Power distribution network photovoltaic scheduling method, device and equipment and storage medium

Abstract

The invention discloses a power distribution network photovoltaic scheduling method, apparatus and device, and a storage medium. The method comprises the steps of constructing an objective function of active power loss and a constraint condition corresponding to the objective function; solving the target function under the constraint condition to obtain a scheduling scheme of an on-load tap changing transformer and a capacitor bank in the power distribution network when the active power loss of the power distribution network is minimum; the photovoltaic generation and load level in each region is measured according to a region controller of a preset deep neural network, and then the optimal reactive power of each region is calculated; the method comprises the steps of generating a reactive power scheduling scheme for photovoltaic reactive power adjustment, generating a photovoltaic scheduling scheme of the power distribution network according to a scheduling scheme of an on-load tap changer and a capacitor bank in the power distribution network and the reactive power scheduling scheme for photovoltaic reactive power adjustment, and scheduling the photovoltaic in the power distribution network according to the photovoltaic scheduling scheme. The photovoltaic scheduling efficiency of the power distribution network can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a photovoltaic dispatching method, device, equipment and storage medium for a distribution network. Background Art

[0002] With the development of renewable energy, active distribution networks have changed from passive to active. However, the uncertainty of photovoltaic systems in distribution networks and the number of users will affect the original voltage level in the power grid, so the voltage of the distribution network needs to be adjusted. The existing technology mainly adopts voltage reactive control. The method of controlling voltage is first to collect the flow distribution information in the distribution network, and then adjust the tap of the on-load tap-changing transformer and change the number of capacitor groups according to the power distribution to achieve the optimal voltage distribution, thereby solving the problem of voltage balance in the network. Voltage reactive control is divided into centralized voltage reactive control, local voltage reactive control and distributed voltage reactive control. Usually, the above schemes are combined and applied to improve the voltage distribution while reducing the power loss between nodes.

[0003] Existing technologies mostly use voltage and reactive power control methods. Local voltage and reactive power control technology proposes a local photovoltaic reduction strategy based on droop to address the problem of voltage control violations. Although it solves the problem of overvoltage, there is insufficient coordination between different areas, resulting in a local optimal situation. Centralized reactive power voltage control technology can execute optimal power flow by collecting information from the node network. Although it can make the distribution network as a whole optimal, it has the problem of requiring a lot of computing power and communication resources to collect complex information on the distribution network. Distributed voltage and reactive power control technology proposes a multi-agent strategy for network losses. Although it effectively alleviates the problem of power loss, it has the problems of slow algorithm speed and many iterations. Summary of the invention

[0004] The present invention provides a distribution network photovoltaic dispatching method, device, equipment and storage medium to solve the technical problems of slow speed and high number of iterations of the existing voltage and reactive power control algorithm.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a photovoltaic scheduling method for a distribution network, comprising:

[0006] Acquire photovoltaic prediction data and load prediction data of the distribution network, and construct an objective function of active power loss and constraint conditions corresponding to the objective function based on the photovoltaic prediction data and the load prediction data, with the goal of minimizing the active power loss of the distribution network;

[0007] Solving the objective function under the constraint conditions to obtain a scheduling scheme for on-load tap-changing transformers and capacitor banks in the distribution network when the active power loss of the distribution network is minimized;

[0008] According to the types of photovoltaics and loads in the distribution network, the distribution network is divided into several areas, and the level of photovoltaic generation and load in each area is measured according to the regional controller of the preset deep neural network, and then the optimal reactive power of each area is calculated according to the level of photovoltaic generation and load;

[0009] Matching a corresponding power allocation strategy according to the type of photovoltaic power generation, and then generating a reactive power dispatching scheme for photovoltaic power generation for reactive power regulation according to the power allocation strategy and the optimal reactive power of each area;

[0010] According to the scheduling scheme of the on-load tap-changing transformer and capacitor bank in the distribution network and the reactive power scheduling scheme for photovoltaic reactive power regulation, a photovoltaic scheduling scheme for the distribution network is generated, and then the photovoltaic in the distribution network is scheduled according to the photovoltaic scheduling scheme.

[0011] As a preferred solution, the objective function of the active power loss is:

[0012]

[0013] Among them, l ij,t is the square of the current in branch ij at time t, r ij is the resistance value of the line, l ij,t r ij is the loss on branch ij, and the sum of the losses of branch ij in time range T is the power loss of the network; t It is the gear position of the on-load tap-changing transformer; and is the reactive power generated by the photovoltaic at time t.

[0014] As a preferred solution, the constraints corresponding to the objective function include: power balance constraint, bus voltage upper and lower limit constraints, photovoltaic power generation reactive power capacity constraint, reactive balance constraint, switching time limit constraint of on-load tap-changing transformer and capacitor bank, and on-load tap-changing transformer voltage constraint;

[0015] The power balance constraint is:

[0016]

[0017] Among them, P PVj,t and P Lj,t are the active power generated by the photovoltaic and load at time t, Q CBj,t and Q Lj,t is the reactive power generated by the capacitor bank and the load at time t, P ij,t and Q ij,t is the active power and reactive power on branch ij;

[0018] The bus voltage upper and lower limits are constrained as follows:

[0019]

[0020] in, and are the upper and lower limits of the square of the bus voltage respectively;

[0021] The photovoltaic power generation reactive power capacity constraint and reactive power balance constraint are:

[0022]

[0023] in, and are the apparent power and active power generated by photovoltaics, respectively;

[0024] The switching time limit constraint of the on-load tap-changing transformer and capacitor bank is:

[0025]

[0026] Among them, c ik,t is the state of the kth capacitor at time step t in the Volt / Var control optimization process of the distribution network, cap ik,max is the maximum value of the capacitor;

[0027] The voltage constraint of the on-load tap-changing transformer is:

[0028] v 1,t =(V s +tap t ΔV T ) 2 ;

[0029] Among them, V s and ΔV T They are the change in transformer input voltage and transformer output voltage respectively.

[0030] As a preferred solution, the optimal reactive power of each area is calculated by the following formula:

[0031]

[0032] in, is the power loss on line ij, is the power loss at the line end mn, is the power loss of all branch networks.

[0033] As a preferred solution, the regional controller includes: an input layer, a hidden layer and an output layer;

[0034] The input and output of the zone controller are:

[0035]

[0036] in, is the active power generated by PV in the jth region, is the reactive power generated by the load in the jth area;

[0037] The input and output of the hidden layer are:

[0038]

[0039] Among them, w1 and b1 are the weight and bias matrix of the first hidden layer respectively, h1 is the output, p is the number of layers, out is the predicted total reactive power, and μ(x)=max{0,x} is the ReLU activation function.

[0040] As a preferred solution, when training the regional controller, the corresponding loss function is:

[0041]

[0042] Among them, the parameter λ is equal to 0.001, N b is the number of samples selected as the mini-batch size, the subscript s is the s-th training, and W is the set of weights and biases.

[0043] As a preferred solution, the power allocation strategy includes: loss reduction mode, fair power sharing mode and voltage safety mode;

[0044] The loss reduction mode is:

[0045]

[0046] in, and are the hourly optimal reactive power of photovoltaic power in the first and second phases, y t is the total reactive power output of the PV in the jth zone obtained by the well-trained DNN;

[0047] The fair power sharing model is:

[0048]

[0049] Among them, N PV is the total number of PVs in a region;

[0050] The voltage safety mode is:

[0051]

[0052] Among them, V C is the voltage measurement value at the critical bus, S VQ (i,c) is the voltage reactive sensitivity between buses i and c.

[0053] Based on the above embodiment, another embodiment of the present invention provides a photovoltaic dispatching device for a distribution network, including: an objective function construction module, an objective function solving module, an optimal reactive power calculation module, a reactive power dispatching scheme generation module and a photovoltaic dispatching module;

[0054] The objective function building module is used to obtain photovoltaic prediction data and load prediction data of the distribution network, and to build an objective function of active power loss and constraint conditions corresponding to the objective function based on the photovoltaic prediction data and load prediction data with the goal of minimizing the active power loss of the distribution network;

[0055] The objective function solving module is used to solve the objective function under the constraint conditions to obtain a scheduling scheme for on-load tap-changing transformers and capacitor banks in the distribution network when the active power loss of the distribution network is minimized;

[0056] The optimal reactive power calculation module is used to divide the distribution network into several areas according to the types of photovoltaics and loads in the distribution network, and measure the level of photovoltaic generation and load in each area according to the regional controller of the preset deep neural network, and then calculate the optimal reactive power of each area according to the level of photovoltaic generation and load;

[0057] The reactive power dispatching scheme generating module is used to match the corresponding power allocation strategy according to the type of photovoltaic, and then generate a reactive power dispatching scheme for photovoltaic reactive power regulation according to the power allocation strategy and the optimal reactive power of each area;

[0058] The photovoltaic scheduling module is used to generate a photovoltaic scheduling plan for the distribution network according to the scheduling plan of the on-load tap-changing transformer and capacitor group in the distribution network and the reactive power scheduling plan for photovoltaic reactive power regulation, and then schedule the photovoltaic in the distribution network according to the photovoltaic scheduling plan.

[0059] Based on the above embodiments, another embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the distribution network photovoltaic scheduling method described in the above invention embodiments is implemented.

[0060] Based on the above embodiments, another embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the distribution network photovoltaic scheduling method described in the above invention embodiment.

[0061] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0062] The present invention provides a photovoltaic dispatching method for a distribution network, which obtains photovoltaic prediction data and load prediction data of the distribution network, and constructs an objective function of active power loss and constraint conditions corresponding to the objective function according to the photovoltaic prediction data and the load prediction data, with the goal of minimizing the active power loss of the distribution network; solves the objective function under the constraint conditions to obtain a dispatching scheme of on-load tap-changing transformers and capacitor banks in the distribution network when the active power loss of the distribution network is minimized; divides the distributed network of the distribution network into a plurality of regions according to the types of photovoltaics and loads in the distribution network, and measures the generation and load levels of photovoltaics in each region according to a regional controller of a preset deep neural network, and then calculates the optimal reactive power of each region according to the generation and load levels of photovoltaics; matches the corresponding power allocation strategy according to the type of photovoltaics, and then generates a reactive power dispatching scheme for photovoltaics to perform reactive power regulation according to the power allocation strategy and the optimal reactive power of each region; generates a photovoltaic dispatching scheme for the distribution network according to the dispatching scheme of on-load tap-changing transformers and capacitor banks in the distribution network and the reactive power dispatching scheme for photovoltaics to perform reactive power regulation, and then dispatches the photovoltaics in the distribution network according to the photovoltaic dispatching scheme. The present invention coordinates the reactive power distribution of the distribution network and reduces the real-time power loss between lines to achieve the optimal power flow of the power system. At the same time, a deep neural network is used to train the controller of each area, which can reduce the number of iterations of the calculation process during information exchange and improve the photovoltaic dispatching efficiency of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 It is a flow chart of a photovoltaic dispatching method for a distribution network provided by an embodiment of the present invention;

[0064] Figure 2 It is a schematic diagram of photovoltaic scheduling in each area;

[0065] Figure 3 It is a schematic diagram of the network structure;

[0066] Figure 4 It is a structural schematic diagram of a photovoltaic dispatching device for a distribution network provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0068] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0069] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0070] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0071] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0072] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0073] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0074] Embodiment 1

[0075] Please refer to Figure 1 , is a flow chart of a photovoltaic dispatching method for a distribution network provided by an embodiment of the present invention, comprising the following specific steps:

[0076] S1. Acquire photovoltaic prediction data and load prediction data of the distribution network, and construct an objective function of active power loss and constraint conditions corresponding to the objective function based on the photovoltaic prediction data and load prediction data, with the goal of minimizing the active power loss of the distribution network;

[0077] Preferably, the objective function of the active power loss is:

[0078]

[0079] Among them, l ij,t is the square of the current in branch ij at time t, r ij is the resistance value of the line, l ij,t r ij is the loss on branch ij, and the sum of the losses of branch ij in time range T is the power loss of the network; t It is the gear position of the on-load tap-changing transformer; and is the reactive power generated by the photovoltaic at time t.

[0080] Preferably, the constraints corresponding to the objective function include: power balance constraint, bus voltage upper and lower limit constraints, photovoltaic power generation reactive power capacity constraint, reactive power balance constraint, switching time limit constraint of on-load tap-changing transformer and capacitor bank, and on-load tap-changing transformer voltage constraint;

[0081] The power balance constraint is:

[0082]

[0083] Among them, P PVj,t and P Lj,t are the active power generated by the photovoltaic and load at time t, Q CBj,t and QLj,t is the reactive power generated by the capacitor bank and the load at time t, P ij,t and Q ij,t is the active power and reactive power on branch ij;

[0084] The bus voltage upper and lower limits are constrained as follows:

[0085]

[0086] in, and are the upper and lower limits of the square of the bus voltage respectively;

[0087] The photovoltaic power generation reactive power capacity constraint and reactive power balance constraint are:

[0088]

[0089] in, and are the apparent power and active power generated by photovoltaics, respectively;

[0090] The switching time limit constraint of the on-load tap-changing transformer and capacitor bank is:

[0091]

[0092] Among them, c ik,t is the state of the kth capacitor at time step t in the Volt / Var control optimization process of the distribution network, cap ik,max is the maximum value of the capacitor;

[0093] The voltage constraint of the on-load tap-changing transformer is:

[0094] v 1,t =(V s +tap t ΔV T ) 2 ;

[0095] Among them, V s and ΔV T They are the change in transformer input voltage and transformer output voltage respectively.

[0096] Specifically, in order to alleviate the problems of voltage violation and reactive power balance in active distribution networks, the present invention proposes a two-stage photovoltaic consumption method by combining centralized VCC and distributed VCC schemes based on the reactive power voltage control (VCC) method. Figure 2, which is a schematic diagram of PV scheduling in each region. First, the central controller (CC) in the active distribution network predicts the photovoltaic (PV) and load data in advance, and executes the optimal power flow (OPF) in the distribution network based on the data to achieve real-time scheduling of on-load tap-changing transformers (OLTC) and capacitor banks (CBs); then the distributed network is divided into different regions according to the different types of PV and load, and the zone controller (ZC) based on the deep neural network (DNN) measures the level of PV generation and load, and calculates the optimal reactive power of each region; then, three different power allocation strategies are designed to schedule reactive power regulation for PV to achieve PV safety and near-optimal power sharing; finally, the mixed integer second-order cone programming model (MISOCP) is programmed on GAMS and solved by the CPLEX solver; the ZC model based on DNN is programmed on PyCharm and solved with "TensorFlow".

[0097] The specific steps include:

[0098] Step 1: Use CC for day-ahead optimization. Based on PV forecast and load forecast, the voltage profile is optimized by real-time dispatching of reactive power of OLTCs and CBs in the distribution network, while also reducing power losses.

[0099] (1) Objective function:

[0100] The OPF model is determined by using centralized VCC technology to solve the uncertainty of PV and load and maintain its robustness. The objective function of minimizing active power loss is shown as follows:

[0101]

[0102] In the formula, l ij,t represents the square of the current in branch ij at time t, r ij represents the resistance value of the line, so the loss on branch ij is l ij,t r ij , the total loss of branch ij in time range T is the power loss of the network; t Indicates the gear position of OLTC; and It represents the reactive power generated by PV at time t. The reason for deriving the OPF model is that the real-time scheduling of OLTC and CBs is based on centralized VCC optimization, and the scheduling of PV can be achieved by using a shorter snapshot on the day.

[0103] (2) OPF model:

[0104] The nonlinear AC power flow equation is relaxed into a quadratic cone programming (SOCP) form to simplify the calculation process while maintaining the global optimal solution characteristics of the problem. The power balance constraint and the relaxation of the original OPF constraint are shown as follows:

[0105]

[0106] Where P PVj,t and P Lj,t They represent the active power generated by PV and load at time t respectively; Q CBj,t and Q Lj,t is the reactive power generated by CBs and loads at time t; P ij,t and Q ij,t is the active and reactive power on branch ij.

[0107] The bus voltage upper and lower limit constraints are as follows:

[0108]

[0109] In the formula, and They represent the upper and lower limits of the square of the bus voltage respectively;

[0110] Equivalence The convex relaxation is performed as follows:

[0111]

[0112] In the formula, v j,t Represents the voltage of the bus at time t.

[0113] The PV generation reactive power capability constraints and reactive power balance constraints at different locations are as follows:

[0114]

[0115] In the formula, and It is the apparent power and active power generated by PV.

[0116] (3) Switching time limit of OLTC and CBs and OLTC voltage constraint:

[0117] In order to extend the service life of OLTC and CBs, their daily switching time is limited. The specific constraints are as follows:

[0118]

[0119] In the formula, c ik,t is the state of the kth capacitor at time step t during the Volt / Var control optimization process of the distribution network; cap ik,maxExpressed as the maximum value of the capacitor.

[0120] The square constraint of the voltage at the tap position OLTC is as follows:

[0121] v 1,t =(V s +tap t ΔV T ) 2 ;

[0122] Where V s , ΔV T It is expressed as the change in transformer input voltage and transformer output voltage, which are set to 1.0pu and 0.01pu respectively.

[0123] Since the above formula contains quadratic terms and absolute values, its constraints are still nonlinear constraints, so it is necessary to convert them into linear constraints to facilitate the solver to solve them. Some 0-1 variables are replaced equivalently, and the nonlinear constraints are converted into linear constraints so that they can be solved by the solver. The converted linear constraints are as follows:

[0124] Q CBi,t =∑ k c ik,t q ik,t ;

[0125]

[0126] In the formula, a ik ,t is an auxiliary variable; g ik ,t represents the state change of the kth CBs between two time steps; σ t is an integer variable between -10 and 10, which is the change of OLTC from time t to t+1; Indicates the highest level of OLTC tap; d k,t To indicate tap t Auxiliary binary variables.

[0127] S2. Solving the objective function under the constraint conditions to obtain a dispatching scheme for on-load tap-changing transformers and capacitor banks in the distribution network when the active power loss of the distribution network is minimized;

[0128] CC is used for day-ahead optimization. Based on PV forecast and load forecast, VCC is used to coordinate the reactive power of OLTC, CBs and PV inverters, optimize the voltage level in the distributed network, and minimize power loss. By calculating the power flow at multiple time points, the system can minimize the active power loss in the power grid while maintaining the voltage of each node within a safe range while satisfying the constraints. The relationship between the power flow model and the objective function can be seen as: the objective function provides the direction of optimization (i.e., reducing losses), and the power flow model is a means to achieve this goal. The objective function can be solved under the constraints to obtain the scheduling plan of the on-load tap-changing transformer and capacitor bank in the distribution network when the active power loss of the distribution network is minimized.

[0129] S3. Divide the distribution network into several areas according to the types of photovoltaics and loads in the distribution network, and measure the photovoltaic generation and load levels in each area according to the regional controller of the preset deep neural network, and then calculate the optimal reactive power of each area according to the photovoltaic generation and load levels;

[0130] Preferably, the optimal reactive power of each area is calculated by the following formula:

[0131]

[0132]

[0133] in, is the power loss on line ij, is the power loss at the line end mn, is the power loss of all branch networks.

[0134] Preferably, the regional controller comprises: an input layer, a hidden layer and an output layer;

[0135] The input and output of the zone controller are:

[0136]

[0137] in, is the active power generated by PV in the jth region, is the reactive power generated by the load in the jth area;

[0138] The input and output of the hidden layer are:

[0139]

[0140] Among them, w1 and b1 are the weight and bias matrix of the first hidden layer respectively, h1 is the output, p is the number of layers, out is the predicted total reactive power, and μ(x)=max{0,x} is the ReLU activation function.

[0141] Preferably, when training the region controller, the corresponding loss function is:

[0142]

[0143] Among them, the parameter λ is equal to 0.001, N b is the number of samples selected as the mini-batch size, the subscript s is the s-th training, and W is the set of weights and biases.

[0144] Step 2: Use a data-driven distributed VCC approach to reduce the number of information exchanges when regulating voltage in different regions and reduce the computational burden. First, derive the minimum power loss, and then design a DNN-based ZC and multi-mode reactive power strategy to balance the power of the active distribution network.

[0145] (1) Analyze and derive the method to minimize power loss, determine the input and output of ZC in different areas, and obtain the optimal reactive power distribution;

[0146] The linearized power flow model has higher accuracy and can simplify the power loss equation while ignoring the active and reactive losses on the line. Assuming that the bus voltage V j For 1.0pu, the further simplified power loss equation is as follows:

[0147]

[0148] In the formula, Represents the power loss on line ij.

[0149] The power loss on the branch mn (m=n-1) at the end of the feeder and the power loss on all branch networks are shown as follows:

[0150]

[0151] In the formula represents the power loss at the line end mn; Represents the power loss of all branch networks.

[0152] The optimal solution of the branch network loss in the above formula can be derived from the first-order optimality condition, as shown in the following formula:

[0153]

[0154] Through a simplified grid power flow model, the relationship between the optimal reactive power of the PV system and the reactive power demand of the load is derived. According to the equation of power loss on the branch network, the optimal reactive power of the PV can be tracked from the reactive power trend of the load.

[0155] (2) Use DNN to train ZC to reduce the computational burden of information exchange between regions.

[0156] ZC consists of an input layer, several hidden layers, and an output layer connected in sequence. The following is the input (x) and output (y) of the jth ZC:

[0157]

[0158] In the formula represents the active power generated by PV in the jth region; Represents the reactive power generated by the load in the jth area. The optimal solution of VCC is obtained by selecting the sum of the optimal PV reactive power in the jth area. In order to apply the average load curve of an area, the load curves of the previous area, this area, and the next area can be input, while excluding the PV bus voltage and reactive power at time t-1, avoiding the instability of the reactive power-voltage curve and simplifying the offline training process of the DNN.

[0159] The first hidden layer is directly connected to the input layer, and its input and output are shown as follows:

[0160]

[0161] Where w1 and b1 are the weight and bias matrices of the first hidden layer, h1 is the output, p is the number of layers, out is the predicted total reactive power, and μ(x)=max{0,x} is the ReLU activation function.

[0162] Offline training minimizes the loss f loss As shown below:

[0163]

[0164] In the formula, the parameter λ is equal to 0.001, N b is the number of samples selected as the mini-batch size; the subscript s represents the sth training; W is the set of weights and biases. Considering the high fitting accuracy and simple network structure of MLP, the present invention trains and models ZC through DNN of multi-layer perceptron (MLP), reduces the number of iterations of power flow calculation, and quickly obtains the optimal solution in the process of ZC information exchange.

[0165] S4, matching the corresponding power allocation strategy according to the type of photovoltaic, and then generating a reactive power dispatching scheme for photovoltaic reactive power regulation according to the power allocation strategy and the optimal reactive power of each area;

[0166] Preferably, the power allocation strategy includes: a loss reduction mode, a fair power sharing mode and a voltage safety mode;

[0167] The loss reduction mode is:

[0168]

[0169] in, and are the hourly optimal reactive power of photovoltaic power in the first and second phases, y t is the total reactive power output of the PV in the jth zone obtained by the well-trained DNN;

[0170] The fair power sharing model is:

[0171]

[0172] Among them, N PV is the total number of PVs in a region;

[0173] The voltage safety mode is:

[0174]

[0175] Among them, V C is the voltage measurement value at the critical bus, S VQ (i,c) is the voltage reactive sensitivity between buses i and c.

[0176] (3) Based on the obtained optimal reactive power distribution, three different modes of reactive power allocation are proposed to maintain voltage safety and mitigate potential voltage violations, thus meeting the operation requirements of the active distribution network under different conditions.

[0177] Mode 1: Loss Reduction Mode (LMR)

[0178] LMR is mainly used to optimize the reactive power output of large photovoltaic systems. When the power loss is reduced, when it belongs to [V min ,V max ] to maintain voltage margin. The optimal power ratio distributes the reactive power of PV as shown below:

[0179]

[0180] In the formula, and is the hourly optimal reactive power of PV in the first and second phases; t is the total reactive power output of the PV in the jth zone obtained by the well-trained DNN.

[0181] Mode 2: Fair Power Sharing Mode (FPM)

[0182] This mode is aimed at the user's own photovoltaic system. At PVs in different locations, FPM is used to achieve the balance of PV reactive power, as shown in the following formula:

[0183]

[0184] Where N PV The total number of PVs in a region.

[0185] Mode 3: Voltage Safety Mode (VSM)

[0186] VSM is enabled when the voltage approaches or exceeds the safety threshold, giving priority to ensuring that the voltage is within the safety range and preventing voltage overruns. VSM mitigates voltage violations by reducing reactive power. The reactive power of utility-scale PV is expressed as follows:

[0187]

[0188] Where V C Voltage measurement at the critical bus; S VQ (i,c) is the voltage reactive sensitivity between buses i and c.

[0189] S5. Generate a photovoltaic scheduling plan for the distribution network according to the scheduling plan of the on-load tap-changing transformer and capacitor bank in the distribution network and the reactive power scheduling plan for photovoltaic reactive power regulation, and then schedule the photovoltaic in the distribution network according to the photovoltaic scheduling plan.

[0190] Step 3: Use CPLEX to solve the MISOCP model and use "TensorFlow" to solve the DNN-based ZC offline training model. The model input and output algorithms are shown in Table 1. The test system is based on an improved IEEE33 node network. Figure 3 is a schematic diagram of the network structure. Figure 3It can be seen that the IEEE33 node network is divided into three areas: industrial (Z1), residential (Z2) and commercial (Z3). There are two types of PVs, OLCT and CBs, on the buses in different areas. The parameters of the VCC equipment are shown in Table 2. All loads are constant power models with a certain power factor, which are used to verify the effectiveness of VCC. In the test, key nodes are selected, the range of safe working voltage and bus voltage is set, and the DNN is tested offline to reasonably schedule OLTC and CBs. In order to verify the advantages of the method adopted by the present invention, three methods are used for comparison:

[0191] Method 1 (traditional centralized VCC): Traditional centralized VCC can mitigate voltage violations and dispatch OLTCs and CBs for reactive power regulation, but it is only suitable for long-snapshot VCC equipment;

[0192] Method 2 (combining centralized and distributed VCC): obtain the optimal voltage-reactive relationship of PV in OPF and adjust the voltage curve, but its performance depends largely on the accuracy of predicting PV and load in advance;

[0193] Method 3 (optimal solution for real-time VCC): global optimization can be achieved, but a large amount of computing resources are required to complete the optimization in a short time, so it is difficult to achieve in reality.

[0194] Table 3 lists the comparison results of various methods. It can be seen from the table that the voltage violation (MV) of the present invention is zero and the power loss (PL) is reduced, and a certain voltage margin can be left, which proves that the method of the present invention can effectively balance the relationship between voltage safety and power loss.

[0195]

[0196]

[0197] Table 1 Model input and output algorithm

[0198] Device parameter Quantity limit / type Location OLTC ±10×1％ 4 1 CB 1,2 3×50kVar 6 16，22 PV 1,7,8 300kVA Utility-Scale Photovoltaics 2，20，25 PV 2,3,9,10 300kVA User Photovoltaic 6，9，28，33 PV 4,5,6 500kVA Utility-Scale Photovoltaics 14，17，18

[0199] Table 2 Parameters of VCC devices

[0200]

[0201]

[0202] Table 3 Comparison results of three methods

[0203] Embodiment 2

[0204] Please refer to Figure 4, is a schematic diagram of the structure of a photovoltaic dispatching device for a distribution network provided by an embodiment of the present invention, the device comprising: an objective function building module, an objective function solving module, an optimal reactive power calculation module, a reactive power dispatching scheme generation module and a photovoltaic dispatching module;

[0205] The objective function building module is used to obtain photovoltaic prediction data and load prediction data of the distribution network, and to build an objective function of active power loss and constraint conditions corresponding to the objective function based on the photovoltaic prediction data and load prediction data with the goal of minimizing the active power loss of the distribution network;

[0206] The objective function solving module is used to solve the objective function under the constraint conditions to obtain a scheduling scheme for on-load tap-changing transformers and capacitor banks in the distribution network when the active power loss of the distribution network is minimized;

[0207] The optimal reactive power calculation module is used to divide the distribution network into several areas according to the types of photovoltaics and loads in the distribution network, and measure the level of photovoltaic generation and load in each area according to the regional controller of the preset deep neural network, and then calculate the optimal reactive power of each area according to the level of photovoltaic generation and load;

[0208] The reactive power dispatching scheme generating module is used to match the corresponding power allocation strategy according to the type of photovoltaic, and then generate a reactive power dispatching scheme for photovoltaic reactive power regulation according to the power allocation strategy and the optimal reactive power of each area;

[0209] The photovoltaic scheduling module is used to generate a photovoltaic scheduling plan for the distribution network according to the scheduling plan of the on-load tap-changing transformer and capacitor group in the distribution network and the reactive power scheduling plan for photovoltaic reactive power regulation, and then schedule the photovoltaic in the distribution network according to the photovoltaic scheduling plan.

[0210] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0211] Those skilled in the art can clearly understand that, for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0212] Embodiment 3

[0213] Accordingly, an embodiment of the present invention provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor implements the distribution network photovoltaic scheduling method described in the above-mentioned embodiment of the invention when executing the computer program.

[0214] The electronic device may be a computing device such as a desktop computer, a notebook, a palm computer, a cloud server, etc. The device may include, but is not limited to, a processor and a memory.

[0215] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the device, and various interfaces and lines are used to connect various parts of the entire device.

[0216] Embodiment 4

[0217] Accordingly, an embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the distribution network photovoltaic scheduling method described in the above-mentioned embodiment of the invention.

[0218] The memory can be used to store the computer program, and the processor realizes various functions of the device by running or executing the computer program stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0219] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When the computer program is executed by the processor, the steps of each method embodiment described above can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0220] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A photovoltaic dispatching method for a distribution network, characterized in that: include: Acquire photovoltaic prediction data and load prediction data of the distribution network, and construct an objective function of active power loss and constraint conditions corresponding to the objective function based on the photovoltaic prediction data and the load prediction data, with the goal of minimizing the active power loss of the distribution network; Solving the objective function under the constraint conditions to obtain a scheduling scheme for on-load tap-changing transformers and capacitor banks in the distribution network when the active power loss of the distribution network is minimized; According to the types of photovoltaics and loads in the distribution network, the distribution network is divided into several areas, and the level of photovoltaic generation and load in each area is measured according to the regional controller of the preset deep neural network, and then the optimal reactive power of each area is calculated according to the level of photovoltaic generation and load; Matching a corresponding power allocation strategy according to the type of photovoltaic power generation, and then generating a reactive power dispatching scheme for photovoltaic power generation for reactive power regulation according to the power allocation strategy and the optimal reactive power of each area; According to the scheduling scheme of the on-load tap-changing transformer and capacitor bank in the distribution network and the reactive power scheduling scheme for photovoltaic reactive power regulation, a photovoltaic scheduling scheme for the distribution network is generated, and then the photovoltaic in the distribution network is scheduled according to the photovoltaic scheduling scheme.

2. The photovoltaic dispatching method for distribution network according to claim 1, characterized in that: The objective function of the active power loss is: Among them, l ij,t is the square of the current in branch ij at time t, r ij is the resistance value of the line, l ij,t r ij is the loss on branch ij, and the sum of the losses of branch ij in time range T is the power loss of the network; t It is the gear position of the on-load tap-changing transformer; and is the reactive power generated by the photovoltaic at time t.

3. The photovoltaic dispatching method for distribution network according to claim 2, characterized in that: The constraints corresponding to the objective function include: power balance constraint, bus voltage upper and lower limit constraints, photovoltaic power generation reactive power capacity constraint, reactive power balance constraint, switching time limit constraint of on-load tap-changing transformer and capacitor bank, and on-load tap-changing transformer voltage constraint; The power balance constraint is: Among them, P PVj,t and P Lj,t are the active power generated by the photovoltaic and load at time t, Q CBj,t and Q Lj,t is the reactive power generated by the capacitor bank and the load at time t, P ij,t and Q ij,t is the active power and reactive power on branch ij; The bus voltage upper and lower limits are constrained as follows: in, and are the upper and lower limits of the square of the bus voltage respectively; The photovoltaic power generation reactive power capacity constraint and reactive power balance constraint are: in, and are the apparent power and active power generated by photovoltaics, respectively; The switching time limit constraint of the on-load tap-changing transformer and capacitor bank is: Among them, c ik,t is the state of the kth capacitor at time step t in the Volt / Var control optimization process of the distribution network, cap ik,max is the maximum value of the capacitor; The voltage constraint of the on-load tap-changing transformer is: v 1,t =(V s +tap t ·ΔV T ) 2 ; Among them, V s and ΔV T They are the change in transformer input voltage and transformer output voltage respectively.

4. The photovoltaic dispatching method for distribution network according to claim 3, characterized in that: The optimal reactive power for each area is calculated using the following formula: in, is the power loss on line ij, is the power loss at the line end mn, is the power loss of all branch networks.

5. The photovoltaic dispatching method for distribution network according to claim 4, characterized in that: The regional controller comprises: an input layer, a hidden layer and an output layer; The input and output of the zone controller are: in, is the active power generated by PV in the jth region, is the reactive power generated by the load in the jth area; The input and output of the hidden layer are: Among them, w1 and b1 are the weight and bias matrix of the first hidden layer respectively, h1 is the output, p is the number of layers, out is the predicted total reactive power, and μ(x)=max{0,x} is the ReLU activation function.

6. The photovoltaic dispatching method for distribution network according to claim 5, characterized in that: When training the region controller, the corresponding loss function is: Among them, the parameter λ is equal to 0.001, N b is the number of samples selected as the mini-batch size, the subscript s is the s-th training, and W is the set of weights and biases.

7. The photovoltaic dispatching method for distribution network according to claim 1, characterized in that: The power allocation strategies include: loss reduction mode, fair power sharing mode and voltage security mode; The loss reduction mode is: in, and are the hourly optimal reactive power of photovoltaic power in the first and second phases, y t is the total reactive power output of the PV in the jth zone obtained by the well-trained DNN; The fair power sharing model is: Among them, N PV is the total number of PVs in a region; The voltage safety mode is: Among them, V C is the voltage measurement value at the critical bus, S VQ (i,c) is the voltage reactive sensitivity between buses i and c.

8. A photovoltaic dispatching device for a distribution network, characterized in that: include: Objective function building module, objective function solving module, optimal reactive power calculation module, reactive power dispatching scheme generation module and photovoltaic dispatching module; The objective function building module is used to obtain photovoltaic prediction data and load prediction data of the distribution network, and to build an objective function of active power loss and constraint conditions corresponding to the objective function based on the photovoltaic prediction data and load prediction data with the goal of minimizing the active power loss of the distribution network; The objective function solving module is used to solve the objective function under the constraint conditions to obtain a scheduling scheme for on-load tap-changing transformers and capacitor banks in the distribution network when the active power loss of the distribution network is minimized; The optimal reactive power calculation module is used to divide the distribution network into several areas according to the types of photovoltaics and loads in the distribution network, and measure the level of photovoltaic generation and load in each area according to the regional controller of the preset deep neural network, and then calculate the optimal reactive power of each area according to the level of photovoltaic generation and load; The reactive power dispatching scheme generating module is used to match the corresponding power allocation strategy according to the type of photovoltaic, and then generate a reactive power dispatching scheme for photovoltaic reactive power regulation according to the power allocation strategy and the optimal reactive power of each area; The photovoltaic scheduling module is used to generate a photovoltaic scheduling plan for the distribution network according to the scheduling plan of the on-load tap-changing transformer and capacitor group in the distribution network and the reactive power scheduling plan for photovoltaic reactive power regulation, and then schedule the photovoltaic in the distribution network according to the photovoltaic scheduling plan.

9. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the photovoltaic dispatching method for a distribution network as claimed in any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the distribution network photovoltaic scheduling method according to any one of claims 1 to 7.

Images