Energy storage distribution robust optimization configuration method considering flexibility and uncertainty

Abstract

The invention discloses an energy storage distribution robust optimization configuration method considering flexibility and uncertainty, and belongs to the field of novel power system energy storage optimization configuration. The method comprises the steps of collecting wind and light historical data of a wind and light unit in a power system, and constructing a wind and light output uncertainty set according to the wind and light historical data and based on a mixed norm; constructing a target function according to the energy storage equivalent annual investment cost and the annual operation cost of the power system, and obtaining a constraint condition; establishing an energy storage distribution robust optimization configuration model based on scene probability driving through the objective function, the constraint condition and the wind and light output uncertainty set; and solving the energy storage distribution robust optimization configuration model through a column sum constraint generation algorithm to obtain an optimal energy storage planning scheme, and performing energy storage of the power system according to the optimal energy storage planning scheme. The energy storage configuration scheme of the optimal configuration model is low in conservative property and can give consideration to both flexibility and economy.

Description

Technical Field

[0001] The present invention relates to the field of power system flexibility and energy storage optimization configuration, and in particular to a method for optimizing energy storage distribution configuration taking flexibility and uncertainty into consideration. Background Art

[0002] Over the past half century of social development, electricity demand has grown rapidly. The large-scale use of traditional primary energy sources has led to environmental pollution and the gradual depletion of traditional primary energy resources. To address the challenges posed by carbon emissions, it is necessary to promote energy structure and industrial transformation, reduce dependence on fossil fuels, and develop clean and reliable energy sources. However, renewable energy sources such as wind power and photovoltaic power generation are significantly affected by external weather conditions, resulting in high uncertainty and intermittent output. Traditional thermal power generators, due to their slow regulation response and high cost, often struggle to effectively cope with the random output fluctuations of wind and photovoltaic power. This can lead to real-time power imbalances in the power system and potentially cause a series of problems such as ineffective investment and curtailed wind and solar power generation.

[0003] The output of wind and solar power in new power systems is highly uncertain, so the optimal allocation of energy storage must take this uncertainty into account. Furthermore, the uncertainty and intermittency of wind and solar power pose challenges to balancing power system flexibility. The flexibility provided by conventional generators is no longer sufficient to meet the flexibility demands of new power systems. Therefore, in the energy transition focused on renewable energy, it is necessary to study system flexibility and uncertainty, and to develop robust optimization methods for energy storage allocation within the power system that account for this flexibility and uncertainty. Summary of the Invention

[0004] Aiming at the problem of optimizing energy storage configuration to improve the flexibility of the power system and the uncertainty of wind and solar power, the present invention proposes a robust optimization configuration method for energy storage distribution that takes flexibility and uncertainty into consideration.

[0005] The present invention adopts the following technical solutions:

[0006] The present invention discloses a method for optimizing configuration of energy storage distribution rods considering flexibility and uncertainty, comprising the following steps:

[0007] 1) Collect historical wind and solar power data of wind and solar power units in the power system, and construct a wind and solar power output uncertainty set based on the historical wind and solar power data and the hybrid norm;

[0008] 2) constructing an objective function based on the annual investment cost and annual operating cost of energy storage in the power system, and using energy storage investment and construction constraints, thermal power unit operation constraints, power balance constraints, energy storage operation constraints, power flow constraints, wind and solar output constraints, and flexibility demand matching coefficient constraints as constraints; establishing a scenario probability-driven energy storage distributed robust optimization configuration model based on the objective function, the constraints, and the wind and solar output uncertainty set;

[0009] 3) Solving the energy storage distribution robust optimization configuration model through a column and constraint generation algorithm to obtain an optimal energy storage planning scheme, and performing energy storage for the power system according to the optimal energy storage planning scheme.

[0010] Furthermore, the specific steps of step 3) are as follows:

[0011] Initialize the lower bound, upper bound and number of iterations;

[0012] According to the current number of iterations, the main problem of the energy storage distribution and rod optimization configuration model is solved to obtain the optimal solution. And update the current lower bound to in, is the decision vector for the main problem; η * The relaxation vector of the main problem; A is the constraint matrix of the main problem; then solve the sub-problems of the energy storage distribution and optimization model, obtain the probability distribution under the worst case, and update the current upper bound;

[0013] If the iteration termination condition is met, the iteration is terminated, the probability distribution under the worst case scenario is returned, and the optimal energy storage planning scheme is obtained; wherein, the iteration termination condition is that the difference between the current upper bound and the current lower bound is less than a preset error; if the iteration termination condition is not met, the current number of iterations is updated according to a preset rule, and the main problem and subproblems of the distributed robust programming model are solved respectively according to the current number of iterations to obtain the probability distribution under the worst case scenario, and the current lower bound and the current upper bound are updated until the iteration termination condition is met to obtain the optimal energy storage planning scheme.

[0014] Compared with the prior art, the present invention has the following beneficial effects:

[0015] This invention proposes a robust optimization method for energy storage allocation that considers flexibility and uncertainty. System verification results show that the invention's construction of a wind and solar power output uncertainty set based on a hybrid norm can realistically depict wind and solar power uncertainty scenarios, thereby improving the economic efficiency of energy storage configuration. The energy storage configuration scheme of the proposed model is less conservative than traditional robust optimization, with a flexibility demand matching coefficient of 0.972, balancing flexibility and economy. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a model framework diagram of the energy storage distribution rod optimization configuration method of the present invention;

[0017] Figure 2 This is a flow chart for solving the energy storage distribution rod optimization configuration model of the present invention. DETAILED DESCRIPTION

[0018] The present invention will be further described and illustrated below in conjunction with specific embodiments. The embodiments are merely illustrative of the present disclosure and do not limit its scope. The technical features of the various embodiments of the present invention may be combined accordingly, provided that there is no conflict between them.

[0019] The present invention aims to find the most economically optimal configuration of energy storage units in a power system. The configuration includes the energy storage unit's connection location within the power system, its energy storage capacity, and its power. In this embodiment, the power system includes an energy storage unit (to be configured), a wind and solar power unit, a thermal power unit, and power loads.

[0020] The present invention considers the flexibility and uncertainty of the energy storage distribution rod optimization configuration method, and its specific implementation method is as follows Figure 1 As shown, the following steps are included:

[0021] Step 1: Construct the wind and solar power output uncertainty set. Collect historical wind and solar power data from wind and solar power units in the power system, and construct the wind and solar power output uncertainty set based on the historical wind and solar power data and the hybrid norm.

[0022] There is a certain deviation between the actual probability distribution P of wind power and the initial distribution P0 formed by clustering historical wind power data. The L1 norm is used to represent the total uncertainty of the probability of wind power output scene. ∞ The norm is used to represent the maximum fluctuation degree of uncertainty in wind and solar power output scenarios.

[0023] In a specific embodiment of the present invention, constructing a wind power output uncertainty set based on wind power historical data and based on a hybrid norm includes: clustering a limited number of typical wind power scenes and a probability distribution of each typical wind power scene in multiple historical scenes according to wind power historical data; and according to the probability distribution of each typical wind power scene, and through L1-norm and L ∞ The -norm takes into account the uncertainty of wind and solar power and constructs the wind and solar power output uncertainty set.

[0024] The wind and solar power output uncertainty set includes the probability distribution uncertainty set based on L1-norm and the probability distribution uncertainty set based on L ∞ -norm probability distribution of an uncertain set.

[0025] Based on L1-norm and L ∞The mathematical expressions of the uncertainty set of the probability distribution of the -norm are:

[0026]

[0027]

[0028] Among them, Ω1 is the probability distribution uncertainty set based on L1-norm; ||*||1 represents L1-norm; γ1 and γ ∞ are L1-norm and L ∞ The uncertainty in the -norm reflects the degree of fluctuation of the probability distribution; s is the probability of occurrence of the sth typical scenery scene in the scenery scene set; represents the probability of the sth typical scenery scene occurring at the initial moment; K is the number of typical scenery scenes in the typical scenery scene set; Ω ∞ Based on L ∞ -norm probability distribution of an uncertain set.

[0029] The L1-norm is used to represent the total uncertainty of the probability of wind and solar output scenarios, while L ∞ -norm is used to express the maximum fluctuation degree of uncertainty in wind and solar output scenario. ∞ The confidence level of the probability distribution of the -norm satisfies the following formula:

[0030]

[0031] Where Pr{·} is the probability that the inequality holds true; and are L1-norm and L ∞ -norm uncertainty probability confidence. Let Thus, the uncertainty of L1-norm γ1, L ∞ -Uncertainty of the norm γ ∞ , the number of historical scenery scenes Z, the number of typical scenery scenes K, and α1 and α ∞ The relationship between them is:

[0032]

[0033] From the above formula, we can see that when there is enough historical data of wind and solar power output, the uncertainty γ1 and γ ∞ The value of approaches 0, and the constructed wind and solar power output uncertainty set will approach the actual distribution of wind and solar power output scenarios.

[0034] Step 2: Construct a distributed robust optimization model for energy storage. Construct an objective function based on the annual investment cost and annual operating cost of energy storage in the power system, and use energy storage configuration construction constraints, thermal power unit operation constraints, power balance constraints, energy storage operation constraints, power flow constraints, wind and solar output constraints, and flexibility demand matching coefficient constraints as constraints. Establish a scenario probability-driven distributed robust optimization model for energy storage using this objective function, these constraints, and the wind and solar output uncertainty set.

[0035] 2.1) Objective Function

[0036] The energy storage distribution and optimization model takes the minimum sum of the annual investment cost and annual operating cost of the power system's energy storage as the objective function, and its mathematical expression is:

[0037]

[0038] Where C is the sum of the annual investment cost of energy storage and the annual operating cost of the power system, which is the annual comprehensive cost of the power system; C inv is the annual investment cost of energy storage in the power system; C oper is the daily operating cost of the power system; σ is the equivalent factor of the annual investment cost, which is 365; p z is the probability distribution of the initial scene of scenery; Ω is the set of probability distributions of the initial scene of scenery.

[0039] 2.1.1) Annual investment cost of energy storage

[0040] Annual investment cost of energy storage in the power system C inv The expression is:

[0041]

[0042] Where α is the energy storage cost coefficient; k is the node number; C inv is the annual investment cost of energy storage; Ω ESS is the energy storage set to be configured; r is the energy storage discount rate; y is the number of years of energy storage discount; C ESS and C p is the unit cost of energy storage capacity and power to be configured; E k P is the energy storage capacity to be configured on node k; k is the energy storage power to be configured on node k.

[0043] 2.1.2) Daily operating costs of the power system

[0044] Daily operating cost of the power system C oper The expression is:

[0045] C oper =Coperation,G +C lack

[0046] Among them, C operation,G is the operating cost of the thermal power unit; C lack Penalty for lack of flexibility.

[0047] Operating cost of thermal power unit C operation,G The formula is:

[0048]

[0049] Among them, g is the number of thermal power unit; G is the set of thermal power unit numbers; T is the maximum time; t is the time number; r is the segment number; R is the maximum segment number; p r is the slope corresponding to the rth segment after linearization; q r is the intercept corresponding to the rth segment after linearization; is the output value of the thermal power unit in the rth segment at time t;

[0050] Lack of flexibility penalty C lack The expression is:

[0051]

[0052] Among them, Q c is the flexibility penalty coefficient; and are the maximum wind power and photovoltaic power in the power system respectively; P w and P v is the wind power and photovoltaic power actually absorbed in the power system; Δt is the unit operation time step.

[0053] 2.2) Matrix form of the energy storage distribution and rod optimization configuration model

[0054] After organizing the constructed energy storage optimization configuration model, it can be seen that the model is a min-max-min three-layer distributed blue-rod optimization problem. The outermost layer min gives the energy storage capacity and power configuration plan with the lowest annual comprehensive cost. The inner layer max-min simulates the operation of the units in the system under the given investment decision in the first stage and calculates the output of each unit with the lowest annual comprehensive cost of the system under the worst probability distribution. After sorting out the variables and constraints at different stages of the model, the matrix form of the energy storage distributed blue-rod optimization configuration model is obtained as follows:

[0055]

[0056] The matrix form of the model constraints is:

[0057] stAx≥b

[0058] Gy s ≥g,s=1,2,…,K

[0059] Ex+Fy s ≥d,s=1,2,…,K

[0060] Uy s +Vξ s =w,s=1,2,…,K

[0061] Where: represents the transpose of the annual investment cost of energy storage in matrix form; x represents the first-stage decision variable in the model consisting of the capacity and power of energy storage configuration; X is the set of investment decision variables for energy storage; ξ s is the sth typical output scenario in the clustered scenario set; y s represents the output plan of the thermal power unit in the second stage, the charge and discharge power of the energy storage system and other decision variables; Y represents the set of decision variables in the second stage; Ax ≥ b represents the constraint condition of the decision variables in the first stage; Gy s ≥g represents the constraint condition of the second-stage decision variables; Ex+Fy s ≥d represents the coupling condition of the decision variables in the first and second stages; Represents the equality constraints on the second-stage decision variables under scenario s. The annual investment cost of configuring energy storage; represents the expected value of the operating cost corresponding to the worst-case distribution under scenario s, given the energy storage planning scheme in the first phase. The energy storage investment cost and system operating cost together constitute the overall objective function of the model.

[0062] Step 3: Model solution. Solve the energy storage distribution robust optimization configuration model through the column and constraint generation algorithm to obtain the optimal energy storage planning scheme, and then perform energy storage of the power system according to the optimal energy storage planning scheme, such as Figure 2 shown.

[0063] The specific steps are as follows:

[0064] a) Initialization and model construction. Based on historical scene clustering, the probability distribution of typical scenery scenes at the initial moment is obtained. Then construct the energy storage distribution and rod optimization configuration model considering flexibility.

[0065] b) Parameter setting: Set the lower bound LB = -∞, the upper bound UB = +∞, and the number of iterations m = 1.

[0066] c) Solve the main problem. Solve the main problem to get the optimal investment decision solution At the same time, the lower bound is updated. The updated lower bound is in, is the decision vector for the main problem; η * is the relaxation vector of the main problem; LB' represents the lower bound at the current moment. When the lower bound is updated for the first time, LB' is equal to LB.

[0067] d) Solve the subproblem. Fix the variables in the first stage Solve the subproblem, get the probability distribution in the worst case, and update the upper bound. The updated upper bound is Among them, S represents the set of typical scenery scenes; UB' represents the upper bound at the current moment. When the upper bound is updated for the first time, UB' is equal to UB.

[0068] e) Feasibility check and iterative update. Check whether the subproblem has an optimal solution. If not, add 1 to the number of iterations at the current moment, and then add a variable to the subproblem. And add feasible cut set condition Gy s ≥g, And return to step c); wherein m' represents the number of iterations at the current moment, that is, when the number of iterations is updated for the first time, m' is equal to m;

[0069] If the subproblem has an optimal solution, check whether the difference between the upper and lower bounds at the current moment is less than or equal to ε (that is, UB'-LB'≤ε). If not, add 1 to the number of iterations at the current moment, and then add the new variable to the main problem. and the corresponding constraints Gy s ≥g, and And return to step c), where η represents the sub-problem optimization objective function value, Indicates the probability of the sth scenario in the m'+1th iteration; if yes, the algorithm iteration process ends and the energy storage configuration result is output.

[0070] The present invention is verified using IEEE39 nodes. The example includes 10 thermal power units. The total capacity of thermal power units in the power system is 7367MW. The wind and solar units are connected to 16 nodes, and their installed capacity is 800MW. The photovoltaic power station is connected to 37 nodes, and its installed capacity is 300MW. The maximum load of the power system is 6254MW. The unit power cost of the energy storage unit to be connected is 2.2×105$ / MW, the unit capacity cost is 3.3×105$ / MWh, and the discount rate is 0.08. Through the method described in the present invention, a wind and solar scene with a historical data scale of Z=500 is generated and reduced to K=4. The initial probability of the generated typical wind and solar scene is shown in Table 1. Based on the initial probability distribution of the typical wind and solar scene, an uncertain set of wind and solar output is constructed, and a robust optimization configuration model for energy storage distribution considering flexibility and uncertainty is solved.

[0071] Table 1 Probability of each scene of scenery

[0072]

[0073] Keeping the basic parameters of the IEEE 39-bus system and other uncertain parameters constant, we used stochastic programming and traditional robust optimization methods to configure energy storage and conducted a comparative analysis. The energy storage configuration options and annual comprehensive costs calculated based on the above model are shown in Table 2.

[0074] Table 2 Energy storage configuration results of different optimization methods

[0075]

[0076] From the analysis of the energy storage configuration results of different optimization methods in Table 2, it can be found that the energy storage configuration scheme obtained based on the robust optimization model is too conservative, and the flexibility demand matching coefficient is the largest. The energy storage configuration scheme obtained by random optimization does not take into account the uncertainty of wind and solar power, and the decision made is better in terms of economy, but less conservative. The energy storage capacity and power results obtained by the distributed robust optimization configuration method proposed in the present invention are moderate, and the flexibility demand matching coefficient is 0.972, which is between random optimization and robust optimization. Therefore, the energy storage distributed robust optimization configuration model proposed in the present invention combines the historical output data of wind and solar power, constructs an uncertain set of wind and solar power output, and performs energy storage configuration under the worst probability distribution. The obtained energy storage configuration scheme has a certain degree of robustness while achieving a better balance between the economy and conservatism of investment decisions.

[0077] The above-described embodiments merely illustrate several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. Persons skilled in the art will readily appreciate that variations and modifications may be made without departing from the scope of the present invention, all of which fall within the scope of protection of the present invention.

Claims

1. A method for optimizing the configuration of energy storage distribution rods considering flexibility and uncertainty, characterized in that: The following steps are involved: 1) Collect historical wind and solar power data of wind and solar power units in the power system, and construct a wind and solar power output uncertainty set based on the historical wind and solar power data and the hybrid norm; 2) constructing an objective function based on the annual investment cost and annual operating cost of energy storage in the power system, and using energy storage investment and construction constraints, thermal power unit operation constraints, power balance constraints, energy storage operation constraints, power flow constraints, wind and solar output constraints, and flexibility demand matching coefficient constraints as constraints; establishing a scenario probability-driven energy storage distributed robust optimization configuration model based on the objective function, the constraints, and the wind and solar output uncertainty set; 3) Solving the energy storage distribution robust optimization configuration model through a column and constraint generation algorithm to obtain an optimal energy storage planning scheme, and performing energy storage for the power system according to the optimal energy storage planning scheme.

2. The energy storage distribution rod optimization configuration method according to claim 1, characterized in that: The power system includes energy storage units, wind and solar units, thermal power units and power loads.

3. The energy storage distribution rod optimization configuration method according to claim 1, characterized in that: In step 1), the step of constructing a wind / solar output uncertainty set based on wind / solar historical data and a hybrid norm includes: Clustering a limited number of typical scenery scenes and a probability distribution of each of the typical scenery scenes in multiple historical scenes according to the historical scenery data; According to the probability distribution of each typical scene, and through L1-norm and L ∞ The -norm takes into account the uncertainty and fluctuation of the wind and solar scene and constructs the wind and solar output uncertainty set.

4. The energy storage distribution rod optimization configuration method according to claim 3, characterized in that: In step 1), the wind and solar power output uncertainty set includes the probability distribution uncertainty set based on L1-norm and the probability distribution uncertainty set based on L ∞ -norm probability distribution uncertainty set; the probability distribution uncertainty set Ω1 based on the L1-norm is: Where P is the true probability distribution of historical wind data; P0 is the initial distribution formed by clustering historical wind data; ||*||1 represents the L1-norm; γ1 is the uncertainty of the L1-norm; K is the number of typical wind scenes in the set of typical wind scenes; p s is the probability of occurrence of the sth typical scenery scene in the scenery scene set; is the probability of the sth typical scenery scene occurring at the initial moment; Z is the number of historical scenery scenes; α1 is the uncertainty probability confidence of the L1-norm; The L-based ∞ -norm probability distribution uncertain set Ω ∞ for: Among them, ||*|| ∞ Indicates L ∞ -norm; γ ∞ For L ∞ -Uncertainty of the norm; α ∞ For L ∞ -norm uncertainty probability confidence.

5. The energy storage distribution rod optimization configuration method according to claim 1, characterized in that: In step 2), the objective function is constructed with the minimum sum of the annual investment cost of energy storage and the annual operating cost of the power system as the goal; the expression of the objective function is: Among them, C inv is the annual investment cost of energy storage; C oper is the daily operating cost of the power system; σ is the equivalent factor of the annual investment cost; p z is the probability distribution of the initial scene of scenery; Ω is the set of probability distributions of the initial scene of scenery.

6. The energy storage distribution rod optimization configuration method according to claim 5, characterized in that: The annual investment cost of energy storage is C inv The expression is: Where α is the energy storage cost coefficient; k is the node number; Ω ESS is the energy storage set to be configured; r is the energy storage discount rate; y is the number of years of energy storage discount; C ESS is the unit cost of the energy storage capacity to be configured; C p is the unit cost of the energy storage power to be configured; E k is the energy storage capacity to be configured on node k; P k is the energy storage power to be configured on node k.

7. The energy storage distribution rod optimization configuration method according to claim 5, characterized in that: The daily operating cost of the power system is C oper The expression is: C oper =C operation,G +C lack Among them, C operation,G is the operating cost of the thermal power unit; C lack Penalty for lack of flexibility.

8. The energy storage distribution rod optimization configuration method according to claim 6, characterized in that: The operating cost of the thermal power unit C operation,G The expression is: Among them, g is the number of thermal power unit; G is the set of thermal power unit numbers; T is the maximum time; t is the time number; r is the segment number; R is the maximum segment number; p r is the slope corresponding to the rth segment after linearization; q r is the intercept corresponding to the rth segment after linearization; is the output value of the thermal power unit in the rth segment at time t; Lack of flexibility penalty C lack The expression is: Among them, Q c is the flexibility penalty coefficient; is the maximum wind power in the power system; P w is the wind power actually absorbed in the power system; is the maximum photovoltaic power in the power system; P v is the photovoltaic power actually consumed in the power system; Δt is the unit operation time step.

9. The energy storage distribution rod optimization configuration method according to claim 1, characterized in that: The specific steps of step 3) are as follows: Initialize the lower bound, upper bound and number of iterations; According to the current number of iterations, the main problem of the energy storage distribution and rod optimization configuration model is solved to obtain the optimal solution. And update the current lower bound to in, is the decision vector for the main problem; η * The relaxation vector of the main problem; A is the constraint matrix of the main problem; then solve the sub-problems of the energy storage distribution and optimization model, obtain the probability distribution under the worst case, and update the current upper bound; If the iteration termination condition is met, the iteration is terminated, the probability distribution under the worst case scenario is returned, and the optimal energy storage planning scheme is obtained; wherein, the iteration termination condition is that the difference between the current upper bound and the current lower bound is less than a preset error; if the iteration termination condition is not met, the current number of iterations is updated according to a preset rule, and the main problem and subproblems of the distributed robust programming model are solved respectively according to the current number of iterations to obtain the probability distribution under the worst case scenario, and the current lower bound and the current upper bound are updated until the iteration termination condition is met to obtain the optimal energy storage planning scheme.

Images