CN-122026458-A - Distributed energy storage system scheduling method and system

CN122026458ACN 122026458 ACN122026458 ACN 122026458ACN-122026458-A

Abstract

The invention discloses a dispatching method and a dispatching system of a distributed energy storage system, wherein the dispatching method is used for realizing quantifiable dynamic balance between peak clipping and valley filling and improving power supply reliability, and comprises the steps of obtaining an equivalent continuous load curve based on random production simulation, determining rated power and rated capacity of the energy storage system by taking a target reliability index threshold value of power shortage probability or electric quantity shortage expected value as constraint, constructing a dynamic balance optimization model taking load stabilization, time-of-use electricity price cost and reliability risk into consideration simultaneously by taking the energy storage charging and discharging power of each period as decision variables, carrying out global search on the dynamic balance optimization model by adopting a self-adaptive particle swarm algorithm driven by reliability, carrying out local refinement on global optimal solution obtained by a particle swarm by adopting an interior point method, and outputting an optimal charging and discharging strategy. The invention can realize dynamic balance of distributed energy storage in peak clipping and valley filling and improving power supply reliability.

Inventors

WANG DAOJING
XIE YUGUANG
CHEN JIE
DING MING
LI XIANG
MA XIAOBING
MA TAO
GUO XINPING
LI JINZHONG
ZHANG HONG
LI ZHE
MA WEI

Assignees

国网安徽省电力有限公司淮北供电公司

Dates

Publication Date: 20260512
Application Date: 20251230

Claims (10)

1. The distributed energy storage system scheduling method is characterized by being used for realizing quantifiable dynamic balance between peak clipping and valley filling and improving power supply reliability, and comprising the following steps of: acquiring an equivalent continuous load curve based on random production simulation, and determining rated power and rated capacity of an energy storage system by taking a target reliability index threshold of power shortage probability or expected value of power shortage as constraint; taking the energy storage charging and discharging power of each period as a decision variable, and constructing a dynamic balance optimization model which simultaneously considers load stabilization, time-of-use electricity price cost and reliability risk; and carrying out global search on the dynamic balance optimization model by adopting a reliability-driven self-adaptive particle swarm algorithm, and carrying out local refinement on a global optimal solution obtained by the particle swarm by adopting an interior point method to output an optimal charge-discharge strategy.
2. The method of claim 1, wherein the energy storage system power rating and capacity rating are determined by applying the following formula: ; Or (b) ; In the formula, As a function of the scale objective of the energy storage system, 、 Each of which is a two-way trade-off coefficient, For the rated power of the energy storage system, For the rated capacity of the energy storage system, To be configured with The probability of a shortage of electric power at the time, Is a threshold value thereof; To configure for When the electric quantity is less than the expected value, Is a threshold value thereof.
3. The distributed energy storage system scheduling method of claim 1, wherein the dynamic balance optimization model is expressed by the following formula: ; In the formula, In order to dynamically balance the optimization model, 、、 Respectively, are the weight coefficients of the multiple targets, In order for the load variance to be the same, Is that The net load at the time of the cycle, Is that The equivalent charge-discharge power at the time of the cycle, For the cost of the time-sharing electricity price, Is a reliability risk.
4. The method for dispatching a distributed energy storage system according to claim 1, wherein the reliability risk comprises a power shortage summary overrun penalty, a power shortage expected overrun penalty or a power shortage expected conditional risk value, and a charge-discharge power constraint, an SOC dynamic constraint, an SOC upper and lower limit constraint and a cycle end SOC regression constraint SOC (T) =soc (0) are set, wherein T is a dispatching cycle.
5. The method of claim 1, wherein performing a global search on the dynamic balance optimization model using a reliability-driven adaptive particle swarm algorithm comprises: Expressing a particle code as a whole section of scheduling sequence, and acquiring the speed and the position of the particle; defining particle population diversity and reliability violation degree; And the inertia weight and the learning factor of the self-adaptive particle swarm algorithm are adjusted in a combined mode according to the diversity of the particle swarm and the degree of reliability violation.
6. The method of claim 5, wherein the population diversity and reliability violations are formulated as follows: ; ; In the formula, Is the first The diversity of the particle population is characterized in the next iteration, As a total population of particles, Is the first At the time of iteration, the first The entire schedule sequence of the individual particles, Is the first The entire scheduling sequence at the time of the iteration, Is a two-norm; is the first The degree of reliability violations at each iteration, Is the first Summary of power deficiency at the time of the iteration, To target the power deficiency summary as a reliability index threshold, Is the first The amount of power at the time of the iteration is less than the desired value, To target a reliability index threshold for a power shortage expected value.
7. The method of claim 6, wherein the inertia weights and learning factors of the adaptive particle swarm algorithm are jointly adjusted according to the diversity of the particle swarm and the degree of reliability violations, and are expressed by the following formula: ; ; ; In the formula, Is the first The inertial weights of the number of iterations, 、 Respectively minimum and maximum inertial weights, Is the base of a natural logarithmic function, Is the exponential decay coefficient of the inertia weight, For the maximum number of iterations to be performed, The weighting coefficients are fed back for reliability violations, Is the first The degree of reliability violation for the next iteration, 、 Two learning factors in the adaptive particle swarm algorithm respectively, 、 Respectively are learning factors Is set to be the maximum value, the minimum value, 、 Respectively are learning factors Is set to be the maximum value, the minimum value, Is the first And (5) particle population diversity characterization during the next iteration.
8. The method of claim 1, wherein the adaptive particle swarm algorithm performs feasible-domain projection repair on the particles after each position update to ensure that the power and the SOC satisfy the constraint throughout, wherein the feasible-domain projection repair is expressed by the following formula: ; In the formula, For a pair of periods The repair value after the upper limit and the lower limit constraint processing is carried out on the charge state, To repair the state of charge resulting from the particle location update prior to repair, Is the lower limit and the upper limit of the charge state, For periods where a feasibility fix is required.
9. The method of claim 1, wherein locally refining the global optimal solution obtained by the particle swarm by an interior point method comprises writing constraints of a dynamic balance optimization model In the form, a global optimal solution obtained by a particle swarm is used as an initial point, an obstacle function is constructed, and a local optimal feasible solution is obtained by iteratively updating obstacle factors, wherein the obstacle function is expressed by the following formula: ; In the formula, As a function of the obstacle, Is the position vector of the particle The corresponding original objective function value is used to determine, As a factor of the obstacle, Is the position vector of the particle Is the first of (2) The number of constraints to be applied to the system, Is natural logarithm.
10. A system for applying the distributed energy storage system scheduling method of any one of claims 1-9, for achieving a quantifiable dynamic balance between peak clipping and valley filling and improving power supply reliability, comprising: the energy storage rated parameter determining module is used for obtaining an equivalent continuous load curve based on random production simulation and determining rated power and rated capacity of the energy storage system by taking a target reliability index threshold value of the power shortage probability or the expected value of the power shortage as constraint; the dynamic balance optimization module is used for constructing a dynamic balance optimization model which simultaneously considers the load stabilization, the time-of-use electricity price cost and the reliability risk by taking the energy storage charging and discharging power of each period as a decision variable; And the optimal charge-discharge strategy module is used for carrying out global search on the dynamic balance optimization model by adopting a reliability-driven self-adaptive particle swarm algorithm, and then carrying out local refinement on a global optimal solution obtained by the particle swarm by adopting an interior point method, so as to output an optimal charge-discharge strategy.

Description

Distributed energy storage system scheduling method and system Technical Field The invention relates to the technical field of energy storage optimization of power systems, in particular to a dispatching method and system of a distributed energy storage system. Background Along with the load fluctuation and the renewable energy permeability improvement of the distribution network, the traditional capacity expansion-dependent mode is difficult to achieve both economy and power supply quality. The battery energy storage is an important means for peak clipping and valley filling because of the rapid charge and discharge capability. However, the existing method has the problems that (1) the reliability index and the peak clipping and valley filling target are not modeled uniformly from the economical efficiency or load stabilization, so that the scheme is easy to fail in a fault or supply and demand shortage period, (2) part of researches are conducted, although intelligent optimization algorithms such as particle swarm and the like are introduced, the main improvement is concentrated on linear decrease of inertia weight or weight adjustment based on fuzzy/neural network, the reliability index is not used as a mechanism for adjusting driving signals directly by algorithm parameters, and an infeasible solution is easy to generate under the strong constraint of SOC (3) the problem that the convergence accuracy is insufficient often caused by a single heuristic algorithm under high-dimensional continuous variable, and the feasibility and the optimality are difficult to ensure simultaneously. Disclosure of Invention The invention aims to solve the technical problem that the current energy storage system scheduling does not consider the problem of realizing quantifiable dynamic balance between peak clipping and valley filling and improving the power supply reliability. In order to solve the technical problems, the invention provides the following technical scheme: A distributed energy storage system scheduling method for achieving quantifiable dynamic balance between peak clipping and valley filling and improving power supply reliability, comprising: acquiring an equivalent continuous load curve based on random production simulation, and determining rated power and rated capacity of an energy storage system by taking a target reliability index threshold of power shortage probability or expected value of power shortage as constraint; taking the energy storage charging and discharging power of each period as a decision variable, and constructing a dynamic balance optimization model which simultaneously considers load stabilization, time-of-use electricity price cost and reliability risk; and carrying out global search on the dynamic balance optimization model by adopting a reliability-driven self-adaptive particle swarm algorithm, and carrying out local refinement on a global optimal solution obtained by the particle swarm by adopting an interior point method to output an optimal charge-discharge strategy. In this embodiment, the rated power and rated capacity of the energy storage system are determined, and the following formula is applied: ; Or (b) ; In the formula,As a function of the scale objective of the energy storage system,、Each of which is a two-way trade-off coefficient,For the rated power of the energy storage system,For the rated capacity of the energy storage system,To be configured withThe probability of a shortage of electric power at the time,Is a threshold value thereof; To configure for When the electric quantity is less than the expected value,Is a threshold value thereof. In this embodiment, the dynamic balance optimization model is expressed by the following formula: ; In the formula, In order to dynamically balance the optimization model,、、Respectively, are the weight coefficients of the multiple targets,In order for the load variance to be the same,Is thatThe net load at the time of the cycle,Is thatThe equivalent charge-discharge power at the time of the cycle,For the cost of the time-sharing electricity price,Is a reliability risk. In this embodiment, the reliability risk includes an electric power deficiency summary overrun penalty, an electric power deficiency expected overrun penalty, or an electric power deficiency expected condition risk value, and sets a charge-discharge power constraint, an SOC dynamic constraint, an SOC upper and lower limit constraint, and a cycle end SOC regression constraint SOC (T) =soc (0), where T is a scheduling period. In this embodiment, performing global search on the dynamic balance optimization model by adopting a reliability-driven adaptive particle swarm algorithm includes: Expressing a particle code as a whole section of scheduling sequence, and acquiring the speed and the position of the particle; defining particle population diversity and reliability violation degree; And the inertia weight and the learning factor of the self-adaptive particle swarm a