Multi-element load access optimization method and system in high-reliability power distribution network planning

Abstract

The invention discloses a multi-element load access optimization method and system in high-reliability power distribution network planning, and relates to the technical field of power systems, and the method comprises the steps: carrying out the demand analysis of a load, dividing a multi-element load into a plurality of hierarchies based on the power demand, the use characteristic and the time sequence characteristic, and building a coupling relation model between the hierarchies; acquiring operation data of a load point in real time, processing the data, and extracting load data; and solving the coupling relation model based on the load data, constructing an objective function, and calculating optimal load access configuration by adopting an optimization algorithm. According to the multi-element load access optimization method in high-reliability power distribution network planning provided by the invention, the mutual influence between load points is comprehensively reflected by establishing the multi-dimensional coupling relation model, and more reasonable load access configuration is realized. The advantages of the particle swarm optimization in global search and the high efficiency of the genetic algorithm in local optimization are fully utilized, the calculation efficiency and the result precision of the optimization process are improved, and the defect that a traditional single algorithm is prone to falling into local optimization is overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a method and system for optimizing multi-load access in high-reliability distribution network planning. Background Art

[0002] With the rapid development of smart grid technology, load management is becoming increasingly important in power systems. Effective load management can ensure the stable operation of the power grid, optimize energy distribution, and improve the economy of the system. However, traditional load management methods mainly rely on preset load scheduling plans and prediction models based on historical data, which have certain limitations and cannot cope with the complex changes and diversified demands of loads in the power grid.

[0003] First, traditional methods often lack real-time and flexibility. Load scheduling plans are usually carried out when load demand is relatively stable, and cannot quickly respond to sudden load fluctuations in the power grid. For example, during peak load periods, demand in certain areas may increase significantly, and traditional scheduling methods often cannot be adjusted in time, which may lead to problems such as grid overload, equipment damage, or power outages.

[0004] Secondly, traditional methods fail to fully consider the coupling relationship between load points. Load points in the power grid are usually not independent, but affect each other. Factors such as power demand, usage differences, and temperature between load points interact with each other to form a complex coupling relationship. However, traditional load management methods usually ignore these coupling relationships, resulting in the failure to achieve coordinated optimization between loads during load scheduling, thereby failing to fully improve energy utilization efficiency.

[0005] In addition, traditional load optimization methods are mostly based on rules or simple linear optimization models, which usually ignore the time-varying nature of load changes and the complex system coupling. In practical applications, these methods cannot adapt to the rapid changes in load fluctuations, nor can they perform multi-dimensional optimization, resulting in low scheduling efficiency and possibly failure to maximize energy benefits.

[0006] Finally, the calculation process in the traditional method may lead to resource waste. Since the decision of load dispatch is based on a fixed plan and lacks a dynamic adjustment mechanism, some load points of the power grid may be over-dispatched or not dispatched in time, resulting in energy waste or unbalanced load of the power grid.

[0007] The present invention overcomes these defects of the traditional methods by introducing real-time data collection, load coupling relationship modeling and optimization algorithm, and provides a more intelligent, efficient and dynamic load optimization solution. Summary of the invention

[0008] In view of the above-mentioned problems, the present invention is proposed.

[0009] Therefore, the technical problems solved by the present invention are: the existing load scheduling methods lack real-time performance, flexibility and cannot consider the complex coupling relationship between load points, as well as how to dynamically optimize the load access configuration based on real-time data.

[0010] In order to solve the above technical problems, the present invention provides the following technical solutions: a multi-load access optimization method in high-reliability distribution network planning, comprising:

[0011] Conduct demand analysis on loads, divide multiple loads into multiple levels based on power demand, usage characteristics and timing characteristics, and establish a coupling relationship model between levels; collect operating data of load points in real time, process the data, and extract load data; solve the coupling relationship model based on load data, construct the objective function, and use the optimization algorithm to calculate the best load access configuration.

[0012] As a preferred solution of the multi-load access optimization method in the high-reliability distribution network planning described in the present invention, wherein: the coupling relationship model is expressed as:

[0013]

[0014] Among them, C ij (t) represents the coupling relationship between load point i and load point j, which changes with time t, σ(#) represents the activation function, The summation symbol indicates the accumulation of different coupling characteristics. k Coupling gain coefficient, the weight of the influence of the kth load characteristic on the coupling relationship, is the exponential decay term of the power demand difference, which indicates the influence of the load point power difference on the coupling relationship. α indicates the sensitivity of the influence of the control power demand difference on the coupling relationship. β(U i -U j ) 2 is the quadratic term of the difference in usage characteristics, which indicates the influence of the difference in usage characteristics of the load point on the coupling relationship. β indicates the influence of the difference in usage characteristics on the coupling relationship. γlog(|T i (t)-T j (t)|+1) is the temperature difference logarithm, which indicates the influence of load point temperature difference on the coupling relationship, γ indicates the influence intensity of control temperature difference on the coupling relationship, |P i (t)-P j (t)| represents the power demand difference between load point i and load point j, P i (t) and P j (t) represent the power demand of load point i and load point j at time t, respectively, |U i -U j| represents the difference in usage characteristics between load point i and load point j, U i and U j Respectively represent the usage characteristics of load point i and load point j, |T i (t)-T j (t)| represents the real-time temperature difference between load point i and load point j, T i (t) and T j (t) represent the temperature of load point i and load point j at time t respectively.

[0015] As a preferred solution of the multi-load access optimization method in the high-reliability distribution network planning described in the present invention, the operating data includes power demand parameters, usage characteristic parameters and temperature parameters.

[0016] As a preferred solution of the multi-load access optimization method in the high-reliability distribution network planning described in the present invention, the data processing includes performing data cleaning operations, removing outliers and noise data, and filling missing values; synchronizing data from different sources according to timestamps; and standardizing the data.

[0017] As a preferred solution of the multi-load access optimization method in the high-reliability distribution network planning described in the present invention, wherein: the load data includes the power demand difference |P i (t)-P j (t) |, usage characteristics difference (U i -U j ) 2 and the temperature difference log(|T i (t)-T j (t)|+1).

[0018] As a preferred solution of the multi-load access optimization method in the high-reliability distribution network planning described in the present invention, wherein: the objective function is expressed as,

[0019]

[0020] Among them, P i (t) represents the power demand of load point i at time t, represents the maximum capacity of load point i, R i represents the load capacity of load point i, w 1 ,w 2 ,w 3 They represent weight coefficients respectively.

[0021] As a preferred solution of the multi-load access optimization method in the high-reliability distribution network planning described in the present invention, wherein: the calculation of the optimal load access configuration includes initializing a particle swarm, each particle represents a load access configuration, randomly assigning a position and a speed to each particle, representing the power demand configuration of different load points, and calculating the objective function value of each particle;

[0022] Based on the particle swarm algorithm, a global search is performed. The particle updates its position in the solution space according to the objective function value, the particle's historical optimal position, and the global optimal position. The particle position is adjusted by the speed update formula, and the optimal solution is searched through multiple iterations.

[0023] Based on the global search of particle swarm, a genetic algorithm is used to locally optimize the global optimal solution of the particle swarm. By selecting particles with higher fitness as parents, crossover and mutation operations are performed to generate new individuals and optimize the load access configuration.

[0024] According to the objective function value of the new individual optimized by the genetic algorithm, the individual optimal solution and the global optimal solution of the particle swarm are updated. If the objective function value of the offspring individual is better than the current optimal solution, the corresponding optimal solution is updated;

[0025] When the particle swarm algorithm and the genetic algorithm run alternately to the set maximum number of iterations, or the objective function value reaches the preset convergence standard, the algorithm execution is terminated and the global optimal solution is output as the optimal load access configuration;

[0026] Output the final optimal load access configuration, including the power demand of each load point at different time points, to meet the optimization objectives, maximize energy benefits and optimize the load access process.

[0027] Another object of the present invention is to provide a multi-load access optimization system in high-reliability distribution network planning, which can solve the problems of load scheduling lag, lack of collaborative optimization, and neglect of complex coupling relationships between loads, energy waste and grid overload in existing load scheduling methods by constructing a multi-load access optimization system in high-reliability distribution network planning.

[0028] To solve the above technical problems, the present invention provides the following technical solutions: a multi-load access optimization system in high-reliability distribution network planning, comprising: a demand analysis module, used to perform demand analysis on loads, divide the multi-loads into multiple levels based on power demand, usage characteristics and timing characteristics, and establish a coupling relationship model between the levels; a feature extraction module, used to collect operating data of load points in real time, process the data, and extract load data; an optimization solution module, used to solve the coupling relationship model based on load data, construct an objective function, and use an optimization algorithm to calculate the optimal load access configuration.

[0029] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the multi-load access optimization method in high-reliability distribution network planning are implemented as described above.

[0030] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for optimizing multi-load access in high-reliability distribution network planning as described above.

[0031] Beneficial effects of the present invention: The multi-load access optimization method in high-reliability distribution network planning provided by the present invention establishes a multi-dimensional coupling relationship model to fully reflect the mutual influence between load points, thereby achieving a more reasonable load access configuration. The optimization strategy combining the particle swarm algorithm with the genetic algorithm is adopted, which makes full use of the advantages of the particle swarm algorithm in global search and the high efficiency of the genetic algorithm in local optimization, improves the computational efficiency and result accuracy of the optimization process, and avoids the defect of the traditional single algorithm that is prone to fall into the local optimum. This combined optimization method not only speeds up the convergence speed, but also ensures the global optimality of the optimization solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0033] Figure 1 An overall flow chart of a method for optimizing multi-load access in high-reliability distribution network planning provided by one embodiment of the present invention.

[0034] Figure 2 This is an overall structural diagram of a multi-load access optimization system in high-reliability distribution network planning provided by the second embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0036] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0037] Example 1

[0038] Reference Figure 1 , is an embodiment of the present invention, and provides a multi-load access optimization method in high-reliability distribution network planning, comprising:

[0039] Step 1: Analyze the load demand, divide the multi-level loads into multiple levels based on power demand, usage characteristics and timing characteristics, and establish a coupling relationship model between the levels;

[0040] Step 2: Collect the operating data of the load point (the load point refers to the current, voltage, power and other operating data of each electrical equipment, device or load port in the power system) in real time, process the data and extract the load data;

[0041] Step 3: Solve the coupling relationship model based on load data, construct the objective function, and use the optimization algorithm to calculate the optimal load access configuration.

[0042] The coupling relationship model is expressed as:

[0043]

[0044] Among them, C ij (t) represents the coupling relationship between load point i and load point j, which changes with time t, σ(#) represents the activation function, The summation symbol indicates the accumulation of different coupling characteristics. k Coupling gain coefficient, the weight of the influence of the kth load characteristic on the coupling relationship, is the exponential decay term of the power demand difference, which indicates the influence of the load point power difference on the coupling relationship. α indicates the sensitivity of the influence of the control power demand difference on the coupling relationship. β(U i -U j ) 2 is the quadratic term of the difference in usage characteristics, which indicates the influence of the difference in usage characteristics of the load point on the coupling relationship. β indicates the influence of the difference in usage characteristics on the coupling relationship. γlog(|T i (t)-T j (t)|+1) is the temperature difference logarithm, which indicates the influence of load point temperature difference on the coupling relationship, γ indicates the influence intensity of control temperature difference on the coupling relationship, |P i (t)-P j(t)| represents the power demand difference between load point i and load point j, P i (t) and P j (t) represent the power demand of load point i and load point j at time t, respectively, |U i -U j | represents the difference in usage characteristics between load point i and load point j, U i and U j Respectively represent the usage characteristics of load point i and load point j, |T i (t)-T j (t)| represents the real-time temperature difference between load point i and load point j, T i (t) and T j (t) represent the temperature of load point i and load point j at time t respectively.

[0045] First of all, the purpose of load demand analysis is to fully understand the basic characteristics of each load point, so as to provide accurate input for subsequent coupling relationship modeling. Load demand analysis includes the following aspects:

[0046] Power demand is one of the most basic characteristics of a load point, which determines the power demand of the load point on the power grid. Based on historical data or actual measurements, the power demand of the load can be divided into multiple levels such as high, medium, and low. For example, a high-power load (such as industrial production equipment) may require continuous high-power supply, while a low-power load (such as small household appliances) may consume less power in a short period of time.

[0047] The purpose of classifying power demand is to identify which loads have a greater impact on the grid during operation and which loads have a smaller impact. In the model, the coupling relationship of high-power loads will be stronger, while the coupling relationship of low-power loads may be weaker.

[0048] Usage characteristics analysis considers the specific usage scenarios or types of loads. For example, the power demand patterns of industrial loads, commercial loads, and residential loads are quite different.

[0049] The characteristics of loads are divided according to their uses. For industrial loads and residential loads, although the power requirements may be similar, their load patterns and power consumption cycles are very different. These different loads need to be classified according to their uses and the possible coupling relationship strength between them needs to be determined.

[0050] Time series characteristics analysis mainly considers the changes in load in the time dimension. For example, the consumption pattern and demand of load will change differently in different time periods (such as peak period and valley period). Time series characteristics may also be related to factors such as seasonal changes and holidays.

[0051] By analyzing the time series characteristics of the load, we can predict the change rules of the load, and use these rules to help predict future load demand. The time series characteristics analysis provides time information for the establishment of coupling relationships and also affects the mutual influence between load points in different time periods.

[0052] According to the above demand analysis results, load points will be divided into different levels, and the level division mainly considers the following factors:

[0053] Power demand level: High, medium and low power demands are divided into high power layer, medium power layer and low power layer respectively. Loads with higher power demand usually have a greater impact on the power system, so they need to be given priority in the access configuration.

[0054] The loads are divided into different levels according to their specific uses (such as industry, residential, and commercial). The loads at each level have different power consumption characteristics, which affect their coupling strength with other loads.

[0055] Different load timing characteristics categories are divided according to the timing characteristics of the load. For example, some loads may consume more power during the peak hours of the day and less power at night.

[0056] These three factors will determine the tier attribution of each load point, and each load point will eventually be classified into a comprehensive tier (for example, high-power and industrial-use loads belong to a high tier).

[0057] The coupling relationship describes the degree of mutual influence between different load points. In load demand analysis, the power demand, usage characteristics and timing characteristics of the load points will affect the coupling relationship between them. The stronger the coupling relationship, the greater the mutual influence between these load points during operation, and vice versa.

[0058] The core of coupling relationship modeling is to use certain mathematical methods (such as distance function, weighted function, etc.) to represent the mutual influence between load points based on power demand, usage characteristics and timing characteristics. In this model, three key characteristics will determine the strength of the coupling relationship between load points.

[0059] The operation data includes power demand parameters, usage characteristic parameters and temperature parameters.

[0060] The data processing includes performing data cleaning operations, removing outliers and noise data, and filling missing values; synchronizing data from different sources according to timestamps; and standardizing data.

[0061] The load data includes the power demand difference |P i (t)-P j (t) |, usage characteristics difference (Ui -U j ) 2 and the temperature difference log(|T i (t)-T j (t)|+1).

[0062] The objective function is expressed as,

[0063]

[0064] Among them, P i (t) represents the power demand of load point i at time t, represents the maximum capacity of load point i, R i represents the load capacity of load point i, w 1 ,w 2 ,w 3 They represent weight coefficients respectively.

[0065] The calculating of the optimal load access configuration includes initializing a particle swarm, each particle representing a load access configuration, randomly assigning a position and a speed to each particle to represent the power demand configuration of different load points, and calculating the objective function value of each particle;

[0066] Based on the particle swarm algorithm, a global search is performed. The particle updates its position in the solution space according to the objective function value, the particle's historical optimal position, and the global optimal position. The particle position is adjusted by the speed update formula, and the optimal solution is searched through multiple iterations.

[0067] Based on the global search of particle swarm, a genetic algorithm is used to locally optimize the global optimal solution of the particle swarm. By selecting particles with higher fitness as parents, crossover and mutation operations are performed to generate new individuals and optimize the load access configuration.

[0068] According to the objective function value of the new individual optimized by the genetic algorithm, the individual optimal solution and the global optimal solution of the particle swarm are updated. If the objective function value of the offspring individual is better than the current optimal solution, the corresponding optimal solution is updated;

[0069] When the particle swarm algorithm and the genetic algorithm run alternately to the set maximum number of iterations, or the objective function value reaches the preset convergence standard, the algorithm execution is terminated and the global optimal solution is output as the optimal load access configuration;

[0070] Output the final optimal load access configuration, including the power demand of each load point at different time points, to meet the optimization objectives, maximize energy benefits and optimize the load access process.

[0071] In the multi-load access optimization method in the high-reliability distribution network planning of the present invention, step 1 aims to comprehensively analyze the needs of each load point in the power system, divide it into multiple levels based on power demand, usage characteristics and timing characteristics, and establish a coupling relationship model between the levels on this basis. First, the power demand analysis of the load is carried out, and the power demand of the load is divided into three levels: high power, medium power and low power through in-depth research on historical data or real-time measurement data. High-power loads, such as industrial production equipment, usually require continuous and stable high-power supply, which has a significant impact on the load capacity and stability of the power grid; medium-power loads, such as commercial facilities, have relatively stable power demand, but may fluctuate during specific periods; low-power loads, such as small household appliances, have low power demand and large volatility, and have relatively small overall impact on the power grid. The purpose of this level division is to identify load points that have a greater impact on the power grid and give them a higher weight coefficient A in the subsequent coupling relationship model. k , thereby ensuring that these critical load points are given priority during the optimization configuration process.

[0072] Subsequently, the usage characteristics analysis is conducted, and the loads are divided into different categories such as industrial loads, commercial loads, and residential loads according to the specific usage scenarios or types of the loads. Industrial loads usually have higher power requirements and relatively fixed power consumption patterns, commercial loads are more flexible in power consumption patterns, and residential loads show greater power consumption volatility and periodicity. The usage characteristics analysis further refines the load hierarchy division by identifying the similarities and differences between load points. Load points with the same purpose have similar power consumption patterns and show stronger mutual influence in the coupling relationship model. Therefore, the usage characteristics difference term β(U i -U j ) 2 It reflects the intensity of their mutual influence. If the usage is very different, the value of this item will increase, resulting in a weakened coupling relationship.

[0073] Next, the time series characteristics analysis is carried out, mainly considering the change law of load in the time dimension, including the fluctuation pattern of load demand in different time periods (such as peak period and valley period) and the impact of seasonal changes on load demand. By analyzing the load during peak period, valley period and seasonal load, the change law of load can be predicted, and the time series characteristic difference term γlog(|T i (t)-T j (t)|+1) reflects its influence on the coupling relationship. The greater the difference in timing characteristics, the more inconsistent the demand changes of the load points in time, and the weaker the coupling relationship, and vice versa. This analysis not only affects the demand fluctuations of a single load point, but also determines the degree of mutual influence between load points in different time periods.

[0074] Based on the analysis results of the above power demand, usage characteristics and timing characteristics, the load points are divided into multiple levels, each level represents a group of load points with similar characteristics. This hierarchical division provides structured information for the construction of the coupling relationship model, so that the mutual influence between load points at different levels can be accurately quantified.

[0075] Through the above comprehensive coupling relationship model, the degree of mutual influence between different load levels can be quantified. The coupling relationship between load points with high power, high usage similarity and high timing similarity is strong, reflecting that they have significant mutual influence in the operation of the power grid and their access configuration needs to be given priority; while the coupling relationship between load points with low power, large usage differences or large timing differences is weak, allowing them to obtain lower priority in the access configuration or be appropriately adjusted. This model can not only accurately reflect the dynamic interaction between load points, but also ensure that the calculation results of the coupling relationship are within a reasonable range through the nonlinear transformation mechanism σ(·), avoiding the problem of over-coupling or under-coupling.

[0076] Through the above steps, the present invention applies the load demand analysis results to the construction of the coupling relationship model. Through the multi-dimensional analysis of power demand, usage characteristics and timing characteristics, the coupling relationship model can fully capture the complex interactions between load points, and then in the subsequent load access configuration optimization process, provide a more accurate and efficient optimization solution, and significantly improve the operating efficiency and stability of the power system.

[0077] The ratio of power demand to maximum capacity is given by the power demand P at load point i at time t. i (t) and the maximum capacity C of the load point max The design purpose of this item is to control the power demand of the load point to optimize within the maximum capacity range. i When (t) approaches the maximum capacity of the load point, the value of this item will become larger, thereby guiding the system optimization algorithm to maintain the power demand of the load point below the maximum capacity, thereby avoiding overload operation. Weight coefficient w 1 Further control the degree of influence of this item on the objective function. If the power demand of the load point is too large, the system will take corresponding measures to limit the power output to ensure system stability.

[0078] Power demand and load point load capacity R i The square term is designed to adjust the sensitivity of the load capacity of the load point. The introduction of the square form makes the relationship between the power demand of the load point and the load capacity nonlinear. When the power demand of the load point P i (t) relative to load capacity R iWhen it is large, the value of this item increases rapidly, forcing the optimization algorithm to consider the load capacity limit of the load point when calculating the load access configuration to avoid excessive power demand at the load point. 2 Determines the degree of influence of this item on the optimization process. In practical applications, if the load capacity of some load points is low, you can avoid over-reliance on these load points by adjusting the weight of this item.

[0079] By ∑ j≠i C ij (t).P j (t) is used to measure the mutual coupling influence between load point i and other load points. Coupling relationship C ij (t) is calculated through the coupling relationship model and represents the degree of mutual influence between load points i and j. The purpose of this item is to ensure that the interaction between load points is reasonably considered to avoid imbalance between load points due to overly concentrated load access configuration. Weight coefficient w 3 It further determines the contribution of the coupling relationship to the objective function. If the coupling relationship between the load points is strong, the value of this item will become larger, prompting the optimization algorithm to consider the mutual influence between the load points when configuring the load access, ensuring that the power demand between the load points can be coordinated and reasonable. This item is particularly important for global optimization, avoiding local optimization and ensuring the balance and stability of the overall operation of the system.

[0080] Among them, the first item controls the power demand of the load point not to exceed its maximum capacity, ensuring the stable operation of each load point; the second item controls the matching of the power demand of the load point with the load capacity to avoid instability caused by excessive load; the third item considers the coupling relationship between load points to ensure the coordination of the global load access configuration and avoid local imbalance. The combination of these items enables the objective function to not only optimize the independent operation status of each load point, but also take into account the interaction between different load points, thereby achieving the effect of global optimization.

[0081] Compared with the prior art, the traditional load access configuration optimization usually only constrains the power demand and load capacity of the load point, but ignores the coupling relationship between the load points. The present invention further improves the accuracy of load access configuration optimization and the overall stability of the system by introducing coupling relationship terms. In particular, when considering the complex interactions between load points, the objective function can dynamically adjust the optimization results to adapt to the mutual influence between different load points, avoiding the local optimal solution problem that may occur in traditional methods.

[0082] In addition, the design of nonlinear terms in the objective function (such as square and logarithmic terms) also increases the complexity of the optimization problem, so that the objective function can not only handle simple linear constraints, but also adapt to more complex load demand and load capacity relationships. This innovation improves the adaptability and flexibility of the objective function in practical applications.

[0083] Specifically, the position of the particle represents the load access configuration. The dimension of each particle is equal to the number of load points. The position and velocity of the particle are initialized in the following way:

[0084] Particle position P i (t), represents the power demand of the i-th load point.

[0085] Particle velocity V i (t) represents the velocity of the particle and is used to control the update of the position.

[0086] The initial values ​​of position and velocity are determined by the possible range of load points.

[0087] At each iteration, the particle's position is updated according to its velocity:

[0088] P i (t+1)=P i (t)+V i (t+1)

[0089] Among them, V i (t+1) is the updated speed.

[0090] The velocity update formula combines the inertia term, the individual experience term, and the global experience term:

[0091] V i (t+1)=w·V i (t)+c 1 ·rand 1 ·(P best -P i (t))+c 2 ·rand 2 ·(G best -P i (t))

[0092] Among them, w is the inertia weight, which controls the range of particle search, c 1 ,c 2 is the learning factor, rand 1 ,rand 2 is a random factor, P best is the best historical position of the particle, G best is the global optimal position.

[0093] For the current position of each particle, its fitness is calculated through the objective function:

[0094]

[0095] Update the particle's best position and the global best position: By calculating the objective function value, update the best position of each particle and the global best position, including P best is the optimal position of the individual, G best is the best solution among all particles.

[0096] The algorithm stops when the predetermined maximum number of iterations is reached or the objective function value converges.

[0097] The genetic algorithm will perform local optimization based on the preliminary solution given by the particle swarm algorithm to further reduce the value of the objective function. The genetic algorithm is particularly suitable for generating new load access configurations through crossover and mutation, and improving the quality of the solution in local search.

[0098] The initial population consists of a set of load access configurations (i.e., the output of the particle swarm algorithm), and each individual is a configuration, which represents the power demand of a load point.

[0099] The fitness of each individual is evaluated by calculating the objective function value. The fitness function is:

[0100]

[0101] Among them, fitness(i) represents the fitness value of the i-th individual. The larger the fitness value, the better the individual performs during the optimization process. Represents the objective function value of the ith individual. The smaller the value, the closer the solution of the individual is to the optimal solution.

[0102] Individuals with high fitness will have a higher probability of being selected into the next generation.

[0103] A single-point crossover operation is used to exchange the genes of two individuals to generate an offspring configuration. The crossover point is determined by random selection. The individuals after the crossover operation are evaluated on the objective function to select individuals with better fitness:

[0104] offspring(i)=crossover(P parent1 (i),P parent2 (i))

[0105] Among them, offspring(i) represents the configuration of the i-th offspring individual, represents the new solution generated by the crossover operation, and P parent1 (i) represents the power demand of the i-th load point of parent individual 1, P parent22(i) represents the power demand of the i-th load point of parent individual 2.

[0106] The mutation operation increases the diversity of the population by fine-tuning the genes of individuals (such as adjusting parameters such as power requirements and load capacity). The expression of the mutation operation is:

[0107] P mutated (i) = P original (i)+ΔP

[0108] Among them, P mutated (i) represents the power demand of the i-th load point after the change, P original (i) represents the power demand of the i-th load point before the mutation, and ΔP represents the mutation amount, which indicates the random adjustment of the power demand.

[0109] Calculate the objective function value of each generation and select individuals with higher fitness to enter the next generation.

[0110] The stopping condition of the genetic algorithm is similar to that of the particle swarm algorithm. The algorithm stops when the maximum number of iterations is reached or the objective function converges.

[0111] Traditional load optimization methods usually only consider a single factor, such as the power demand and capacity limit of the load point. These methods use simple weight distribution and linear relationships to configure load access, ignoring the complex coupling relationship between load points, the interaction of multi-dimensional factors such as time series characteristics, resulting in the optimization results only having local optimality and being unable to fully cope with the complex situations that occur in actual power grid operation. Most existing load optimization methods only focus on simple factors such as power demand or load capacity, ignoring dynamic factors such as the mutual influence between power grid load points and time series fluctuations. Although the optimization methods that ignore these factors perform well in static environments, they cannot cope with the mutual coupling and dynamic changes between different load points in the power grid, which can easily lead to uneven load distribution, even overload or unstable operation. Traditional methods mostly rely on simple linear or empirical models, and their optimization targets are often local, lacking a precise grasp of overall system optimization. The nonlinear relationship, coupling effect and multiple constraints between power grid load points often make it difficult for traditional methods to obtain comprehensive and accurate optimization solutions.

[0112] The method of this embodiment comprehensively considers multiple factors (such as power demand, load capacity, temperature, usage characteristics, etc.) and optimizes based on an accurate coupling relationship model, which can fully reflect the complex interaction effects between load points in the power grid. By adopting a comprehensive objective function, the demand and mutual relationship of each load point can be fully quantified, thereby avoiding the singleness and limitations of traditional methods. Through this multi-dimensional optimization, the present invention can effectively cope with the variable working conditions of the power grid and improve the stability and energy efficiency of the system.

[0113] Traditional load optimization methods are usually calculated based on static data and lack real-time adjustment capabilities. In actual power grid operation, factors such as load demand, equipment status, and environmental changes are dynamically changing. Traditional methods fail to consider how to adjust load access configuration based on these real-time information, resulting in a lack of timeliness and flexibility in optimization results. Many traditional load optimization methods rely only on historical data or prediction models for static optimization, ignoring the impact of real-time changing factors. The change of power grid load is dynamic. For example, factors such as climate, time period, and equipment failure can cause fluctuations in load demand, but the optimization results of traditional methods often cannot be adjusted in time when these fluctuations occur, resulting in the inadaptability of load access configuration. Traditional methods usually perform a one-time load optimization calculation at the beginning and then no longer adjust. This design lacking real-time monitoring and adjustment makes it impossible for traditional methods to flexibly respond to emergencies and load changes in actual operation, which easily leads to unstable power grid operation or waste of resources.

[0114] The method of this embodiment can continuously monitor key parameters such as grid load and equipment status during operation by introducing a real-time data feedback mechanism, and realize dynamic adjustment of load access configuration. Whenever the load demand changes, the real-time feedback mechanism can immediately adjust the load configuration according to the latest operating data to ensure that the grid always maintains the best state under different load change scenarios. This dynamic adjustment capability enables the present invention to effectively respond to fluctuations and emergencies in load point demand in the grid, avoids the lack of flexibility caused by static optimization in traditional methods, and improves the responsiveness and stability of the grid.

[0115] Traditional optimization algorithms often rely on a single optimization method (such as particle swarm optimization, genetic algorithm, etc.). Their limitation is that it is difficult to take into account the dual requirements of global search and local fine adjustment. Although global search can better avoid local optimality, it may not be as fine as local optimization, and vice versa. Many existing methods rely on a single optimization algorithm. Although the particle swarm algorithm (PSO) can perform global search, its ability to explore fine solutions is relatively weak. The genetic algorithm (GA) has strong local optimization capabilities, but it is easy to fall into local optimality when the search space is large. Using only one of the methods often cannot take into account the needs of global and local optimization. In traditional methods, particle swarm optimization and genetic algorithm can often only solve one aspect of the problem respectively - either find a rough global optimal solution or further optimize through local search, but the combination of the two often has the risk of inefficiency and unsatisfactory results.

[0116] The method of this embodiment combines the particle swarm algorithm with the genetic algorithm to give full play to the advantages of both, ensuring global search capabilities while being able to perform local fine optimization. The particle swarm algorithm is used to perform a global search to locate the potential optimal solution area, and then the genetic algorithm is used for fine adjustment to optimize the load access configuration. This combined optimization can improve the calculation accuracy and efficiency while ensuring the global optimum, avoiding the problems of low optimization accuracy or long calculation time caused by a single algorithm in traditional methods. The combined optimization method of the present invention breaks the limitations of a single optimization algorithm, and through the mutual complementation of the global and local, it not only improves the optimization effect of the load access configuration, but also optimizes the resource utilization in the calculation process, providing a more efficient and accurate solution.

[0117] Step 4: According to the results of step 3, based on the optimized load access configuration and real-time operation data, perform real-time monitoring and dynamic adjustment of load access configuration. Specifically, it includes the following sub-steps:

[0118] In step 3, the particle swarm algorithm and genetic algorithm have determined the power demand configuration of each load point at different time points as the preliminary optimization result. At this time, the system will continue to monitor the status of each load point in real time and obtain various real-time data related to the load point, including but not limited to power demand, load change, voltage, current, temperature, frequency and other operating parameters. The system obtains the latest status data of the load point in real time through high-precision sensors and data acquisition devices, and uploads it to the data processing module for analysis.

[0119] Based on real-time data collection, various sensor data are processed through data fusion technology to ensure data accuracy and consistency. For example, Kalman filtering or weighted averaging can be used to process possible noise and outliers. After data fusion, the system evaluates the operating status of each load point, taking into account the following factors: the deviation between load demand and actual power, load fluctuation amplitude, voltage and current stability, and the matching degree between equipment operating temperature and load.

[0120] By using machine learning algorithms (such as long short-term memory networks (LSTM)) to analyze and predict historical data trends, the system can predict the changing trend of load demand in the future time period based on the operating data of the current load point and historical load demand. This process not only takes into account the demand fluctuations of each load point, but also combines the global load distribution characteristics of the power grid. The prediction results will serve as decision inputs to help the system warn of possible future load fluctuations in advance and optimize load access configuration.

[0121] After acquiring real-time data and completing load demand forecasting, the system will adjust the load access configuration in real time based on feedback control mechanisms (such as PID control, model predictive control MPC). This process is not just a simple adjustment of the power of a certain load point, but also requires comprehensive adjustments based on the balance of the overall load. For example, if the actual power demand of a load point exceeds the predetermined value, the system will dynamically adjust according to the priority of the load point, which may involve power regulation of adjacent load points, scheduling of energy storage equipment, or auxiliary access to renewable energy.

[0122] Under the feedback control mechanism, the system will further optimize by combining the results of the particle swarm algorithm and the genetic algorithm. Specifically, based on the global optimal configuration obtained by the particle swarm algorithm optimization, the system first adjusts the power demand configuration of the load point to ensure global load balance; then, the genetic algorithm is used to make local adjustments to the particle swarm optimization results to further improve the power utilization efficiency and system stability of the local load point. The local optimization part of the genetic algorithm selects particles with higher fitness (representing solutions with better load configuration) for crossover and mutation operations, thereby generating new load configuration individuals, optimizing the load point configuration and reducing power loss.

[0123] The adjustment process also involves multiple rule engines. For example, if the temperature of a load point exceeds the set threshold, the power demand of the load point will be reduced first to ensure the safe operation of the equipment; if the grid voltage fluctuates greatly, voltage balance adjustment will be performed to avoid grid instability caused by unbalanced load access. Through these rule systems, combined with real-time data, adjustment plans are automatically generated to ensure that load access configuration maximizes resource utilization while ensuring system stability.

[0124] After adjusting the load access configuration, the system will continue to monitor the adjusted load points and evaluate the adjustment effect in real time. If the adjusted load configuration fails to achieve the expected optimization effect, the system will continue to make feedback adjustments until the load access configuration reaches the optimal state. The system will also periodically check the load access configuration and re-optimize the access configuration based on the new operating data to ensure long-term optimization effect.

[0125] After the adjustment is completed, the system will output the final optimal load access configuration and update the optimized results to the load scheduling module. This configuration includes the power demand of each load point in the future period and ensures the global load balance of the power grid and maximizes the system efficiency. The optimization results will be used for the calculation of the next round of load access configuration, forming a closed-loop optimization and adjustment process.

[0126] Compared with the simple regulation or passive response in the existing technology, this solution realizes the forward-looking optimization of load access configuration by combining real-time data with load demand forecast. This combination of forecast and adjustment enables the load access plan to be dynamically adjusted before uncertain factors occur, improving the intelligence and stability of the system.

[0127] The existing technology usually uses a single algorithm to optimize load configuration. Through the combination of particle swarm algorithm and genetic algorithm, a two-layer optimization structure is formed. The particle swarm algorithm is used to search for the optimal solution of load configuration globally, and the genetic algorithm makes local adjustments to the global optimal solution, making the optimization process more refined. While ensuring the global optimal solution, it can also make local adjustments according to actual conditions, solving the problem that traditional methods cannot take into account both global and local optimization.

[0128] The present invention introduces a variety of adaptive adjustment strategies, including PID control, model predictive control, etc. These control strategies can not only respond to the current load point status, but also self-adjust according to future trends and historical data. Through the intelligent decision-making mechanism, the system can actively adjust the load, reduce the problem of over-adjustment or lagging adjustment, and make the load access configuration more efficient and stable.

[0129] Real-time data collection and dynamic adjustment mechanism enable the system to optimize load access configuration in time according to the real-time operation status and load changes of the power grid, thus enhancing the stability and flexibility of the power grid. By continuously monitoring the power demand, voltage, temperature and other parameters of the load point, the system can dynamically adjust the load configuration to ensure the balance and efficient operation of the power grid under different operating conditions.

[0130] Example 2

[0131] Reference Figure 2 , which is an embodiment of the present invention, provides a multi-load access optimization system in high-reliability distribution network planning, including:

[0132] The demand analysis module 100 is used to analyze the demand of the load, divide the multi-level load into multiple levels based on power demand, usage characteristics and timing characteristics, and establish a coupling relationship model between the levels;

[0133] The feature extraction module 200 is used to collect the operation data of the load point in real time, process the data, and extract the load data;

[0134] The optimization solution module 300 is used to solve the coupling relationship model based on the load data, construct the objective function, and calculate the best load access configuration using an optimization algorithm.

[0135] Example 3

[0136] An embodiment of the present invention is different from the first two embodiments in that:

[0137] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0138] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0139] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0140] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0141] Embodiment 4

[0142] An embodiment of the present invention provides an optimization method for multi-source load access in high-reliability distribution network planning. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0143] In a certain smart grid laboratory, a simulation experiment on an optimization method for multi-source load access in high-reliability distribution network planning was carried out. The experimental environment simulated a medium-sized power grid system with 50 load points, which covered various categories such as industrial electricity, commercial electricity, and residential electricity. The experimental equipment included high-precision power meters, temperature and humidity sensors, data acquisition devices, and computing servers equipped with advanced processors. During the experiment, the power demand, usage characteristics, and temperature data of each load point were collected in real time, and the collection frequency was set to once per second. The experiment lasted for 60 minutes. The experiment was carried out under different load fluctuation scenarios, including peak hours, valley hours, and sudden load changes, to comprehensively evaluate the performance of the optimization method in actual operation.

[0144] The traditional load optimization method first manually analyzes the load demand, and divides the 50 load points into 3 levels according to the power demand, usage characteristics, and time series characteristics: industrial load, commercial load, and residential load. Then, empirical formulas are used to estimate the coupling relationship between load points, ignoring the dynamic interaction effects. In the data collection stage, only the real-time power demand, usage characteristics, and temperature data are recorded, without data cleaning, synchronization, or standardization processing. The optimization configuration calculation uses the linear programming method, and the goal is to minimize the total energy consumption and reduce the system load. Due to the complex load coupling relationship not being considered, each load point is independently adjusted during the optimization process, resulting in uneven load distribution and low energy utilization efficiency. Finally, the output load access configuration plan performs poorly in terms of energy efficiency and load balance.

[0145] In contrast, the method of the present invention uses an automated algorithm to divide the load points into multiple levels based on real-time collected data during the load demand analysis stage, and establishes a complex coupling relationship model. During the data processing process, data cleaning is performed to remove outliers and noise, fill in missing values, synchronize data from different sources according to timestamps, and perform standardization to ensure data quality. In the optimization calculation stage, a multidimensional objective function is constructed, and the power demand, load capacity and load coupling relationship are comprehensively considered. The particle swarm algorithm is used for global search to quickly locate the optimal configuration area; then the genetic algorithm is combined for local optimization, and the load configuration is finely adjusted to improve the optimization accuracy. The experimental results are shown in Table 1.

[0146] Table 1 Comparison of experimental results

[0147] Experimental protocol Traditional methods Method of the present invention Power demand fluctuation range (kW) 12.8 6.4 Maximum power demand error (%) 15.6 4.3 Energy efficiency improvement (%) 6.2 18.7 Load access stability (%) 70.5 96.8 Optimization time (seconds) 150 90

[0148] Referring to Table 1, the method of this embodiment significantly improves the accuracy and comprehensiveness of load access configuration by introducing automated load demand analysis and multi-dimensional coupling relationship models. This improvement enables the system to more effectively identify the complex interactions between load points, reduces the fluctuation range of power demand, enhances the balancing ability of the power grid load, and improves the stability of the overall system. The combination of particle swarm algorithm and genetic algorithm not only speeds up the positioning speed of the optimal solution, but also significantly reduces the power demand error through fine adjustment, avoiding the overload and energy waste problems caused by ignoring the load coupling relationship in traditional methods. The design of multi-dimensional objective function further improves the energy utilization efficiency, ensuring that the system can maximize energy benefits and reduce energy consumption during the optimization process. At the same time, the shortening of optimization time shows the advantage of the method of the present invention in computing efficiency, which can respond to real-time load changes more quickly and improve the real-time adjustment capability of the system.

[0149] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A multi-load access optimization method in high reliability distribution network planning, characterized in that: include: Conduct demand analysis on loads, divide multiple loads into multiple levels based on power demand, usage characteristics and timing characteristics, and establish a coupling relationship model between levels; Collect the operation data of the load point in real time, process the data and extract the load data; The coupling relationship model is solved based on load data, the objective function is constructed, and the optimal load access configuration is calculated using the optimization algorithm.

2. The method for optimizing multi-load access in high-reliability distribution network planning according to claim 1, characterized in that: The coupling relationship model is expressed as: Among them, C ij (t) represents the coupling relationship between load point i and load point j, which changes with time t, σ(·) represents the activation function, The summation symbol indicates the accumulation of different coupling characteristics. k Coupling gain coefficient, the weight of the influence of the kth load characteristic on the coupling relationship, is the exponential decay term of the power demand difference, which indicates the influence of the load point power difference on the coupling relationship. α indicates the sensitivity of the influence of the control power demand difference on the coupling relationship. β(U i -U j ) 2 is the quadratic term of the difference in usage characteristics, which indicates the influence of the difference in usage characteristics of the load point on the coupling relationship. β indicates the influence of the difference in usage characteristics on the coupling relationship. γlog(|T i (t)-T j (t)|+1) is the temperature difference logarithm, which indicates the influence of load point temperature difference on the coupling relationship, γ indicates the influence intensity of control temperature difference on the coupling relationship, |P i (t)-P j (t)| represents the power demand difference between load point i and load point j, P i (t) and P j (t) represent the power demand of load point i and load point j at time t, respectively, |U i -U j | represents the difference in usage characteristics between load point i and load point j, U i and U j Respectively represent the usage characteristics of load point i and load point j, |T i (t)-T j (t)| represents the real-time temperature difference between load point i and load point j, T i (t) and T j (t) represent the temperature of load point i and load point j at time t respectively.

3. The method for optimizing multi-load access in high-reliability distribution network planning according to claim 2, characterized in that: The operation data includes power demand parameters, usage characteristic parameters and temperature parameters.

4. The method for optimizing multi-load access in high-reliability distribution network planning according to claim 3, characterized in that: The data processing includes performing data cleaning operations, removing outliers and noise data, and filling missing values; synchronizing data from different sources according to timestamps; and standardizing data.

5. The method for optimizing multi-load access in high-reliability distribution network planning according to claim 4, characterized in that: The load data includes the power demand difference |P i (t)-P j (t) |, usage characteristics difference (U i -U j ) 2 and the temperature difference log(|T i (t)-T j (t)|+1).

6. The method for optimizing multi-load access in high-reliability distribution network planning according to claim 5, characterized in that: The objective function is expressed as, Among them, P i (t) represents the power demand of load point i at time t, represents the maximum capacity of load point i, R i represents the load capacity of load point i, w1, w2, w3 represent weight coefficients respectively.

7. The method for optimizing multi-load access in high-reliability distribution network planning according to claim 6, characterized in that: The calculating of the optimal load access configuration includes initializing a particle swarm, each particle representing a load access configuration, randomly assigning a position and a speed to each particle to represent the power demand configuration of different load points, and calculating the objective function value of each particle; Based on the particle swarm algorithm, a global search is performed. The particle updates its position in the solution space according to the objective function value, the particle's historical optimal position, and the global optimal position. The particle position is adjusted by the speed update formula, and the optimal solution is searched through multiple iterations. Based on the global search of particle swarm, a genetic algorithm is used to locally optimize the global optimal solution of the particle swarm. By selecting particles with higher fitness as parents, crossover and mutation operations are performed to generate new individuals and optimize the load access configuration. According to the objective function value of the new individual optimized by the genetic algorithm, the individual optimal solution and the global optimal solution of the particle swarm are updated. If the objective function value of the offspring individual is better than the current optimal solution, the corresponding optimal solution is updated; When the particle swarm algorithm and the genetic algorithm run alternately to the set maximum number of iterations, or the objective function value reaches the preset convergence standard, the algorithm execution is terminated and the global optimal solution is output as the optimal load access configuration; Output the final optimal load access configuration, including the power demand of each load point at different time points, to meet the optimization objectives, maximize energy benefits and optimize the load access process.

8. A system using the multi-load access optimization method in high-reliability distribution network planning as claimed in any one of claims 1 to 7, characterized in that: include: A demand analysis module (100) is used to perform demand analysis on loads, divide multiple loads into multiple levels based on power demand, usage characteristics and timing characteristics, and establish a coupling relationship model between the levels; A feature extraction module (200) is used to collect the operation data of the load point in real time, process the data, and extract the load data; The optimization solution module (300) is used to solve the coupling relationship model based on the load data, construct the objective function, and use the optimization algorithm to calculate the best load access configuration.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for optimizing multi-load access in high-reliability distribution network planning described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for optimizing multi-load access in high-reliability distribution network planning according to any one of claims 1 to 7 are implemented.

Images