Search

CN-122026389-A - Micro-grid group coordinated scheduling and energy management method and system

CN122026389ACN 122026389 ACN122026389 ACN 122026389ACN-122026389-A

Abstract

S1, acquiring source side output prediction data and load side demand prediction data of a plurality of micro grids forming the micro grid group under a plurality of time sections, and acquiring current charge state and maximum charge and discharge power data of an energy storage system in each micro grid; S2, generating a multi-scene simulation data set describing the uncertainty of the new energy output and the uncertainty of the load fluctuation based on the source side output prediction data and the load side demand prediction data. According to the invention, uncertainty of new energy output and load fluctuation is depicted by constructing the multi-scene simulation data set, and a Latin hypercube sampling combined with synchronous return reduction method is adopted to generate a typical operation scene, so that the optimal scheduling model can fully consider statistical characteristics of prediction errors, and the adaptability of a scheduling plan to actual operation conditions is effectively improved.

Inventors

  • HE JIANRONG
  • WANG YONG

Assignees

  • 合邦电力科技有限公司

Dates

Publication Date
20260512
Application Date
20260410

Claims (10)

  1. 1. The micro-grid group coordination scheduling and energy management method is characterized by comprising the following steps of: S1, acquiring source side output prediction data and load side demand prediction data of a plurality of micro-grids forming a micro-grid group under a plurality of time sections, and acquiring current charge state and maximum charge and discharge power data of an energy storage system in each micro-grid; S2, based on source side output prediction data and load side demand prediction data, a pre-constructed error probability distribution model of new energy output and load fluctuation is combined, and a multi-scene simulation data set describing the uncertainty of the new energy output and the uncertainty of the load fluctuation is generated; S3, establishing a distributed optimization scheduling model which aims at minimizing the overall running cost of the micro-grid group and comprises internal power balance constraint, tie line power exchange constraint and charge and discharge power constraint of each micro-grid; s4, carrying out iterative solution on the distributed optimization scheduling model by adopting a distributed optimization algorithm, and independently solving a local sub-optimization problem and updating a local energy storage output plan and a tie line exchange power plan in each iterative process on the basis of local information and boundary exchange power predicted values transmitted by adjacent micro-grids through a communication network; S5, stopping iteration when the exchange power planning values at the public connection points are agreed between the adjacent micro-grids and the global residual error of two adjacent iterations is smaller than a preset threshold value, and outputting the energy storage charging and discharging power time sequence instruction of each micro-grid and the power exchange time sequence instruction of the adjacent micro-grid; S6, generating execution control signals of the distributed power supplies and the energy storage converters in the micro-grids according to the energy storage charging and discharging power time sequence instruction and the power exchange time sequence instruction, and generating economic settlement reports among the main bodies in the micro-grid group based on deviation electric quantity data between actual operation results and the time sequence instruction.
  2. 2. The method of coordinated scheduling and energy management of a microgrid population according to claim 1, wherein the step of generating a multi-scenario simulation dataset in S2 comprises: Acquiring an error sequence of the output predicted data of the inner side and an error sequence of the demand predicted data of the inner side of each micro-grid in a history contemporaneous period, and fitting to obtain normal distribution parameters or non-parameter nuclear density distribution functions obeyed by the predicted errors; Based on normal distribution parameters or non-parameter kernel density distribution functions, carrying out layered sampling in a prediction error space by using a Latin hypercube sampling method to generate a plurality of error scenes representing prediction deviation; Respectively overlapping error scenes on a reference curve of source side output prediction data and load side demand prediction data to generate a plurality of initial prediction scenes; and reducing a plurality of initial prediction scenes by synchronous back substitution and subtraction, eliminating scenes with close probability distances, and reserving a set number of typical scenes and the corresponding occurrence probability thereof.
  3. 3. The method for coordinated scheduling and energy management of micro-grid groups according to claim 1, wherein the objective function of the distributed optimization scheduling model established in S3 is specifically represented as a sum of running costs of each micro-grid, and the running costs include fuel costs of distributed power generation units in each micro-grid, transaction costs of purchasing electricity to adjacent micro-grids, and cyclic aging costs according to energy storage charging and discharging depth conversion.
  4. 4. The method for coordinated scheduling and energy management of micro-grid clusters according to claim 1, wherein the step S4 of performing iterative solution by using a distributed optimization algorithm specifically comprises: In each iteration, each micro-grid firstly calculates a local state change evaluation index based on the local current energy storage charge state change rate and the net load prediction fluctuation rate; when the local state change evaluation index exceeds a preset trigger threshold, the micro-grid activates a communication module to broadcast state information containing a local marginal cost mapping value and a boundary exchange power request value to an adjacent micro-grid; after receiving the state information, the adjacent micro-grids perform feasibility verification on the boundary exchange power request value according to the self-operation constraint, and return correction information containing the local marginal cost mapping value until the difference of the marginal cost mapping values between the adjacent micro-grids is smaller than a preset consistency deviation threshold; The marginal cost mapping value is a Lagrange multiplier corresponding to a tie line exchange power variable in the local sub-optimization problem of each micro-grid or a scalar value obtained by normalizing marginal cost of all local schedulable resources, and when the micro-grid contains various adjustment resources, the marginal cost mapping value takes a weighted average value of marginal cost of each resource, and the weight of the weighted average value is in direct proportion to the available adjustment capacity of each resource at the current moment.
  5. 5. The method for coordinated scheduling and energy management of micro-grid groups according to claim 1, wherein before outputting the stored energy charging and discharging power timing command of each micro-grid in S5, the method further comprises a dynamic feasible region correction step: Predicting the state of charge value of each micro-grid energy storage system at the end moment of a dispatching cycle according to the energy storage charging and discharging plan obtained by solving in the current iteration; when the state of charge value touches a preset state of charge safety boundary, dynamically tightening the feasible region range of the tie line power exchange constraint in the next round of iteration according to the degree of the state of charge deviating from the safety boundary; restarting an iterative process of the distributed optimization algorithm based on the tightened feasible region range until the output energy storage charging and discharging plan meets the charge state safety boundary requirement and the convergence condition is met; The dynamic tightening operation specifically comprises the steps of reducing the upper power limit of the micro-grid for transmitting power to the adjacent micro-grid according to a preset first tightening coefficient ratio if the predicted end state of charge value is higher than a preset upper safety limit, reducing the upper power limit of the micro-grid for receiving power from the adjacent micro-grid according to a preset second tightening coefficient ratio if the predicted end state of charge value is lower than a preset lower safety limit, wherein the size of the first tightening coefficient or the second tightening coefficient is in direct proportion to the absolute value of the state of charge deviated from the safety limit, and the feasible region range after tightening is not empty.
  6. 6. The method for coordinated scheduling and energy management of a micro-grid cluster according to claim 1, wherein the convergence condition for determining that the iteration is stopped in S5 is that a global residual is smaller than a preset threshold, wherein the global residual is defined as a euclidean norm variation of all micro-grid tie-line exchange power planning values in two adjacent iterations, namely , Wherein the method comprises the steps of For the total number of links to be used, And Respectively the first The tie line is at the first Secondary and tertiary Power plan values in the multiple iterations.
  7. 7. The method for coordinated scheduling and energy management of micro-grid groups according to claim 1, wherein the step of generating the economic settlement report in S6 includes counting the deviation electric quantity between the actual running power curve of each micro-grid and the time sequence command output in S5 in the scheduling period, and based on a preset deviation electric quantity allocation rule, allocating the total deviation electric quantity among the micro-grid main bodies and generating the economic settlement report containing the deviation electric charge settlement list among the main bodies in combination with the internal purchase electricity agreement price, wherein the deviation electric quantity allocation rule is one of allocating according to the planned total electric quantity proportion of each micro-grid in the scheduling period or allocating according to the inverse proportion of the available adjustment capacity of each micro-grid in the deviation occurrence period.
  8. 8. A micro grid cluster coordinated scheduling and energy management system for implementing the method of any one of claims 1 to 7, comprising: The data acquisition and prediction module is used for acquiring source side output prediction data and load side demand prediction data of a plurality of micro-grids forming the micro-grid group, and current charge state and maximum charge and discharge power data of an energy storage system in each micro-grid; The uncertainty analysis module is connected with the data acquisition and prediction module, and is used for generating a multi-scene simulation data set for describing the uncertainty of the new energy output and the uncertainty of the load fluctuation according to the received source-side output prediction data and the load-side demand prediction data and by combining a pre-constructed error probability distribution model of the new energy output and the load fluctuation; the optimization model construction module is used for establishing a distributed optimization scheduling model which aims at minimizing the overall running cost of the micro-grid group and comprises internal power balance constraint, tie line power exchange constraint and charge and discharge power constraint of the energy storage system; The distributed collaborative solving module is respectively connected with the uncertainty analysis module and the optimization model construction module, receives the multi-scene simulation data set and loads the distributed optimization scheduling model, adopts a distributed optimization algorithm to carry out iterative solving on the distributed optimization scheduling model, and in each iterative process, the module issues solving tasks to each micro-grid through a communication network and receives a local energy storage output plan and a tie line exchange power plan fed back by each micro-grid until the exchange power plan values at a public connection point are agreed between adjacent micro-grids and the global residual error of two adjacent iterations is smaller than a preset threshold value, and finally outputs an energy storage charge and discharge power time sequence instruction of each micro-grid and a power exchange time sequence instruction of the adjacent micro-grid; the instruction generation and settlement module is connected with the distributed collaborative solving module, generates execution control signals of the distributed power supply and the energy storage converter in each micro-grid according to the energy storage charging and discharging power time sequence instruction and the power exchange time sequence instruction, and generates economic settlement reports among all the main bodies in the micro-grid group according to deviation electric quantity data between an actual operation result and the time sequence instruction.
  9. 9. The micro grid cluster coordinated scheduling and energy management system of claim 8, wherein the uncertainty analysis module comprises: The error statistics unit is used for obtaining an error sequence of the endogenous side output prediction data and an error sequence of the load side demand prediction data of each micro-grid in the historical synchronization period, and fitting to obtain normal distribution parameters or non-parameter nuclear density distribution functions obeyed by the prediction errors; The scene sampling generation unit is connected with the error statistics unit, performs layered sampling in a prediction error space by using a Latin hypercube sampling method based on normal distribution parameters or non-parameter kernel density distribution functions to generate a plurality of error scenes representing prediction deviation, and respectively overlaps the error scenes on reference curves of source side output prediction data and load side demand prediction data to generate a plurality of initial prediction scenes; The scene reduction unit is connected with the scene sampling generation unit, reduces a plurality of initial prediction scenes through synchronous back substitution and subtraction, eliminates scenes with similar probability distances, reserves a set number of typical scenes and the occurrence probability corresponding to the typical scenes, and sends the typical scenes and probability data thereof to the distributed collaborative solving module.
  10. 10. The micro grid group coordinated scheduling and energy management system according to claim 8, wherein the distributed collaborative solving module specifically performs, in an iterative solving process using a distributed optimization algorithm: In each iteration, each micro-grid firstly calculates a local state change evaluation index based on the local current energy storage charge state change rate and the net load prediction fluctuation rate; when the local state change evaluation index exceeds a preset trigger threshold, the micro-grid activates a communication module to broadcast state information containing a local marginal cost mapping value and a boundary exchange power request value to an adjacent micro-grid; after receiving the state information, the adjacent micro-grids perform feasibility verification on the boundary exchange power request value according to the self-operation constraint, and return correction information containing the local marginal cost mapping value until the difference of the marginal cost mapping values between the adjacent micro-grids is smaller than a preset consistency deviation threshold; The marginal cost mapping value is a Lagrange multiplier corresponding to a tie line exchange power variable in the local sub-optimization problem of each micro-grid or a scalar value obtained by normalizing marginal cost of all local schedulable resources, and when the micro-grid contains various adjustment resources, the marginal cost mapping value takes a weighted average value of marginal cost of each resource, and the weight of the weighted average value is in direct proportion to the available adjustment capacity of each resource at the current moment.

Description

Micro-grid group coordinated scheduling and energy management method and system Technical Field The invention relates to the technical field of micro-grid energy management and coordination control, in particular to a micro-grid group coordination scheduling and energy management method and system. Background Along with the large-scale access of the distributed photovoltaic power generation at the user side, a single micro-grid gradually evolves to a micro-grid group form of multi-micro-grid interconnection, the micro-grid group is electrically interconnected through a public connection point, and the complementary mutual utilization of new energy power generation and the sharing utilization of energy storage resources can be realized in a larger range, however, the coordinated operation of the micro-grid group faces multiple technical challenges, firstly, the photovoltaic power generation has obvious intermittence and fluctuation, the output prediction has larger errors, the single micro-grid is difficult to independently stabilize the power fluctuation, secondly, the micro-grids have competition relationship and cooperate in the operation process, so that global optimization scheduling is realized on the premise of protecting the operation privacy of each micro-grid, and thirdly, the energy storage system is used as a flexible adjustment resource, and the charge and discharge strategy of the energy storage system directly influences the operation economy and the service life of the micro-grid group. The method for managing the energy of the micro-grid disclosed in the prior art mainly adopts a centralized optimization framework, a central controller collects operation data of all the micro-grids and uniformly solves an optimization model, and the method has the obvious defects that the centralized framework requires each micro-grid to report complete operation data and prediction information to the central controller, so that the operation privacy of each micro-grid cannot be effectively protected, when the number of the micro-grids is increased or the network topology is changed, the calculation burden of the central controller is exponentially increased, the solving time of the optimization model is greatly prolonged, the real-time requirement of on-line scheduling is difficult to meet, the reliability and bandwidth requirement of the centralized method on the communication network are higher, once the central controller or a backbone communication network fails, the whole micro-grid group loses the coordination control capability, in addition, the prior art generally adopts a simple deterministic prediction or robust optimization method when the uncertainty of new energy output is processed, the former is difficult to deal with power unbalance caused by prediction deviation, and the operation economy is often sacrificed due to over high conservation. Disclosure of Invention The invention aims to provide a method and a system for coordinated dispatching and energy management of a micro-grid group, which are used for solving the problems in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the micro-grid group coordination scheduling and energy management method comprises the following steps: S1, acquiring source side output prediction data and load side demand prediction data of a plurality of micro-grids forming a micro-grid group under a plurality of time sections, and acquiring current charge state and maximum charge and discharge power data of an energy storage system in each micro-grid; S2, based on source side output prediction data and load side demand prediction data, a pre-constructed error probability distribution model of new energy output and load fluctuation is combined, and a multi-scene simulation data set describing the uncertainty of the new energy output and the uncertainty of the load fluctuation is generated; S3, establishing a distributed optimization scheduling model which aims at minimizing the overall running cost of the micro-grid group and comprises internal power balance constraint, tie line power exchange constraint and charge and discharge power constraint of each micro-grid; s4, carrying out iterative solution on the distributed optimization scheduling model by adopting a distributed optimization algorithm, and independently solving a local sub-optimization problem and updating a local energy storage output plan and a tie line exchange power plan in each iterative process on the basis of local information and boundary exchange power predicted values transmitted by adjacent micro-grids through a communication network; S5, stopping iteration when the exchange power planning values at the public connection points are agreed between the adjacent micro-grids and the global residual error of two adjacent iterations is smaller than a preset threshold value, and outputting the energy storage charging and dischargi