Power distribution network energy storage double-layer planning method considering demand response under different photovoltaic permeability

Abstract

The invention relates to a power distribution network energy storage bilevel planning method considering demand response under different photovoltaic permeability. The method comprises the following steps: step 1, power distribution network flexibility resource modeling: flexible supply of the power distribution network mainly comes from flexible resources such as an energy storage device of each node, an interruptible load of a demand side and a superior main network; 2, constructing a demand response model, wherein demand response means that a user adjusts a power utilization mode through price information or an incentive mechanism, and the demand response comprises an interruptible load in a price type demand response and an incentive demand response; step 3, power distribution network energy storage double-layer planning; 4, example analysis is carried out; the method has the advantages that upper and lower layer correlation modeling is adopted, the bilayer planning model is converted into a single layer, the solving efficiency is improved, and the economical efficiency and flexibility of the power distribution network are improved.

Description

Technical Field

[0001] This invention belongs to the field of power distribution network technology, specifically relating to a two-layer planning method for energy storage in power distribution networks that considers demand response under different photovoltaic penetration rates. Background Technology

[0002] To help achieve the "dual carbon" goals, distributed photovoltaic (PV) power, with its advantages of sustainability, low pollution, and abundant resources, will be widely integrated into the distribution network. However, the mismatch between load and PV output timing leads to significant fluctuations in the net load of the distribution network, causing severe curtailment and load shedding, reducing stability, and drastically increasing the demand for distribution network flexibility. Energy storage systems (ESS), as a key technology for distribution networks, have advantages such as rapid response, bidirectional switching, and flexible installation locations. They can effectively mitigate the impact of PV output fluctuations on the distribution network, promote PV consumption, and improve the economic efficiency of distribution network operation. In addition, demand-side response load is an important controllable resource within the distribution network; reasonable control can reduce load peaks and alleviate peak loads. Using electricity to ensure the safe and reliable operation of the distribution network, the development of energy storage technology and the implementation of demand-side response have provided the distribution network with stronger adjustability and ensured its safe and stable operation. Currently, most solutions for distribution network energy storage planning models use intelligent algorithms to solve the two-layer model separately, iterating through the upper and lower layers. However, this method may lead to poor convergence of the upper layer and a long solution time. Therefore, it is necessary to provide a two-layer planning method for distribution network energy storage considering demand response under different photovoltaic penetration rates, which adopts upper and lower layer correlation modeling, converts the two-layer planning model into a single layer, improves solution efficiency, and enhances the economy and flexibility of the distribution network. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a two-layer planning method for energy storage in distribution networks with different photovoltaic penetration rates that adopts upper and lower layer correlation modeling, converts the two-layer planning model into a single layer, improves solution efficiency, and enhances the economy and flexibility of the distribution network.

[0004] The objective of this invention is achieved as follows: a two-layer planning method for energy storage in distribution networks considering demand response under different photovoltaic penetration rates, the method comprising the following steps:

[0005] Step 1: Distribution network flexibility resource modeling: The supply of distribution network flexibility mainly comes from energy storage devices at each node, interruptible loads on the demand side, and flexibility resources such as the upstream main grid;

[0006] Step 2: Construct a demand response model: Demand response refers to users adjusting their electricity consumption patterns based on price information or incentive mechanisms, including price-based demand response and interruptible loads in incentive-based demand response;

[0007] Step 3: Two-tier planning for energy storage in the distribution network;

[0008] Step 4: Case analysis.

[0009] The distribution network flexibility resource modeling in step 1 specifically involves the following calculation formulas for energy storage, interruptible loads, and the main grid's flexibility adjustment capabilities: In the formula: P represents the upward and downward adjustment capabilities of the ESS at node i at time t, respectively. ESS,i E ESS,i These are the rated power and rated capacity of the ESS at node i, respectively; Let be the power of the ESS at node i at time t; Store the energy of the ESS at node i at time t; η represents the lower limit of the ESS charge at node i; η represents the ESS charge / discharge efficiency. These represent the upward and downward adjustment capabilities provided by the interruptible load at node i at time t; v j,t This represents the interruption status of the interruptible load at node i at time t, where 1 indicates interruption and 0 indicates non-interruption. Let be the interruption amount of the interruptible load at node i at time t; These represent the upward and downward regulation capabilities provided by the upstream power grid during time period t; ΔP grid,max P represents the maximum ramping power of the upstream main network. grid,max This is the maximum transmission power of the upstream main network; This refers to the active power transmitted by the main network at any given time.

[0010] The construction of the demand response model in step 2 includes the following steps:

[0011] Step 2.1: Price-based demand response model;

[0012] Step 2.2: Interruptible load response model.

[0013] The price-based demand response model in step 2.1 specifically refers to the following: Price-based demand response influences users' electricity consumption patterns through electricity prices, and the degree of user response to electricity prices is typically described using the demand elasticity coefficient. in In the formula, e ii e is the self-elasticity coefficient; ij The mutual elasticity coefficient; These are the electricity prices at times i and j before the implementation of demand response; c i c j These are the electricity prices at time i and j after the implementation of demand response, respectively. q iThese represent electricity consumption before and after the electricity price response; dividing a day into 24 hours, the elasticity coefficient matrix is ​​represented as follows: The electricity consumption change matrix after the electricity price change is represented as follows: The change in user electricity consumption after the change in electricity price is as follows: After the electricity price change, the user's electricity consumption is:

[0014]

[0015] The interruptible load response model in step 2.2 is as follows: its response characteristics are shown in the following formula: ① Interruption capacity constraint: In the formula: ① The upper and lower limits of the interruptible load interruption capacity at node i are respectively; ② Interruption count constraint: Where: N CL The maximum number of interruptible load cycles within a single cycle; ③ Maximum interruption time constraint: In the formula: The maximum interruption time for the interruptible load at node i.

[0016] The two-tier planning of distribution network energy storage in step 3 includes the following steps:

[0017] Step 3.1: Upper-level planning layer model: including upper-level objective function and upper-level constraints;

[0018] Step 3.2: Lower-level planning layer model: including the lower-level objective function and lower-level constraints;

[0019] Step 3.3: Two-layer model conversion.

[0020] In step 3.1, the upper-level objective function of the upper-level planning model is: minC Total =C I +C O (13), where: C Total For the overall cost of the distribution network; C I For ESS investment costs; C O This refers to the operating costs of the power distribution network.

[0021] The energy storage investment cost C I for: In the formula, Ω ESS For the set of ESS nodes to be installed; c e c p These are the unit capacity and power costs of ESS, respectively; E ESS,i P ESS,i Here, C represents the rated capacity and rated power of the ESS installed at node i, respectively; and the distribution network operating cost C. O for: In the formula, C Loss 、C CL 、C Grid 、C FL 、C Carbon are the energy storage operation and maintenance cost, network loss cost, interruptible load response cost, power purchase cost, flexibility deficiency penalty cost, and carbon emission penalty cost respectively; is the cost per unit charge and discharge of the ESS; N is the number of operating days of the distribution network in a year; Φ is the set of typical operating scenarios; p s is the probability of scenario s occurring; is the charge and discharge power of the ESS at node i at time t in scenario s; c loss is the unit network loss cost; is the active power loss of line ij at time t in scenario s; Ω Line is the set of lines; c cl is the unit cost of interruptible load; Ω CL is the set of interruptible load nodes; is the interruption amount of the interruptible load at node i at time t in scenario s; is the power purchase price in period t of scenario s, which is the time-of-use electricity price; is the power purchase amount from the superior power grid at time t in scenario s; c up 、c doen are the upward and downward flexibility deficiency penalty coefficients respectively; are the load shedding amount and light curtailment amount at time t in scenario s respectively; c carbon is the unit carbon emission penalty cost; γ grid is the unit carbon emission intensity of the superior power grid.

[0022] The upper-layer constraint conditions of the upper-layer planning layer model in step 3.1 include: ① Installation constraint of the nodes where the ESS is to be installed: In the formula, P ESS,max 、P ESS,min are the upper and lower limits of the rated power of the ESS to be planned respectively; E ESS,max 、E ESS,min are the upper and lower limits of the rated capacity of the ESS to be planned respectively; ② Photovoltaic penetration constraint: In the formula, Ω PV is the node where the distributed photovoltaic is installed; is the installed capacity of the distributed photovoltaic at node i; α is the photovoltaic penetration rate; is the load peak.

[0023] The lower-layer objective function of the lower-layer planning layer model in step 3.2 aims to minimize the operating cost: An optimization model is established using ESS charging and discharging power, interruption status of interruptible loads, and electricity price as decision variables. The lower-level constraints include ESS operation constraints, price-based demand response constraints, interruptible load constraints, active power injection constraints from the upper-level main grid, PV operation constraints, power flow constraints, and system security constraints.

[0024] The beneficial effects of this invention are as follows: This invention provides a two-layer planning method for energy storage in distribution networks considering demand response under different photovoltaic penetration rates. In practice, this invention analyzes the flexibility of the distribution network based on photovoltaic penetration rate, and uses principal component analysis to reduce the dimensionality of a large amount of historical photovoltaic load data, retaining the main characteristics of the data while simplifying the data structure, thereby improving clustering efficiency. Then, a hierarchical clustering algorithm is used to obtain typical photovoltaic load scenarios, and on this basis, a two-layer planning model for energy storage considering interruptible loads and price-based demand response is established. By performing correlation modeling on the upper and lower layers of the model, it is converted into a single-layer model for solution. Finally, the effectiveness of the proposed planning method is verified using the IEEE 33-node distribution system as an example. This invention has the advantages of using correlation modeling on the upper and lower layers, converting the two-layer planning model into a single layer, improving solution efficiency, and improving the economy and flexibility of the distribution network. Attached Figure Description

[0025] Figure 1 This is the two-level programming model of the present invention.

[0026] Figure 2 This is a flowchart of the present invention.

[0027] Figure 3 This is the IEEE 33 distribution network topology diagram of the present invention.

[0028] Figure 4 This is the equivalent load curve diagram of Scheme 2 of the present invention.

[0029] Figure 5 The equivalent load curves for schemes 2 and 3 of the present invention are shown.

[0030] Figure 6 This is a flowchart of the hybrid optimization algorithm of the present invention. Detailed Implementation

[0031] The present invention will now be further described with reference to the accompanying drawings.

[0032] Example 1

[0033] like Figure 1-6 As shown, a two-layer planning method for energy storage in distribution networks considering demand response is presented under different photovoltaic penetration rates. The method includes the following steps:

[0034] Step 1: Distribution network flexibility resource modeling: The supply of distribution network flexibility mainly comes from energy storage devices at each node, interruptible loads on the demand side, and flexibility resources such as the upstream main grid;

[0035] In this embodiment, step 1 specifically involves the following calculation formulas for energy storage, interruptible loads, and the main grid's flexibility adjustment capabilities: In the formula: P represents the upward and downward adjustment capabilities of the ESS at node i at time t, respectively. ESS,i E ESS,i These are the rated power and rated capacity of the ESS at node i, respectively; Let be the power of the ESS at node i at time t; Store the energy of the ESS at node i at time t; η represents the lower limit of the ESS charge at node i; η represents the ESS charge / discharge efficiency. These represent the upward and downward adjustment capabilities provided by the interruptible load at node i at time t; v j,t This represents the interruption status of the interruptible load at node i at time t, where 1 indicates interruption and 0 indicates non-interruption. Let be the interruption amount of the interruptible load at node i at time t; These represent the upward and downward regulation capabilities provided by the upstream power grid during time period t; ΔP grid,max P represents the maximum ramping power of the upstream main network. grid,max This is the maximum transmission power of the upstream main network; This refers to the active power transmitted by the main network at any given time.

[0036] Step 2: Construct a demand response model: Demand response refers to users adjusting their electricity consumption patterns based on price information or incentive mechanisms, including price-based demand response and interruptible loads in incentive-based demand response;

[0037] In this embodiment, step 2 includes the following steps:

[0038] Step 2.1: Price-based demand response model;

[0039] Specifically, price-based demand response influences users' electricity consumption patterns through electricity prices, enabling peak shaving and valley filling, promoting photovoltaic power consumption, and improving the stability of the distribution network. The degree to which users respond to electricity prices is typically described using the demand elasticity coefficient. in In the formula, e ii e is the self-elasticity coefficient; ij The mutual elasticity coefficient; These are the electricity prices at times i and j before the implementation of demand response; c i c jThese are the electricity prices at time i and j after the implementation of demand response, respectively. q i These represent electricity consumption before and after the electricity price response; dividing a day into 24 hours, the elasticity coefficient matrix is ​​represented as follows: The electricity consumption change matrix after the electricity price change is represented as follows: The change in user electricity consumption after the change in electricity price is as follows: After the electricity price change, the user's electricity consumption is:

[0040]

[0041] Step 2.2: Interruptible load response model.

[0042] Specifically, incentive-based demand response only considers interruptible loads, which are mainly targeted at large industrial users. These interruptible loads can respond to signals from the dispatching department to reduce or interrupt part of the load in the event of load peaks or system failures. Their response characteristics are shown in the following formula: ① Interruption capacity constraint: In the formula: ① The upper and lower limits of the interruptible load interruption capacity at node i are respectively; ② Interruption count constraint: Where: N CL The maximum number of interruptible load cycles within a single cycle; ③ Maximum interruption time constraint: In the formula: The maximum interruption time for the interruptible load at node i.

[0043] Step 3: Two-tier planning for energy storage in the distribution network;

[0044] Step 4: Case analysis.

[0045] This invention presents a two-layer planning method for energy storage in distribution networks considering demand response under different photovoltaic penetration rates. In practice, this invention analyzes the flexibility of the distribution network based on photovoltaic penetration rate and uses principal component analysis to reduce the dimensionality of a large amount of historical photovoltaic load data, preserving the main characteristics of the data while simplifying the data structure, thereby improving clustering efficiency. Then, a hierarchical clustering algorithm is used to obtain typical photovoltaic load scenarios, and on this basis, a two-layer energy storage planning model considering interruptible loads and price-based demand response is established. By performing correlation modeling on the upper and lower layers of the model, it is converted into a single-layer model for solution. Finally, the effectiveness of the proposed planning method is verified using the IEEE 33-node distribution system as an example. This invention has the advantages of using correlation modeling on the upper and lower layers, converting the two-layer planning model into a single layer, improving solution efficiency, and enhancing the economy and flexibility of the distribution network.

[0046] Example 2

[0047] like Figure 1-6As shown, a two-layer planning method for energy storage in distribution networks considering demand response is presented under different photovoltaic penetration rates. The method includes the following steps:

[0048] Step 1: Distribution network flexibility resource modeling: The supply of distribution network flexibility mainly comes from energy storage devices at each node, interruptible loads on the demand side, and flexibility resources such as the upstream main grid;

[0049] Step 2: Construct a demand response model: Demand response refers to users adjusting their electricity consumption patterns based on price information or incentive mechanisms, including price-based demand response and interruptible loads in incentive-based demand response;

[0050] Step 3: Two-tier planning for energy storage in the distribution network;

[0051] In this embodiment, traditional distribution network planning, without considering operation, focuses too much on a single operating scenario, leading to low equipment utilization and redundant investment. Therefore, planning should consider multiple scenarios and operating strategies. Since the modeling approach of distribution network planning-operation dual-layer optimization has a hierarchical characteristic, and the hierarchical model can clearly describe the optimization content of the planning and operation layers, the upper planning layer of this invention optimizes the energy storage planning scheme with the goal of minimizing the overall cost of the distribution network and passes the planning scheme to the lower layer. The lower operation layer, given the planning scheme, formulates operating strategies for typical scenarios with the goal of minimizing operating costs, comprehensively considering controllable means such as distributed photovoltaics, energy storage, and demand-side response, and feeds back the optimization results from the lower layer to the upper layer, thereby further guiding the formulation of the optimal planning scheme. The structural block diagram of the energy storage dual-layer planning model is as follows: Figure 1 As shown;

[0052] Step 3 includes the following steps:

[0053] Step 3.1: Upper-level planning layer model: including upper-level objective function and upper-level constraints;

[0054] The objective function at the upper level is: minC Total =C I +C O (13), where: C Total For the overall cost of the distribution network; C I For ESS investment costs; C O The operating cost of the distribution network; ① Energy storage investment cost C I for: In the formula, Ω ESS For the set of ESS nodes to be installed; c e c p These are the unit capacity and power costs of ESS, respectively; E ESS,i P ESS,i ① The rated capacity and rated power of the ESS installed at node i, respectively; ② Distribution network operating cost CO for: In the formula, C Loss C CL C Grid C FL C Carbon These are energy storage operation and maintenance costs, grid loss costs, interruptible load response costs, electricity purchase costs, insufficient flexibility penalty costs, and carbon emission penalty costs; The cost per unit charge / discharge of the ESS; N is the number of operating days in a year for the distribution network, taken as 365 days; Φ is the set of typical operating scenarios; p s Let s be the probability of scenario s occurring. c represents the charging and discharging power of the ESS at node i at time t in scene s; loss Cost per unit of network loss; Let Ω represent the active power loss of line ij at time t in scenario s; Line For the set of lines; c cl Unit cost of interruptible load; Ω CL A set of interruptible load nodes; Let be the interruptible load at node i at time t in scenario s; Let be the electricity purchase price during time period t in scenario s, and be the time-of-use electricity price. c represents the amount of electricity purchased from the upstream power grid at time t in scenario s; up c doen These are the penalty coefficients for insufficient upward and downward flexibility, respectively. These represent the load shedding amount and light wastage amount at time t in scenario s, respectively; c carbon Penalty costs per carbon emission unit; γ grid Carbon emission intensity of the superior power grid unit;

[0055] Upper-level constraints include: ① Installation constraints of the ESS node to be installed: In the formula, P ESS,max P ESS,min These are the upper and lower limits of the rated power of the ESS to be planned; E ESS,max E ESS,min These represent the upper and lower limits of the rated capacity of the planned ESS (Energy Storage System); ② Photovoltaic penetration rate constraint: To analyze the operation of the distribution network under different photovoltaic penetration rates, a constraint is imposed on the photovoltaic penetration rate. By solving the model, energy storage planning schemes and operating conditions under different penetration rates are obtained, providing some guidance for the gradual promotion of distributed photovoltaics in the region. In the formula, Ω PV For installation nodes of distributed photovoltaic systems; Let be the installed capacity of the distributed photovoltaic system at node i; α be the photovoltaic penetration rate. This represents the peak load.

[0056] Step 3.2: Lower-level planning layer model: including the lower-level objective function and lower-level constraints;

[0057] The lower-level objective function aims to minimize the running cost: An optimization model is established using ESS charging and discharging power, interruption status of interruptible loads, and electricity price as decision variables. The lower-level constraints include ESS operation constraints, price-based demand response constraints, interruptible load constraints, active power injection constraints from the upper-level main grid, PV operation constraints, power flow constraints, and system security constraints.

[0058] Among them, ①ESS operational constraints: In the formula, SOC i,s,t SOC i,s,t+1 Let represent the charge state of the ESS at node i at times t and t+1 in scene s; SOC i,s,0 SOC i,s,24 These represent the initial and final states of charge of the ESS at node i in scene s during its runtime; SOC i,max SOC i,min These are the upper and lower limits of the ESS state of charge at node i, respectively; ② Price-based demand response constraint: The total load should remain unchanged before and after the implementation of price-based demand response, i.e.: In the formula, The total load of the distribution network at time t represents the scenarios before and after the implementation of price-based demand response; to ensure user satisfaction with electricity consumption, user satisfaction constraints are added. In the formula, S min Δq represents the minimum user satisfaction level. i,s Let q represent the change in user electricity consumption after implementing a price-based demand response at time i in scenario s; i,s ③ The original user power consumption at time i in scenario s; ④ Interruptible load constraints: as shown in equations (16)-(18); ⑤ Active power constraints injected by the upper-level main network: In the formula, ⑤ PV operation constraints: (This refers to the active power injected into the distribution network by the upstream main grid at time t in scenario s.) This represents the actual power generation of the photovoltaic system at node i at time t in scenario s; ⑥ Power flow constraint: The nonlinear terms in the DisFlow power flow model are transformed using a second-order cone relaxation transformation, as shown in the following equation:

[0059] In the formula, AL(j,:) and AL(:,j) are the sets of branch nodes with starting and ending nodes j, respectively; P ij,s,t Q ij,s,tLet r represent the active and reactive power of line ij at time t in scenario s; ij x ij These represent the resistance and reactance of line ij, respectively; P j,s,t Q j,s,t These represent the active and reactive power injected into node j at time t in scenario s, respectively. These represent the active and reactive power of the load at node j after the price-based demand response at time t in scenario s; These represent the active and reactive power injected by the main network at time t in scenario s at node j, respectively. Let be the square of the current of line ij at time t in scenario s; ⑦ The square of the voltage of line ij at time t in scenario s; ⑦ System safety constraints: In the formula, U i,max U i,min These are the upper and lower limits of the voltage at node i, respectively; I ij,s,t This is the safe current for the circuit.

[0060] Step 3.3: Two-layer model conversion.

[0061] In this embodiment, specifically: the above-mentioned two-layer model models the planning layer and the operation layer separately. The upper planning layer passes the planning scheme to the lower operation layer. The lower operation layer optimizes the scheduling based on the given planning scheme and feeds the results back to the upper model. The two-layer energy storage planning model is solved by iteratively solving the upper and lower layers. The two-layer model can be represented by the simplified formula (29): In the formula, F and f are the objective functions of the upper and lower level models, respectively; G(·) and g(·) are the equality constraints of the upper and lower level models, respectively; H(·) and h(·) are the inequality constraints of the upper and lower level models, respectively; the specific contents are described above.

[0062] However, the iterative solution method involving upper and lower layers is inefficient and struggles to obtain the global optimum. Therefore, to effectively solve the two-layer model, this invention transforms it into a single-layer model by performing relational modeling on the upper and lower layers. The specific operation process is as follows: In the formula, It is a 0-1 variable; it is 1 if ESS is installed at node i, and 0 otherwise.

[0063] Equation (30) links the energy storage installation constraint (16) and the energy storage operation constraint (19) at the planning layer, while the remaining constraints remain unchanged; if A value of 0 indicates that ESS is not installed. ESS,i E ESS,i Both are 0, and the charging and discharging power is also 0; if The value is 1, and P is restricted by equation (30). ESS,i E ESS,iThe upper and lower limits, and through P ESS,i E ESS,i The relevant constraints affecting the operation layer are such as equations (19) and (25); this is equivalent to the upper layer in the two-layer model passing the energy storage planning scheme to the lower layer. Therefore, equation (30) and the constraints of the lower layer model together constitute the constraints of the single-layer model; the transformed model can be represented by the simplified formula (31): In the formula, G(·), g(·), H(·), and h(·) are the constraints of the transformed single-layer model. The transformed model has some nonlinear terms, which can be solved directly by calling the Gurobi solver, or linearized first and then solved by calling the solver.

[0064] Step 4: Case analysis.

[0065] In this embodiment, specifically: based on the above analysis, the steps of the proposed distribution network energy storage planning method are as follows, and the flowchart is as follows: Figure 2 As shown:

[0066] Step 1: Collect historical photovoltaic power output and load data, as well as relevant parameters of the distribution network energy storage planning model;

[0067] Step 2: Dimensionality reduction of historical photoelectric load data is performed using principal component analysis, and typical photoelectric load scenarios are obtained using hierarchical clustering algorithm;

[0068] Step 3: Establish a two-layer planning model for energy storage in the distribution network: the upper planning layer optimizes the energy storage planning scheme with the goal of minimizing the overall cost; the lower operation layer formulates the distribution network operation strategy under typical scenarios with the goal of minimizing the operating cost.

[0069] Step 4: Convert the two-layer planning model of distribution network energy storage into a single-layer model through the association and unification method, and solve it using the Gurobi solver;

[0070] Step 5: Adjust the photovoltaic installed capacity to achieve the specified penetration rate, and solve for each of them one by one;

[0071] Step 6: Output the distribution network energy storage planning scheme and other relevant data results, and analyze and compare them.

[0072] ① Model Parameter Description: To verify the effectiveness of the proposed energy storage planning model, simulation verification was conducted on a standard IEEE 33-node distribution system. Distributed photovoltaic (PV) systems were connected in batches at nodes 10, 15, 19, and 31 to achieve PV penetration rates of approximately 40%, 60%, and 80% in the distribution network, respectively. Interruptible loads with interruptible capacity limits of 40kW, 50kW, 60kW, and 70kW were set at nodes 5, 11, 18, and 30, respectively. The distribution network topology is as follows: Figure 3 As shown;

[0073] The basic load of the IEEE 33-node distribution system is 3715kW + j2300kvar. When solving the model, the upper and lower limits of the rated power and rated capacity of the energy storage system need to be given. Considering the cost and regulation capability of energy storage, it is not advisable to set them too high or too low. Therefore, the total number of ESS nodes that can be installed is set to 2-3. The rated power range of the individual ESS to be planned is [40,100]kW, the rated capacity range is [80,500]kW·h, the charge and discharge efficiency is 95%, the initial state of charge is 0.2, and the upper and lower limits of the state of charge are 0.9 and 0.1, respectively. The relevant parameters of an individual ESS can be adjusted according to the actual situation.

[0074] Cost-related parameters are shown in Table 1. The electricity purchase price adopts the time-of-use electricity price, as shown in Table 2. The price elasticity demand matrix consists of self-elasticity coefficient and mutual elasticity coefficient. It can be obtained by mathematical calculation and SPSS software to estimate relevant parameters based on the load and electricity price data before and after the response. The price elasticity demand matrix adopted in this invention is shown in Table 3.

[0075] Table 1 Parameter Settings

[0076]

[0077] Table 2 Time-of-use Electricity Prices

[0078]

[0079] Table 3 Price Demand Elasticity Matrix

[0080]

[0081] ② Example Setup and Result Analysis: To verify the economic efficiency and effectiveness of the proposed energy storage planning model, ESS optimization configuration was performed on the distribution network under different photovoltaic penetration rates. To better analyze the impact of demand response on the ESS optimization configuration results, ESS optimization configuration was performed under three conditions: considering single demand response, multiple demand responses, and not considering demand response. In summary, the following four schemes were set up for comparative analysis: 1) Scheme 1: Without considering flexible resources, calculate the various costs of the distribution network under photovoltaic penetration rates of 40%, 60%, and 80%; 2) Scheme 2: Considering ESS but not demand response load, calculate the various costs of the distribution network and the ESS planning scheme under photovoltaic penetration rates of 40%, 60%, and 80%; 3) Scheme 3: Considering both ESS and interruptible load, obtain the various costs of the distribution network and the ESS planning scheme under a photovoltaic penetration rate of 80% by solving a bi-level programming model; 4) Scheme 4: Based on Scheme 3, considering price-based demand response, obtain the various costs of the distribution network and the ESS planning scheme under a photovoltaic penetration rate of 80% by solving a bi-level programming model.

[0082] The optimized ESS configuration results obtained from Scheme 1 and Scheme 2 are shown in Table 4, and the various costs of the distribution network are shown in Table 5. As can be seen from Table 5, with the gradual increase in the penetration rate of distributed photovoltaics, the power purchase cost and carbon emission penalty cost of the distribution network in Scheme 1 are gradually decreasing. This shows that increasing the installed capacity of photovoltaics can help achieve the dual carbon goals. However, since the photovoltaic installation locations are concentrated in a few designated nodes and reverse power flow is not allowed, the phenomenon of curtailment and load shedding occurs. The insufficient flexibility penalty costs of the distribution network at distributed photovoltaic penetration rates of 40%, 60%, and 80% are RMB 109,800, RMB 1,212,100, and RMB 1,906,100, respectively. The insufficient flexibility penalty cost increases continuously with the increase in photovoltaic penetration rate, and the phenomenon of insufficient distribution network flexibility is aggravated. A certain amount of flexibility resources should be allocated to reduce the amount of curtailment and load shedding in the distribution network.

[0083] In Scheme 2, the investment cost of ESS (Emerging Power Supply) increases continuously with the increase of distributed photovoltaic penetration rate, indicating that more ESS need to be configured to improve the flexibility and regulation capability of the distribution network under high penetration rate conditions. Compared with Scheme 1, Scheme 2 increases the investment cost of ESS by RMB 335,600, RMB 1,067,500, and RMB 1,521,200 respectively. However, after configuring ESS, the network loss cost, electricity purchase cost, carbon emission penalty cost, and comprehensive cost of the distribution network are all reduced. Among them, the comprehensive cost is reduced by RMB 45,300, RMB 41,300, and RMB 307,400 respectively, and the insufficient flexibility penalty cost is reduced by RMB 166,300, RMB 891,100, and RMB 1,441,000 respectively, representing reductions of 88.27%, 73.52%, and 75.60%. Therefore, it can be seen that... Configuring ESS can promote photovoltaic (PV) integration, reduce grid losses, and improve the flexibility and economy of the distribution network. As shown in Table 5, the installation locations of energy storage are similar under the three PV penetration rates, mainly around nodes 18, 20, and 31. The rated capacity and rated power differ. The PV installed capacity and penetration rate of the distribution network should gradually increase. If the trend of PV penetration rate in the distribution network in the future can be known, energy storage configuration can be arranged in advance. Distribution network planners can refer to the energy storage planning results under different PV penetration rates to determine the appropriate ESS installation location and reserve a certain amount of installation space. The rated power and rated capacity of the ESS can be adjusted to adapt to the changes in PV penetration rate.

[0084] Table 4. ESS Optimization Configuration Results for Scheme 2

[0085]

[0086]

[0087] Table 5. Cost Comparison of Option 1 and Option 2

[0088]

[0089] Table 6 shows the cost calculation results for schemes 1 to 4 under a photovoltaic penetration rate of 80%, and Table 7 shows the optimized ESS configuration results. Table 6 shows that as the types of flexible resources considered increase, the insufficient flexibility penalty cost and overall cost of schemes 1 to 4 gradually decrease, and the ESS investment cost of schemes 2 to 4 also gradually decreases. From the planning results, scheme 2 configures ESS at nodes 17, 21, and 32; scheme 3 configures ESS at nodes 17, 21, and 32; and scheme 4 configures ESS at nodes 17, 20, and 31. The configuration nodes, installed rated capacity, and rated power differ among the three schemes. Since scheme 1 does not consider flexible resources, its grid loss cost, electricity purchase cost, insufficient flexibility penalty cost, carbon emission penalty cost, and overall cost are the highest among the four schemes. The insufficient flexibility penalty... The initial cost of Scheme 1 is 1.9061 million yuan, with a significant amount of curtailed solar power. Scheme 2, which includes an Energy Storage System (ESS), requires an investment cost of 1.5212 million yuan, but other costs are reduced. Compared to Scheme 2, Scheme 3 introduces interruptible loads to participate in the dispatch and operation of the distribution network. Although it requires paying an interruptible load response cost of 396,100 yuan, its investment cost and insufficient flexibility penalty cost are reduced by 57,700 yuan and 70,800 yuan respectively, resulting in a comprehensive cost reduction of 337,400 yuan, demonstrating better economic efficiency and flexibility. Scheme 4, based on Scheme 3, introduces price-based demand-side response, further reducing energy storage investment costs, electricity purchase costs, and insufficient flexibility penalty costs. Among these, the insufficient flexibility penalty cost is the lowest at 292,300 yuan. In conclusion, considering demand response loads in ESS planning leads to a more reasonable and economical ESS planning scheme.

[0090] Table 6 Cost Comparison of Options 1-4

[0091]

[0092] Table 7 ESS Optimization Configuration Results for Schemes 2-4

[0093]

[0094] The equivalent net load curve of Scheme 3 in Scenario 5 is as follows: Figure 4 As shown, curve 1 is the net load curve, which is the total load minus the total photovoltaic output; curve 2 is the equivalent load curve taking into account curtailment; curve 3 is the equivalent load curve taking into account curtailment and energy storage charging and discharging; curve 4 is the equivalent load curve taking into account curtailment, energy storage charging and discharging, and interruptible load; from Figure 4It can be seen that curtailment only occurs during the period of peak photovoltaic output from December 12th to 16th, and scenario 5 has the highest photovoltaic output among all scenarios. Overall, curtailment is within an acceptable range. By comparing curves 2 and 3, it can be seen that adjusting the charging and discharging power of energy storage can smooth the equivalent load curve. After taking into account the interruptible load response, curve 4 further smooths the equivalent load curve, and plays an overall role in peak shaving and valley filling.

[0095] The equivalent net load curves for Schemes 2 and 3 in Scenario 5 are as follows: Figure 5 As shown, curve one is the net load curve; curves two and three are the equivalent load curves under schemes 2 and 3, respectively; from Figure 5 It can be seen that, compared with the net load, the equivalent load curves of Scheme 2 and Scheme 3 have been greatly improved. The flexibility of energy storage and interruptible load can smooth out load fluctuations. The hourly fluctuation of the equivalent load curve under Scheme 3 is slightly smaller than that under Scheme 2. The effect is more obvious during the period from 8 to 11 when the photovoltaic starts to generate power. The process of the equivalent load from high point to low point is smoother.

[0096] ③ Comparison of Two-Level Solution Algorithms: To verify the superiority of the proposed unified solution method for upper and lower level correlation, a hybrid optimization algorithm combining particle swarm optimization and second-order cone programming was used to iteratively solve the two-level programming model for comparison. The flowchart is shown below. Figure 6 As shown;

[0097] The solution algorithms in this chapter and the hybrid optimization algorithm were used to calculate the various costs of the distribution network and the solution time in Scheme 2 with 80% photovoltaic penetration. The results are shown in Table 8, and the ESS planning scheme is shown in Table 9.

[0098] Table 8. Solution results of different algorithms

[0099]

[0100]

[0101] Table 9. ESS Optimization Configuration Results for Different Algorithms

[0102]

[0103] Table 8 shows that different solving algorithms yield different costs and solution times. The hybrid optimization algorithm results in slightly higher ESS investment costs and overall costs than the algorithm proposed in this paper, while other costs are slightly lower. Table 9 shows that the ESS planning schemes obtained by the two algorithms are quite similar, with little difference. From the perspective of planning cost, the ESS planning model aims to minimize overall cost, and the algorithm proposed in this paper is slightly better. In terms of solution time, the algorithm proposed in this paper takes 491 seconds, which is significantly lower than the 3584 seconds of the hybrid optimization algorithm. This is because the two-layer modeling method requires continuous iteration between upper and lower layers, which takes more time. Compared with the hybrid optimization algorithm, the method of converting the two-layer model to a single-layer model and using the Gurobi solver to solve the single-layer model does not require iterative loops between upper and lower layers, thus improving the solution efficiency.

[0104] Practical Application: Taking the Yichuan Northeast Grid as an example, when the photovoltaic penetration rate is 80% and only the distribution network considers distributed energy storage (DESS), the simulation calculation shows that the planned DESS capacity is 760.35, 1247.93, and 1549 kWh, and the energy storage power is 380.12, 628.27, and 811.53 kW.

[0105] When considering distribution network energy storage planning under demand-side response with a photovoltaic penetration rate of 80%, simulation calculations yielded planned DESS capacities of 660.75, 1147.83, and 1379 kWh, with energy storage powers of 330.3, 553.2, and 650.25 kW. Compared to considering only the distribution network with distributed energy storage, the overall cost was reduced by RMB 437,900, RMB 479,600, and RMB 553,500 respectively, accounting for approximately 5.3% of the total investment.

[0106] In summary, ① as photovoltaic penetration increases, the penalty cost for insufficient distribution network flexibility also increases, necessitating the allocation of more flexibility resources; the ESS investment costs for photovoltaic penetration rates of 40%, 60%, and 80% are RMB335,600, RMB1,067,500, and RMB1,521,200 respectively, with the penalty cost for insufficient flexibility decreasing by RMB166,300, RMB891,100, and RMB1,441,000 respectively, representing reductions of 88.27%, 73.52%, and 75.60%; ② comprehensively considering demand-side response in distribution network energy storage planning can achieve... A more reasonable and economical ESS planning scheme improves the economy and flexibility of the distribution network; the investment cost of ESS considering single and two types of demand response is reduced by RMB 57,700 and RMB 116,900 respectively, and the equivalent load curve considering demand response is smoother, which can mitigate the fluctuation of net load; ③ The method of upper and lower layer correlation modeling is used to convert the two-layer planning model of distribution network energy storage into a single layer, and then call the Gurobi solver for solution; compared with the method of solving the two-layer model, the upper and lower layer iterative loops are not required during the solution, which improves the solution efficiency.

[0107] This invention presents a two-layer planning method for energy storage in distribution networks considering demand response under different photovoltaic penetration rates. In practice, this invention analyzes the flexibility of the distribution network based on photovoltaic penetration rate and uses principal component analysis to reduce the dimensionality of a large amount of historical photovoltaic load data, preserving the main characteristics of the data while simplifying the data structure, thereby improving clustering efficiency. Then, a hierarchical clustering algorithm is used to obtain typical photovoltaic load scenarios, and on this basis, a two-layer energy storage planning model considering interruptible loads and price-based demand-side response is established. By performing correlation modeling between the upper and lower layers of the model, it is converted into a single-layer model for solution. This invention comprehensively considers demand-side response in distribution network energy storage planning, resulting in a more reasonable and economical ESS planning scheme, improving the economy and flexibility of the distribution network. This invention has the advantages of using correlation modeling between the upper and lower layers, converting the two-layer planning model into a single layer, improving solution efficiency, and enhancing the economy and flexibility of the distribution network.

Claims

1. A two-layer planning method for energy storage in distribution networks considering demand response under different photovoltaic penetration rates, characterized by: The method comprises the following steps: Step 1: Distribution network flexibility resource modeling: The supply of distribution network flexibility mainly comes from the energy storage devices at each node, the interruptible load on the demand side, and the upper main grid and other flexibility resources; Step 2: Construct a demand response model: Demand response refers to users adjusting their electricity usage through price information or incentive mechanisms, including price-based demand response and interruptible loads in incentive-based demand response; Step 3: Two-tier planning of energy storage for distribution networks; Step 4: Case analysis.

2. The method for planning energy storage in a distribution network considering demand response under different photovoltaic penetration rates according to claim 1, characterized in that: The distribution network flexibility resource modeling in step 1 is specifically as follows: the calculation formula for energy storage, interruptible load and flexibility regulation capacity of the main network is as follows: Where: are the upward and downward adjustment capabilities of ESS at node i at time t; P ESS,i 、E ESS,i are the rated power and rated capacity of ESS at node i respectively; is the power of ESS at node i at time t; is the power stored in ESS at node i at time t; is the lower limit of ESS power at node i; η is the charging and discharging efficiency of ESS; are the upward and downward regulation capabilities provided by the interruptible load at node i at time t; v j,t is the interruption status of the interruptible load at node i at time t, interruption is 1, otherwise it is 0; is the interruption amount of the interruptible load at node i at time t; are the upward and downward regulation capabilities provided by the upper power grid in period t; ΔP grid,max is the maximum climbing power of the upper main grid; P grid,max is the maximum transmission power of the upper main network; P t grid The active power transmitted by the main network at any time.

3. The dual-layer planning method for energy storage in distribution networks considering demand response under different photovoltaic penetration rates as claimed in claim 1, characterized in that: The construction of the demand response model in step 2 includes the following steps: Step 2.1: Price-based demand response model; Step 2.2: Interruptible load response model.

4. The dual-layer planning method for energy storage in distribution networks considering demand response under different photovoltaic penetration rates as claimed in claim 3, characterized in that: The price-based demand response model in step 2.1 is specifically as follows: price-based demand response affects the user's electricity consumption through electricity prices, and the user's response to electricity prices is usually described by the demand elasticity coefficient: in In the formula, e ii is the elastic coefficient; e ij is the mutual elastic coefficient; are the electricity prices at time i and time j before implementing demand response; c i 、c j are the electricity prices at time i and time j after the implementation of demand response; q i are the electricity consumption before and after the electricity price response respectively; if a day is divided into 24 hours, the elasticity coefficient matrix can be expressed as: The electricity quantity change matrix after the electricity price changes is expressed as: The change in user electricity consumption after the electricity price changes is: After the electricity price changes, the user's electricity consumption is:

5. The method for planning energy storage in a distribution network considering demand response under different photovoltaic penetration rates according to claim 3, characterized in that: The interruptible load response model in step 2.2 is specifically as follows: Its response characteristics are shown in the following formula: ① Interruption capacity constraint: Where: are the upper and lower limits of the interruption capacity of the interruptible load at node i respectively; ② Constraint on the number of interruptions: Where: N CL The maximum number of interruptions that can be made within a cycle; ③ Maximum interruption time constraint: Where: is the maximum interruption time of the interruptible load at node i.

6. The dual-layer planning method for energy storage in distribution networks considering demand response under different photovoltaic penetration rates as claimed in claim 1, characterized in that: The two-tier planning of energy storage for the distribution network in step 3 includes the following steps: Step 3.1: Upper planning layer model: including upper objective function and upper constraint conditions; Step 3.2: Lower planning layer model: including lower objective function and lower constraint conditions; Step 3.3: Two-layer model conversion.

7. The method for planning energy storage in a distribution network considering demand response under different photovoltaic penetration rates according to claim 6, characterized in that: The upper objective function of the upper planning layer model in step 3.1 is: minC Total =C I +C O (13), where: C Total is the comprehensive cost of the distribution network; C I is the investment cost of ESS; C O The operating cost of the distribution network.

8. The method for planning energy storage in a distribution network considering demand response under different photovoltaic penetration rates according to claim 7, characterized in that: The energy storage investment cost C I for: In the formula, Ω ESS is the set of nodes to be installed in ESS; c e 、c p are ESS unit capacity and power cost respectively; E ESS,i , P ESS,i are the rated capacity and rated power of ESS installed at node i; the distribution network operation cost C O for: In the formula, C Loss , C CL , C Grid , C FL , C Carbon They are energy storage operation and maintenance costs, network loss costs, interruptible load response costs, electricity purchase costs, insufficient flexibility penalty costs, and carbon emission penalty costs; is the unit charge and discharge cost of ESS; N is the number of operating days of the distribution network in a year; Φ is the typical operating scenario set; p s is the probability of scene s appearing; is the charging and discharging power of ESS at node i at time t in scenario s; c loss is the unit network loss cost; is the active power loss of line ij at time t in scenario s; Ω Line is the line set; c cl is the unit cost of interruptible load; Ω CL is a set of interruptible load nodes; is the interruption amount of the interruptible load at node i at time t in scenario s; is the electricity purchase price in time period t in scenario s, and is the time-of-use electricity price; is the amount of electricity purchased from the upper power grid at time t in scenario s; c up 、c doen are the upward and downward flexibility deficiency penalty coefficients respectively; are the load shedding amount and light abandonment amount at time t in scene s, respectively; c carbon Penalty cost for carbon emission unit; γ grid It is the carbon emission intensity of the parent grid unit.

9. The method for planning energy storage in a distribution network considering demand response under different photovoltaic penetration rates according to claim 8, characterized in that: The upper-level constraints of the upper-level planning model in step 3.1 include: ① ESS installation node installation constraints: Where P ESS,max , P ESS,min are the upper and lower limits of the rated power of the ESS to be planned; E ESS,max 、E ESS,min They are the upper and lower limits of the rated capacity of the ESS to be planned; ② PV penetration rate constraint: In the formula, Ω PV It is the installation node of distributed photovoltaic; is the installed capacity of distributed photovoltaic at node i; α is the photovoltaic penetration rate; is the peak load.

10. The method for planning energy storage in a distribution network considering demand response under different photovoltaic penetration rates according to claim 6, characterized in that: The lower objective function of the lower planning layer model in step 3.2 is to minimize the operating cost target: An optimization model is established with ESS charging and discharging single power, interruption of interruptible load and electricity price as decision variables; the lower-level constraints include ESS operation constraints, price-type demand response constraints, interruptible load constraints, upper main grid active power injection constraints, PV operation constraints, power flow constraints and system safety constraints.

Images