CN-115619143-B - Two-stage regional comprehensive energy system distribution robust economic scheduling method

CN115619143BCN 115619143 BCN115619143 BCN 115619143BCN-115619143-B

Abstract

The invention relates to a two-stage regional comprehensive energy system distribution robust economic dispatching method, which is an uncertainty for describing wind power probability distribution, and constructing a wind power probability distribution fuzzy set by taking the Wasserstein distance as a measure between the empirical distribution and the true distribution. To cope with possible line transmission overload, unit output out-of-limit problems, joint opportunity constraints are applied to limit the influence of wind power on line transmission, unit output and the like within a predetermined safety level. The model is converted into a mixed integer linear programming problem by using a linear decision rule and a linear increment method. The method effectively improves the energy supply limitation of a single system when dealing with uncertain wind power through multi-energy complementation, solves the problem of low calculation efficiency caused by nonlinearity and non-convexity of a model, can coordinate and consider economy and robustness, and can provide schemes with different risks for a decision maker by adjusting measurement distance and confidence.

Inventors

LIU HUI
FAN ZHENGGANG
XIE HAIMIN
WANG NI
MA SIYU
HUANG LIDONG

Assignees

广西大学

Dates

Publication Date: 20260512
Application Date: 20221013

Claims (7)

1. The two-stage regional comprehensive energy system distribution robust economic scheduling method is characterized by comprising the following steps of: (1) Reading comprehensive energy system data and wind power data; (2) Constructing a probability distribution fuzzy set based on Wasserstein distance according to wind power historical data; (3) Constructing a distributed robust joint opportunity constraint based on a Wasserstein fuzzy set, and converting the distributed robust joint opportunity constraint into a finite-dimensional deterministic constraint set by applying a conditional risk value approximation and Bonferroni conservation approximation; (4) Establishing a regional comprehensive energy forward system day scheduling model considering wind power prediction errors; (5) Establishing a real-time scheduling model of the regional comprehensive energy system, and further forming a two-stage distributed robust economic scheduling model considering joint opportunity constraint; (6) Converting the model into a mixed integer second order cone programming problem by using a linear decision rule and a linear increment method; (7) Solving to obtain an economic dispatching scheme of the regional comprehensive energy system taking economy and robustness into consideration; the general form of the joint opportunity constraint in step (3) is as follows: , wherein: an index representing the energy device or transmission line, Is the total amount of energy devices or transmission lines, Is a predefined confidence level; then, the joint opportunity constraints are divided into by Bonferroni conservative approximation Individual opportunity constraints with confidence levels of : , The above equation is converted into a worst case conditional risk value approximation: , Thus, according to the strong dual theory, it is possible to obtain: 。
2. the two-stage regional integrated energy system distribution robust economic dispatch method of claim 1, wherein the integrated energy system comprises a thermodynamic system, an electric system, a natural gas system.
3. The two-stage regional integrated energy system distribution robust economic dispatch method of claim 1, wherein step (2) is implemented by: for an uncertainty wind power deviation The historical data is Uncertainty wind power deviation True distribution of (3) Can be approximately expressed as Wherein Representative uncertainty variable history sample Dirac measure (Dirac measure), and when In the time-course of which the first and second contact surfaces, Infinitely close to true distribution I.e., as the historical data sample size increases, And true distribution The distance between them becomes smaller and smaller, so that in order to introduce probability distribution information of uncertainty variables, a description can be built up using historical data And (3) with A fuzzy set of distances between; Wasserstein distance The definition of (2) is as follows: , wherein: Representing probability distribution And A distance therebetween; To at the same time Any possible norm form thereon; As an uncertainty random variable And Is used to determine the joint probability distribution of (1), And Respectively uncertainty random variables And Is arranged in the distribution of the edges of the sheet, Representing a set of polyhedrons All probability measures of the uncertainty variable supported above; further, the fuzzy set constructed based on Wasserstein metric distance has the following form: , wherein: represents the radius constant of the Wasserstein sphere.
4. The two-stage regional integrated energy system distribution robust economic dispatch method of claim 1, wherein the objective function of the day-ahead dispatch model in step (4) is as follows: , wherein: The method is the output cost of the traditional generator set, the gas generator set and the cogeneration unit; is the cost of natural gas; representing the planned daily output of each unit; planning gas consumption for the day ahead; , respectively representing the upper standby power and the lower standby power of the unit, the cost is respectively , ; The output of the unit is adjusted to be the cost; and the adjustment quantity of the unit under the prediction deviation probability distribution of wind power is considered.
5. The two-stage regional integrated energy system distribution robust economic dispatch method of claim 1, wherein the objective function of the real-time dispatch model in step (5) is as follows: , wherein: Is the load shedding amount; represents the air discarding quantity; The power is output for each machine set in the daytime; Therefore, the two-stage regional comprehensive energy system distribution robust economic dispatch model is as follows: (1) (2) (3) (4) (5) (6), Wherein the objective function (1) is to minimize the first stage running cost and the desired cost caused by the energy adjustment; Is a decision variable comprising unit output and reserve capacity; Representing a wind power containing uncertainty probability distribution Equations (2) - (4) give the constraint of the first stage, and the real-time process is represented by equations (5) and (6), wherein The variable coefficient of expression (5).
6. The two-stage regional integrated energy system distribution robust economic scheduling method according to claim 1, wherein the step (5) converts a model into a mixed integer second order cone programming problem by using a linear decision rule and a linear increment method, and the converted model objective function is as follows: , wherein: 、 And Is an auxiliary variable.
7. The two-stage regional integrated energy system distribution robust economic dispatch method of claim 2, wherein the Weymouth equation simulating gas flow in the natural gas system constraint is as follows: , to increase the calculation speed, the gas flow equation is linearized by piecewise linear programming techniques: 。

Description

Two-stage regional comprehensive energy system distribution robust economic scheduling method Technical Field The invention relates to the technical field of power systems, in particular to a two-stage regional comprehensive energy system distribution robust economic scheduling method. Background The comprehensive energy system breaks through barriers among energy supply of different energy systems through interconnection and complementation of various energy sources, realizes high-efficiency utilization of energy sources, and has huge practical value. However, the strong uncertainty of renewable energy sources such as wind power and the like after large-scale access brings unprecedented challenges to the economic dispatch of the comprehensive energy system. In addition, in the traditional optimization method for solving the uncertainty problem, accurate probability distribution information of wind power output is difficult to obtain through stochastic programming, and robust optimization based on uncertainty variable boundary information can lead to relatively conservative decision results or overhigh cost. Therefore, distributed robust optimization taking into account random variable probability uncertainty is becoming of increasing interest because it can effectively solve the problem of decision-making too optimistic or too conservative. The construction of a typical distribution robust fuzzy set is based on first-order, second-order moment information (mean, covariance) and higher-order moment information of wind power, but different distributions may have the same moment information, which makes it difficult for a decision to determine the probability distribution of wind power in the worst case. In the fuzzy set based on the distance measurement, the widely applied Kullback-Leibler divergence fuzzy set can be derived by using data only when the real distribution is supported on a limited set, but the real distribution of wind power generation is continuous. In contrast, wasserstein-based contains all (continuous or discrete) probability distributions that are sufficiently close to discrete empirical distributions that they exhibit good performance in terms of both limited sample assurance and confidence set. Disclosure of Invention The invention provides a two-stage distributed robust economic scheduling method considering joint opportunity constraint, which aims at the economic scheduling problem of an regional comprehensive energy system under wind power uncertainty based on the advantages of the distributed robust optimization method and joint opportunity constraint in the aspect of processing the renewable energy uncertainty such as wind power and the like, and ensures the economic and reliable operation of the regional comprehensive energy system. The invention adopts the following technical scheme: a two-stage regional comprehensive energy system distribution robust economic scheduling method comprises the following steps: (1) Reading comprehensive energy system data and wind power data; (2) Constructing a probability distribution fuzzy set based on Wasserstein distance according to wind power historical data; (3) Constructing a distributed robust joint opportunity constraint based on a Wasserstein fuzzy set, and converting the distributed robust joint opportunity constraint into a finite-dimensional deterministic constraint set by applying a conditional risk value approximation and Bonferroni conservation approximation; (4) Establishing a regional comprehensive energy forward system day scheduling model considering wind power prediction errors; (5) Establishing a real-time scheduling model of the regional comprehensive energy system, and further forming a two-stage distributed robust economic scheduling model considering joint opportunity constraint; (6) Converting the model into a mixed integer second order cone programming problem by using a linear decision rule and a linear increment method; (7) And solving to obtain an economic dispatching scheme of the regional comprehensive energy system taking the economical efficiency and the robustness into consideration. The comprehensive energy system comprises a thermodynamic system, an electric power system and a natural gas system. The step (2) is realized by the following method: for an uncertainty wind power deviation Its historical data isUncertainty wind power deviationThe true distribution P of (2) can be approximately expressed asWherein the method comprises the steps ofA Dirac measure representing an uncertainty variable history sample ζ k, and when N → infinity, P N is infinitely close to the true distribution P, i.e. the distance between P N and the true distribution P becomes smaller as the history data sample size is larger, therefore, in order to introduce probability distribution information of an uncertainty variable, a fuzzy set describing the distance between P N and P can be established by using the history data; the definition o