Multi-time scale reactive power dispatching method and system for distribution network containing controllable photovoltaics

By constructing the photovoltaic power generation inverter model and the multi-time scale optimization strategy model of the distribution network, and using two-stage robust optimization technology, the problem of failure to consider the uncertainty of photovoltaic power generation in the existing technology is solved, and better distribution is achieved Grid operation adaptability and grid stability.

CN115313504BActive Publication Date: 2024-11-22国网甘肃省电力公司陇南供电公司 +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210043778.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-01-14
Publication Date
2024-11-22
Estimated Expiration
2042-01-14

AI Technical Summary

Technical Problem

The existing active distribution network optimization operation technology failed to effectively consider the uncertainty impact brought by photovoltaic power generation, and it was difficult to adapt to the actual distribution network operation. The optimization strategy is based on the recent prediction curve, making it difficult to cope with the operation strategy of photovoltaic power generation inverter. .

Method used

The control strategy for the photovoltaic power generation inverter model is constructed, combined with the multi-time scale optimization strategy model of the distribution network, and adopts two-stage robust optimization technology, considering the switching of the photovoltaic power generation inverter, and establishing the goal of minimum grid loss and minimum voltage deviation The scheduling model is solved by CCG algorithm, and the optimal control strategy and reactive power compensation guide daily energy scheduling.

Benefits of technology

It can better adapt to the actual distribution network operation, reduce grid losses, improve the safety and voltage stability of the power grid, and deal with uncertainty and load fluctuations in photovoltaic power generation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115313504B_ABST
    Figure CN115313504B_ABST
Patent Text Reader

Abstract

The present invention discloses a multi-time scale reactive power dispatching method and system for a distribution network containing controllable photovoltaics. According to the day-ahead dispatching result, an intraday optimization model is established with the goal of minimizing voltage deviation to obtain the optimal PV control strategy and PV reactive compensation amount. The uncertainty impact brought by photovoltaic power generation is taken into consideration, and then the CCG algorithm is used to solve the above model. The two-stage robust optimization technology is used to solve the active distribution network optimization strategy under the worst case, and the operation strategy of the photovoltaic power generation inverter is considered to better adapt to the actual distribution network operation. This scheme combines traditional voltage regulation equipment, energy storage elements and photovoltaic power generation operation strategies to propose a two-stage robust optimization dispatching method for active distribution networks to cope with the problems of large voltage deviation, voltage fluctuation and large network loss in the distribution network under the uncertainty of photovoltaic power generation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of energy dispatching, and in particular to a multi-time scale reactive power dispatching method and system for a distribution network containing controllable photovoltaics. Background Art

[0002] my country will vigorously increase the penetration rate of distributed photovoltaic power generation. The form of the distribution network system is undergoing fundamental changes. The architecture of the active distribution network has begun to take shape, and photovoltaic power generation (PV) grid-connected technology has gradually changed from providing only active power to coordinated active and reactive power output. Therefore, how to achieve the coordination of photovoltaic storage technology and reactive power optimization, while maximizing the response to the uncertainty of photovoltaic power generation, is of great practical significance.

[0003] The current active distribution network optimization operation technology has a deterministic model, but the deterministic model does not take into account the uncertainty of photovoltaic power generation, and it is difficult to deal with the uncertainty of photovoltaic power generation; there are random scenarios that simulate the uncertainty of photovoltaic power generation, but the simulation scenario calculation is huge, and its robustness is usually poor. In the existing technology, there is a two-stage robust optimization technology to solve the active distribution network optimization strategy under the worst case, but it does not consider the operation strategy of the photovoltaic power inverter. At the same time, its optimization strategy is based on the day-ahead prediction curve, which is difficult to adapt to the actual distribution network operation.

[0004] In view of this, the object of the present invention is to provide a multi-time-scale reactive power scheduling method and system for a distribution network containing controllable photovoltaics. Summary of the invention

[0005] The technical problem to be solved by the present invention is that the existing active distribution network optimization operation technology has a deterministic model, but does not take into account the uncertainty impact brought by photovoltaic power generation, and it is difficult to cope with the uncertain impact of photovoltaic power generation. The optimization strategy is based on the day-ahead prediction curve, which is difficult to adapt to the actual distribution network operation; the purpose of the present invention is to provide a multi-time scale reactive power scheduling method and system for a distribution network containing controllable photovoltaics to solve this technical problem.

[0006] The present invention is achieved through the following technical solutions:

[0007] This scheme provides a multi-time scale reactive power dispatching method for a distribution network including controllable photovoltaics, including the following steps:

[0008] Step 1: Construct the control strategy of the photovoltaic power inverter model;

[0009] Step 2: Construct a multi-time scale optimization strategy model for the distribution network, wherein the multi-time scale optimization strategy model for the distribution network takes minimizing network loss as the objective function of the day-ahead scheduling stage and minimizing voltage deviation as the objective function of the intraday scheduling stage; and the multi-time scale optimization strategy model for the distribution network considers the switching of the photovoltaic power generation inverter model to construct the constraint conditions of the objective function;

[0010] Step 4: Solve the multi-time scale optimization strategy model of the distribution network based on the two-stage robust optimization model to obtain the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive power compensation;

[0011] Step 5: Guide daily energy scheduling based on the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive power compensation.

[0012] Working principle of this scheme: The present invention first gives four control models of active-reactive power of photovoltaic inverters, and then establishes a two-stage robust day-ahead optimization model of the distribution network with the goal of minimizing network loss. According to the day-ahead scheduling results, with the goal of minimizing voltage deviation, an intraday optimization model is established to obtain the optimal PV control strategy and PV reactive compensation. Considering the uncertainty of photovoltaic power generation, the CCG algorithm is then used to solve the above model; the two-stage robust optimization technology is used to solve the active distribution network optimization strategy under the worst conditions, and the operation strategy of photovoltaic power inverter is considered to better adapt to the actual distribution network operation; this scheme combines traditional voltage regulation equipment, energy storage components and photovoltaic power generation operation strategies to propose a two-stage robust optimization scheduling method for active distribution networks to deal with the problems of large voltage deviation, voltage fluctuation and large network loss in the distribution network under the uncertainty of photovoltaic power generation.

[0013] Furthermore, the control strategies of the photovoltaic power generation inverter model include: no control mode strategy, reactive power control mode strategy (Reactive power control, RPC), active power curtailment mode strategy (active power curtailment, APC) and optimal inverter dispatch strategy (optimal inverter dispatch, OID).

[0014] Furthermore, the strategy that does not adopt the control method (strategy a) is:

[0015] The reactive power control strategy (strategy b) is:

[0016] The active power limitation strategy (strategy c) is:

[0017] The inverter optimization scheduling strategy (strategy d) is:

[0018] Where P PV is the maximum output of photovoltaic power; S G is the apparent power of the photovoltaic inverter; θ is the power factor angle of the photovoltaic inverter, Q G The maximum value of reactive power generated by the inverter; Q pv_out and P pv_out is the actual output of photovoltaic reactive power and the actual output of photovoltaic active power, Q pv_out,max It is the upper limit of photovoltaic output.

[0019] Robust optimization (RO) is actually an optimization strategy based on the day-ahead forecast curves of load and photovoltaic power generation. However, the distribution network is always affected by sudden fluctuations in photovoltaic power generation or load in actual operation. Therefore, this paper proposes a multi-time scale and multi-device coordinated optimization strategy to cope with the uncertainty of photovoltaic power generation and load while maintaining the safe voltage of the power grid and minimizing network losses.

[0020] Day-ahead dispatch stage: Day-ahead dispatch aims to obtain SVC, ESS, OLTC and SC dispatch strategies through robust optimization of RO. In this stage, PV power generation and load fluctuations are predicted within a limited time interval T, and upper and lower limits of changes are set. They can change randomly between boundaries. Different from the classic multi-time scale optimization scheduling method, this scheme considers the mode switching of photovoltaic inverters while adding traditional VVC devices to the scheduling in the day-ahead scheduling stage; in the robust optimization RO coordination model of this stage, VVC devices and inverter modes are optimized as control variables. It is worth noting that VVC devices are compensation operations under the worst conditions, and photovoltaic inverters are also the mode under the worst conditions. This scheme selects every 4 hours as a scheduling cycle, and uses robust optimization RO to obtain the scheduling strategy of VVC devices; however, only the scheduling strategy of the first hour is retained, and the strategies of the remaining 3 hours are discarded; and in the day-ahead scheduling stage, VVC devices will be affected by the mode switching of photovoltaic inverters. However, in the actual operation of the distribution network, photovoltaic power generation will operate according to the actual operation curve, so the strategy selection of photovoltaic inverters will also be discarded and will be reselected in the intraday scheduling.

[0021] Intraday scheduling stage: The day-ahead operation of the VVC equipment is a strategy made under the worst conditions of the forecast curve. In the intraday stage, the PV inverter makes a scheduling plan in a shorter period of time to stabilize the system voltage level and cope with sudden fluctuations in PV power generation and load.

[0022] Furthermore, the objective function of the day-ahead scheduling phase with the minimum network loss is expressed as:

[0023] Where f1 is the distribution network loss; I ij,t is the current value of branch ij at time t; T is the number of time periods in the dispatch cycle; N is the number of nodes in the distribution network; Δt is the dispatch interval;

[0024] Taking the minimum voltage deviation as the objective function of the intraday scheduling stage as follows:

[0025] Where V ref The reference voltage is 1.0 pu. The control variables in the intraday stage are the control mode and reactive power output of the photovoltaic inverter to adjust the photovoltaic inverter.

[0026] Furthermore, the constraints include: radial distribution network branch power flow constraints, node voltage constraints, shunt capacitor group constraints, energy storage device operation constraints, static VAR compensation device constraints, on-load voltage regulator constraints, and photovoltaic inverter control mode switching constraints;

[0027] The radial distribution network branch power flow constraint is: Figure 2 The Distflow power flow model of the radial distribution network shown is as follows: For node j in the distribution network, there is:

[0028]

[0029]

[0030]

[0031]

[0032]

[0033] In the formula, σ(j) represents the branch set with node j as the first node, π(j) represents the branch set with node j as the last node; P ij,t and Q ij,t They represent the active power and reactive power flowing through branch ij respectively; l ij,t 、r ij,t and x ij,t Respectively represent the square of the current, resistance and reactance of branch ij; V j,t , Q j,t and P j,t are the voltage amplitude, injected reactive power and active power of node j respectively; P g,j,t and Q g,j,tP is the active power and reactive power generated by the distributed power source connected to node j at time t; ch,j,t P is the charge of the energy storage element connected to node j at time t; dch,j,t is the discharge amount of the energy storage element connected to node j at time t; Q sc,j,t is the output of the reactive power compensation capacitor group connected to node j at time t; Q svc,j,t is the output of the SVC connected to node j at time t; It represents a branch circuit without an on-load voltage regulator installed; Represents the branch circuit where the on-load voltage regulator is installed; H ij is the transformation ratio of the on-load voltage regulator on branch ij;

[0034] The node voltage constraint is:

[0035] In the formula, and are the minimum voltage and maximum voltage of node i respectively;

[0036] The branch current constraint is:

[0037] In the formula, is the maximum current of branch ij;

[0038] The parallel capacitor bank constraint is:

[0039] Q sc,j,t =K sc,j,t ΔQ sc (12)

[0040]

[0041] Shunt capacitors are discrete reactive power compensation devices, where k sc,j,t is the number of capacitors switched on and off at node j at time t, ΔQ sc is the reactive power compensation of a single capacitor, and is the minimum value of the switched capacitor group at node j and the maximum value of the switched capacitor group;

[0042] The parallel capacitor group can only be switched in groups, so the model of the parallel capacitor group (13) is a discrete optimization model, which is converted into a 0-1 integer programming model by the following formula:

[0043]

[0044] In the formula, k is determined according to the specific number of installed switchable capacitor groups. is the upper limit of the number of switchable capacitor groups at node j; in order to protect the service life of the switchable capacitor group SC, the number of switching times of the switchable capacitor group SC in one day is limited.

[0045] ε 1j (t)-ε 2j (t)K sc,j ≤K sc,j (t+1)-K sc,j (t)≤ε 1j (t)K sc,j -ε 2j (t)

[0046] where ε 1j (t),ε 2j (t)∈{0,1} (15)

[0047] 0≤ε 1j (t)+ε 2j (t)≤1 (16)

[0048]

[0049] In the formula, ε 1j (t), ε 2j (t) is a binary auxiliary variable, is the maximum switching times of the switchable capacitor bank SC;

[0050] The energy storage device operation constraints are:

[0051]

[0052]

[0053]

[0054] D ch,j,t +D dch,j,t ≤1 (21)

[0055]

[0056] Formula (17) represents the power constraint of the energy storage system ESS, E j,t is the power of the energy storage system connected to node j at time t; P ch,j,t and P dch,j,t are the charging power and discharging power of the corresponding energy storage system ESS in period t respectively; η ch and η dch They are charging efficiency and discharging efficiency respectively. To ensure that the electrical efficiency has the same regulation capability in the new dispatching cycle, the initial energy storage of each dispatching cycle must be the same.

[0057] and are the upper limit of charging power and discharging power of the energy storage system ESS at node j respectively; since the charging and discharging of the energy storage system cannot be performed at the same time, D ch,j,t and D dch,j,t Two 0-1 variables indicate that the energy storage system ESS can only be in one of the three states of charging, discharging and neither charging nor discharging during the period t;

[0058] is the upper limit of the energy storage system ESS at node j. Considering the service life of the energy storage device, its actual charging and discharging range is set to 20%-100%;

[0059] The static VAR compensation device is constrained as follows:

[0060] In the formula, and are the upper and lower limits of the static VAR compensation device SVC; Q svc,t is the actual output of the static VAR compensation device SVC; t is the current time period;

[0061] The OLTC constraint is that the OLTC can only operate in one gear, so the ratio of the OLTC is is a discrete optimization variable, n is the number of gears of the on-load voltage regulator; therefore, By introducing 0-1 binary variables and the Big M method, it is transformed into a 0-1 programming problem;

[0062]

[0063]

[0064]

[0065] In the formula, m ij,k,t and h j,k,t are auxiliary variables for transforming the problem; the constraints in equation (25) are two complementary conditions, taking the optimal condition of the objective function. M is an arbitrarily large real number.

[0066] In a dispatching cycle, the number of gear adjustments of the on-load voltage regulator will be limited, that is,

[0067]

[0068] Where N T.max is the maximum number of transformer switching times in a dispatching cycle; n is the number of transformer gears.

[0069] The photovoltaic inverter control mode switching constraint is: the photovoltaic inverter needs to switch modes, so the control strategy needs to be transformed; the transformation from not adopting the control mode strategy to adopting the reactive power control mode strategy:

[0070] P pv_out =P PV (28)

[0071]

[0072] -Q pv_out,max z g ≤Q pv_out ≤Q pv_out,max z g (30)

[0073] Q pv_out,max z g is a non-convex bilinear term consisting of a continuous variable and a binary variable, so the big M method is used to reformulate formula (30);

[0074]

[0075] Conversion from non-control strategy to inverter optimization scheduling strategy:

[0076] P PV (1-z f )≤P pv_out ≤P PV (32)

[0077]

[0078] -P pv_out z f ≤Q pv_out ≤P pv_out z f (34)

[0079] Similarly, P pv_out z f Re-expressed using the Big M method is;

[0080]

[0081] -G pv_out ≤Q pv_out ≤G pv_out (36)

[0082]

[0083] The PV inverter installed at a certain node can independently select the control strategy, and its PQ capability can switch between strategy a and strategy d or strategy a and strategy b. For the inverter with the ability to switch between the strategy without adopting the control mode and the strategy adopting the reactive power control mode, z g =0 means the inverter does not use the control mode strategy. g =1 means that the inverter selects the reactive power control strategy to operate; for the inverter with four strategy switching capabilities, z f =0 means the inverter does not use the control mode strategy. f =1 means that the inverter optimization scheduling strategy is selected for the inverter.

[0084] A further optimization scheme is that the constraints of the objective function of the day-ahead scheduling stage include: formula (2), formula (3), formula (5), formula (29), formula (31), formula (32), formula (35) and formula (37); the constraints of the objective function of the intraday scheduling stage include: formula (2), formula (3), formula (5), formula (11), formula (28), formula (29), formula (30), formula (31), formula (35) and formula (37).

[0085] The traditional two-stage robust optimization of distribution networks only considers the load and photovoltaic active output capacity as uncertainty sets, and does not take the strategy switching capability of PV inverters into consideration in the optimization. In fact, the operation strategy of PV inverters will vary according to the actual situation of the distribution network. Therefore, it is particularly important to establish a two-stage robust optimization model that considers both the uncertainty of photovoltaic output and the switching of photovoltaic operation strategies; specifically, the binary variable Z used to select the PV inverter is added to the first-stage variables of RO, and it is always subject to the newly added column constraints during the RO solution process.

[0086] A further optimization scheme is that the two-stage robust optimization model is:

[0087]

[0088] Z={z j ∈{0,1},z g ∈{0,1}}

[0089] In the formula,

[0090] A further optimization scheme is that step 4 includes the following sub-steps:

[0091] Step 4.1: Divide the robust optimization solution into the main problem and sub-problems, which are expressed as the first stage and the second stage respectively.

[0092] Step 4.2: Add the binary variable Z used to select the PV inverter to the first-stage variables of the robust optimization, and make it always subject to the newly added column constraints during the robust optimization solution process to obtain a two-stage robust optimization model;

[0093] Step 4.3: Use the CCG algorithm to iteratively solve the main problem and sub-problems until the difference between the lower bound of the main problem and the upper bound of the sub-problem is less than the convergence criterion, and the problem is solved; the strong duality theorem and the outer approximation method are used to transform the sub-problem into a solvable linear programming problem to obtain the optimal solution for the second stage.

[0094] Standard form of two-stage robust optimization model Cy is essentially a "min-max-min" three-layer robust optimization RO problem; the solution to the "min-max-min" robust optimization RO problem can be divided into a main problem and a sub-problem, which can be expressed as two stages respectively; then the CCG (column constraint generation) algorithm is used to iteratively solve the main and sub-problems until the difference between the lower limit of the main problem and the upper limit of the sub-problem is less than the convergence criterion, and the problem is solved. The first stage of RO is a "min" problem; the second stage is a "max-min" problem. This scheme uses the strong duality theorem and the outer approximation method (OA decomposition method) to transform the "max-min" problem into a solvable linear programming problem, thereby obtaining the optimal solution for the second stage.

[0095] A further optimization scheme is that the first stage includes inequality constraints, PV switching constraints and equality constraints, and the second stage includes inequality constraints and inequality constraints; the inequality constraints of the first stage correspond to equations (10), (11), (13), (14), (15), (16), (17), (19), (20), (21), (22), (27) and (36); the PV switching constraints of the first stage correspond to equations (31), (32) and (37); the equality constraints of the first stage correspond to equations (5), (6), (7), (12), (18), (24), (25), (26) and (28); the inequality constraints of the second stage correspond to equation (23); the second-order cone constraints of the second stage correspond to equations (9), (29) and (35).

[0096] The present solution also provides a multi-time-scale reactive power dispatching system for a distribution network including controllable photovoltaics, which is constructed according to the above method and includes: a control strategy construction module, a model construction module, a calculation module and a guidance module; the control strategy construction module is used to construct a control strategy for a photovoltaic power generation inverter model; the model construction module is used to construct a multi-time-scale optimization strategy model for a distribution network, the multi-time-scale optimization strategy model for a distribution network takes minimizing network loss as the objective function of the day-ahead dispatching stage and minimizing voltage deviation as the objective function of the intraday dispatching stage; the model construction module is also used to consider the switching of the photovoltaic power generation inverter model to construct the constraint conditions of the objective function; the calculation module is used to solve the multi-time-scale optimization strategy model for a distribution network based on a two-stage robust optimization model to obtain the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive compensation amount; the guidance module is used to guide daily energy dispatching based on the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive compensation amount.

[0097] In order to cope with the uncertainty of photovoltaic power generation and load, this scheme establishes a multi-time scale robust optimization model of active distribution network. The first stage of robust optimization in the day-ahead stage is the optimization strategy of OLTC, SC, PV inverter reactive compensation and ESS, and the second stage is the optimization strategy of SVC. The uncertainty is the fluctuation of active-reactive coordinated photovoltaic output and load. The scheduling results of OLTC, SC, SVC and ESS in the first stage are retained to the intraday stage, and the operation strategy and reactive compensation of PV inverter are re-optimized in the intraday stage. This model dispatches various fast and slow-moving equipment in the distribution network in a targeted manner, coordinates PV inverter strategies, and fully guarantees the safe operation of the distribution network.

[0098] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0099] The multi-time-scale reactive power dispatching method and system of the distribution network containing controllable photovoltaics provided by the present invention establishes an intraday optimization model based on the day-ahead dispatching result and takes the minimum voltage deviation as the goal to obtain the optimal PV control strategy and PV reactive compensation amount, taking into account the uncertainty impact brought by photovoltaic power generation, and then adopts the CCG algorithm to solve the above model; the two-stage robust optimization technology is used to solve the active distribution network optimization strategy under the worst case, and the operation strategy of the photovoltaic power generation inverter is considered, which can better adapt to the actual distribution network operation; this scheme combines traditional voltage regulation equipment, energy storage components and photovoltaic power generation operation strategies to propose a two-stage robust optimization dispatching method for active distribution networks to cope with the problems of large voltage deviation, voltage fluctuation and large network loss in the distribution network under the uncertainty of photovoltaic power generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0101] Figure 1 It is a flow chart of a multi-time-scale reactive power dispatching method for a distribution network including controllable photovoltaics;

[0102] Figure 2 Schematic diagram of four control strategies for PV inverters;

[0103] Figure 3 This is a schematic diagram of the Distflow power flow model;

[0104] Figure 4 Schematic diagram of the multi-time scale optimization model framework;

[0105] Figure 5 The CCG method solution flow chart for two-stage robust optimization;

[0106] Figure 6 This is a schematic diagram of the IEEE33 node system of Example 3;

[0107] Figure 7 This is a predicted fluctuation curve diagram of photovoltaic power and load in Example 3;

[0108] Figure 8 This is a schematic diagram of the day-ahead decision of SC and OLTC in Example 3;

[0109] Fig. 9 Schematic diagram of the day-ahead decision of SVC and ESS in Example 3;

[0110] Fig.10 Schematic diagram of the day-ahead decision of the PV inverter in Example 3;

[0111] Fig.11 Schematic diagram of reactive power input of the photovoltaic inverter in Example 3;

[0112] Fig.12 This is a schematic diagram of intraday decision-making of a PV inverter in Example 3;

[0113] Fig.13 This is a schematic diagram of the scheduling result of the PV inverter in Example 3;

[0114] Fig.14 Schematic diagram of network loss comparison in three scenarios in Example 3;

[0115] Fig.15Schematic diagram of voltage comparison of three scenarios in Example 3;

[0116] Fig.16 Schematic diagram of network loss comparison of two optimization methods in Example 3;

[0117] Fig.17 Schematic diagram of voltage comparison of two optimization methods in Example 3. DETAILED DESCRIPTION

[0118] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0119] Example 1

[0120] This embodiment provides a multi-time scale reactive power dispatching method for a distribution network including controllable photovoltaics, such as Figure 1 As shown, the steps include: Step 1: constructing a control strategy for a photovoltaic power generation inverter model;

[0121] Step 2: Construct a multi-time scale optimization strategy model for the distribution network, wherein the multi-time scale optimization strategy model for the distribution network takes minimizing network loss as the objective function of the day-ahead scheduling stage and minimizing voltage deviation as the objective function of the intraday scheduling stage; and the multi-time scale optimization strategy model for the distribution network considers the switching of the photovoltaic power generation inverter model to construct the constraint conditions of the objective function;

[0122] Step 4: Solve the multi-time scale optimization strategy model of the distribution network based on the two-stage robust optimization model to obtain the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive power compensation;

[0123] Step 5: Guide daily energy scheduling based on the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive power compensation.

[0124] like Figure 2 As shown, the control strategies of the photovoltaic power generation inverter model include: no control mode strategy, reactive power control mode strategy, active power limitation mode strategy and inverter optimization scheduling strategy.

[0125] The strategy of not adopting the control mode is expressed as:

[0126] The reactive power control strategy is expressed as:

[0127] The active power limiting strategy is expressed as:

[0128] The inverter optimization scheduling strategy is expressed as:

[0129] Where P PV is the maximum output of photovoltaic power; S G is the apparent power of the photovoltaic inverter; θ is the power factor angle of the photovoltaic inverter, Q G The maximum value of reactive power generated by the inverter; Q pv_out and P pv_out is the actual output of photovoltaic reactive power and the actual output of photovoltaic active power, Q pv_out,max It is the upper limit of photovoltaic output.

[0130] Robust optimization (RO) is actually an optimization strategy based on the day-ahead forecast curve of load and photovoltaic power generation. However, in actual operation, the distribution network is always affected by sudden fluctuations in photovoltaic power generation or load. Therefore, this paper proposes a multi-time scale multi-device coordinated optimization strategy to cope with the uncertainty of photovoltaic power generation and load while maintaining the safe voltage of the power grid and minimizing network losses. The optimization strategy framework is as follows Figure 4 shown.

[0131] Day-ahead dispatching stage: Day-ahead dispatching aims to obtain SVC, ESS, OLTC and SC dispatching strategies through RO; in this stage, PV power generation and load fluctuations are predicted within a limited time interval T, and upper and lower limits of changes are set. They can vary randomly between boundaries.

[0132] Different from the classic multi-time scale optimization scheduling method, this scheme considers the mode switching of PV inverters while adding traditional VVC equipment to the scheduling during the day-ahead scheduling stage.

[0133] Intraday dispatch stage: The day-ahead operation of the VVC equipment is a strategy made under the worst conditions of the forecast curve. In the intraday stage, the PV inverter makes a dispatch plan in a shorter period of time to stabilize the system voltage level and cope with sudden fluctuations in PV power generation and load.

[0134] The objective function of the day-ahead scheduling phase with the network loss minimized is expressed as:

[0135] Where f1 is the distribution network loss; I ij,t is the current value of branch ij at time t; T is the number of time periods in the dispatch cycle; N is the number of nodes in the distribution network; Δt is the dispatch interval;

[0136] Taking the minimum voltage deviation as the objective function of the intraday scheduling stage as follows:

[0137] Where V ref is the reference voltage value, take 1.0pu.

[0138] The constraints include: radial distribution network branch power flow constraints, node voltage constraints, shunt capacitor group constraints, energy storage device operation constraints, static VAR compensation device constraints, on-load voltage regulator constraints and photovoltaic inverter control mode switching constraints;

[0139] The branch power flow constraint of the radial distribution network is:

[0140] Distflow power flow model Figure 3 As shown, for node j in the distribution network, there is:

[0141]

[0142]

[0143]

[0144]

[0145]

[0146] In the formula, σ(j) represents the branch set with node j as the first node, π(j) represents the branch set with node j as the last node; P ij,t and Q ij,t They represent the active power and reactive power flowing through branch ij respectively; l ij,t 、r ij,t and x ij,t Respectively represent the square of the current, resistance and reactance of branch ij; V j,t , Q j,t and P j,t are the voltage amplitude, injected reactive power and active power of node j respectively; P g,j,t and Q g,j,t P is the active power and reactive power generated by the distributed power source connected to node j at time t; ch,j,t P is the charge of the energy storage element connected to node j at time t; dch,j,t is the discharge amount of the energy storage element connected to node j at time t; Q sc,j,t is the output of the reactive power compensation capacitor group connected to node j at time t; Q svc,j,t is the output of the SVC connected to node j at time t; It represents a branch circuit without an on-load voltage regulator installed; Represents the branch circuit where the on-load voltage regulator is installed; H ij is the transformation ratio of the on-load voltage regulator on branch ij;

[0147] The node voltage constraint is:

[0148] In the formula, and are the minimum voltage and maximum voltage of node i respectively;

[0149] The branch current constraint is:

[0150] In the formula, is the maximum current of branch ij;

[0151] The parallel capacitor bank constraint is: Q sc,j,t =K sc,j,t ΔQ sc (12)

[0152]

[0153] In the formula, k sc,j,t is the number of capacitors switched on and off at node j at time t, ΔQ sc is the reactive power compensation of a single capacitor, and is the minimum value of the switched capacitor group at node j and the maximum value of the switched capacitor group;

[0154] The parallel capacitor group can only be switched in groups, so the model of the parallel capacitor group (13) is a discrete optimization model, which is converted into a 0-1 integer programming model by the following formula:

[0155] In the formula, k is determined according to the specific number of installed switchable capacitor groups. is the upper limit of the number of switchable capacitor groups at node j;

[0156] ε 1j (t)-ε 2j (t)K sc,j ≤K sc,j (t+1)-K sc,j (t)≤ε 1j (t)K sc,j -ε 2j (t)

[0157] where ε 1j (t),ε 2j (t)∈{0,1} (15)

[0158] 0≤ε 1j (t)+ε 2j (t)≤1 (16)

[0159]

[0160] In the formula, ε 1j (t), ε 2j (t) is a binary auxiliary variable, is the maximum switching times of the switchable capacitor bank;

[0161] The energy storage device operation constraints are:

[0162]

[0163]

[0164]

[0165] D ch,j,t +D dch,j,t ≤1 (21)

[0166]

[0167] Formula (17) represents the power constraint of the energy storage system, E j,t is the power of the energy storage system connected to node j at time t; P ch,j,t and P dch,j,t are the charging power and discharging power of the corresponding energy storage system in period t respectively; η ch and η dch They are charging efficiency and discharging efficiency respectively. To ensure that the electrical efficiency has the same regulation capability in the new dispatching cycle, the initial energy storage of each dispatching cycle must be the same. and are the upper limit of charging power and discharging power of the energy storage system at node j respectively; since the charging and discharging of the energy storage system cannot be performed at the same time, D ch,j,t and D dch,j,t Two 0-1 variables represent that the energy storage system can only be in one of the three states of charging, discharging and neither charging nor discharging during the period t; E j max is the upper limit of the energy storage system at node j. Considering the service life of the energy storage device, its actual charging and discharging range is set to 20%-100%;

[0168] The static VAR compensation device is constrained as follows:

[0169] In the formula, and are the upper and lower limits of the static VAR compensation device respectively; Q svc,t is the actual output of the static VAR compensation device; t is the current time period;

[0170] The on-load voltage regulator is constrained as follows: the on-load voltage regulator can only operate in one of the gears, so the ratio of the on-load voltage regulator is a discrete optimization variable, n is the number of gears of the on-load voltage regulator; therefore, By introducing 0-1 binary variables and the Big M method, it is transformed into a 0-1 programming problem;

[0171]

[0172]

[0173]

[0174] In the formula, m ij,k,t and h j,k,t are auxiliary variables for transforming the problem; the constraints in formula (25) are two complementary conditions, which are the conditions for optimizing the objective function; M is an arbitrarily large real number

[0175] In a dispatching cycle, the number of gear adjustments of the on-load voltage regulator will be limited, that is,

[0176]

[0177] Where N T.max is the maximum number of transformer switching times in a dispatching cycle; n is the number of transformer gears.

[0178] The photovoltaic inverter control mode switching constraints are:

[0179] The photovoltaic inverter needs to switch modes, so the control strategy needs to be transformed;

[0180] Conversion from a model without control mode to a model with reactive power control mode:

[0181] P pv_out =P PV (28)

[0182]

[0183] -Q pv_out,max z g ≤Q pv_out ≤Q pv_out,max z g (30)

[0184] Q pv_out,max zg is a non-convex bilinear term consisting of a continuous variable and a binary variable, so the big M method is used to reformulate formula (30);

[0185]

[0186] Conversion from the non-control mode model to the inverter optimization scheduling model:

[0187] P PV (1-z f )≤P pv_out ≤P PV (32)

[0188]

[0189] -P pv_out z f ≤Q pv_out ≤P pv_out z f (34)

[0190] Similarly, P pv_out z f Re-expressed using the Big M method is;

[0191]

[0192] -G pv_out ≤Q pv_out ≤G pv_out (36)

[0193]

[0194] For an inverter that has the ability to switch between a non-control strategy and a reactive power control strategy, g =0 means the inverter does not use the control mode strategy. g =1 means that the inverter selects the reactive power control strategy to operate; for the inverter with four strategy switching capabilities, z f =0 means the inverter does not use the control mode strategy. f =1 means that the inverter optimization scheduling strategy is selected for the inverter.

[0195] The constraints of the objective function in the day-ahead scheduling stage include: equation (2), equation (3), equation (5), equation (29), equation (31), equation (32), equation (35), and equation (37); the constraints of the objective function in the intraday scheduling stage include: equation (2), equation (3), equation (5), equation (11), equation (28), equation (29), equation (30), equation (31), equation (35), and equation (37).

[0196] The traditional two-stage robust optimization of distribution networks only considers the load and photovoltaic active output capacity as uncertainty sets, and does not take the strategy switching capability of PV inverters into consideration in the optimization. In fact, the operation strategy of PV inverters will vary according to the actual situation of the distribution network. Therefore, it is particularly important to establish a two-stage robust optimization model that considers both the uncertainty of photovoltaic output and the switching of photovoltaic operation strategies. Specifically, the binary variable Z used to select the PV inverter is added to the first-stage variables of RO, and it is always subject to the newly added column constraints during the RO solution process;

[0197] Therefore, the two-stage robust optimization model is:

[0198]

[0199] Z={z j ∈{0,1},z g ∈{0,1}}

[0200] In the formula,

[0201] In the formula, and The upper and lower limits of load fluctuation.

[0202] Step 4 includes the following sub-steps:

[0203] Step 4.1: Divide the robust optimization solution into the main problem and sub-problems, which are expressed as the first stage and the second stage respectively.

[0204] Step 4.2: Add the binary variable Z used to select the PV inverter to the first-stage variables of the robust optimization, and make it always subject to the newly added column constraints during the robust optimization solution process to obtain a two-stage robust optimization model;

[0205] Step 4.3: Use the CCG algorithm to iteratively solve the main problem and sub-problems until the difference between the lower bound of the main problem and the upper bound of the sub-problem is less than the convergence criterion, and the problem is solved; the strong duality theorem and the outer approximation method are used to transform the sub-problem into a solvable linear programming problem to obtain the optimal solution for the second stage.

[0206] like Figure 5As shown, the inequality constraint Ax≤b of the first stage corresponds to equations (10), (11), (13), (14), (15), (16), (17), (19), (20), (21), (22), (27) and (36); the PV switching constraint zx≤Z; z∈{0,1} of the first stage corresponds to equations (31), (32) and (37); the equality constraint Bx+Iy+Ju=c of the first stage corresponds to equations (5), (6), (7), (12), (18), (24)-(26) and (28). The inequality constraint Fy≤f of the second stage corresponds to equation (23); the second-order cone constraint ||Ky||2≤hy of the second stage corresponds to equations (9), (29) and (35).

[0207] Example 2

[0208] This embodiment is based on a multi-time-scale energy dispatching system for a distribution network including controllable photovoltaics, and is constructed according to the method of Embodiment 1, including: a control strategy construction module, a model construction module, a calculation module and a guidance module; the control strategy construction module is used to construct a control strategy for a photovoltaic power generation inverter model; the model construction module is used to construct a multi-time-scale optimization strategy model for a distribution network, and the multi-time-scale optimization strategy model for a distribution network takes minimizing network loss as the objective function of the day-ahead dispatching stage and minimizing voltage deviation as the objective function of the intraday dispatching stage; the model construction module is also used to consider the switching of the photovoltaic power generation inverter model to construct the constraint conditions of the objective function; the calculation module is used to solve the multi-time-scale optimization strategy model for a distribution network based on a two-stage robust optimization model to obtain the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive compensation; the guidance module is used to guide daily energy dispatching based on the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive compensation.

[0209] Example 3

[0210] This embodiment selects Figure 6The simulation calculation is performed in the improved IEEE33 node system shown in the figure. The voltage of the root node is set to 1.0pu, and the voltage of the other nodes is between 0.9-1.1; the maximum current of each branch is 400A; 7 groups of capacitors are installed at nodes 16 and 31, respectively, with reactive capacity of 75kVar and 80kVar for each group; 1 group of energy storage batteries is installed at node 18, with a capacity of 600kWh, a charging and discharging power of 120kW, and a charging and discharging efficiency of 90%; PV inverters that can switch between strategies a and b are installed at node 9, PV inverters that can switch between strategies a and d are installed at node 18, and PV inverters with photovoltaic inverter control strategies a and c are installed at nodes 25 and 33. Assuming that the maximum installed capacity of photovoltaic is 200kW and the apparent power of the inverter is 105% of the installed capacity, the maximum reactive power it can output is 96kVar. SVC is installed at node 16 and node 33, with a compensation capacity of -300kVar to 300kVar.

[0211] Load and PV forecast curves such as Figure 7 As shown, the load fluctuation range is between [0.85-1.15], and the photovoltaic fluctuation range is between [0.85-1.15]. Based on Matlab platform coding, the Gurobi solver is called by Yalmip on a computer with a configuration of 2.30GHz 16.0GB and a working environment of Windows 10 system to solve the example;

[0212] Day-ahead dispatching stage: The day-ahead dispatching stage firstly performs two-stage robust optimization based on the uncertainty interval of PV active power and load demand, and considers the control strategy of PV inverter at the same time, and obtains the gear positions of OLTC and SC and the dispatching decisions of PV inverter, ESS and SVC under the PV inverter control strategy at this stage. The decision results are as follows: Figure 8 Fig. 9 and Fig.10 As shown; it is worth noting that the day-ahead scheduling decision is based on a 4-hour cycle and only the decision obtained in the first hour is retained. Since the PV inverter will reselect the scheduling strategy during the day to deal with some emergencies, the day-ahead scheduling strategy of the PV inverter will also be abandoned.

[0213] Intraday dispatch stage: In the intraday dispatch stage, the PV inverter mode and reactive output will be dispatched based on the strategies of the SVC, ESS, OLTC and SC scheduled on the day-ahead basis and on the short-term forecast (time scale of 15 minutes) of the given PV output and load demand. Taking the time from AM10:00 to AM11:00 as an example, the reactive output of the PV inverter is as follows: Fig.11As shown; the PV inverters of nodes 25 and 33 are set to strategy a and strategy c respectively, so their output reactive power is 0.

[0214] The PV inverter control strategy for scheduling 96 points during the day is as follows Fig.12 As shown in Figure 1, in the day-ahead dispatch, the PV inverter control of node 9 is operated in strategy b, while in the intraday dispatch, the PV inverter of node 18 is operated in strategy d (maximum control range); the active output P and reactive output Q capabilities of the four PV inverters (PV9, PV18, PV25 and PV33) in one day are as follows: Fig.13 shown.

[0215] In order to verify the effectiveness of intraday control of PV inverters, two scenarios are set up for comparison with the scenario of PV inverters participating in intraday scheduling. The first scenario is that the PV inverter does not participate in control during the intraday stage, and its control strategy still participates in the distribution network operation according to the control strategy generated by the day-ahead scheduling. The second scenario is a scenario where no PV inverter participates in the distribution network operation. The optimization results of these two scenarios are compared with the results of PV inverters participating in intraday scheduling. Fig.14 and Fig.15 shown.

[0216] It can be seen from the scheduling results that after adding PV inverters to the intraday scheduling stage, the system network loss is significantly reduced, the voltage is significantly increased, and the power quality is significantly improved, which verifies the effectiveness of the multi-time scale robust optimization of the distribution network with controllable photovoltaics proposed in this paper.

[0217] This embodiment uses Monte Carlo simulation to analyze the robustness of multi-time scale robust optimization under uncertain conditions. 200 daily scenarios are randomly generated using normal distribution. The standard deviation of load and photovoltaic power generation is 5% in the scenario generation. The worst scenario is then extracted, and its scheduling result is added to the intraday forecast curve for comparison with the multi-time scale robust optimization result. The grid loss and voltage comparison results are shown in Figure 2. Fig.16 and Fig.17 The network loss, average voltage value and voltage deviation are shown in Table 1.

[0218] Table 1 Comparison of two methods

[0219] TABLE 1 The results of two comparison methods

[0220]

[0221] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A multi-time scale reactive power dispatching method for a distribution network including controllable photovoltaics, characterized in that: Includes steps: Step 1: constructing a control strategy for a photovoltaic power generation inverter model; the control strategy for the photovoltaic power generation inverter model includes: no control mode strategy, reactive power control mode strategy, active power limitation mode strategy and inverter optimization scheduling strategy; The strategy of not adopting the control mode is expressed as: The reactive power control strategy is expressed as: The active power limiting strategy is expressed as: The inverter optimization scheduling strategy is expressed as: Where P PV is the maximum output of photovoltaic power; S G is the apparent power of the photovoltaic inverter; Q pv_out and P pv_out is the actual output of photovoltaic reactive power and the actual output of photovoltaic active power, Q pv_out,max The upper limit of photovoltaic output; Step 2: Construct a multi-time scale optimization strategy model for the distribution network, wherein the multi-time scale optimization strategy model for the distribution network takes minimizing network loss as the objective function of the day-ahead scheduling stage and minimizing voltage deviation as the objective function of the intraday scheduling stage; and the multi-time scale optimization strategy model for the distribution network considers the switching of the photovoltaic power generation inverter model to construct the constraint conditions of the objective function; The objective function of the day-ahead scheduling phase with the network loss minimized is expressed as: Where f1 is the distribution network loss; I ij,t is the current value of branch ij at time t; T is the number of time periods in the dispatch cycle; N is the number of nodes in the distribution network; Δt is the dispatch interval; Taking the minimum voltage deviation as the objective function of the intraday scheduling stage as follows: Where V ref is the reference voltage value, take 1.0pu; The constraints include: radial distribution network branch flow constraints, node voltage constraints, branch current constraints, parallel capacitor group constraints, energy storage device operation constraints, static VAR compensation device constraints, on-load voltage regulator constraints and photovoltaic inverter control mode switching constraints; The constraint conditions of the objective function in the day-ahead dispatching stage include: radial distribution network branch power flow constraint and photovoltaic inverter control mode switching constraint; The constraints of the objective function in the intraday dispatch stage include: radial distribution network branch power flow constraints, branch current constraints and photovoltaic inverter control mode switching constraints; Step 3: Solve the multi-time scale optimization strategy model of the distribution network based on the two-stage robust optimization model to obtain the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive power compensation; Step 4: Guide daily energy scheduling based on the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive power compensation.

2. The multi-time scale reactive power dispatching method for a distribution network containing controllable photovoltaics according to claim 1 is characterized in that: The two-stage robust optimization model is: From={from j ∈{0,1},from g ∈{0,1}} In the formula, 3. The multi-time scale reactive power dispatching method for a distribution network containing controllable photovoltaics according to claim 2 is characterized in that: Step 3 includes the following sub-steps: Step 3.1: Divide the robust optimization solution into the main problem and sub-problems, which are expressed as the first stage and the second stage respectively. Step 3.2: Add the binary variable Z used to select the PV inverter to the first-stage variables of the robust optimization, and make it always subject to the newly added column constraints during the robust optimization solution process to obtain a two-stage robust optimization model; Step 3.3: Use the CCG algorithm to iteratively solve the main problem and sub-problems until the difference between the lower bound of the main problem and the upper bound of the sub-problem is less than the convergence criterion, and the problem is solved; The strong duality theorem and the outer approximation method are used to transform the subproblem into a solvable linear programming problem to obtain the optimal solution of the second stage.

4. The multi-time scale reactive power dispatching method for a distribution network containing controllable photovoltaics according to claim 3 is characterized in that: The first stage includes inequality constraints, PV switching constraints, and equality constraints, and the second stage includes inequality constraints and inequality constraints; The inequality constraints in the first stage correspond to node voltage constraints, branch current constraints, shunt capacitor group constraints, energy storage device operation constraints, on-load voltage regulator constraints, and photovoltaic inverter control mode switching constraints; The PV switching constraints of the first stage correspond to the photovoltaic inverter control mode switching constraints; The equation constraints of the first stage correspond to radial distribution network branch power flow constraints, parallel capacitor group constraints, energy storage device operation constraints, on-load voltage regulator constraints and photovoltaic inverter control mode switching constraints; The inequality constraints of the second stage correspond to the static VAR compensation device constraints; The second-order cone constraints in the second stage correspond to the radial distribution network branch power flow constraints and the photovoltaic inverter control mode switching constraints.

5. A multi-timescale reactive power dispatching system for a distribution network including controllable photovoltaics, constructed according to the method described in any one of claims 1 to 4, characterized in that: include: Control strategy building module, model building module, calculation module and guidance module; The control strategy building module is used to build a control strategy for a photovoltaic power generation inverter model; The model building module is used to build a multi-time scale optimization strategy model for the distribution network, wherein the multi-time scale optimization strategy model for the distribution network takes minimizing network loss as the objective function of the day-ahead scheduling phase and minimizing voltage deviation as the objective function of the intraday scheduling phase; the model building module is also used to consider the switching of the photovoltaic power generation inverter model to build the constraint conditions of the objective function; The calculation module is used to solve the distribution network multi-time scale optimization strategy model based on the two-stage robust optimization model to obtain the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive compensation amount; The guidance module is used to guide daily energy scheduling based on the optimal photovoltaic power generation control strategy and photovoltaic power generation reactive compensation amount.