阅读说明：本技术 一种考虑动态充放电效率的电池储能系统调度方法 (A kind of battery energy storage system dispatching method considering dynamic efficiency for charge-discharge ) 是由 童博 赵勇 文乐 张宝锋 于 2019-08-22 设计创作，主要内容包括：一种考虑动态充放电效率的电池储能系统调度方法,包括以下步骤：A、获取电池储能系统基本信息,对充放电效率特性曲线进行分段线性化；B、获取电池储能系统运行限制、初末电量、预测分时电价或价格乘子等调度信息；C、建立调度模型；D、计算电池储能系统各时段存储电量的动态上下限；E、动态规划反向递推,获得各时段需要反复求解的核心子问题的优化模型；F、利用核心子问题优化模型的几何特征,从下到上依次求解各水平子区域中最优路径；G、求解水平子区域内的最优路径；H、处理最优路径跳变；I、动态规划正向递推,计算各时段的最优充放电电量及功率决策；J、随时间发展进入下一时段,更新步骤B中的调度信息,执行步骤C。本方法为电池储能系统的优化运行提供了一种有效途径,具有工程实用价值。(A kind of battery energy storage system dispatching method considering dynamic efficiency for charge-discharge, comprising the following steps: A, obtain battery energy storage system essential information, piece-wise linearization is carried out to efficiency for charge-discharge characteristic curve；B, battery energy storage system run-limiting, just last electricity, the prediction scheduling informations such as tou power price or price multiplier are obtained；C, scheduling model is established；D, the dynamic bound of battery energy storage system day part storing electricity is calculated；E, Dynamic Programming reverse recursion obtains the Optimized model for the core subproblem that day part needs to solve repeatedly；F, using the geometrical characteristic of core subproblem Optimized model, optimal path in each horizontal subregion is successively solved from top to bottom；G, the optimal path in horizontal subregion is solved；H, processing optimal path jump；I, Dynamic Programming forward recursive calculates the optimal charge/discharge electricity amount and power decision of day part；J, development enters subsequent period at any time, updates the scheduling information in step B, executes step C.This method provides a kind of effective way for the optimization operation of battery energy storage system, has engineering practical value.)

1. a kind of battery energy storage system dispatching method for considering dynamic efficiency for charge-discharge, it is characterised in that: can with this method The optimization operational decisions for formulating battery energy storage system, improve the performance driving economy of battery energy storage system, this method includes following step It is rapid:

A, the limitation of capacity, dynamic efficiency for charge-discharge, charge-discharge electric power and the battery energy storage system capacity of battery energy storage system are determined Bound, according to the required precision of scheduling select suitable segments to charge-discharge electric power-relationship between efficiency of battery energy storage system into Row piece-wise linearization；

B, the target storing electricity when initial storage electricity of acquisition battery energy storage system, finishing scheduling, day part electricity volume Constraint and timesharing forecasted electricity market price or equivalent value multiplier；

C, the scheduling model of battery energy storage system is established, the optimization aim of scheduling model is to formulate battery energy storage system day part to fill Discharge power decision maximizes total gene-ration revenue of battery energy storage system, the objective function of scheduling model are as follows:

In formula, T is scheduling slot sum, and t is scheduling slot number, λ_tMultiply for the forecasted electricity market price or equivalent value of t period Son, q_tFor the charge/discharge electricity amount of the battery energy storage system of t period, discharges for positive value, be charged as negative value, p (q_t) the t period The charge-discharge electric power of battery energy storage system is q_tUnitary piecewise linear function, δ be scheduling slot length；

The constraint condition of scheduling model includes:

1) the electric quantity balancing constraint of battery energy storage system

In formula, e_tFor the storing electricity of t period end battery energy storage system, L_tFor consolidating for known t period battery energy storage system Setting loss power loss amount is negative value；

2) electricity volume constrains

In formula,WithThe respectively electricity volume bound of t period battery energy storage system；

3) the charge-discharge electric power constraint of battery energy storage system

In formula, P^maxAnd P^minThe respectively upper physical limit of battery energy storage system discharge power and charge power, function p are dull Piecewise linear function, charge-discharge electric power constraint (4) can be converted to the constraint of discharge and recharge:

In formula, function p' is the inverse function of function p, p'(P^max) and p'(P^min) it is respectively battery energy storage system discharge electricity amount and to fill The upper physical limit of power consumption；

4) electricity SOC is constrained

In formula, E^maxAnd E^minThe respectively bound of the storing electricity of battery energy storage system；

5) the first last Constraint of battery energy storage system

e₀=E⁰,e_T=E^T (7)

In formula, E⁰And E^TTarget electricity when being battery energy storage system initial time electricity and finishing scheduling respectively, e₀And e_TRespectively For the storing electricity of the battery energy storage system of scheduling start time and T period Mo；

By (1), (2), (3), (5), (6), (7) formula constitutes the scheduling model of battery energy storage system；

D, the dynamic bound constraint for calculating battery energy storage system day part storing electricity, according to the last battery energy storage system tune of scheduling Target storing electricity E at the end of degree^T(2) and (5) are constrained with battery energy storage system day part storing electricity bound, by from T Period starts reverse recursion, calculates the dynamic bound of battery energy storage system day part storing electricity are as follows:

In formula, j is time variable,WithThe t period end battery energy storage system day part that respectively reverse-direction derivation obtains is deposited The dynamic bound of reserve of electricity, E^TThe target storing electricity of battery energy storage system, L when for finishing scheduling_jFor the known jth period The dead loss electricity of battery energy storage system is negative value, p'(P^max) and p'(P^min) it is respectively battery energy storage system discharge electricity amount With the upper physical limit of charge capacity,WithThe respectively bound of jth period battery energy storage system electricity volume；

According to the initial quantity of electricity E of battery energy storage system⁰With battery energy storage system day part storing electricity bound constraint (2) and (5), the period carries out forward recursive since scheduling, calculates the dynamic bound of battery energy storage system day part storing electricity Are as follows:

In formula, j is time variable,WithThe respectively positive battery energy storage system day part for deriving the t period Mo obtained The dynamic bound of storing electricity, E⁰The storing electricity of battery energy storage system, L when starting for scheduling_jFor known jth period electricity The dead loss electricity of pond energy-storage system, p'(P^max) and p'(P^min) it is respectively battery energy storage system discharge electricity amount and charge capacity Upper physical limit,WithThe respectively bound of jth period battery energy storage system electricity volume；

Forward direction is derived and reverse-direction derivation result combines, calculates the dynamic bound of the storing electricity of battery energy storage system are as follows:

In formula,Withe _tRespectively calculate the dynamic of battery energy storage system day part storing electricity of the t period of acquisition or more Limit, in conjunction with (6) formula, obtains the bound of new battery energy storage system day part storing electricity are as follows:

In formula, e_tFor the storing electricity of t period end battery energy storage system, E^maxAnd E^minThe respectively storage of battery energy storage system The physics bound of electricity；

By equivalent transformation, by (1), (2), (3), (5), (7), (13) formula constitutes the scheduling model of battery energy storage system；

E, the Dynamic Programming reverse recursion stage obtains according to dynamic programming principle

In formula, f_t(e_t) be from t-th of scheduling slot end to finishing scheduling when maximum return, set f_T(E_T)=0, by battery energy storage (14) are brought in the electric quantity balancing constraint (2) of system into, obtain the Dynamic Programming cost-to-go for needing to solve repeatedly in each period Function:

(15) formula is by (2) in solution procedure, and (3), (5), (15) formula is defined as dynamic and advised by the constraint of (7) and (13) formula Reverse recursion stage center center problem is drawn, the Optimized model of the core subproblem is by (2), (3), (5), (7), (13) and (15) Formula is constituted；

F, substitution of variable is carried out to (15) formula, by decision content q_t+1It is expressed as x, e_tIt is expressed as y, maximum return f_t(e_t) it is expressed as h (x), while by f_t+1(e_t-L_t+1-q_t+1) translation L_t+1It obtains function g (y-x), Dynamic Programming reverse recursion stage center center is asked The Optimized model equivalent transformation of topic is；

In formula, a and b respectively indicate the t+1 period discharge and recharge q determined jointly by (3) and (5) formula_t+1Bound, c and d Respectively indicate the t period end battery energy storage system day part storing electricity e determined jointly by (7) and (13) formula_tBound, According to (2) formula, α and β actually respectively indicate battery energy storage system in the bound of t+1 period end storing electricity；

According to the geometric meaning of (16) formula, domain is split into several sub-districts by the segmentation bend line of function h (x) and g (y-x) Domain is able to demonstrate that optimal value must be taken in boundary in each parallelogram or similar to the subregion of parallelogram , functional value is all linear change relative to parameter x, i.e., straight line where boundary can only about the segmentation turning point of functional value It obtains in the endpoint of parallelogram subregion, is ranked up the endpoint of each parallelogram subregion according to each endpoint y value, The feasible zone of problem is divided into N number of horizontal subregion by the horizontal line where adjacent endpoint, is successively sought from lower n=1 to upper n=N Look for optimal path in each horizontal subregion；

G, horizontal subregion optimal path solves, and since n=1, there is 90 ° of vertical lines of L item and 45 ° of oblique lines in horizontal subregion, Corresponding is the segmentation bend line of function h (x) He g (y-x), solves this these segmentation bend line and horizontal subregion or more Boundary intersection functional value is respectively z_l(y_u) and z_l(y_d), find lower boundary functional value maximum point z_l'(y_d) and corresponding path l', Judge the functional value z of these paths and coboundary intersection point_l'(y_u) whether be all coboundary intersection point functional values maximum point, if It is not then to illustrate the jump for occurring optimal path in the area, executes step H；If it is, the paths are this water Maximum path in flat subregion, judges whether n > N, if it is, t period core subproblem, which solves process, to terminate, executes Step I continues to solve the optimal path for closing on subregion above the horizontal subregion if it is not, then setting n=n+1；

H, optimal path jump refers to that the maximum point of the functional value of lower boundary is risen by initial optimal path, and in certain turnover Point at jump to an other optimal path continue rise then reach the maximum point of upper boundary function value, using l' as benchmark Path initializes l'=l'_s, s=1, calculate earliest with the intersection point and slope of l' intersecting paths, obtain its intersecting paths l'_s+1's Turning point coordinate (y_s+1, z_l's+1(y_s+1)) and l'_s+1With the functional value z of coboundary intersection point_l's+1(y_u), judge z_l's+1(y_u) be No is coboundary maximum value, if it is not, s=s+1 is enabled, it will be by l'_s+1As new reference path, H is re-executed；If so, Then export each turning point coordinate (y_s+1, z_l's+1(y_s+1)) and z_l'(y_d),z_l's(y_u) value, obtain optimal path l'_s, s=1, 2 ..., S, S are optimal path sum, which terminates, and judge whether n > N, if it is, t Period core subproblem, which solves process, to terminate, and executes step I and continues to solve the horizontal subregion if it is not, then setting n=n+1 The optimal path for closing on subregion of top；

I, t=t-1 is set, if judgement t > 0, returns to step E, continues the core subproblem for solving the t period, if t=0, Execute step J；

J, the Dynamic Programming forward recursive stage enablesTo t=1,2 ..., T is successively calculated:

In formula, q_t ^*For the optimal charge/discharge electricity amount of battery energy storage system t period,For the battery energy storage system of t period Optimal storage electricity, L_tFor the dead loss electricity of known t period battery energy storage system；By (3)-when (17) formula of solution (6) formula constrains, according to dynamic programming principle, solve (3)-(6) and maximization problem acquisition that (17) formula defines from the 1st period To the discharge and recharge of the battery energy storage system of T periodAnd corresponding charge-discharge electric powerIt is exactly battery energy storage The optimal solution of system call problem；

J, development enters subsequent period at any time, when updating the initial storage electricity of battery energy storage system, finishing scheduling in step B Target storing electricity, the constraint of day part electricity volume and timesharing forecasted electricity market price or equivalent value multiplier, continue to execute step Rapid C.

Technical field

The present invention relates to battery energy storage technical field more particularly to a kind of battery energy storage systems for considering dynamic efficiency for charge-discharge System dispatching method.

Background technique

Renewable energy constantly quickly accesses power grid, and uncertainty proposes newly the control and management of electric system Challenge.In recent years, energy storage has become new industrial hot spot.By the extensive use of energy-storage system, power can not only be stabilized Fluctuation improves power quality, enhances network system reliability of operation.

Under normal circumstances, the efficiency for charge-discharge of energy storage is with charge-discharge electric power dynamic change.Such as lead-acid battery and lithium electricity Energy-storage system state-of-charge (SOC) in electric discharge in pond is in 20% to 95%, and voltage is not substantially within this range, when electric discharge Become, efficiency can decrease to some degree with the increase of discharge current.Charging process is considered as the inverse process of electric discharge, efficiency change Similar with electric discharge, when deviateing declared working condition, loss increases, and efficiency reduces.

Existing battery energy storage system dispatching method is usually the prediction data according to following one day at present, formulates the fortune of energy storage Row plan, is modified according to set amount when actual motion deviates operational plan, optimizes the on-road efficiency of battery energy storage system. But be generally used general business optimization software when solving the scheduling model of battery energy storage system at present and solved, it lacks Rare specific aim, adaptable dedicated derivation algorithm.

Therefore, there is an urgent need to have a kind of tailor-made algorithm that can be directed to scheduling model self structure characteristic can be realized at present The Efficient Solution of scheduling problem.

Summary of the invention

The object of the present invention is to provide a kind of battery energy storage system dispatching method for considering dynamic efficiency for charge-discharge, the present invention Method provides a kind of effective way for the optimization operation of battery energy storage system, has engineering practical value.

To achieve the above object, the technical solution adopted by the present invention the following steps are included:

A kind of battery energy storage system dispatching method considering dynamic efficiency for charge-discharge can formulate battery storage with this method The optimization operational decisions of energy system, improve the performance driving economy of battery energy storage system, method includes the following steps:

A, capacity, dynamic charge-discharge electric power and relationship between efficiency, charge-discharge electric power limitation and the electricity of battery energy storage system are determined Pond energy storage system capacity bound selects suitable segments to the charge and discharge electric work of battery energy storage system according to the required precision of scheduling Rate-relationship between efficiency carries out piece-wise linearization；

B, target storing electricity, day part when obtaining initial storage electricity, the finishing scheduling of battery energy storage system are surfed the Internet Constraint and timesharing forecasted electricity market price or equivalent value multiplier；

C, the scheduling model of battery energy storage system is established, when the optimization aim of scheduling model is that formulation battery energy storage system is each Section charge-discharge electric power decision, maximizes total gene-ration revenue of battery energy storage system, the objective function of scheduling model are as follows:

In formula, T is scheduling slot sum, and t is scheduling slot number, λ_tFor the t period forecasted electricity market price or wait potency It is worth multiplier, q_tFor the charge/discharge electricity amount of the battery energy storage system of t period, discharges for positive value, be charged as negative value, p (q_t) t when The charge-discharge electric power of the battery energy storage system of section is q_tUnitary piecewise linear function, δ be scheduling slot length；

The constraint condition of scheduling model includes:

1) the electric quantity balancing constraint of battery energy storage system

In formula, e_tFor the storing electricity of t period end battery energy storage system, L_tFor known t period battery energy storage system Dead loss electricity, be negative value；

2) electricity volume constrains

In formula,WithThe respectively electricity volume bound of t period battery energy storage system；

3) the charge-discharge electric power constraint of battery energy storage system

In formula, P^maxAnd P^minThe respectively upper physical limit of battery energy storage system discharge power and charge power, function p are single The piecewise linear function of tune, charge-discharge electric power constraint (4) can be converted to the constraint of discharge and recharge:

In formula, function p' is the inverse function of function p, p'(P^max) and p'(P^min) it is respectively battery energy storage system discharge electricity amount With the upper physical limit of charge capacity；

4) electricity SOC is constrained

In formula, E^maxAnd E^minThe respectively bound of the storing electricity of battery energy storage system；

5) the first last Constraint of battery energy storage system

e₀=E⁰,e_T=E^T (7)

In formula, E⁰And E^TTarget electricity when being battery energy storage system initial time electricity and finishing scheduling respectively, e₀And e_T Respectively dispatch the storing electricity of the battery energy storage system of start time and T period Mo；

By (1), (2), (3), (5), (6), (7) formula constitutes the scheduling model of battery energy storage system；

D, the dynamic bound constraint for calculating battery energy storage system day part storing electricity, according to scheduling last battery energy storage system Target storing electricity E when finishing scheduling of uniting^T(2) and (5) are constrained with battery energy storage system day part storing electricity bound, are passed through The reverse recursion since the T period calculates the dynamic bound of battery energy storage system day part storing electricity are as follows:

In formula, j is time variable,WithWhen the t period end battery energy storage system of respectively reverse-direction derivation acquisition is each The dynamic bound of section storing electricity, E^TThe target storing electricity of battery energy storage system, L when for finishing scheduling_jFor known jth The dead loss electricity of period battery energy storage system is negative value, p'(P^max) and p'(P^min) it is respectively battery energy storage system electric discharge The upper physical limit of electricity and charge capacity,WithThe respectively bound of jth period battery energy storage system electricity volume；

According to the initial quantity of electricity E of battery energy storage system⁰(2) are constrained with battery energy storage system day part storing electricity bound (5), the period carries out forward recursive since scheduling, calculates the dynamic bound of battery energy storage system day part storing electricity Are as follows:

In formula, j is time variable,WithThe respectively positive battery energy storage system for deriving the t period Mo obtained is each The dynamic bound of period storing electricity, E⁰The storing electricity of battery energy storage system, L when starting for scheduling_jWhen for known jth The dead loss electricity of section battery energy storage system, p'(P^max) and p'(P^min) it is respectively battery energy storage system discharge electricity amount and charging The upper physical limit of electricity,WithThe respectively bound of jth period battery energy storage system electricity volume；

Forward direction is derived and reverse-direction derivation result combines, calculates the dynamic bound of the storing electricity of battery energy storage system Are as follows:

In formula,And e_tIt respectively calculates in the dynamic of battery energy storage system day part storing electricity of the t period of acquisition Lower limit obtains the bound of new battery energy storage system day part storing electricity in conjunction with (6) formula are as follows:

In formula, e_tFor the storing electricity of t period end battery energy storage system, E^maxAnd E^minRespectively battery energy storage system The physics bound of storing electricity；

By equivalent transformation, by (1), (2), (3), (5), (7), (13) formula constitutes the scheduling mould of battery energy storage system Type；

E, the Dynamic Programming reverse recursion stage obtains according to dynamic programming principle

In formula, f_t(e_t) be from t-th of scheduling slot end to finishing scheduling when maximum return, set f_T(E_T)=0, will be electric (14) are brought in the electric quantity balancing constraint (2) of pond energy-storage system into, obtain the Dynamic Programming cost- for needing to solve repeatedly in each period To-go function:

(15) formula is by (2) in solution procedure, and (3), (5), (15) formula is defined as moving by the constraint of (7) and (13) formula State plans reverse recursion stage center center problem, and the Optimized model of the core subproblem is by (2), (3), (5), (7), (13) and (15) formula is constituted；

F, substitution of variable is carried out to (15) formula, by decision content q_t+1It is expressed as x, e_tIt is expressed as y, maximum return f_t(e_t) indicate For h (x), while by f_t+1(e_t-L_t+1-q_t+1) translation L_t+1It obtains function g (y-x), core in the Dynamic Programming reverse recursion stage The Optimized model equivalent transformation of subproblem is；

In formula, a and b respectively indicate the t+1 period discharge and recharge q determined jointly by (3) and (5) formula_t+1Bound, c The t period end battery energy storage system day part storing electricity e determined jointly by (7) and (13) formula is respectively indicated with d_tUp and down Limit, according to (2) formula, α and β actually respectively indicate battery energy storage system in the bound of t+1 period end storing electricity；

According to the geometric meaning of (16) formula, domain is split into several sons by the segmentation bend line of function h (x) and g (y-x) Region is able to demonstrate that optimal value must be in boundary in each parallelogram or similar to the subregion of parallelogram It obtains, functional value is all linear change relative to parameter x, i.e. segmentation turning point of the straight line where boundary about functional value It can obtain in the endpoint of parallelogram subregion, be arranged the endpoint of each parallelogram subregion according to each endpoint y value The feasible zone of problem is divided into N number of horizontal subregion by sequence, the horizontal line where adjacent endpoint, successively from lower n=1 to upper n=N Find optimal path in each horizontal subregion；

G, horizontal subregion optimal path solves, and since n=1, has 90 ° of vertical lines of L item and 45 ° in horizontal subregion tiltedly Line, corresponding is the segmentation bend line of function h (x) He g (y-x), solves this these segmentation bend line and horizontal subregion Up-and-down boundary intersection point functional value is respectively z_l(y_u) and z_l(y_d), find lower boundary functional value maximum point z_l'(y_d) and corresponding road Diameter l' judges the functional value z of these paths and coboundary intersection point_l'(y_u) whether be all coboundary intersection point functional values maximum Point executes step H if it is not, then the jump of optimal path occurs in the area in explanation；If it is, the paths It is the maximum path in this horizontal subregion, judges whether n > N, if it is, t period core subproblem solves process knot Beam executes step I if it is not, then setting n=n+1 and continues to solve the optimal road for closing on subregion above the horizontal subregion Diameter；

H, optimal path jump refers to that the maximum point of the functional value of lower boundary is risen by initial optimal path, and at certain Jumped at turning point an other optimal path continue rise then reach the maximum point of upper boundary function value, using l' as Reference path initializes l'=l'_s, s=1, calculate earliest with the intersection point and slope of l' intersecting paths, obtain its intersecting paths l'_s+1Turning point coordinate (y_s+1, z_l's+1(y_s+1)) and l'_s+1With the functional value z of coboundary intersection point_l's+1(y_u), judge z_l's+1 (y_u) it whether is coboundary maximum value, if it is not, s=s+1 is enabled, it will be by l'_s+1As new reference path, H is re-executed； If it is, exporting each turning point coordinate (y_s+1, z_l's+1(y_s+1)) and z_l'(y_d),z_l's(y_u) value, obtain optimal path l'_s, s =1,2 ..., S, the horizontal subregion optimal path solution terminate, and judge whether n > N, if it is, t period core is asked Topic, which solves process, to be terminated, and is executed step I and is continued to solve above the horizontal subregion and close on son if it is not, then setting n=n+1 The optimal path in region；

I, t=t-1 is set, if judgement t > 0, returns to step E, continues the core subproblem for solving the t period, if t =0, execute step J；

J, the Dynamic Programming forward recursive stage enablesTo t=1,2 ..., T is successively calculated:

In formula, q_t ^*For the optimal charge/discharge electricity amount of battery energy storage system t period,For the battery energy storage system of t period The optimal storage electricity of system, L_tFor the dead loss electricity of known t period battery energy storage system；Solve (17) formula when by (3)-(6) formula constraint, according to dynamic programming principle, solve (3)-(6) and maximization problem acquisition that (17) formula defines from the 1st The discharge and recharge decision of the battery energy storage system of period to T periodAnd corresponding charge-discharge electric powerIt is exactly The optimal solution of battery energy storage system scheduling problem；

J, development enters subsequent period at any time, updates the initial storage electricity of battery energy storage system in step B, scheduling knot Target storing electricity, the constraint of day part electricity volume and timesharing forecasted electricity market price or equivalent value multiplier when beam, continue to hold Row step C.

Compared to the prior art compared with the present invention has following advantage:

The present invention has carried out point the nonlinear efficiency for charge-discharge characteristic curve of battery energy storage system when establishing model Section linearisation, establishes mixed integer linear programming model, more more accurate than traditional model；In addition, solving MIXED INTEGER Generally all it can only just can be carried out Efficient Solution using commercial packages such as CPLEX when linear programming model, first choice of the present invention passes through Step D calculates the dynamic bound constraint of battery energy storage system day part storing electricity, reduces the feasible zone range of understanding, reduces Solve the time；Secondly, the geometry characteristic for the core subproblem for needing to solve repeatedly for Dynamic Programming in step F, In Proposed in step G and step H can Efficient Solution, while can fully ensure that the fining of the numerical stability of calculating process Algorithm, so that business software can be substituted when solving this kind of battery energy storage system scheduling problem.

Detailed description of the invention

Fig. 1 is entire dispatching algorithm flow chart.

Fig. 2 is cell discharge efficiency-power curve.

Fig. 3 is the geometry feature of dynamic programming algorithm center center problem.

Fig. 4 is processing " optimal path jump " submethod schematic diagram.

Fig. 5 is processing " optimal path jump " submethod algorithm flow chart.

Specific embodiment

The present invention is intended to provide a kind of battery energy storage system dispatching method for considering dynamic efficiency for charge-discharge characteristic, it can basis The market guidance of battery energy storage self-characteristic, operation constraint and prediction formulates battery energy storage system operation in following a period of time Optimizing decision improves running efficiency of system and operation income so that battery energy storage system be instructed to run.

The embodiment and whole flow process of the dispatching method in 1 pair of invention are described further with reference to the accompanying drawing.

A kind of battery energy storage system dispatching method considering dynamic efficiency for charge-discharge can formulate energy storage electricity with this method The optimization operational decisions of cell system, improve the performance driving economy of battery energy storage system, method includes the following steps:

A, the limitation of capacity, dynamic efficiency for charge-discharge, charge-discharge electric power and the battery energy storage system of battery energy storage system are determined Capacity bound carries out piece-wise linearization, root to charge-discharge electric power-relationship between efficiency figure of battery energy storage system referring to attached drawing 2 Suitable segments is selected according to required precision.

B, target storing electricity, day part when obtaining initial storage electricity, the finishing scheduling of battery energy storage system are surfed the Internet Constraint and timesharing forecasted electricity market price or equivalent value multiplier.

C, the scheduling model of battery energy storage system is established, when the optimization aim of scheduling model is that formulation battery energy storage system is each Section charge-discharge electric power decision, maximizes total gene-ration revenue of battery energy storage system, the objective function of model are as follows:

In formula, T is scheduling slot sum, and t is scheduling slot number, λ_tFor the forecasted electricity market price or equivalent value of t period Multiplier, q_tFor the charge/discharge electricity amount of the battery energy storage system of t period, discharges for positive value, be charged as negative value, p (q_t) the t period Battery charging and discharging power, be q_tUnitary piecewise linear function, δ be scheduling slot length.

Scheduling model constraint condition includes:

1) the electric quantity balancing constraint of battery energy storage system

E in formula_tFor the storing electricity of t period end battery energy storage system, L_tFor known t period battery energy storage system Storage dead loss electricity, be negative value.

2) electricity volume constrains

In formulaWithThe respectively electricity volume bound of t period battery energy storage system.

3) the charge-discharge electric power constraint of battery energy storage system

P in formula^maxAnd P^minThe respectively upper physical limit of battery energy storage system discharge power and charge power, function p are single The piecewise linear function of tune, charge-discharge electric power constraint (4) can be converted to the constraint of discharge and recharge:

Function p' is the inverse function of function p, p'(P in formula^max) and p'(P^min) it is respectively battery energy storage system discharge electricity amount With the upper physical limit of charge capacity.

4) electricity (SOC) constrains

E in formula^maxAnd E^minThe respectively bound of the storing electricity of battery energy storage system.

5) the first last Constraint of battery energy storage system

e₀=E⁰,e_T=E^T (7)

E in formula⁰And E^TTarget electricity when being battery energy storage system initial time electricity and finishing scheduling respectively, e₀And e_T Respectively dispatch the storing electricity of the battery energy storage system of start time and T period Mo.

By (1), (2), (3), (5), (6), (7) formula constitutes the scheduling model of battery energy storage system.

D, the dynamic bound constraint for calculating battery energy storage system day part storing electricity, according to scheduling last battery energy storage system Target storing electricity E when finishing scheduling of uniting^T(2) and (5) are constrained with battery energy storage system day part storing electricity bound, are passed through The reverse recursion since the T period calculates the dynamic bound of battery energy storage system day part storing electricity are as follows:

J is time variable in formula,WithThe respectively t period end battery energy storage system day part of reverse-direction derivation acquisition The dynamic bound of storing electricity, E^TThe target storing electricity of battery energy storage system, L when for finishing scheduling_jWhen for known jth The dead loss electricity of section battery energy storage system, p'(P^max) and p'(P^min) it is respectively battery energy storage system discharge electricity amount and charging The upper physical limit of electricity,WithThe respectively bound of jth period battery energy storage system electricity volume.

According to the initial quantity of electricity E of battery energy storage system⁰(2) are constrained with battery energy storage system day part storing electricity bound (5), the period carries out forward recursive since scheduling, calculates the dynamic bound of battery energy storage system day part storing electricity Are as follows:

J is time variable in formula,WithWhen the battery energy storage system of the respectively positive t period Mo for deriving acquisition is each The dynamic bound of section storing electricity, E⁰The storing electricity of battery energy storage system, L when starting for scheduling_jFor the known jth period The dead loss electricity of battery energy storage system is negative value, p'(P^max) and p'(P^min) it is respectively battery energy storage system discharge electricity amount With the upper physical limit of charge capacity,WithThe respectively bound of jth period battery energy storage system electricity volume.

Forward direction is derived and reverse-direction derivation result combines, calculates the dynamic bound of the storing electricity of battery energy storage system Are as follows:

In formulaWithe _tIt respectively calculates in the dynamic of battery energy storage system day part storing electricity of the t period of acquisition Lower limit obtains the bound of new battery energy storage system day part storing electricity in conjunction with (6) formula are as follows:

In formula, e_tFor the storing electricity of t period end battery energy storage system, E^maxAnd E^minRespectively battery energy storage system The physics bound of storing electricity；

By equivalent transformation, by (1), (2), (3), (5), (7), (13) formula constitutes the scheduling mould of battery energy storage system Type.

E, the Dynamic Programming reverse recursion stage obtains according to dynamic programming principle

F in formula_t(e_t) be from t-th of scheduling slot end to finishing scheduling when maximum return.Set f_T(E_T)=0, by battery (14) are brought in the electric quantity balancing constraint (2) of energy-storage system into, obtain the Dynamic Programming cost- for needing to solve repeatedly in each period To-go function:

(15) formula is by (2) in solution procedure, and (3), (5), (15) formula is defined as moving by the constraint of (7) and (13) formula State plans reverse recursion stage center center problem, and the Optimized model of the core subproblem is by (2), (3), (5), (7), (13) and (15) formula is constituted.

F, substitution of variable is carried out to (15) formula, by decision content q_t+1It is expressed as x, e_tIt is expressed as y, maximum return f_t(e_t) indicate For h (x), while by f_t+1(e_t-L_t+1-q_t+1) translation L_t+1It obtains function g (y-x), core in the Dynamic Programming reverse recursion stage The Optimized model equivalent transformation of subproblem are as follows:

In formula, a and b respectively indicate the t+1 period discharge and recharge q determined jointly by (3) and (5) formula_t+1Bound, c The t period end battery energy storage system day part storing electricity e determined jointly by (7) and (13) formula is respectively indicated with d_tUp and down Limit, according to (2) formula, α and β actually respectively indicate battery energy storage system in the bound of t+1 period end storing electricity.

Referring to attached drawing 3, according to the geometric meaning of core subproblem Optimized model, the segmentation bend line of function h (x) is x= x_θ, θ=1,2,3... Φ, Φ are that function h (x) is segments, 90 ° of numerical value chain-dotted lines in corresponding diagram；The segmentation of g (y-x) turns Broken line is y-x=ρ_ν, v=1,2,3.....V, V are the segments of function g (y-x), the empty oblique line of 45 ° in corresponding diagram.Function h (x) and domain is split into several subregions by the segmentation bend line of g (y-x), in each parallelogram or is similar to parallel The subregion of quadrangle is able to demonstrate that optimal value must be obtained in boundary, and functional value is all linearly to become relative to parameter x Change, i.e., straight line where boundary can only be obtained about the segmentation turning point of functional value in the endpoint of parallelogram subregion, root The endpoint of each parallelogram subregion is ranked up according to each endpoint y value, the horizontal line where adjacent endpoint can by problem Row domain is divided into N number of horizontal subregion, and optimal path in each horizontal subregion is successively found from lower n=1 to upper n=N.

G, horizontal subregion optimal path solves, and referring to attached drawing 4, since n=1,90 ° of L item is shared in horizontal subregion Vertical line and 45 ° of oblique lines, corresponding is the segmentation bend line of function h (x) He g (y-x), solves this these segmentation bend line Z is distinguished with horizontal subregion up-and-down boundary intersection point functional value_l(y_u) and z_l(y_d), find lower boundary functional value maximum point z_l'(y_d) And corresponding path l', judge the functional value z of these paths and coboundary intersection point_l'(y_u) it whether is all coboundary intersection point functions The maximum point of value executes step H if it is not, then the jump of optimal path occurs in the area in explanation；If it is, The paths are the maximum paths in this horizontal subregion, judge whether n > N, if it is, t period core subproblem is asked Solution process terminates, and executes step I and continues to solve above the horizontal subregion and close on subregion if it is not, then setting n=n+1 Optimal path.

H, optimal path jump refers to that the maximum point of the functional value of lower boundary is risen by initial optimal path, and at certain Jumped at turning point an other optimal path continue rise then reach the maximum point of upper boundary function value；Referring to attached drawing 5, the thought for handling optimal path " jump " method is illustrated by a simple case: upper following in recording level subregion respectively The functional value of boundary and all l paths intersection points is respectively z_l(y_u) and z_l(y_d), it can obtain corresponding in zoy plane upper pathway l Linear function z_l(y), from the maximum a point of lower boundary functional value, the l before intersecting with other paths₁For optimal path, and When itself and path l₃Intersection is after b point, due to l₃Climbing speed be greater than former optimal path l₁, therefore the optimal road since b point Diameter becomes l₃, similarly become l again in c point optimal path₄Until reaching the maximum d point of upper boundary function value；Detailed calculation process ginseng See attached drawing 4, using l' as reference path, initializes l'=l'_s, s=1, calculate earliest with the intersection point and slope of l' intersecting paths, Obtain its intersecting paths l'_s+1Turning point coordinate (y_s+1, z_l's+1(y_s+1)) and l'_s+1With the functional value of coboundary intersection point z_l's+1(y_u), judge z_l's+1(y_u) it whether is coboundary maximum value, if it is not, s=s+1 is enabled, it will be by l'_s+1As new base Quasi- path, re-executes H；If it is, exporting each turning point coordinate (y_s+1, z_l's+1(y_s+1)) and z_l'(y_d),z_l's(y_u) value, Obtain optimal path l'_s, s=1,2 ..., S, which, which solves, terminates, judge whether n > N, if so, Then t period core subproblem solves process and terminates, and executes step I and continues to solve the level if it is not, then setting n=n+1 The optimal path for closing on subregion above subregion.

I, t=t-1 is set, if judgement t > 0, returns to step E, continues the core subproblem for solving the t stage, if t =0, execute step J；

J, the Dynamic Programming forward recursive stage enablesTo t=1,2 ..., T is successively calculated:

In formula, q_t ^*For the optimal charge/discharge electricity amount of battery energy storage system t period,For the battery energy storage system of t period The optimal storage electricity of system, L_tFor the dead loss electricity of known t period battery energy storage system；Solve (17) formula when by (3)-(6) formula constraint, according to dynamic programming principle, solve (3)-(6) and maximization problem acquisition that (17) formula defines from the 1st The discharge and recharge decision of the battery energy storage system of period to T periodAnd corresponding charge-discharge electric powerIt is exactly The optimal solution of battery energy storage system scheduling problem.

J, development enters subsequent period at any time, updates the initial storage electricity of battery energy storage system in step B, scheduling knot Target storing electricity, the constraint of day part electricity volume and timesharing forecasted electricity market price or equivalent value multiplier when beam, continue to hold Row step C.