Power distribution network resource adaptive scheduling method, system, device and medium

Abstract

The invention relates to the technical field of power distribution networks, in particular to a power distribution network resource self-adaptive scheduling method, system and device and a medium. Using a support vector machine algorithm to construct an agent model representing a nonlinear coupling relationship between the novel power factor real-time parameters and the initial data set of the power distribution network; decomposing a multivariable interaction effect in the agent model by using a global sensitivity analysis method to obtain a sensitivity index; adaptively correcting the sampling space of the agent model according to the sensitivity index to obtain an optimized agent model; and on the basis of the optimization agent model and the real-time monitoring data of the power distribution network, generating a power distribution network resource adaptive scheduling strategy by adopting a particle swarm optimization algorithm, thereby carrying out optimal configuration on the power distribution network resources in the novel power factor access scene. According to the method, efficient and accurate optimal configuration of the power distribution network resources in a novel power element access scene is realized, and the operation performance and stability of the power distribution network are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution networks, and particularly to a method, system, device, and medium for adaptive scheduling of distribution network resources. Background Art

[0002] With the rapid development of the new power system, the operating environment of the distribution network has changed significantly. The wide access of new power elements such as distributed power sources, electric vehicles, and new energy storage in the distribution network has transformed the operating characteristics of the distribution network from the traditional single-way power supply mode into a complex interactive power system with multi-source collaboration. This transformation has greatly enriched the operating scenarios of the distribution network, but at the same time has posed severe challenges to the real-time regulation ability of the power grid, especially in scenarios such as node load balancing, new energy consumption capacity, and rapid fault response.

[0003] However, most of the existing technologies rely on static linear models or single-variable analysis methods, and it is difficult to accurately capture the non-linear coupling relationship between new power elements such as distributed power sources, electric vehicles, and new energy storage and the multi-dimensional performance of the distribution network. This limitation makes it difficult to provide accurate and reliable decision-making support in the face of complex and changing operating scenarios, thus affecting the stable operation and efficient management of the power grid. In addition, the existing scheduling systems lack the support of a multi-dimensional sensitivity analysis framework and it is difficult to establish a mapping relationship between the degree of element influence and the scheduling decision. Such defects make it difficult for the existing scheduling systems to achieve adaptive optimization in a dynamic environment in the face of the high uncertainty brought by new power elements, restricting the improvement of the overall operating efficiency and reliability of the distribution network.

[0004] In summary, the existing technologies are difficult to provide reliable scheduling support in complex and changing operating scenarios. Especially under the multi-dimensional influence of new power elements, the real-time performance and accuracy of the scheduling system are severely restricted. Therefore, there is an urgent need to provide a method for adaptive scheduling of distribution network resources to construct a quantitative analysis system that can analyze the non-linear coupling mechanism of multi-elements and provide refined decision-making basis for intelligent scheduling. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides a method, system, device, and medium for adaptive scheduling of distribution network resources.

[0006] In a first aspect, the present invention provides a method for adaptive scheduling of distribution network resources, the method comprising the following steps: Collect the operating state data of the distribution network, and preprocess the operating state data of the distribution network to obtain an initial data set of the distribution network; According to the initial data set of the distribution network and the real-time parameters of new power elements, use the support vector machine algorithm to construct a surrogate model representing the non-linear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network; Decompose the multivariate interaction effects in the surrogate model using the global sensitivity analysis method, and quantitatively obtain the sensitivity indicators of each new power element to the initial dataset of the distribution network; According to the sensitivity indicators, adaptively correct the sampling space of the surrogate model with an iterative update strategy to obtain an optimized surrogate model; Based on the optimized surrogate model and the real-time monitoring data of the distribution network, use the particle swarm optimization algorithm to generate an adaptive scheduling strategy for the distribution network resources; Optimize the allocation of the distribution network resources in the scenario of the access of new power elements according to the adaptive scheduling strategy of the distribution network resources.

[0007] In a further implementation, the step of constructing a surrogate model that characterizes the non-linear coupling relationship between the real-time parameters of new power elements and the initial dataset of the distribution network using the support vector machine algorithm according to the initial dataset of the distribution network and the real-time parameters of new power elements includes: Align the initial dataset of the distribution network and the real-time parameters of new power elements according to the time stamp to form an initial sampling space dataset; According to the initial sampling space dataset, identify the key distribution network features affecting the real-time parameters of new power elements through the support vector machine algorithm, and generate a low-dimensional sample point data set; Use the radial basis kernel function to map and transform the low-dimensional sample point data set into a high-dimensional feature space to obtain high-dimensional feature vectors; In the high-dimensional feature vectors, fit the non-linear coupling relationship between the real-time parameters of new power elements and the initial dataset of the distribution network through the support vector regression function to obtain an initial regression model in the high-dimensional feature space; Taking the minimization of the prediction error as the optimization objective, solve the initial regression model through the sequential minimal optimization algorithm to obtain the optimal parameters of the regression model; Construct a surrogate model that characterizes the non-linear coupling relationship between the real-time parameters of new power elements and the initial dataset of the distribution network according to the optimal parameters of the regression model.

[0008] In a further implementation, the step of decomposing the multivariate interaction effects in the surrogate model using the global sensitivity analysis method and quantitatively obtaining the sensitivity indicators of each new power element to the initial dataset of the distribution network includes: Based on the global sensitivity analysis method, decompose the new power elements in the surrogate model to obtain the first-order sensitivity components caused by the independent action of a single new power element, the second-order sensitivity components generated by the superposition of any two new power elements, and the high-order sensitivity components generated by the coupling of at least three new power elements; Add the first-order sensitivity component, the second-order sensitivity component, and the high-order sensitivity component of each new power element to obtain the total effect sensitivity of the corresponding new power element to the fluctuation of the distribution network operation state; Calculate the variance contribution ratio of the first-order sensitivity component of each new power element to the total effect sensitivity to obtain the main effect contribution rate of a single new power element; Normalize the main effect contribution rate to form the sensitivity index of each new power element to the initial data set of the distribution network.

[0009] In a further implementation, the surrogate model includes a benchmark probability term for the normal operation state of the distribution network under the condition of no access perturbation of any new power element, a first-order independent influence term representing the independent influence of a single new power element on the distribution network state classification probability, a second-order superposition effect term representing the influence of the superposition effect between any two new power elements on the distribution network state classification probability, and a high-order interaction term representing the influence of the coupling effect of at least three new power elements on the distribution network state classification probability.

[0010] In a further implementation, the step of adaptively correcting the sampling space of the surrogate model according to the sensitivity index by an iterative update strategy to obtain an optimized surrogate model includes: Define the feasible region of the new power element according to the physical constraint conditions of the real-time parameters of the new power element, and use the Latin hypercube sampling method to uniformly select an initial sample point set within the feasible region of the new power element; Input the initial sample point set into the surrogate model for iterative optimization, and extract the current optimal solution from the output result of the surrogate model in each round of iterative optimization; Take the current optimal solution as the center and determine the core sampling interval of the current round of iteration according to a preset contraction ratio; Based on the sensitivity index of each new power element to the initial data set of the distribution network, screen out the high-sensitivity full-scale sampling interval and the low-sensitivity sampling interval within the core sampling interval of the current round of iteration; Compress the low-sensitivity sampling interval according to the exponential decay law to form a low-sensitivity compressed sampling interval; Expand the high-sensitivity full-scale sampling interval with the current optimal solution as the center to obtain a high-sensitivity expanded interval, and take the union of the high-sensitivity expanded interval and the low-sensitivity compressed sampling interval to form a non-uniform dynamic sampling space; Calculate the prediction error of the surrogate model within the non-uniform dynamic sampling space, and when the prediction error of the surrogate model reaches a preset error threshold, optimize and update the surrogate model according to the non-uniform dynamic sampling space to obtain an optimized surrogate model.

[0011] In a further embodiment, the steps of generating an adaptive scheduling strategy for distribution network resources based on the optimized surrogate model and real-time monitoring data of the distribution network include: Based on the sensitivity index corresponding to the optimized surrogate model, a high-dimensional decision space is constructed within the feasible region of new power elements; Several particles are randomly generated within the high-dimensional decision space, and the learning factors of the particles are initialized according to the movement rate and current position of each particle within the high-dimensional decision space; Based on the learning factors of the particles, the particle swarm optimization algorithm is used to update the spatial positions of each particle within the high-dimensional decision space; In each iteration process, the potential solution spacing index is calculated based on the average value of the spatial position differences of the particles in the real-time parameter dimensions of each new power element, and the learning factors of the particles are adaptively adjusted according to the potential solution spacing index to obtain the latest learning factors; Calculate the relative error between the predicted value output by the optimized surrogate model and the real-time monitoring data of the distribution network, and monitor the global optimal position change rate in multiple consecutive iterations; When the relative error is less than the preset relative error threshold and the global optimal position change rate in multiple consecutive iterations is less than the preset position change threshold, stop the iterative optimization process and output the current global optimal parameter solution; The radial basis kernel function is used to map the global optimal parameter solution to a high-dimensional space for nonlinear transformation to obtain an adaptive scheduling strategy for distribution network resources.

[0012] In a further embodiment, the learning factor is used to control the learning step size and direction of the particles during the search process.

[0013] In a second aspect, the present invention provides an adaptive scheduling system for distribution network resources, and the system includes: A data acquisition module, configured to collect the operating state data of the distribution network and preprocess the operating state data of the distribution network to obtain an initial data set of the distribution network; A model construction module, configured to use the support vector machine algorithm to construct a surrogate model characterizing the nonlinear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network according to the initial data set of the distribution network and the real-time parameters of new power elements; A variable decomposition module, configured to decompose the multivariable interaction effect in the surrogate model by using the global sensitivity analysis method and quantitatively obtain the sensitivity indexes of each new power element to the initial data set of the distribution network; A space optimization module, configured to adaptively correct the sampling space of the surrogate model according to the sensitivity indexes by using an iterative update strategy to obtain an optimized surrogate model; A strategy generation module, configured to generate an adaptive scheduling strategy for distribution network resources based on the optimized surrogate model and real-time monitoring data of the distribution network by using a particle swarm optimization algorithm; An optimization configuration module, configured to optimize the configuration of distribution network resources in a new power element access scenario according to the adaptive scheduling strategy for distribution network resources.

[0014] In a third aspect, the present invention further provides a computer device, including a processor and a memory. The processor is connected to the memory. The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the computer device executes the steps of implementing the above method.

[0015] In a fourth aspect, the present invention further provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of implementing the above method are realized.

[0016] The present invention provides an adaptive scheduling method, system, device and medium for distribution network resources. The method collects operation status data of the distribution network and preprocesses the operation status data of the distribution network to obtain an initial data set of the distribution network; according to the initial data set of the distribution network and real-time parameters of new power elements, a surrogate model characterizing the non-linear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network is constructed by using a support vector machine algorithm; the multi-variable interaction effect in the surrogate model is decomposed by using a global sensitivity analysis method, and the sensitivity index of each new power element to the initial data set of the distribution network is quantitatively obtained; according to the sensitivity index, the sampling space of the surrogate model is adaptively corrected with an iterative update strategy to obtain an optimized surrogate model; based on the optimized surrogate model and real-time monitoring data of the distribution network, an adaptive scheduling strategy for distribution network resources is generated by using a particle swarm optimization algorithm; the distribution network resources in a new power element access scenario are optimized according to the adaptive scheduling strategy for distribution network resources. Compared with the prior art, this method analyzes the non-linear coupling relationship between new power elements and the distribution network through technologies such as sensitivity analysis and particle swarm optimization algorithm, realizes the efficient, accurate and adaptive optimization configuration of distribution network resources in a new power element access scenario, and significantly improves the operation performance and stability of the distribution network. Description of the Drawings

[0017] Figure 1 is a schematic flowchart of an adaptive scheduling method for distribution network resources provided by an embodiment of the present invention; Figure 2 is a comparison diagram of prediction results of a support vector machine algorithm training set provided by an embodiment of the present invention; Figure 3 is a comparison diagram of prediction results of a support vector machine algorithm test set provided by an embodiment of the present invention; Figure 4 It is a block diagram of a distribution network resource adaptive scheduling system provided by an embodiment of the present invention; Figure 5 It is a schematic structural diagram of a computer device provided by an embodiment of the present invention.

[0018] Explanation of reference numerals: 101, data acquisition module; 102, model construction module; 103, variable decomposition module; 104, space optimization module; 105, policy generation module; 106, optimization configuration module. Specific embodiments

[0019] The embodiments of the present invention will be specifically described below in conjunction with the accompanying drawings. The given embodiments are only for illustrative purposes and should not be construed as limiting the present invention. The accompanying drawings are for reference and illustration only and do not constitute a limitation to the protection scope of the present invention's patent, because many changes can be made to the present invention without departing from the spirit and scope of the present invention.

[0020] Reference Figure 1 , an embodiment of the present invention provides a distribution network resource adaptive scheduling method, as Figure 1 shown, the method includes the following steps: S1. Collect the operation status data of the distribution network, and preprocess the operation status data of the distribution network to obtain an initial data set of the distribution network.

[0021] In this embodiment, the acquisition period is determined according to the actual operation conditions of the distribution network and the data analysis requirements, and multi-dimensional distribution network operation status data is collected in real time from the distribution network monitoring system during the acquisition period. The distribution network operation status data includes distribution network operation indicators, distribution network carrying capacity indicators, and distribution network power supply support capacity indicators. Among them, the distribution network operation indicators may include substation load rate, line load rate, and line loss rate. Among them, the substation load rate represents the ratio of the actual power supply of the substation to the rated power supply within a preset acquisition period, which is used to measure the operation efficiency and load utilization of the substation; the line load rate represents the ratio of the actual power of the line to the rated power, which is used to measure the utilization rate and operation status of the line; the line loss rate represents the proportion of the grid transmission loss in the total power supply. The distribution network carrying capacity indicators may include the new energy power generation accommodation rate. The new energy power generation accommodation rate represents the proportion of the actual accommodated power of new energy in the total power generation, which is an important indicator for measuring the effective utilization degree of new energy power and reflects the accommodation capacity of new energy power in the power grid. The distribution network power supply support capacity indicators may include physical parameters such as the amount of fault load loss and power supply interruption time. The amount of fault load loss represents the total load that is disconnected due to a fault, and the power supply interruption time represents the cumulative duration required for fault recovery. In this embodiment, preprocessing operations are performed on the collected distribution network operation status data to obtain an initial distribution network dataset. Among them, the preprocessing operations include data cleaning and data standardization processing. Data cleaning includes checking whether there are missing values in the dataset and deleting duplicate values in the dataset.

[0022] At the same time, this embodiment obtains the real-time operation parameters of new power elements to form real-time parameters of new power elements. The real-time parameters of new power elements can include multiple dimensions such as distributed power generation output, electric vehicle charging load, energy storage charge and discharge power, and microgrid operation mode. Among them, distributed power generation can reduce the substation and line load rates during the normal operation of the distribution network, effectively reducing line losses; in terms of carrying capacity, the access of distributed power generation may reduce the new energy accommodation rate, and even exacerbate the phenomenon of abandoned wind and light in extreme cases; in terms of power supply support capacity, distributed power generation can reduce the amount of fault load loss and shorten the power outage time during a fault, enhancing the overall flexibility of the system.

[0023] The disordered charging of electric vehicle charging loads during the normal operation of the distribution network can increase the load rate and line losses; in terms of carrying capacity, if the electric vehicle charging load does not match the new energy power generation time sequence, it will weaken the new energy accommodation and increase the peak shaving pressure on the traditional power grid at the same time; in terms of power supply support capacity, the vehicle-to-grid interaction technology of electric vehicles helps to reduce fault losses and improve the response speed.

[0024] The energy storage charging and discharging power can effectively reduce the load rates of substations and lines through peak shaving and valley filling strategies during the normal operation of the distribution network, and reduce the line losses caused by tidal current fluctuations. In terms of the carrying capacity, the new energy storage can improve the penetration rate of distributed power sources and the consumption rate of new energy. In terms of the power supply support capacity, its fast response characteristics can significantly reduce the load loss and power outage time during faults, and significantly enhance the anti-disturbance ability of the distribution network.

[0025] The microgrid operation mode can relieve the load pressure on the superior substations and lines and reduce the network losses through the optimized dispatching of internal power sources during the normal operation of the distribution network. In terms of the carrying capacity, the independent operation characteristics of the microgrid may weaken the main grid's control ability over new power elements. Especially when multiple microgrids send power to the main grid simultaneously, it may cause the penetration rate in local areas to exceed the standard, leading to voltage fluctuations or misoperation of protection. In terms of the power supply support capacity, the island operation ability of the microgrid can minimize the scope of fault impact and significantly shorten the power supply restoration time.

[0026] S2. According to the initial distribution network dataset and the real-time parameters of new power elements, use the support vector machine algorithm to construct a surrogate model that characterizes the non-linear coupling relationship between the real-time parameters of new power elements and the initial distribution network dataset.

[0027] In some embodiments, the step of constructing a surrogate model that characterizes the non-linear coupling relationship between the real-time parameters of new power elements and the initial distribution network dataset according to the initial distribution network dataset and the real-time parameters of new power elements using the support vector machine algorithm includes: Align the initial distribution network dataset and the real-time parameters of new power elements according to the time stamp to form an initial sampling space dataset; According to the initial sampling space dataset, use the support vector machine algorithm to identify the key distribution network features that affect the real-time parameters of new power elements, and generate a low-dimensional sample point data set; Use the radial basis kernel function to map and transform the low-dimensional sample point data set to a high-dimensional feature space to obtain high-dimensional feature vectors; In the high-dimensional feature vectors, fit the non-linear coupling relationship between the real-time parameters of new power elements and the initial distribution network dataset through the support vector regression function to obtain an initial regression model in the high-dimensional feature space; Taking the minimization of the prediction error as the optimization goal, solve the initial regression model through the sequential minimal optimization algorithm to obtain the optimal parameters of the regression model; Construct a surrogate model that characterizes the non-linear coupling relationship between the real-time parameters of new power elements and the initial distribution network dataset according to the optimal parameters of the regression model.

[0028] Specifically, in this embodiment, the initial distribution network data set and the real-time parameters of new power elements are aligned according to the time stamp, and the aligned data are combined into an initial sampling space data set containing synchronous time points to ensure the consistency of the data in the time dimension. Among them, for the missing time points, this embodiment can perform interpolation processing or ignore them according to the actual situation. After alignment, the initial sampling space data set contains the data of the initial distribution network data set and the real-time parameters of new power elements at the same time point. Then, this embodiment uses the recursive feature elimination method of the support vector machine to analyze the correlation between the initial distribution network data set and the real-time parameters of new power elements, and selects the key distribution network features with a correlation greater than the preset correlation threshold, so as to identify the key distribution network features that have a significant impact on the real-time parameters of new power elements. These key distribution network features can include the substation load rate, line load rate, line loss rate, etc. This embodiment uses the selected key distribution network features as the input of the low-dimensional sample points and the real-time parameters of new power elements as the output to generate a low-dimensional sample point data set.

[0029] Next, this embodiment uses the radial basis kernel function as the mapping function. The radial basis kernel function can map the low-dimensional sample point data to a high-dimensional feature space, solve the problem of linear inseparability in the low-dimensional space, and thus can better capture the complex non-linear relationship between the new power elements and the distribution network. Specifically, this embodiment maps each sample point in the low-dimensional sample point data set to a high-dimensional feature space through the radial basis kernel function to obtain high-dimensional feature vectors. These high-dimensional feature vectors contain the feature representations of the original low-dimensional data in the high-dimensional space and can more comprehensively reflect the relationship between the new power elements and the distribution network. In this embodiment, in the high-dimensional feature space, the support vector regression function is used to fit the non-linear coupling relationship between the real-time parameters of new power elements and the initial distribution network data set. Among them, the support vector regression function can be expressed as: In the formula, y is the support vector regression function, which is used to fit the non-linear coupling relationship between the real-time parameters of new power elements and the multi-dimensional state of the distribution network; is the predicted output value representing the support vector regression function; is the weight vector; is the non-linear mapping function from the low-dimensional space to the high-dimensional space; b is the bias term.

[0030] By fitting the support vector regression function, this embodiment can find the hyperplane in the high-dimensional feature space that can best describe the relationship between the new power elements and the distribution network. At the same time, in order to determine the weight vector and the bias term, this embodiment minimizes the prediction error and the model complexity (by the norm of the weight vector w Taking the balance between the complexity and fitting accuracy of the model as the optimization goal, while introducing a penalty factor and a relaxation factor to control the complexity of the model and the degree of fitting to the training data, the objective function of the optimization problem is as follows: In the formula, Q is the objective function; is the weight vector; C is the penalty factor, which is used to balance the model complexity and the training error; is the lower bound relaxation factor for dealing with the negative direction; is the upper bound relaxation factor for dealing with the positive direction, which allows the sample points to exceed the error tolerance range of the insensitive zone; is the squared norm of the weight vector. In the objective function, by minimizing to control the complexity of the model, prevent overfitting, and reflect the overall strength of the model's sensitivity to features; m is the number of training samples.

[0031] The constraint conditions of the optimization problem are: In the formula, is the input feature vector of the i-th sample; is the high-dimensional feature vector after mapping the sample ; b is the bias term; is the true output value of the i-th sample; is the insensitive zone of the loss function, which defines the allowable boundary of the prediction error, that is, the prediction error within the range is considered acceptable and not penalized.

[0032] In this embodiment, the sequential minimal optimization (SMO) algorithm is used to iteratively solve the constructed optimization problem. The SMO algorithm is a quadratic programming solution algorithm that can select two variables for optimization successively, thereby gradually approaching the global optimal solution. In each iteration, the SMO algorithm selects two Lagrange multipliers for optimization until the convergence condition is satisfied. After solving, the optimal weight vector and the optimal bias term b of the regression model are obtained. In this embodiment, the optimal parameters of the regression model obtained by solving are substituted into the support vector regression function to obtain the final surrogate model. This surrogate model can characterize the nonlinear coupling relationship between the real-time parameters of new power elements and the initial dataset of the distribution network. Based on the initial dataset of the distribution network and the real-time parameters of new power elements, it predicts the multi-dimensional index changes of the distribution network, providing a basis for subsequent sensitivity analysis and decision-making.

[0033] In this embodiment, the proxy model is used to quantify the non - linear coupling relationship between new power elements and the operation state of the distribution network. The mathematical terms of the proxy model include a benchmark probability term for the normal operation state of the distribution network under the condition of no access disturbance of any new power elements, a first - order independent influence term representing the independent influence of a single new power element (such as distributed power generation output or energy storage charge - discharge power) on the distribution network state classification probability, a second - order superposition effect term representing the influence of the superposition effect between any two new power elements (such as electric vehicle charging load and micro - grid operation mode) on the distribution network state classification probability, and a high - order interaction term representing the influence of the coupling effect of at least three new power elements (such as distributed power source, energy storage, and electric vehicle) on the distribution network state classification probability. Through the superposition of the benchmark probability term and multi - order effect terms in this embodiment, the dynamic influence of new power elements acting independently and synergistically on the operation state of the distribution network can be comprehensively characterized.

[0034] S3. Use the global sensitivity analysis method to decompose the multi - variable interaction effect in the proxy model, and quantify the sensitivity index of each new power element to the initial data set of the distribution network.

[0035] In some embodiments, the steps of using the global sensitivity analysis method to decompose the multi - variable interaction effect in the proxy model and quantify the sensitivity index of each new power element to the initial data set of the distribution network include: Based on the global sensitivity analysis method, decompose the new power elements in the proxy model for variance decomposition to obtain a first - order sensitivity component caused by the independent action of a single new power element, a second - order sensitivity component generated by the superposition action of any two new power elements, and a high - order sensitivity component generated by the coupling effect of at least three new power elements; Add the first - order sensitivity component, the second - order sensitivity component, and the high - order sensitivity component of each new power element to obtain the total effect sensitivity of the corresponding new power element to the fluctuation of the distribution network operation state; Calculate the variance contribution ratio of the first - order sensitivity component of each new power element to the total effect sensitivity to obtain the main effect contribution rate of a single new power element; Normalize the main effect contribution rate to form the sensitivity index of each new power element to the initial data set of the distribution network.

[0036] Specifically, to quantify the influence degree of new power elements on the multi - dimensional capabilities of the distribution network, in this embodiment, based on the non - linear fitting relationship represented by the proxy model, the Sobol global sensitivity analysis method is used to decompose the multi - dimensional interaction effect of new power elements on the distribution network state, and the non - linear relationship output by the proxy model is decomposed into a benchmark probability term, an independent influence term, and an interaction effect term, which is specifically expressed in the following form: In the formula, is the probability model of the non - linear fitting relationship between the new power elements and the multi - dimensions of the distribution network, that is, it quantifies the variance contribution of the multi - dimensional state of the distribution network affected by the new power elements; is the baseline probability term, which represents the probability of the normal operation of the distribution network when no new power elements are connected; is the first - order independent influence term, which characterizes a single new power element (such as the output of distributed power sources) on the independent influence component of the classification probability of the distribution network state; is the second - order superposition effect term, which characterizes the influence component of the superposition effect between any two new power elements and on the classification probability of the distribution network state; n is the number of new power elements; is the high - order interaction term, which characterizes the influence component of the superposition effect among at least three new power elements on the classification probability of the distribution network state; is the i - th new power element; is the j - th new power element; is the n - th new power element.

[0037] In this embodiment, the total variance is decomposed into the contributions of each order term, and the decomposition form of the total variance is expressed as: In the formula, is the total variance, which represents the total contribution degree of all new power elements acting alone and in superposition to the multi - dimensional state fluctuation of the distribution network; is the integral of the surrogate model output, that is, the total probability contribution of all new power elements acting jointly to the state of the distribution network; is the square value of the baseline probability, which reflects the uncertainty or fluctuation degree of the multi - dimensional state of the distribution network itself without the influence of new power elements; is the n - dimensional real - number space (the value range of the real - time parameters of all new power elements); dX is the differential element.

[0038] Among them, the decomposition terms under the single or superposition items are calculated by multi - dimensional integration. In this embodiment, by summing the integrals of each individual and superposition effect, the contribution part of each new power element and their interaction to the total variance is obtained. Assuming in the range, the decomposition terms under the single or superposition items are specifically expressed as: In the formula, is the multidimensional joint variance term, i.e., the variance contribution of all the new power elements considered to the distribution network status; is the real-time parameter set of all new power elements considered, Each is any element in the real-time parameter set of all new power elements; It is the probability distribution of distribution network status under the coupling of multiple new power elements; are the upper and lower limits of the integration of the nth new power factor variable.

[0039] Therefore, the total variance decomposition formula and the decomposition formula under separate or superimposed terms can be further decomposed into: In the formula, is the first-order sensitivity component, representing a single new power element The degree of influence of independent actions on the multi-dimensional capabilities of the distribution network; is the second-order sensitivity component, representing and The degree of influence of the superposition of the two factors on the multi-dimensional capacity of the distribution network; It is a high-order sensitivity component, which characterizes the influence of multi-factor coupling on the multi-dimensional capacity of the distribution network.

[0040] Therefore, this embodiment decomposes the output variance of the proxy model into first-order sensitivity components, second-order sensitivity components and high-order sensitivity components through the above steps, wherein the first-order sensitivity component represents the variance increment caused by the independent action of a single new power element (such as distributed power output, energy storage charging and discharging power); the second-order sensitivity component represents the additional variance increment caused by the synergistic action of any two new power elements (such as electric vehicle charging load and new energy consumption rate); the high-order sensitivity component represents the complex variance increment caused by the joint action of three or more new power elements, thereby obtaining a set of independent and interactive sensitivity components of each new power element. Then, based on the variance decomposition result, this embodiment quantifies the contribution rate of each new power element to the state fluctuation of the distribution network. Specifically, For each new power element, this embodiment adds its first-order sensitivity component and all second-order and high-order sensitivity components involving the element to obtain the total effect sensitivity of the element to the fluctuation of the distribution network operation state, which reflects the sum of the influence of the new power element on the distribution network under all possible individual and interactive conditions, thereby obtaining the total effect sensitivity set of each new power element. Next, this embodiment divides the first-order sensitivity component of a single new power element by its total effect sensitivity to obtain the main effect contribution rate of the single new power element. The main effect contribution rate reflects the contribution ratio of the independent action of a single new power element to the total effect, which represents the contribution ratio of the new power element to the multi-dimensional state fluctuation of the distribution network when it acts alone. The specific mathematical expression is: wherein, is the contribution rate of the main effect of a single new power element acting independently to the total variance.

[0041] In this embodiment, all the contribution rates of the main effects are normalized so that their values are within the range of [0, 1] to eliminate the dimension difference, and the normalized contribution rate of the main effect is defined as the sensitivity index of each new power element to the operation state of the distribution network.

[0042] S4. According to the sensitivity index, adaptively correct the sampling space of the surrogate model with an iterative update strategy to obtain an optimized surrogate model.

[0043] In some embodiments, the step of adaptively correcting the sampling space of the surrogate model with an iterative update strategy according to the sensitivity index to obtain an optimized surrogate model includes: Define the feasible region of the new power element according to the physical constraint conditions of the real-time parameters of the new power element, and use the Latin hypercube sampling method to uniformly select an initial sample point set within the feasible region of the new power element; Input the initial sample point set into the surrogate model for iterative optimization, and extract the current optimal solution from the output result of the surrogate model during each round of iterative optimization; Taking the current optimal solution as the center, determine the core sampling interval of the current round of iteration according to a preset contraction ratio; Based on the sensitivity index of each new power element to the initial data set of the distribution network, screen out the high-sensitivity full-sampling interval and the low-sensitivity sampling interval within the core sampling interval of the current round of iteration; Compress the sampling interval of the low-sensitivity sampling interval according to the exponential decay law to form a low-sensitivity compressed sampling interval; Expand the high-sensitivity full-sampling interval with the current optimal solution as the center to obtain a high-sensitivity expanded interval, and take the union of the high-sensitivity expanded interval and the low-sensitivity compressed sampling interval to form a non-uniform dynamic sampling space; Calculate the prediction error of the surrogate model within the non-uniform dynamic sampling space, and when the prediction error of the surrogate model reaches a preset error threshold, optimize and update the surrogate model according to the non-uniform dynamic sampling space to obtain an optimized surrogate model.

[0044] Specifically, in this embodiment, a multi-dimensional feasible region of real-time parameters of new power elements is defined according to the physical constraint conditions of each new power element. For example, the physical constraint conditions of real-time parameters of new power elements may include the range boundaries of parameters such as the output range of distributed power sources, the power limit of energy storage charging and discharging, and the charging load threshold of electric vehicles. Then, the Latin hypercube sampling method is used to uniformly select an initial sample point set within the feasible region of new power elements. This method can ensure that the sample points are evenly distributed within the feasible region and improve the representativeness of the initial sample points. The initial sample point set evenly covers the typical operation scenarios of each new power element. For example, the initial sample point set can cover operation scenarios such as full charge of energy storage and centralized charging of electric vehicles. In this embodiment, the sampling space is initialized, and enough sample points are selected from the initial sample point set to fit the non-linear relationship between new power elements and the multi-dimensional capabilities of the distribution network, and the initial sample point set is input into the surrogate model for iterative optimization to fit the non-linear relationship between new power elements and the operation state of the distribution network. In each round of iterative optimization process, the current optimal solution (such as the parameter combination that minimizes the line loss rate) is extracted from the output result of the surrogate model as the benchmark for subsequent sampling space adjustment. For example, in the (k - 1)-th round of iteration, the current optimal solution is obtained. .

[0045] Next, in this embodiment, with the optimal solution of the (k - 1)-th round of iteration of the surrogate model as the center, it shrinks towards the upper and lower limits according to a preset ratio (such as 10% of the sampling space size of the previous round) to generate the core sampling interval of the current round, which can be expressed as: In the formula, is the lower limit of the core sampling interval determined by the current optimal solution of the surrogate model in the k-th iteration; is the upper limit of the core sampling interval determined by the current optimal solution of the surrogate model in the k-th iteration; is the current optimal solution in the (k - 1)-th iteration; is the sampling space size in the (k - 1)-th iteration.

[0046] In this embodiment, within the core sampling interval, the highly sensitive full-sampling interval with a sensitivity index higher than the preset sensitivity threshold is expanded centered on the current optimal solution to obtain a highly sensitive expansion interval. For the low-sensitive sampling interval with a sensitivity index lower than the preset sensitivity threshold, the sampling range of the low-sensitive sampling interval in the previous iteration is compressed with a preset exponential decay coefficient (such as 20% compression in each iteration) to form a low-sensitive compressed sampling interval. In this embodiment, the union of the highly sensitive expansion interval and the low-sensitive compressed sampling interval is taken to form a non-uniform dynamic sampling space. Finally, a new set of sample points is regenerated within the non-uniform dynamic sampling space to preferentially cover the highly sensitive parameter region, and the surrogate model parameters are updated based on the new sample points. The prediction error of the surrogate model within the non-uniform dynamic sampling space is calculated through the root mean square error. If the prediction error of the surrogate model within the non-uniform dynamic sampling space reaches the preset error threshold and the error change rate in three consecutive iterations is less than 1%, it is determined to converge. The surrogate model is retrained using the set of sample points within the non-uniform dynamic sampling space to update the model parameters and obtain an optimized surrogate model.

[0047] S5. Based on the optimized surrogate model and the real-time monitoring data of the distribution network, a particle swarm optimization algorithm is used to generate an adaptive scheduling strategy for the distribution network resources.

[0048] S6. Optimally allocate the distribution network resources in the new power element access scenario according to the adaptive scheduling strategy for the distribution network resources.

[0049] In some embodiments, the steps of using a particle swarm optimization algorithm to generate an adaptive scheduling strategy for the distribution network resources based on the optimized surrogate model and the real-time monitoring data of the distribution network include: Based on the sensitivity index corresponding to the optimized surrogate model, a high-dimensional decision space is constructed within the feasible region of the new power element; A number of particles are randomly generated within the high-dimensional decision space, and the learning factors of the particles are initialized according to the movement rate and the current position of each particle in the high-dimensional decision space; Based on the learning factors of the particles, the particle swarm optimization algorithm is used to update the spatial positions of each particle within the high-dimensional decision space; In each iteration process, the potential solution spacing index is calculated according to the average value of the spatial position differences of the particles in each real-time parameter dimension of the new power element, and the learning factors of the particles are adaptively adjusted according to the potential solution spacing index to obtain the latest learning factors; Calculate the relative error between the predicted value output by the optimized surrogate model and the real-time monitoring data of the distribution network, and monitor the global optimal position change rate in multiple consecutive iterations; When the relative error is less than the preset relative error threshold and the change rate of the global optimal position is less than the preset position change threshold in consecutive multiple iterations, stop the iterative optimization process and output the current global optimal parameter solution; Map the global optimal parameter solution to a high-dimensional space using a radial basis kernel function for non-linear transformation to obtain an adaptive scheduling strategy for distribution network resources.

[0050] Specifically, in this embodiment, the parameter sampling range of high-sensitivity elements is expanded according to the sensitivity index to construct a high-dimensional decision space, and the Latin hypercube sampling method is used to randomly generate a particle swarm in the high-dimensional decision space. Each particle represents a potential distribution network resource scheduling strategy. At the same time, in this embodiment, the individual learning factor (exploring the historical optimum) and the social learning factor (tracking the global optimum) of the chestnut are initialized according to the moving speed and current position of the particle in the space. For example, the initial values of the individual learning factor and the social learning factor are 1.5 and 2.0 respectively. The learning factor is used to control the learning step size and direction of the particle in the search process. Among them, the individual learning term is used to control the particle to accelerate and move in the direction of its own historical optimum position, and the social learning term is used to control the particle to accelerate and move in the direction of the current global optimum position. In this embodiment, the current speed is updated by a weighted combination of the inertia weight, the individual learning term and the social learning term, and the coordinates of the particle in the parameter space are adjusted according to the updated speed. For each particle, this embodiment uses the following formula to update its speed and position: In the formula, is the speed of the q-th particle in the d-th dimensional space at the g-th iteration; is the inertia weight, which is used to balance the global search ability and the local search ability of the particle; is the speed of the q-th particle in the d-th dimensional space at the (g - 1)-th iteration; is the individual learning factor, which reflects the learning ability of the particle to its own historical optimum position; is the social learning factor, which reflects the learning ability of the particle to the global optimum position; and are independent random functions between [0, 1]; is the historical optimum position of the q-th particle; is the position of the q-th particle in the d-th dimensional space at the (g - 1)-th iteration; is the global optimum position; is the position of the q-th particle in the d-th dimensional space at the g-th iteration.

[0051] In each iteration process, this embodiment calculates the average value of the position differences of the particle swarm in each parameter dimension to obtain a potential solution spacing index. The calculation formula for the potential solution spacing is as follows: In the formula, is the potential solution spacing; e is the number of particles; h is the spatial dimension; is the value of the q-th particle in the d-th dimensional space; is the average value of all particles in the d-th dimensional space; is the number of samples in the sampling space.

[0052] This embodiment adaptively adjusts the individual learning factor and social learning factor of the particles according to the potential solution spacing index. When the potential solution spacing index is relatively large, the social learning factor is increased to accelerate the global search; when the potential solution spacing index is relatively small, the individual learning factor is decreased to enhance the local optimization ability, and a dynamically adjusted set of learning factor parameters is obtained. This embodiment can calculate the relative error between the predicted value of the surrogate model and the actual monitored value. If the relative error is less than 3% and the change rate of the global optimal position is less than 1% in three consecutive iterations, then the radial basis kernel function is used to map the global optimal parameter solution to a high-dimensional space for nonlinear transformation, enhancing the model's analytical ability for complex coupling effects, and obtaining a high-precision parameter solution set that meets the convergence conditions and a corrected nonlinear relationship model. The corrected nonlinear relationship model can be expressed as: In the formula, is the corrected nonlinear relationship; is the Lagrange multiplier, which is the coefficient obtained by solving the optimization problem during the training process of the support vector machine algorithm, reflecting the contribution of each support vector to the decision function; is the sample corresponding to the optimal solution of the Lagrange multiplier; is the input feature vector of the j-th sample; is the input feature vector of the i-th sample.

[0053] This embodiment can generate corresponding control instructions according to the globally optimal parameter solution. For example, the energy storage system suppresses load fluctuations according to the optimal charging and discharging time sequence; the electric vehicle dynamically adjusts the upper limit of the charging power to guide the charging load to match the peak and valley of the new energy output; the microgrid switches to the off-grid mode to alleviate the main grid voltage fluctuation; the distributed power source adjusts the output priority according to the sensitivity weight to improve the accommodation efficiency. This embodiment transmits the control instructions to each new power element controller through the distribution automation system, and real-time monitors the grid state after the instructions are executed, triggering the iterative update of the strategy. In summary, this embodiment accurately generates a scheduling strategy adapted to the new power element access scenario through dynamic particle swarm optimization and kernel function correction, significantly improving the operation efficiency and stability of the distribution network. Figure 2 and Figure 3 The figure shows the comparison chart of the prediction results of the training set and the test set of the support vector machine algorithm. It can be seen from the results that the support vector machine algorithm has a good fitting effect on the non-linear relationship between the new power elements and the multi-dimensional operation state of the distribution network. Among them, the root mean square error of the training set is 0.9253, and the root mean square error of the test set is 0.9355.

[0054] The embodiment of the present invention provides a method for self-adaptive scheduling of distribution network resources. The method includes collecting distribution network operation state data, preprocessing the distribution network operation state data to obtain an initial data set of the distribution network; constructing a surrogate model characterizing the non-linear coupling relationship between the real-time parameters of the new power elements and the initial data set of the distribution network by using the support vector machine algorithm according to the initial data set of the distribution network and the real-time parameters of the new power elements; decomposing the multi-variable interaction effect in the surrogate model by using the global sensitivity analysis method, and quantitatively obtaining the sensitivity index of each new power element to the initial data set of the distribution network; adaptively correcting the sampling space of the surrogate model according to the sensitivity index by using an iterative update strategy to obtain an optimized surrogate model; generating a self-adaptive scheduling strategy for distribution network resources by using the particle swarm optimization algorithm based on the optimized surrogate model and the real-time monitoring data of the distribution network; and optimizing the configuration of the distribution network resources in the new power element access scenario according to the self-adaptive scheduling strategy of the distribution network resources. Compared with the prior art, this method analyzes the non-linear coupling relationship between the new power elements and the distribution network through technologies such as sensitivity analysis and particle swarm optimization algorithm, realizes the efficient, accurate and self-adaptive optimization configuration of the distribution network resources in the new power element access scenario, and significantly improves the operation performance and stability of the distribution network.

[0055] It should be noted that the size of the serial numbers of the above processes does not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0056] In one embodiment, as Figure 4As shown in the figure, an embodiment of the present invention provides a distribution network resource adaptive scheduling system, and the system includes: A data acquisition module 101, configured to collect operation status data of the distribution network, and preprocess the operation status data of the distribution network to obtain an initial data set of the distribution network; A model construction module 102, configured to construct a surrogate model characterizing the non-linear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network by using a support vector machine algorithm according to the initial data set of the distribution network and the real-time parameters of new power elements; A variable decomposition module 103, configured to decompose the multi-variable interaction effect in the surrogate model by using a global sensitivity analysis method, and quantitatively obtain the sensitivity index of each new power element to the initial data set of the distribution network; A space optimization module 104, configured to adaptively correct the sampling space of the surrogate model with an iterative update strategy according to the sensitivity index to obtain an optimized surrogate model; A strategy generation module 105, configured to generate a distribution network resource adaptive scheduling strategy by using a particle swarm optimization algorithm based on the optimized surrogate model and the real-time monitoring data of the distribution network; An optimization configuration module 106, configured to optimize the configuration of the distribution network resources in the access scenario of new power elements according to the distribution network resource adaptive scheduling strategy.

[0057] For the specific limitations on a distribution network resource adaptive scheduling system, reference can be made to the above limitations on a distribution network resource adaptive scheduling method, which will not be elaborated here. Those of ordinary skill in the art can realize that, combined with the various modules and steps described in the embodiments disclosed in the present application, they can be implemented by hardware, software, or a combination of both. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0058] An embodiment of the present invention provides a distribution network resource adaptive scheduling system. The system collects the operation status data of the distribution network through a data acquisition module, preprocesses the operation status data of the distribution network, and obtains an initial data set of the distribution network. The model construction module constructs a surrogate model representing the non-linear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network by using the support vector machine algorithm according to the initial data set of the distribution network and the real-time parameters of new power elements. The variable decomposition module decomposes the multi-variable interaction effect in the surrogate model by using the global sensitivity analysis method, and quantitatively obtains the sensitivity index of each new power element to the initial data set of the distribution network. The space optimization module adaptively corrects the sampling space of the surrogate model with an iterative update strategy according to the sensitivity index, and obtains an optimized surrogate model. The strategy generation module generates a distribution network resource adaptive scheduling strategy by using the particle swarm optimization algorithm based on the optimized surrogate model and the real-time monitoring data of the distribution network. The optimization configuration module optimizes the configuration of the distribution network resources in the scenario of new power element access according to the distribution network resource adaptive scheduling strategy. Compared with the prior art, the system analyzes the non-linear coupling relationship between new power elements and the distribution network through technologies such as sensitivity analysis and particle swarm optimization algorithm, realizes the efficient, accurate and adaptive optimization configuration of the distribution network resources in the scenario of new power element access, and significantly improves the operation performance and stability of the distribution network.

[0059] Figure 5 A computer device provided by an embodiment of the present invention includes a memory, a processor and a transceiver, which are connected through a bus; the memory is used to store a set of computer program instructions and data, and can transmit the stored data to the processor, and the processor can execute the program instructions stored in the memory to execute the steps of the above method.

[0060] Among them, the memory may include volatile memory or non-volatile memory, or may include both volatile and non-volatile memory; the processor may be a central processing unit, a microprocessor, an application specific integrated circuit, a programmable logic device or a combination thereof. By way of example but not limitation, the above programmable logic device may be a complex programmable logic device, a field programmable gate array, a generic array logic or any combination thereof.

[0061] In addition, the memory may be a physically independent unit or integrated with the processor.

[0062] Those of ordinary skill in the art can understand that Figure 5 the structure shown in

[0063] In one embodiment, the embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above method are implemented.

[0064] In the above embodiment, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from a website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line) or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data storage device such as a server or data center that includes one or more integrated available media. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, an SSD), etc.

[0065] Those skilled in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods.

[0066] The above embodiments only represent several preferred implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art of this technology, without departing from the technical principle of the present invention, several improvements and substitutions can be made, and these improvements and substitutions should also be regarded as the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the protection scope of the claims.

Claims

1. A method for adaptive scheduling of distribution network resources, characterized in that, It includes the following steps: Collect the operation status data of the distribution network, and preprocess the operation status data of the distribution network to obtain the initial data set of the distribution network; According to the initial data set of the distribution network and the real-time parameters of new power elements, use the support vector machine algorithm to construct a surrogate model representing the non-linear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network; Use the global sensitivity analysis method to decompose the multi-variable interaction effects in the surrogate model, and quantitatively obtain the sensitivity indexes of each new power element to the initial data set of the distribution network; According to the sensitivity indexes, adaptively correct the sampling space of the surrogate model with an iterative update strategy to obtain an optimized surrogate model; Based on the optimized surrogate model and the real-time monitoring data of the distribution network, use the particle swarm optimization algorithm to generate an adaptive scheduling strategy for the distribution network resources; Optimize the allocation of the distribution network resources in the new power element access scenario according to the adaptive scheduling strategy of the distribution network resources.

2. The adaptive scheduling method for distribution network resources according to claim 1, characterized in that, The step of using the support vector machine algorithm to construct a surrogate model representing the non-linear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network according to the initial data set of the distribution network and the real-time parameters of new power elements includes: Align the initial data set of the distribution network and the real-time parameters of new power elements according to the time stamp to form an initial sampling space data set; According to the initial sampling space data set, use the support vector machine algorithm to identify the key distribution network features affecting the real-time parameters of new power elements, and generate a low-dimensional sample point data set; Use the radial basis kernel function to map and transform the low-dimensional sample point data set into a high-dimensional feature space to obtain high-dimensional feature vectors; In the high-dimensional feature vectors, fit the non-linear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network through the support vector regression function to obtain an initial regression model in the high-dimensional feature space; Taking the minimization of the prediction error as the optimization goal, solve the initial regression model through the sequential minimal optimization algorithm to obtain the optimal parameters of the regression model; Construct a surrogate model representing the non-linear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network according to the optimal parameters of the regression model.

3. The adaptive scheduling method for distribution network resources according to claim 1, characterized in that, The step of using the global sensitivity analysis method to decompose the multi-variable interaction effects in the surrogate model and quantitatively obtain the sensitivity indexes of each new power element to the initial data set of the distribution network includes: Based on the global sensitivity analysis method, decompose the new power elements in the surrogate model to obtain the first-order sensitivity components caused by the independent action of a single new power element, the second-order sensitivity components generated by the superposition of any two new power elements, and the high-order sensitivity components generated by the coupling of at least three new power elements; Add the first-order sensitivity components, the second-order sensitivity components and the high-order sensitivity components of each new power element to obtain the total effect sensitivity of the corresponding new power element to the operation state fluctuation of the distribution network; Calculate the variance contribution ratio of the first-order sensitivity component of each new power element to the total effect sensitivity to obtain the main effect contribution rate of a single new power element; Normalize the contribution rate of the main effect to form a sensitivity index of each new power element to the initial dataset of the distribution network.

4. The adaptive scheduling method for distribution network resources according to claim 3, characterized in that: The surrogate model includes a benchmark probability term for the normal operating state of the distribution network under the condition of no access disturbance of any new power element, a first-order independent influence term representing the independent influence of a single new power element on the state classification probability of the distribution network, a second-order superposition effect term representing the influence of the superposition effect between any two new power elements on the state classification probability of the distribution network, and a high-order interaction term representing the influence of the coupling effect of at least three new power elements on the state classification probability of the distribution network.

5. The adaptive scheduling method for distribution network resources according to claim 1, wherein, The steps of adaptively correcting the sampling space of the surrogate model with an iterative update strategy according to the sensitivity index to obtain an optimized surrogate model include: Define the feasible region of the new power element according to the physical constraint conditions of the real-time parameters of the new power element, and use the Latin hypercube sampling method to uniformly select a set of initial sample points within the feasible region of the new power element; Input the set of initial sample points into the surrogate model for iterative optimization, and extract the current optimal solution from the output results of the surrogate model during each round of iterative optimization; Taking the current optimal solution as the center, determine the core sampling interval of the current round of iteration according to the preset contraction ratio; Based on the sensitivity index of each new power element to the initial dataset of the distribution network, screen out the high-sensitivity full-sampling interval and the low-sensitivity sampling interval within the core sampling interval of the current round of iteration; Compress the low-sensitivity sampling interval according to the exponential decay law to form a low-sensitivity compressed sampling interval; Expand the high-sensitivity full-sampling interval with the current optimal solution as the center to obtain a high-sensitivity extended interval, and take the union of the high-sensitivity extended interval and the low-sensitivity compressed sampling interval to form a non-uniform dynamic sampling space; Calculate the prediction error of the surrogate model within the non-uniform dynamic sampling space, and when the prediction error of the surrogate model reaches the preset error threshold, optimize and update the surrogate model according to the non-uniform dynamic sampling space to obtain an optimized surrogate model.

6. The adaptive scheduling method for distribution network resources according to claim 5, wherein, The steps of generating an adaptive scheduling strategy for distribution network resources using the particle swarm optimization algorithm based on the optimized surrogate model and the real-time monitoring data of the distribution network include: Construct a high-dimensional decision space within the feasible region of the new power element based on the sensitivity index corresponding to the optimized surrogate model; Randomly generate a number of particles within the high-dimensional decision space, and initialize the learning factor of the particles according to the movement rate and current position of each particle in the high-dimensional decision space; Based on the learning factor of the particles, use the particle swarm optimization algorithm to update the spatial position of each particle within the high-dimensional decision space; During each round of iteration, calculate the potential solution spacing index according to the average value of the spatial position differences of the particles in the dimensions of the real-time parameters of each new power element, and adaptively adjust the learning factor of the particles according to the potential solution spacing index to obtain the latest learning factor; Calculate the relative error between the predicted value output by the optimized surrogate model and the real-time monitoring data of the distribution network, and monitor the global optimal position change rate during multiple consecutive iterations; When the relative error is less than a preset relative error threshold and the change rate of the global optimal position is less than a preset position change threshold in multiple consecutive iterations, stop the iterative optimization process and output the current global optimal parameter solution; Use a radial basis kernel function to map the global optimal parameter solution to a high-dimensional space for non-linear transformation, and obtain an adaptive scheduling strategy for distribution network resources.

7. The adaptive scheduling method for distribution network resources according to claim 6, wherein: The learning factor is used to control the learning step size and direction of the particle during the search process.

8. An adaptive scheduling system for distribution network resources, characterized in that, The system includes: A data acquisition module, configured to collect operation status data of the distribution network and preprocess the operation status data of the distribution network to obtain an initial data set of the distribution network; A model construction module, configured to use a support vector machine algorithm to construct a surrogate model characterizing the non-linear coupling relationship between the real-time parameters of new power elements and the initial data set of the distribution network according to the initial data set of the distribution network and the real-time parameters of new power elements; A variable decomposition module, configured to decompose the multi-variable interaction effect in the surrogate model by using a global sensitivity analysis method and quantitatively obtain the sensitivity index of each new power element to the initial data set of the distribution network; A space optimization module, configured to adaptively correct the sampling space of the surrogate model according to the sensitivity index by using an iterative update strategy to obtain an optimized surrogate model; A strategy generation module, configured to generate an adaptive scheduling strategy for distribution network resources by using a particle swarm optimization algorithm based on the optimized surrogate model and the real-time monitoring data of the distribution network; An optimization configuration module, configured to optimize the configuration of the distribution network resources in the access scenario of new power elements according to the adaptive scheduling strategy for distribution network resources.

9. A computer device, characterized in that: Comprising a processor and a memory, the processor is connected to the memory, the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the computer device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored in the computer-readable storage medium, and when the computer program is run, the method according to any one of claims 1 to 7 is implemented.

Images