Figure 1 shown, the AI-based distributed energy optimal scheduling and management method provided by the embodiments of the present invention includes:

[0068] S1. Collect data from multiple heterogeneous data sources, and perform preprocessing to obtain a data set.

[0069] Specifically, in distributed energy scheduling, the present invention faces heterogeneous data from different sources (such as weather, load, battery status, electricity market price, etc.). The quality, timeliness, and importance of these data sources are different. Therefore, the present invention needs to perform data fusion to ensure the consistency and reliability of the data input into the model. For this purpose, the present invention proposes a dynamic weighted fusion model based on timeliness weight adjustment.

[0070] The present invention first defines the weight w i (t) of each data source, which is closely related to the quality, timeliness, and historical performance of the data source. The weight formula is:

[0071]

[0072] where q i (t) represents the quality factor (including credibility and accuracy) of data source i at time t, α(t) is the timeliness factor (i.e., the deviation between the timestamp of the data source and the current time), w i (t) is the weighting coefficient of this data source, and n is the number of data sources. The data fusion formula is:

[0073]

[0074] where x fused (t) represents the fused data, x i (t) is the original data of the i-th data source, and w i (t) is the corresponding weight. This fusion method can ensure that under different behavior patterns, the different degrees of trust in the data sources are dynamically adjusted, thereby reducing the impact of noise on the subsequent model.

[0075] Furthermore, in the data collection stage, especially sensor data and weather forecast data, there are often noises. To improve the data quality, the present invention adopts a method based on Adaptive Weighted Kalman Filtering (AWKF) to remove noises. This method adaptively adjusts the filtering parameters and dynamically responds to the quality changes of the data sources. The state update equation of the weighted Kalman filter is:

[0076] x filtered (t) = w i (t) · (A · x t-1 + B · u t ) + (1 - w i (t)) · (C · y t )

[0077] Among them, x filtered (t) is the filtered data, A and B are the state transition and control matrices of the Kalman filter, u t is the control input, C is the measurement matrix, y t is the observed value, w i (t) is the weighting coefficient of data source i, which is dynamically adjusted according to data quality.

[0078] The innovation of this filtering method lies in dynamically adjusting the weighting coefficient of the Kalman filter according to the quality of each data source, so that when processing sensor data and weather data, it can adaptively denoise according to the timeliness and reliability of the data source, reducing the problems that traditional Kalman filtering cannot handle timeliness fluctuations and quality differences.

[0079] Furthermore, in order to improve the learning efficiency of the subsequent model and reduce the impact of redundant data on computational performance, the present invention designs a feature selection method based on mutual information maximization and regularization. This method calculates the correlation between each feature and the target variable, and at the same time introduces redundant regularization to screen the most effective feature set. The mutual information calculation formula is:

[0080]

[0081] Among them, I(X i ; Y) represents the mutual information between feature X i and the target variable Y, P(x i ,y) is the joint probability distribution of feature X i and the target variable Y, and P(x i ) and P(y) are their marginal probability distributions respectively.

[0082] In order to reduce the impact of redundant features on the model, the present invention introduces a regularization term:

[0083]

[0084] Among them, λ is the regularization coefficient, and Redundancy(X i ) represents the redundancy of feature X i with other features. In this way, the present invention can screen out those features with high correlation with the target variable and low redundancy, providing efficient input for the training of the subsequent model.

[0085] The final output is a data set that has been intelligently preprocessed, denoised, fused, and feature selected, and these data can be used as high-quality input for the multi-objective scheduling optimization model.

[0086] S2. Construct a multi-objective optimal scheduling model based on the data set to determine the optimal scheduling plan. Among them, the objectives of the multi-objective optimal scheduling model include energy cost, energy efficiency, load balance, and the proportion of green energy used. The optimal scheduling plan includes node power production, consumption, and load information.

[0087] In Step 1: Data collection and intelligent preprocessing, the time-series data x fused (t) has been obtained through fusion and weighting methods. It includes various characteristics of the power system, such as energy consumption, equipment status, load, etc. This data set will be used as the input for this step to construct a multi-objective optimal scheduling model.

[0088] Furthermore, the objective of the present invention is to construct a multi-objective optimal scheduling model to optimize energy cost, energy efficiency, load balance, and the use of green energy simultaneously. Since there are certain conflicts among these objectives, the model needs to make trade-offs. For this purpose, the present invention defines a comprehensive objective function This function is the weighted sum of all objective functions.

[0089]

[0090] Among them, w1(t), w2(t), w3(t), w4(t) are the dynamic weights of the objectives, and the specific values are adjusted according to real-time data and system status. The specific meaning of each objective is as follows:

[0091] Energy cost: Reflects the economic burden of energy procurement.

[0092] Energy efficiency: Evaluates the energy usage efficiency of the system and tries to minimize energy waste.

[0093] Load balance: Reduces system load fluctuations and improves system stability.

[0094] Proportion of green energy used: Promotes the use of green energy and reduces environmental impact.

[0095] Furthermore, in order to adjust the importance of the objectives in real time, the present invention designs a feedback mechanism that automatically adjusts the weights of the objectives according to real-time data (such as market price, equipment health status). The adjustment factor α k (t) is used to control the dynamic weights of the objective function, and the calculation method is based on the system status and external environment (such as market fluctuations):

[0096] w k (t) = α k (t) · f adjust (x fused (t))

[0097] Here, α k(t) is the weight adjustment factor for feedback calculation, and f adjust (x fused (t)) is the adaptive adjustment factor calculated based on x fused (t), which dynamically updates the weight according to real-time energy consumption, load, and device status.

[0098] Furthermore, during the optimization process, multiple constraint conditions need to be considered, which ensure that the scheduling results of the model meet the actual operation requirements of the power system:

[0099] Power balance constraint: Ensure that the total generated power is equal to the demanded power:

[0100]

[0101] Energy storage device constraint: Limit the charging and discharging capabilities of energy storage devices to avoid overcharging or over-discharging:

[0102] 0 ≤ E i (t) ≤ E max (t)

[0103] Proportion of green energy: Ensure that the usage amount of green energy reaches a predetermined proportion:

[0104] G usage (t) ≥ G min

[0105] Furthermore, to efficiently solve the multi-objective optimization problem, the present invention adopts an Adaptive Particle Swarm Optimization (APSO) algorithm, which combines a feedback mechanism to dynamically adjust the search strategy of particles. Specifically, the update rule of particle swarm optimization is as follows:

[0106] v i (t + 1) = w(t)·v i (t) + c1·r1·(p i - x i (t)) + c2·r2·(g - x i (t))

[0107] Wherein, v i (t) is the particle velocity, x i (t) is the particle position, p i and g are respectively the historical best and global best positions of the particle, w(t) is the inertia weight, and r1 and r2 are random numbers. This algorithm can avoid falling into local optimal solutions by adaptively adjusting the search strategy, and effectively balance the weights between multiple objectives.

[0108] Furthermore, through the solution of the optimization algorithm, the finally output scheduling scheme x optimized(t) can effectively balance various objectives and provide an optimized power dispatching scheme. This scheme can not only minimize the energy cost, but also improve the energy efficiency, reduce the load fluctuation, and maximize the use of green energy.

[0109] S3. Obtain the optimized dispatching scheme, design a collaboration metric function, and based on the collaboration metric function, dynamically adjust the collaboration weights between nodes, optimize the energy transmission and sharing strategy among multiple nodes, and obtain the updated dispatching scheme; after obtaining the updated dispatching scheme, for the situation where there may be energy redundancy or shortage between nodes, design a shared optimization objective function to optimize the energy flow.

[0110] The input of this step comes from the output of step 2, that is, the optimized dispatching scheme x optimized (t), which contains information such as power production, consumption, and load balancing of each node. This information will be used to adjust the collaboration mode and energy sharing strategy between nodes. The collaboration metric function is designed to measure the collaboration efficiency between nodes, introduce a dynamic weight adjustment mechanism, and define the collaboration metric function:

[0111]

[0112] where: w ij (t) is the collaboration weight between node i and node j, indicating the degree of collaboration tightness; γ ij (t) is the transmission loss factor, reflecting the energy transmission loss between node i and node j; α ij is the adjustment factor, controlling the impact of transmission loss on collaboration; P i (t) and P j (t) are the power production or consumption of node i and node j respectively. The purpose of this function is to enhance the consideration of transmission loss between nodes, making the collaboration more in line with the actual situation.

[0113] Furthermore, based on the above collaboration metric function, the collaboration weights between nodes are dynamically adjusted according to the following rules:

[0114]

[0115] where β is the adjustment factor, used to control the change of collaboration weight. When the collaboration demand of node i increases, its weight w ij (t) will increase, thus enhancing the collaboration with other nodes.

[0116] Furthermore, combining the dynamic weight and collaboration metric, update the dispatching status of the nodes. The updated dispatching scheme is:

[0117]

[0118] where, P optimized(t) is the optimized scheduling scheme for Step 2, and α is an adjustment factor used to balance the relationship between the optimization scheme and the collaboration requirements. This step updates the scheduling scheme by integrating collaboration information to improve energy utilization efficiency.

[0119] Furthermore, in an environment with multiple energy nodes, there may be situations of energy redundancy or shortage between nodes. Therefore, an energy sharing mechanism is introduced to optimize energy flow. The present invention designs a shared optimization objective function:

[0120]

[0121] where η ij (t) is the sharing efficiency coefficient, representing the efficiency of node i sharing energy with node j, considering the losses during transmission. This function comprehensively considers transmission losses and sharing efficiency, aiming to optimize the energy sharing strategy among multiple nodes.

[0122] Furthermore, in order to further improve the sharing efficiency, a decreasing sharing benefit factor η ij (t) is designed, and its expression is:

[0123]

[0124] where: P share (t) is the amount of electricity shared by node i with node j; P max (t) is the maximum power output capacity of node i; γ(t) is the sharing adjustment factor, dynamically adjusting the sharing capacity. This step optimizes the sharing efficiency coefficient to control the energy transmission between nodes, ensuring the high efficiency and fairness of the sharing process.

[0125] After being optimized by the intelligent collaboration and energy sharing mechanism, each node in the system will obtain the updated scheduling scheme P adjusted (t). This scheme fully considers the collaboration requirements and sharing strategies between nodes, can effectively reduce energy waste, and increase the proportion of green energy used. Through the design of this step, the collaboration and energy sharing between nodes have been significantly improved. In particular, innovative mechanisms such as dynamic weight adjustment, transmission loss factor, and sharing efficiency coefficient are introduced. These schemes provide an optimized path for energy scheduling and sharing in multi - energy systems and can keep the system running efficiently in various dynamic environments.

[0126] S4. Collect the real-time feedback of the updated scheduling scheme. By measuring the difference between the power demand and the actual scheduling power, dynamically adjust the scheduling scheme, and design an adjustment factor according to the difference between the power demand and the actual scheduling power to adjust the node power output. Combine the real-time feedback and the adjustment factor to further update the updated scheduling scheme, and introduce multi-objective optimization to further optimize the scheduling scheme to obtain the final scheduling scheme, and ensure that the final scheduling scheme meets all the constraints, where the constraints include power upper limit, load balance, and transmission loss.

[0127] The input of this step comes from the output of step 3, that is, the optimized scheduling scheme P adjusted (t), as well as the cooperation metric and energy sharing information between nodes. This information will be used to further adjust the scheduling scheme and consider real-time changes.

[0128] Furthermore, introduce a real-time feedback mechanism to dynamically adjust the scheduling scheme based on the current energy demand and system state. To ensure the stability and efficiency of the system, the present invention designs the following feedback adjustment function:

[0129]

[0130] where P demand (t) is the required power of node i at time t; P current (t) is the actual scheduling power of node i at time t; η i (t) is the adjustment factor, which dynamically adjusts the response speed of the feedback. This feedback function is used to measure the difference between the node power demand and the current power, and guide the system to make adjustments according to the difference.

[0131] Furthermore, to adjust the power output of the node according to the feedback signal, the present invention designs an adjustment factor η i (t), and its formula is:

[0132]

[0133] where λ i is the response sensitivity factor of node i, which controls the reaction strength of the system to the demand difference; |P demand (t)-P current (t)| is the absolute difference between the current demand and the actual scheduling. This adjustment factor design allows the system to respond quickly when the load changes greatly, and adjust smoothly when the load is stable.

[0134] Furthermore, based on the real-time feedback and the adjustment factor, the present invention can use the following rules to dynamically update the scheduling scheme:

[0135]

[0136] where P adjusted (t) is the scheduling scheme in step 3; η i (t) is the adjustment factor of the above design; is the feedback signal. This rule updates the scheduling scheme according to the real-time feedback signal, enabling the system to respond to changes in the actual load.

[0137] Furthermore, in order to balance the distribution of energy and further optimize the scheduling scheme, the present invention introduces the constraint conditions of multi-objective optimization. These objectives include minimizing power loss, balancing the grid load, and reducing energy waste. The optimization objectives can be expressed as:

[0138]

[0139] where is the power loss of node i at time t; is the load fluctuation metric of node i at time t; α1 and α2 are weighting coefficients that control the importance of power loss and load fluctuation. This objective function comprehensively considers power loss and load fluctuation during the dynamic adjustment process, thus providing a balanced and efficient solution for the system.

[0140] Furthermore, when performing optimization adjustments, it is necessary to ensure that the scheduling scheme satisfies the following constraint conditions:

[0141] The actual output power of the node cannot exceed its maximum power capacity;

[0142] The load of the power network must be balanced to ensure the stability of the grid;

[0143] The transmission loss between nodes must be considered to avoid overloading operation.

[0144] These constraint conditions can be expressed in the following way:

[0145]

[0146] where P max (t) is the maximum power output capacity of node i; is the power constraint condition of node i.

[0147] The system will perform a feasibility check on the scheduling scheme of each node according to these constraints to ensure that all adjusted schemes are physically feasible.

[0148] After real-time feedback and optimization adjustment, the final scheduling scheme P final(t), this solution has undergone real-time feedback and multi-objective optimization to ensure that all constraints are met. The scheduling strategy for each node will minimize power losses as much as possible, balance load fluctuations, and improve the overall system efficiency at the same time. Through the real-time feedback mechanism and optimization adjustment rules in this step, the system can dynamically adjust the scheduling plan according to the load changes in actual operation and optimize the energy distribution. By designing specific feedback adjustment factors, constraints, and multi-objective optimization goals, the system can adapt to different operating states in real time and maintain stable and efficient energy scheduling.

[0149] Embodiment 2

[0150] As Figure 2 shown, the embodiment of the present invention also provides a distributed energy optimization scheduling and management platform based on AI. The platform includes:

[0151] The heterogeneous data source acquisition module 101 is used to collect data from a variety of heterogeneous data sources and perform preprocessing to obtain a data set;

[0152] The multi-objective optimization module 102 is used to construct a multi-objective optimization scheduling model based on the data set and determine the optimized scheduling plan. Among them, the objectives of the multi-objective optimization scheduling model include energy cost, energy efficiency, load balance, and the proportion of green energy use; the optimized scheduling plan includes node power production, consumption, and load information;

[0153] The collaborative scheduling module 103 is used to obtain the optimized scheduling plan, design a collaborative metric function, and based on the collaborative metric function, dynamically adjust the collaborative weights between nodes, optimize the energy transmission and sharing strategy between multiple nodes, and obtain the updated scheduling plan; after obtaining the updated scheduling plan, in view of the possible energy redundancy or shortage between nodes, design a shared optimization objective function to optimize the energy flow;

[0154] The scheduling optimization module 104 is used to collect the real-time feedback of the updated scheduling plan, dynamically adjust the scheduling plan by measuring the difference between the power demand and the actual scheduling power, and design an adjustment factor according to the difference between the power demand and the actual scheduling power to adjust the node power output. Combine the real-time feedback and the adjustment factor to further update the updated scheduling plan, and introduce multi-objective optimization to further optimize the scheduling plan to obtain the final scheduling plan, and ensure that the final scheduling plan meets all constraints, and the constraints include power upper limit, load balance, and transmission loss.

[0155] In summary, by comprehensively applying multi-objective optimization, intelligent collaboration mechanisms, and big data analysis technologies, the present invention not only optimizes the scheduling and management of distributed energy but also enhances the system's intelligence level and flexibility. It effectively solves multiple problems in the prior art, such as poor flexibility in centralized scheduling, insufficient collaboration among distributed nodes, and single-objective optimization, providing a brand-new solution for achieving more efficient, safe, and environmentally friendly distributed energy management.

[0156] The specific embodiments of this specification have been described above, and other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve the desired results. Additionally, the processes depicted in the drawings do not necessarily have to be performed in the specific order or continuous order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0157] The systems, devices, modules, or units illustrated in the above embodiments can be specifically implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or any combination of these devices.

[0158] For convenience of description, when describing the above devices, they are divided into various units according to their functions and described separately. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0159] Those skilled in the art should understand that the embodiments of this specification can be provided as a method, a system, or a computer program product. Therefore, the embodiments of this specification can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0160] This specification is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the specification. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device generate means for implementing the functions specified in one or more of the flows Figure 1 one or more of the flows and / or blocks Figure 1 one or more of the blocks.

[0161] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, such that the instructions stored in the computer-readable memory produce a manufacture including instruction means that implement the functions specified in one or more of the flows Figure 1 one or more of the flows and / or blocks Figure 1 one or more of the blocks.

[0162] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more of the flows Figure 1 one or more of the flows and / or blocks Figure 1 one or more of the blocks.

[0163] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0164] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.

[0165] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0166] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0167] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0168] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0169] Finally, it should be noted that what is disclosed in the distributed energy optimization scheduling and management platform based on AI disclosed in the embodiments of the present invention is only the preferred embodiments of the present invention, which are only used to illustrate the technical solutions of the present invention, rather than limiting it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. An AI-based distributed energy optimization scheduling and management method, characterized in that The method includes the following steps: Collect data from multiple heterogeneous data sources and perform preprocessing to obtain a data set; Construct a multi-objective optimization scheduling model based on the data set to determine an optimized scheduling plan; among them, the objectives of the multi-objective optimization scheduling model include energy cost, energy efficiency, load balance, and the proportion of green energy used; the optimized scheduling plan includes node power production, consumption, and load information; Obtain the optimized scheduling plan, design a collaboration metric function, and based on the collaboration metric function, dynamically adjust the collaboration weights between nodes, optimize the energy transmission and sharing strategy among multiple nodes, and obtain an updated scheduling plan; after obtaining the updated scheduling plan, design a shared optimization objective function to optimize the energy flow in case of possible energy redundancy or shortage between nodes; Collect real-time feedback of the updated scheduling plan, dynamically adjust the scheduling plan by measuring the difference between the power demand and the actual scheduling power, and design an adjustment factor based on the difference between the power demand and the actual scheduling power to adjust the node power output. Combine the real-time feedback and the adjustment factor to further update the updated scheduling plan, and introduce multi-objective optimization to further optimize the scheduling plan to obtain the final scheduling plan, and ensure that the final scheduling plan meets all constraints, and the constraints include power upper limit, load balance, and transmission loss.

2. The AI-based distributed energy optimization scheduling and management method according to claim 1, wherein The collection of data from multiple heterogeneous data sources includes weather, load, battery status, and electricity market price; The preprocessing includes: Determine the weight of the corresponding data source based on the quality, timeliness, and historical performance of the data source; Perform weighted fusion on different data according to the weight of the corresponding data source to obtain the fused data; Obtain the fused data and use the method based on adaptive weighted Kalman filter to remove noise to obtain the filtered data.

3. The AI-based distributed energy optimization scheduling and management method according to claim 2, characterized in that After obtaining the preprocessed data, based on the feature selection method of maximizing mutual information and regularization, screen out the most effective feature set as the data set; the feature selection method based on maximizing mutual information and regularization is expressed as: Among them, I(X i ; Y) represents the mutual information between feature X i and the target variable Y, P(x i , y) is the joint probability distribution of feature X i and the target variable Y, and P(x i ) and P(y) are their marginal probability distributions respectively; At the same time, introduce the regularization L reg term to reduce the impact of redundant features on the model: where λ is the regularization coefficient, and Redundancy(X i ) represents the redundancy of feature X i with other features.

4. The AI-based distributed energy optimization scheduling and management method according to claim 1, characterized in that The comprehensive objective function of the multi-objective optimal scheduling model is the weighted sum of the objectives of the multi-objective optimal scheduling model, and the weights are adjusted according to real-time data and status: Among them, w1(t), w2(t), w3(t), w4(t) are the dynamic weights of the objectives, and the specific values are adjusted according to real-time data and system status.

5. The AI-based distributed energy optimization scheduling and management method according to claim 4, characterized in that, The construction of the multi-objective optimization scheduling model based on the data set to determine the optimized scheduling plan specifically includes: Design a feedback mechanism for the dynamic weights of the objectives, and the feedback mechanism automatically adjusts the weights of the objectives according to real-time data; among them, the real-time data includes market price and equipment health status; During the optimization process, it is necessary to ensure that the constraints are met, including: power balance constraint, energy storage device constraint, and green energy ratio; Among them, The power balance constraint is used to ensure that the total generated power is equal to the demanded power; The energy storage device constraint is used to limit the charging and discharging capabilities of the energy storage device to avoid overcharging or over-discharging; The green energy ratio is used to ensure that the usage amount of green energy reaches a predetermined ratio; For the comprehensive objective function, an adaptive particle swarm optimization algorithm is adopted, which combines a feedback mechanism to dynamically adjust the search strategy of the particles; among them, the update rule of the adaptive particle swarm optimization algorithm is as follows: v i (t + 1) = w(t)·v i (t) + c1·r1·(p i -x i (t)) + c2·r2·(g - x i (t)) Among them, v i (t + 1) is the updated particle velocity, v i (t) is the particle velocity, x i (t) is the particle position, p i and g are respectively the historical best and global best positions of the particle, w(t) is the inertia weight, and r1 and r2 are random numbers.

6. The AI-based distributed energy optimization scheduling and management method according to claim 1, wherein The collaboration metric function is expressed as: Among them, w ij (t) is the collaboration weight between node i and node j, representing the degree of closeness of collaboration; γ ij (t) is the transmission loss factor, reflecting the energy transmission loss between node i and node j; α ij is the adjustment factor, controlling the impact of transmission loss on collaboration; P i (t) and P j (t) are the electricity production or consumption of node i and node j respectively; Meanwhile, dynamically update the collaboration weight; when the collaboration requirement of node i increases, its weight w ij (t) will increase, thereby enhancing collaboration with other nodes; Update the scheduling scheme of the node by combining the collaboration weight and the collaboration metric function, expressed as: Among them, P optimized (t) is the optimized scheduling scheme, and α is the adjustment factor used to balance the relationship between the optimization scheme and the collaboration requirements.

7. The AI-based distributed energy optimization scheduling and management method according to claim 6, wherein The shared optimization objective function, expressed as: Combined with transmission loss and sharing efficiency Among them, η ij (t) is the sharing efficiency coefficient, which represents the efficiency of node i sharing energy with node j, considering the losses during transmission.

8. The AI-based distributed energy optimization scheduling and management method according to claim 1, characterized in that, The multi-objective optimization is expressed as: Among them, is the power loss of node i at time t; is the load fluctuation metric of node i at time t; α1 and α2 are weighting coefficients that control the importance of power loss and load fluctuation.

9. The AI-based distributed energy optimization scheduling and management method according to claim 1, wherein When performing multi-objective optimization During adjustment, the following constraints need to be satisfied: Among them, P max (t) is the maximum power output capacity of node i; is the power constraint condition of node i, and P new (t) is the final scheduling scheme.

10. An AI-based distributed energy optimization scheduling and management platform, characterized in that the platform Include: The heterogeneous data source acquisition module is used to collect data from multiple heterogeneous data sources, preprocess it, and obtain a data set; The multi-objective optimization module is used to construct a multi-objective optimization scheduling model based on the data set to determine the optimized scheduling scheme; among them, the objectives of the multi-objective optimization scheduling model include energy cost, energy efficiency, load balance, and the proportion of green energy use; the optimized scheduling scheme includes node power production, consumption, and load information; The collaborative scheduling module is used to obtain the optimized scheduling scheme, design a collaboration metric function, dynamically adjust the collaboration weight between nodes based on the collaboration metric function, optimize the energy transmission and sharing strategy between multiple nodes, and obtain the updated scheduling scheme; after obtaining the updated scheduling scheme, in view of the possible energy redundancy or shortage between nodes, design a shared optimization objective function to optimize the energy flow; The scheduling optimization module is used to collect the real-time feedback of the updated scheduling scheme, dynamically adjust the scheduling scheme by measuring the difference between the power demand and the actual scheduling power, design an adjustment factor according to the difference between the power demand and the actual scheduling power to adjust the node power output, combine the real-time feedback and the adjustment factor to further update the updated scheduling scheme, introduce multi-objective optimization, further optimize the scheduling scheme, obtain the final scheduling scheme, and ensure that the final scheduling scheme meets all constraints, and the constraints include power upper limit, load balance, and transmission loss.

Images