Method and device for quantifying voltage-reactive power regulation capability of power distribution network

Abstract

The invention relates to a power distribution network voltage-reactive power regulation capability quantification method and device, and belongs to the technical field of power distribution network voltage control, and the method comprises the following steps: collecting and sorting a large amount of transmission-power distribution network data, analyzing a voltage coordination optimization control mechanism of a transmission-power distribution network, and building a coordination optimization control framework; establishing a voltage optimization control model based on second-order cone programming by taking minimization of power loss as a target and taking safety and economy of operation of the power distribution network as constraints based on a coordinated optimization control framework, and converting and solving voltage-reactive power regulation capability of the power distribution network based on the optimization model; and designing a deep neural network, carrying out offline training based on historical data of the power transmission and distribution network, carrying out online deployment on the trained model, and carrying out real-time quantification of the voltage-reactive power regulation capability of the power distribution network. According to the method, efficient and accurate quantification is realized, and voltage-reactive power support is provided for a power transmission network while the safe and economic operation of the power distribution network is guaranteed.

Description

Technical Field

[0001] The invention relates to a method and a device for quantifying voltage-reactive power regulation capability of a distribution network, and belongs to the technical field of voltage control of distribution networks. Background Art

[0002] With the rapid development of smart grid, distributed power sources with clean energy as the core are being used more and more widely. In recent years, distributed power sources have played an increasingly active role in distribution networks, using the reactive voltage control capability of distributed power sources to achieve voltage regulation of distribution networks, thereby making the grid operate stably and greatly improving the reliability of the grid.

[0003] In modern power systems, renewable energy sources such as wind power and photovoltaics are being connected to the grid in large quantities. As the amount of renewable energy generation increases, the number of centralized generators continues to decrease, thus weakening the voltage-reactive power control capability of the transmission network. In addition, the intermittent and uncertain nature of renewable energy generation leads to rapid and frequent voltage fluctuations and over-limits, posing new challenges to traditional voltage control devices with discrete actions and slow response speeds.

[0004] At the same time, the distribution network is connected to a large number of flexible voltage regulation resources, which can provide voltage support for the transmission network by adjusting the reactive power output of the unit. Therefore, by coordinating the renewable energy in the distribution network and providing the transmission network with flexible and low-cost voltage-reactive power regulation, it will become an important way to solve the voltage safety problem of the power grid. Summary of the invention

[0005] In order to solve the above problems, the present invention proposes a method and device for quantifying the voltage-reactive power regulation capability of a distribution network, which can ensure the safe and economical operation of the distribution network while providing voltage-reactive power support for the transmission network.

[0006] The technical solution adopted by the present invention to solve the technical problem is: In a first aspect, an embodiment of the present invention provides a method for quantifying voltage-reactive power regulation capability of a distribution network, comprising the following steps: Step 1: Collect and organize a large amount of transmission-distribution network data, analyze the voltage coordinated optimization control mechanism of the transmission-distribution network, and establish a coordinated optimization control framework; Step 2: Based on the coordinated optimization control framework, with the goal of minimizing power loss and the safety and economy of distribution network operation as constraints, a voltage optimization control model based on second-order cone programming is established, and the voltage-reactive power regulation capacity of the distribution network is solved based on the optimization model conversion; Step 3: Design a deep neural network, perform offline training based on historical data of the transmission and distribution network, deploy the trained model online, and perform real-time quantification of the voltage-reactive power regulation capability of the distribution network.

[0007] As a possible implementation of this embodiment, the objective function of the voltage optimization control model is: , , , in, c ik,t for t Moment k The state of the capacitors; Q PVj,t yes t Time Node j Reactive power flow of photovoltaic power generation system; Q i1,t Reactive power injection for balancing nodes; P * Loss ,t represents the minimum power loss obtained by solving the optimization problem; l ij yes t Time branch ij The square of the current amplitude flowing through; r ij It is a branch road ij Resistance; δ is a threshold value that represents the additional power loss that the distribution network can tolerate.

[0008] As a possible implementation of this embodiment, the constraint conditions of the voltage optimization control model include: Power flow constraints of distribution network: , , , in, P PVj,t , Q PVj,t yes t Time Node j Active and reactive power flows of photovoltaic power generation systems; P Lj,t , Q Lj,t yes t Time Node j Active power and reactive power load requirements; Q CBj,t yes t Time Node j The reactive power of capacitors; v i,t , v j,t yest Time Node i and nodes j The square of the voltage amplitude; P jk,t , Q jk,t yes t Time flow through the branch jk Active power and reactive power; P ij,t , Q ij,t yes t Time flow through the branch ij The active power and reactive power of x ij , x jk It is a branch road ij and branch roads jk The reactance, l jk,t yes t Time branch jk The square of the current amplitude flowing through; r jk It is a branch road jk Resistance; The node voltage constraints are: , in, V min , V max are the minimum and maximum voltage values; The reactive power regulation capability of photovoltaic inverters and capacitors is: , , in, S PV is the PV inverter capacity, set to 1.1 times the maximum PV power generation; c ik,t for t Moment k The state of the capacitors; q ik,t yes t Moment k The reactive power output of a capacitor.

[0009] As a possible implementation of this embodiment, the voltage-reactive power regulation range of the distribution network is: , in[ Q LV,t , QUV,t ] are the lower and upper limits of voltage-reactive power regulation capability.

[0010] As a possible implementation of this embodiment, the real-time quantification of the voltage-reactive power regulation capability of the distribution network includes: A deep neural network is trained to characterize and learn the functional relationship between multiple electrical inputs and output voltage-reactive power regulation capabilities. The voltage-reactive power regulation range of the distribution network is output through the deep neural network to determine the voltage-reactive power regulation capability.

[0011] As a possible implementation of this embodiment, the input and output features of the deep neural network are: , The input of the deep neural network includes the active power of the photovoltaic system P PVi,t , Load demand P Li,t and Q Li,t , and the voltage at the connection point of the transmission and distribution grid V 1,t , Q LV,t and Q UV,t It is the lower and upper limits of voltage-reactive power regulation capability.

[0012] As a possible implementation of this embodiment, the input-output function of the first hidden layer of the deep neural network is: , in w 1 and b 1 represents the weight and bias connecting the input and the first hidden layer respectively; h 1 is the output of the first hidden layer; μ ( x )=max{0, x} is the ReLU activation function.

[0013] As a possible implementation of this embodiment, the output layer function of the deep neural network is: , in w 6 and b 6 represent the weight and bias connecting the last hidden layer and the output layer respectively.

[0014] As a possible implementation of this embodiment, the loss function of the deep neural network is: , in fis the loss function, which is the mean square error between the neural network prediction and the optimal value; N b is a small batch size; y k It is k training samples, y * k for y k The optimal label of ; W is the set of weights and biases; λ is a hyperparameter of L2 regularization, used to avoid overfitting.

[0015] In a second aspect, an embodiment of the present invention provides a device for quantifying voltage-reactive power regulation capability of a distribution network, comprising: The control framework establishment module is used to collect and organize a large amount of transmission-distribution network data, analyze the voltage coordination optimization control mechanism of the transmission-distribution network, and establish a coordination optimization control framework; The control model building module is used to establish a voltage optimization control model based on second-order cone programming based on the coordinated optimization control framework, with the goal of minimizing power loss and the safety and economy of distribution network operation as constraints, and to solve the voltage-reactive power regulation capacity of the distribution network based on the optimization model conversion; The model training and deployment module is used to design deep neural networks, conduct offline training based on historical data of the transmission and distribution network, deploy the trained model online, and perform real-time quantification of the voltage-reactive power regulation capability of the distribution network.

[0016] As a possible implementation of this embodiment, the objective function of the voltage optimization control model is: , , , in, c ik,t for t Moment k The state of the capacitors; Q PVj,t yes t Time Node j Reactive power flow of photovoltaic power generation system; Q i1,t Reactive power injection for balancing nodes; P * Loss ,t represents the minimum power loss obtained by solving the optimization problem; l ij yes t Time branch ij The square of the current amplitude flowing through; rij It is a branch road ij Resistance; δ is a threshold value that represents the additional power loss that the distribution network can tolerate.

[0017] As a possible implementation of this embodiment, the constraint conditions of the voltage optimization control model include: Power flow constraints of distribution network: , , , in, P PVj,t , Q PVj,t yes t Time Node j Active and reactive power flows of photovoltaic power generation systems; P Lj,t , Q Lj,t yes t Time Node j Active power and reactive power load requirements; Q CBj,t yes t Time Node j The reactive power of capacitors; v i,t , v j,t yes t Time Node i and nodes j The square of the voltage amplitude; P jk,t , Q jk,t yes t Time flow through the branch jk Active power and reactive power; P ij,t , Q ij,t yes t Time flow through the branch ij The active power and reactive power of x ij , x jk It is a branch road ij and branch roads jk The reactance, l jk,t yes t Time Branch jk The square of the current amplitude flowing through; r jk It is a branch road jk Resistance; The node voltage constraints are: , in, V min , V max are the minimum and maximum voltage values; The reactive power regulation capability of photovoltaic inverters and capacitors is: , , in, S PV is the PV inverter capacity, set to 1.1 times the maximum PV power generation; c ik,t for t Moment k The state of the capacitors; q ik,t yes t Moment k The reactive power output of a capacitor.

[0018] As a possible implementation of this embodiment, the voltage-reactive power regulation range of the distribution network is: , in[ Q LV,t , Q UV,t ] are the lower and upper limits of voltage-reactive power regulation capability.

[0019] As a possible implementation of this embodiment, the specific process of the model training and deployment module performing real-time quantification of the voltage-reactive power regulation capability of the distribution network is as follows: A deep neural network is trained to characterize and learn the functional relationship between multiple electrical inputs and output voltage-reactive power regulation capabilities. The voltage-reactive power regulation range of the distribution network is output through the deep neural network to determine the voltage-reactive power regulation capability.

[0020] As a possible implementation of this embodiment, the input and output features of the deep neural network are: , The input of the deep neural network includes the active power of the photovoltaic system P PVi,t , Load demand P Li,t and Q Li,t , and the voltage at the connection point of the transmission and distribution grid V1,t , Q LV,t and Q UV,t It is the lower and upper limits of voltage-reactive power regulation capability.

[0021] As a possible implementation of this embodiment, the input-output function of the first hidden layer of the deep neural network is: , in w 1 and b 1 represents the weight and bias connecting the input and the first hidden layer respectively; h 1 is the output of the first hidden layer; μ ( x )=max{0, x} is the ReLU activation function.

[0022] As a possible implementation of this embodiment, the output layer function of the deep neural network is: , in w 6 and b 6 represent the weight and bias connecting the last hidden layer and the output layer respectively.

[0023] As a possible implementation of this embodiment, the loss function of the deep neural network is: , in f is the loss function, which is the mean square error between the neural network prediction and the optimal value; N b is a small batch size; y k It is k training samples, y * k for y k The optimal label of ; W is the set of weights and biases; λ is a hyperparameter of L2 regularization, used to avoid overfitting.

[0024] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the memory through the bus, and the processor executes the machine-readable instructions to perform the steps of the method for quantifying the voltage-reactive power regulation capability of any distribution network as described above.

[0025] In a fourth aspect, an embodiment of the present invention provides a storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method for quantifying the voltage-reactive power regulation capability of any distribution network as described above are executed.

[0026] The beneficial effects of the technical solution of the embodiment of the present invention are as follows: A method for quantifying the voltage-reactive power regulation capability of a distribution network according to the technical solution of an embodiment of the present invention comprises the following steps: Step 1, collecting and collating a large amount of transmission-distribution network data, analyzing the voltage coordinated optimization control mechanism of the transmission-distribution network, and establishing a coordinated optimization control framework; Step 2, based on the coordinated optimization control framework, with the goal of minimizing power loss and the safety and economy of the distribution network operation as constraints, a voltage optimization control model based on second-order cone programming is established, and the voltage-reactive power regulation capability of the distribution network is solved based on the optimization model conversion; Step 3, designing a deep neural network, performing offline training based on the historical data of the transmission-distribution network, and performing online deployment of the trained model to perform real-time quantification of the voltage-reactive power regulation capability of the distribution network. The present invention can realize the effective quantification of the voltage-reactive power regulation capability under the application of deep learning, realize efficient and accurate quantification in large-scale distribution networks, and provide voltage-reactive power support for the transmission network while ensuring the safe and economical operation of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a flow chart of a method for quantifying voltage-reactive power regulation capability of a distribution network according to an exemplary embodiment; Figure 2 is a structural schematic diagram of a device for quantifying voltage-reactive power regulation capability of a distribution network according to an exemplary embodiment; Figure 3 is a specific implementation flow chart of quantifying voltage-reactive power regulation capability of a distribution network according to an exemplary embodiment; Figure 4 is a voltage optimization control framework diagram of transmission-distribution coordination according to an exemplary embodiment; Figure 5 is a diagram of an offline learning process based on a neural network structure according to an exemplary embodiment; Figure 6 is an improved IEEE 33-node power distribution network diagram shown according to an exemplary embodiment; Figure 7 is a 24-hour photovoltaic, load, and voltage curve test data diagram according to an exemplary embodiment; Figure 8 is a loss function diagram of a neural network training according to an exemplary embodiment; Fig. 9is a sensitivity analysis diagram showing the influence of voltage on reactive power regulation capability according to an exemplary embodiment; Fig.10 is a sensitivity analysis diagram showing the effect of parameter δ on reactive power regulation capability according to an exemplary embodiment; Fig.11 is a diagram showing a comparison of lower limits of voltage-reactive power regulation capabilities of the present invention and a model-based optimization method according to an exemplary embodiment; Fig.12 is a comparison diagram of the upper limits of voltage-reactive power regulation capabilities of the present invention and a model-based optimization method according to an exemplary embodiment; Fig.13 The diagram is a 33-node voltage curve diagram for solving maximum and minimum reactive power according to an exemplary embodiment. DETAILED DESCRIPTION

[0028] In order to more clearly illustrate the technical features of the solution of the present invention, the present invention is described in detail below through specific implementation methods and in conjunction with the accompanying drawings.

[0029] like Figure 1 As shown, a method for quantifying voltage-reactive power regulation capability of a distribution network provided by an embodiment of the present invention includes the following steps: Step 1: Collect and organize a large amount of transmission-distribution network data, analyze the voltage coordinated optimization control mechanism of the transmission-distribution network, and establish a coordinated optimization control framework; Step 2: Based on the coordinated optimization control framework, with the goal of minimizing power loss and the safety and economy of distribution network operation as constraints, a voltage optimization control model based on second-order cone programming is established, and the voltage-reactive power regulation capacity of the distribution network is solved based on the optimization model conversion; Step 3: Design a deep neural network, perform offline training based on historical data of the transmission and distribution network, deploy the trained model online, and perform real-time quantification of the voltage-reactive power regulation capability of the distribution network.

[0030] As a possible implementation of this embodiment, the objective function of the voltage optimization control model is: , , , in, c ik,t for t Moment k The state of the capacitors; Q PVj,t yes t Time Node j Reactive power flow of photovoltaic power generation system;Q i1,t Reactive power injection for balancing nodes; P * Loss ,t represents the minimum power loss obtained by solving the optimization problem; l ij yes t Time branch ij The square of the current amplitude flowing through; r ij It is a branch road ij Resistance; δ is a threshold value that represents the additional power loss that the distribution network can tolerate.

[0031] As a possible implementation of this embodiment, the constraint conditions of the voltage optimization control model include: Power flow constraints of distribution network: , , , in, P PVj,t , Q PVj,t yes t Time Node j Active and reactive power flows of photovoltaic power generation systems; P Lj,t , Q Lj,t yes t Time Node j Active power and reactive power load requirements; Q CBj,t yes t Time Node j The reactive power of capacitors; v i,t , v j,t yes t Time Node i and nodes j The square of the voltage amplitude; P jk,t , Q jk,t yes t Time flow through the branch jk Active power and reactive power; P ij,t , Q ij,t yes t Time flow through the branch ij The active power and reactive power of xij , x jk It is a branch road ij and branch roads jk The reactance, l jk,t yes t Time Branch jk The square of the current amplitude flowing through; r jk It is a branch road jk Resistance; The node voltage constraints are: , in, V min , V max are the minimum and maximum voltage values; The reactive power regulation capability of photovoltaic inverters and capacitors is: , , in, S PV is the photovoltaic inverter capacity, which is set to 1.1 times the maximum photovoltaic power generation; c ik,t for t Moment k The state of the capacitors; q ik,t yes t Moment k The reactive power output of a capacitor.

[0032] As a possible implementation of this embodiment, the voltage-reactive power regulation range of the distribution network is: , in[ Q LV,t , Q UV,t ] are the lower and upper limits of voltage-reactive power regulation capability.

[0033] As a possible implementation of this embodiment, the real-time quantification of the voltage-reactive power regulation capability of the distribution network includes: A deep neural network is trained to characterize and learn the functional relationship between multiple electrical inputs and output voltage-reactive power regulation capabilities. The voltage-reactive power regulation range of the distribution network is output through the deep neural network to determine the voltage-reactive power regulation capability.

[0034] As a possible implementation of this embodiment, the input and output features of the deep neural network are: , The input of the deep neural network includes the active power of the photovoltaic system P PVi,t , Load demand P Li,t and Q Li,t , and the voltage at the connection point of the transmission and distribution grid V 1,t , Q LV,t and Q UV,t It is the lower and upper limits of voltage-reactive power regulation capability.

[0035] As a possible implementation of this embodiment, the input-output function of the first hidden layer of the deep neural network is: , in w 1 and b 1 represents the weight and bias connecting the input and the first hidden layer respectively; h 1 is the output of the first hidden layer; μ ( x )=max{0, x} is the ReLU activation function.

[0036] As a possible implementation of this embodiment, the output layer function of the deep neural network is: , in w 6 and b 6 represent the weight and bias connecting the last hidden layer and the output layer respectively.

[0037] As a possible implementation of this embodiment, the loss function of the deep neural network is: , in f is the loss function, which is the mean square error between the neural network prediction and the optimal value; N b is a small batch size; y k It is k training samples, y * k for y k The optimal label of ; W is the set of weights and biases; λ is a hyperparameter of L2 regularization, used to avoid overfitting.

[0038] like Figure 2As shown, an embodiment of the present invention provides a device for quantifying voltage-reactive power regulation capability of a distribution network, comprising: The control framework establishment module is used to collect and organize a large amount of transmission-distribution network data, analyze the voltage coordination optimization control mechanism of the transmission-distribution network, and establish a coordination optimization control framework; The control model building module is used to establish a voltage optimization control model based on second-order cone programming based on the coordinated optimization control framework, with the goal of minimizing power loss and the safety and economy of distribution network operation as constraints, and to solve the voltage-reactive power regulation capacity of the distribution network based on the optimization model conversion; The model training and deployment module is used to design deep neural networks, conduct offline training based on historical data of the transmission and distribution network, deploy the trained model online, and perform real-time quantification of the voltage-reactive power regulation capability of the distribution network.

[0039] As a possible implementation of this embodiment, the objective function of the voltage optimization control model is: , , , in, c ik,t for t Moment k The state of the capacitors; Q PVj,t yes t Time Node j Reactive power flow of photovoltaic power generation system; Q i1,t Reactive power injection for balancing nodes; P * Loss ,t represents the minimum power loss obtained by solving the optimization problem; l ij yes t Time branch ij The square of the current amplitude flowing through; r ij It is a branch road ij Resistance; δ is a threshold value that represents the additional power loss that the distribution network can tolerate.

[0040] As a possible implementation of this embodiment, the constraint conditions of the voltage optimization control model include: Power flow constraints of distribution network: , , , in, P PVj,t , Q PVj,t yes t Time Node j Active and reactive power flows of photovoltaic power generation systems; P Lj,t , Q Lj,t yes t Time Node j Active power and reactive power load requirements; Q CBj,t yes t Time Node j The reactive power of capacitors; v i,t , v j,t yes t Time Node i and nodes j The square of the voltage amplitude; P jk,t , Q jk,t yes t Time flow through the branch jk Active power and reactive power; P ij,t , Q ij,t yes t Time flow through the branch ij The active power and reactive power of x ij , x jk It is a branch road ij and branch roads jk The reactance, l jk,t yes t Time Branch jk The square of the current amplitude flowing through; r jk It is a branch road jk Resistance; The node voltage constraints are: , in, V min , V max are the minimum and maximum voltage values; The reactive power regulation capability of photovoltaic inverters and capacitors is: , , in, S PV is the photovoltaic inverter capacity, which is set to 1.1 times the maximum photovoltaic power generation; c ik,t for t Moment k The state of the capacitors; q ik,t yes t Moment k The reactive power output of a capacitor.

[0041] As a possible implementation of this embodiment, the voltage-reactive power regulation range of the distribution network is: , in[ Q LV,t , Q UV,t ] are the lower and upper limits of voltage-reactive power regulation capability.

[0042] As a possible implementation of this embodiment, the specific process of the model training and deployment module performing real-time quantification of the voltage-reactive power regulation capability of the distribution network is as follows: A deep neural network is trained to characterize and learn the functional relationship between multiple electrical inputs and output voltage-reactive power regulation capabilities. The voltage-reactive power regulation range of the distribution network is output through the deep neural network to determine the voltage-reactive power regulation capability.

[0043] As a possible implementation of this embodiment, the input and output features of the deep neural network are: , The input of the deep neural network includes the active power of the photovoltaic system P PVi,t , Load demand P Li,t and Q Li,t , and the voltage at the connection point of the transmission and distribution grid V 1,t , Q LV,t and Q UV,t It is the lower and upper limits of voltage-reactive power regulation capability.

[0044] As a possible implementation of this embodiment, the input-output function of the first hidden layer of the deep neural network is: , in w 1 and b 1 represents the weight and bias connecting the input and the first hidden layer respectively; h1 is the output of the first hidden layer; μ ( x )=max{0, x} is the ReLU activation function.

[0045] As a possible implementation of this embodiment, the output layer function of the deep neural network is: , in w 6 and b 6 represent the weight and bias connecting the last hidden layer and the output layer respectively.

[0046] As a possible implementation of this embodiment, the loss function of the deep neural network is: , in f is the loss function, which is the mean square error between the neural network prediction and the optimal value; N b is a small batch size; y k It is k training samples, y * k for y k The optimal label of ; W is the set of weights and biases; λ is a hyperparameter of L2 regularization, used to avoid overfitting.

[0047] like Figure 3 As shown in the figure, the quantification process of the voltage-reactive power regulation capability of the distribution network based on deep learning is divided into three parts: establishing a coordinated optimization control framework for the voltage of the transmission-distribution network, establishing an optimal operation model of the distribution network with the minimum power loss as the goal, and improving the optimization model considering the reactive power interaction constraints at the connection point of the transmission-distribution network.

[0048] like Figure 4 As shown in the figure, the theoretical basis for constructing a coordinated voltage optimization control framework for transmission-distribution networks is as follows: the dispatching decision of the transmission network can affect the optimal operation of the distribution network, and vice versa. On the one hand, the substation node voltage obtained by the transmission network through optimal dispatching will affect the overall voltage distribution of the distribution network, thereby limiting the power regulation capability of the voltage regulation equipment, such as the feasible domain of capacitors and photovoltaic power generation systems. On the other hand, the optimal dispatching results of the distribution network will affect the power interaction at the connection point of the transmission-distribution network and thus affect the optimal operation of the transmission network. The transmission-distribution networks each optimize their operation and exchange voltage and power information at the connection point. Through continuous iteration to meet the convergence conditions, collaborative optimization is achieved.

[0049] The optimization model function is constructed based on power loss, power flow equation, node voltage constraint and reactive power condition capability of flexible voltage regulation resources. In order to minimize power loss while meeting voltage safety constraints, the objective function is the sum of all line power losses: , in: l ij It is a branch road i In time t The node current amplitude at ; r ij It is a branch road ij resistance.

[0050] The Distflow mathematical model used to describe the power flow in the distribution network is: , , , in, P PVj,t , Q PVj,t yes t Time Node j Active power and reactive work of photovoltaic power generation system; P Lj,t , Q Lj,t yes t Time Node j Active power and reactive power load; Q CBj,t yes t Time Node j The reactive power flow of capacitors; v i,t , v j,t yes t Time Node i and nodes j The square of the voltage amplitude; P jk,t , Q jk,t yes t Time flow through the branch jk Active power and reactive power; P ij,t , Q ij,t yes t Time flow through the branch ij The active power and reactive power of x ij , x jkIt is a branch road ij and branch roads jk The reactance, l jk,t yes t Time Branch jk The square of the current amplitude flowing through; r jk It is a branch road jk resistance.

[0051] The node voltage constraint is expressed as: , in, V min , V max are the minimum and maximum voltage values.

[0052] When considering the photovoltaic power generation system inverter and capacitor as controllable devices, the corresponding reactive power regulation capability function is: , , in, S PV is the photovoltaic inverter capacity, which is set to 1.1 times the maximum photovoltaic power generation; c ik,t for t Moment k The state of the capacitors; q ik,t yes t Moment k The reactive power output of a capacitor.

[0053] The voltage-reactive power regulation capability of the distribution network can be obtained by solving the following optimization problem under the constraints of power flow, voltage and reactive power of voltage regulating equipment: , , in, Q i1,t Reactive power injection for balancing nodes; P * Loss ,t represents the minimum power loss obtained by solving the optimization problem; δ is a threshold value, which indicates the additional power loss that the distribution network can tolerate, and can be determined by the transmission-distribution network negotiation. The PV inverter operates in the maximum power point tracking mode, and the system load follows the typical daily load curve, i.e. P PVi,t , P Li,t and Q Li,tIt is a state variable that changes passively. The voltage-reactive power regulation capability is only affected by Q PVi,t and Q CBi,t control.

[0054] The maximum and minimum reactive power range at the transmission-distribution network interface is solved: , in,[ Q LV,t , Q UV,t ] describes the voltage-reactive power regulation capability, that is, the reactive power regulation range at the transmission-distribution network interface. A positive value indicates that the distribution network absorbs inductive reactive power from the transmission network.

[0055] Based on the optimization model, data is generated and a deep learning network is trained to effectively characterize the functional relationship between various electrical inputs and output voltage-reactive power regulation capabilities, thereby achieving efficient and accurate quantification. Q LV,t and Q UV,t , the voltage-reactive power regulation capability can be directly determined. The input of the deep neural network should consider all the key factors that affect the voltage-reactive power regulation capability. The input and output functions of deep learning are: , Among them: Input active power of photovoltaic power generation system P PVi,t , Load demand P Li,t , Q Li,t and the voltage at the connection point of the transmission and distribution network V 1,t Influence. P PVi,t It directly determines the reactive power regulation capability of the photovoltaic inverter; the changing load changes the voltage distribution and power flow of the distribution network; the voltage at the connection point of the transmission-distribution network will affect the voltage safety of the distribution network and may limit its voltage-reactive power regulation capability.

[0056] The maximum and minimum reactive power output of the capacitor is predetermined and remains constant during daily operation, and it is not necessary to include Q in the input. CBi,t .

[0057] like Figure 5 As shown, the input-output function of the first hidden layer of offline deep learning is: , in: w 1 and b1 represents the weight and bias connecting the input and the first hidden layer respectively; h 1 is the output of the first hidden layer; μ ( x )=max{0, x} is the ReLU activation function.

[0058] Offline deep learning output function based on forward propagation: , in: w 6 and b 6 represent the weight and bias connecting the last hidden layer and the output layer respectively.

[0059] Loss function for offline deep learning: , in f is the loss function, which is the mean square error between the neural network prediction and the optimal value; N b is a small batch size; y k It is k training samples, y * k for y k The optimal label of ; W is the set of weights and biases; λ is a hyperparameter of L2 regularization, used to avoid overfitting.

[0060] In the embodiment of the present invention, a transmission-distribution network with a coupled connection relationship is selected, that is, an improved IEEE 33-node distribution network. The specific topology is as follows: Figure 6 The 24-hour photovoltaic, load, and voltage curve test data in the system are shown in Figure 7 shown.

[0061] Consider the following three case analysis scenarios: Scenario 1: Convergence analysis of loss function during training iterations; Scenario 2: Sensitivity analysis of reactive power regulation capability to key input variables; Scenario 3: Accuracy and efficiency verification analysis.

[0062] For case 1, Figure 8 Shows the loss function of deep learning during training iterations. Figure 8 The loss function in the first 30 iterations decreases rapidly and eventually converges to less than 5*10 -4 level. Figure 8The trend of the loss function in shows that deep learning has learned the functional relationship between voltage and power measurements and the voltage-reactive power regulation capability of the distribution network.

[0063] For case 2, a sensitivity analysis was conducted on the key input variables of voltage-reactive power regulation capability. Figure 9-10 The corresponding simulation results are shown using the system operation data at 13:00 as an example. Fig. 9 In, when V 1,t When approaching the upper and lower voltage boundaries, the reactive power regulation capability changes significantly, that is, when V 1,t When it is equal to 1.05 pu, Q LV Increase, when V 1,t When it is equal to 0.99pu, Q UV Reduce (when V 1,t <0.99 pu, there is no feasible solution). V 1,t When t is close to 1.05 pu, the reactive power output of photovoltaic power generation is limited to avoid voltage rise problems, thereby limiting the reactive power support capacity at the transmission-distribution grid interface. Similar analysis also applies to low voltage scenarios. Fig.10 In, with δ increase, Q UV Gradually increasing, Q LV Remain relatively stable and have no significant change trend. Q LV In comparison, the change of network loss has an impact on Q UV Therefore, improving the tolerance of distribution network losses may not necessarily improve the voltage support capacity at the transmission-distribution network interface.

[0064] For case 3, the accuracy and efficiency of the present invention were verified by comparing the present invention with the model-based optimization method. Fig.11 and Fig.12 The performance of the model-based optimization method and the present invention in terms of the voltage regulation capability of the distribution network under 24 hours is compared. Fig.11 In Q LV , the method proposed in this invention successfully tracks the optimal quantization result. Fig.12 In Q UV, the proposed method has high accuracy in most of the time, but the reactive power deviation occurs at 19:00, with a maximum deviation of about 240 kVar, accounting for 8% of the total voltage reactive power regulation capacity. This error may be caused by the interpretability problem of the deep learning method, and its occurrence frequency is low and still within the acceptable range. Fig.13 To solve Q UV and Q LV The voltage curve of 33 nodes in the distribution network (the lowest voltage level) is shown in Figure 2. The results show that all bus voltages are higher than the threshold of 0.95 pu, which meets the voltage safety requirements in the distribution network.

[0065] The present invention is based on deep learning and has high computational efficiency. It only takes 0.0019 seconds to estimate the voltage-reactive power regulation capability of the distribution network. In contrast, the model-based optimization method takes 6.72 seconds to solve the optimization problem and obtain Q UV and Q LV The computational efficiency of the method proposed in the present invention is significantly improved, which is about 3500 times that of the model-based optimization method.

[0066] An electronic device provided by an embodiment of the present invention includes a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the device is running, the processor communicates with the memory via the bus, and the processor executes the machine-readable instructions to perform the steps of the method for quantifying the voltage-reactive power regulation capability of any distribution network as described above.

[0067] Specifically, the above-mentioned memory and processor can be general-purpose memory and processor, which are not specifically limited here. When the processor runs the computer program stored in the memory, it can execute the above-mentioned method for quantifying the voltage-reactive power regulation capability of the distribution network.

[0068] Those skilled in the art will appreciate that the structure of the electronic device does not limit the electronic device and may include more or fewer components than shown in the figure, or combine or split certain components, or arrange the components differently.

[0069] In some embodiments, the electronic device may also include a touch screen that can be used to display a graphical user interface (e.g., a startup interface of an application) and receive user operations on the graphical user interface (e.g., startup operations on an application). The specific touch screen may include a display panel and a touch panel. The display panel may be configured in the form of LCD (Liquid Crystal Display), OLED (Organic Light-Emitting Diode), etc. The touch panel may collect the user's contact or non-contact operations on or near it, and generate pre-set operation instructions, for example, the user uses any suitable object such as a finger, stylus, or accessories on or near the touch panel. In addition, the touch panel may include two parts: a touch detection device and a touch controller. The touch detection device detects the user's touch position and posture, detects the signal brought by the touch operation, and transmits the signal to the touch controller; the touch controller receives the touch information from the touch detection device, converts it into information that the processor can process, and then sends it to the processor, and can receive and execute commands from the processor. In addition, the touch panel can be implemented by various types such as resistive, capacitive, infrared and surface acoustic wave, and any technology developed in the future can also be used to implement the touch panel. Further, the touch panel can cover the display panel, and the user can operate on or near the touch panel covered on the display panel according to the graphical user interface displayed on the display panel. After the touch panel detects the operation on or near it, it is transmitted to the processor to determine the user input, and then the processor provides corresponding visual output on the display panel in response to the user input. In addition, the touch panel and the display panel can be implemented as two independent components or integrated.

[0070] Corresponding to the startup method of the above-mentioned application, an embodiment of the present invention also provides a storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the quantification method of the voltage-reactive power regulation capability of any distribution network are executed.

[0071] The application startup device provided in the embodiment of the present application can be specific hardware on the device or software or firmware installed on the device. The device provided in the embodiment of the present application, its implementation principle and the technical effect produced are the same as those in the aforementioned method embodiment. For the sake of brief description, the parts not mentioned in the device embodiment can refer to the corresponding contents in the aforementioned method embodiment. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above can all refer to the corresponding processes in the aforementioned method embodiment, and will not be repeated here.

[0072] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0073] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of modules is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0074] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0075] In addition, each functional module in the embodiments provided in the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0076] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0077] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0078] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for quantifying the voltage-reactive power regulation capability of a distribution network, characterized in that: The steps include: Step 1: Collect and organize a large amount of transmission-distribution network data, analyze the voltage coordinated optimization control mechanism of the transmission-distribution network, and establish a coordinated optimization control framework; Step 2: Based on the coordinated optimization control framework, with the goal of minimizing power loss and the safety and economy of distribution network operation as constraints, a voltage optimization control model based on second-order cone programming is established, and the voltage-reactive power regulation capacity of the distribution network is solved based on the optimization model conversion; Step 3: Design a deep neural network, perform offline training based on historical data of the transmission and distribution network, deploy the trained model online, and perform real-time quantification of the voltage-reactive power regulation capability of the distribution network.

2. The method for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 1, characterized in that: The objective function of the voltage optimization control model is: ， ， ， in, c ik,t for t Moment k The state of the capacitors; Q PVj,t yes t Time Node j Reactive power flow of photovoltaic power generation system; Q i1,t Reactive power injection for balancing nodes; P * Loss ,t represents the minimum power loss obtained by solving the optimization problem; l ij,t yes t Time branch ij The square of the magnitude of the current flowing through; r ij It is a branch road ij The resistance of δ Represents the additional power loss that the distribution network can tolerate.

3. The method for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 2, characterized in that: The constraints of the voltage optimization control model include: Power flow constraints of distribution network: ， ， ， in, P PVj,t , Q PVj,t yes t Time Node j Active and reactive power flows of photovoltaic power generation systems; P Lj,t , Q Lj,t yes t Time Node j Active power and reactive power load requirements; Q CBj,t yes t Time Node j The reactive power of capacitors; v i,t , v j,t yes t Time Node i and nodes j The square of the voltage amplitude; P jk,t , Q jk,t yes t Time flow through the branch jk Active power and reactive power; P ij,t , Q ij,t yes t Time flow through the branch ij The active power and reactive power of x ij , x jk It is a branch road ij and branch roads jk The reactance, l jk,t yes t Time branch jk The square of the magnitude of the current flowing through; r jk It is a branch road jk The resistance of The node voltage constraints are: ， in, V min , V max are the minimum and maximum voltage values; The reactive power regulation capability of photovoltaic inverters and capacitors is: ， ， in, S PV is the PV inverter capacity; c ik,t for t Moment k The state of the capacitors; q ik,t yes t Moment k The reactive power output of a capacitor.

4. The method for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 1, characterized in that: The voltage-reactive power regulation range of the distribution network is: ， in[ Q LV,t , Q UV,t ] are the lower and upper limits of voltage-reactive power regulation capability.

5. The method for quantifying the voltage-reactive power regulation capability of a distribution network according to any one of claims 1 to 4, characterized in that: The real-time quantification of the voltage-reactive power regulation capability of the distribution network includes: A deep neural network is trained to characterize and learn the functional relationship between multiple electrical inputs and output voltage-reactive power regulation capabilities. The voltage-reactive power regulation range of the distribution network is output through the deep neural network to determine the voltage-reactive power regulation capability.

6. The method for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 5, characterized in that: The input and output features of the deep neural network are: ， The input of the deep neural network includes the active power of the photovoltaic system P PVi,t , Load demand P Li,t and Q Li,t , and the voltage at the connection point of the transmission and distribution grid V 1,t , Q LV,t and Q UV,t It is the lower and upper limits of voltage-reactive power regulation capability.

7. The method for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 5, characterized in that: The input-output function of the first hidden layer of the deep neural network is: ， in w 1 and b 1 represents the weight and bias connecting the input and the first hidden layer respectively; h 1 is the output of the first hidden layer; μ ( x )=max{0, x } is the ReLU activation function.

8. The method for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 5, characterized in that: The output layer function of the deep neural network is: ， in w 6 and b 6 represent the weight and bias connecting the last hidden layer and the output layer respectively.

9. The method for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 5, characterized in that: The loss function of the deep neural network is: ， in f is the loss function, which is the mean square error between the neural network prediction and the optimal value; N b is a small batch size; y k It is k training samples, y * k for y k The optimal label of ; W is the set of weights and biases; λ is the hyperparameter of L2 regularization.

10. A device for quantifying the voltage-reactive power regulation capability of a distribution network, characterized in that: include: The control framework establishment module is used to collect and organize a large amount of transmission-distribution network data, analyze the voltage coordination optimization control mechanism of the transmission-distribution network, and establish a coordination optimization control framework; The control model building module is used to establish a voltage optimization control model based on second-order cone programming based on the coordinated optimization control framework, with the goal of minimizing power loss and the safety and economy of distribution network operation as constraints, and to solve the voltage-reactive power regulation capacity of the distribution network based on the optimization model conversion; The model training and deployment module is used to design deep neural networks, conduct offline training based on historical data of the transmission and distribution network, deploy the trained model online, and perform real-time quantification of the voltage-reactive power regulation capability of the distribution network.

11. The device for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 10, characterized in that: The objective function of the voltage optimization control model is: ， ， ， in, c ik,t for t Moment k The state of the capacitors; Q PVj,t yes t Time Node j Reactive power flow of photovoltaic power generation system; Q i1,t Reactive power injection for balancing nodes; P * Loss ,t represents the minimum power loss obtained by solving the optimization problem; l ij yes t Time Branch ij The square of the current amplitude flowing through; r ij It is a branch road ij The resistance of δ Represents the additional power loss that the distribution network can tolerate.

12. The device for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 11, characterized in that: The constraints of the voltage optimization control model include: Power flow constraints of distribution network: ， ， ， in, P PVj,t , Q PVj,t yes t Time Node j Active and reactive power flows of photovoltaic power generation systems; P Lj,t , Q Lj,t yes t Time Node j Active power and reactive power load requirements; Q CBj,t yes t Time Node j The reactive power of capacitors; v i,t , v j,t yes t Time Node i and nodes j The square of the voltage amplitude; P jk,t , Q jk,t yes t Time flow through the branch jk Active power and reactive power; P ij,t , Q ij,t yes t Time flow through the branch ij The active power and reactive power of x ij , x jk It is a branch road ij and branch roads jk The reactance, l jk,t yes t Time Branch jk The square of the current amplitude flowing through; r jk It is a branch road jk The resistance of The node voltage constraints are: ， in, V min , V max are the minimum and maximum voltage values; The reactive power regulation capability of photovoltaic inverters and capacitors is: ， ， in, S PV is the PV inverter capacity; c ik,t for t Moment k The state of the capacitors; q ik,t yes t Moment k The reactive power output of a capacitor.

13. The device for quantifying the voltage-reactive power regulation capability of a distribution network according to claim 10, characterized in that: The voltage-reactive power regulation range of the distribution network is: ， in[ Q LV,t , Q UV,t ] are the lower and upper limits of voltage-reactive power regulation capability.

14. The device for quantifying the voltage-reactive power regulation capability of a distribution network according to any one of claims 10 to 13, characterized in that: The specific process of the model training and deployment module for real-time quantification of the voltage-reactive power regulation capability of the distribution network is as follows: A deep neural network is trained to characterize and learn the functional relationship between multiple electrical inputs and output voltage-reactive power regulation capabilities. The voltage-reactive power regulation range of the distribution network is output through the deep neural network to determine the voltage-reactive power regulation capability.

15. An electronic device, characterized in that: It includes a processor, a memory and a bus, the memory stores machine-readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, and the processor executes the machine-readable instructions to perform the steps of the method for quantifying the voltage-reactive power regulation capability of the distribution network as described in any one of claims 1-9.

16. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, executes the steps of the method for quantifying the voltage-reactive power regulation capability of a distribution network as claimed in any one of claims 1 to 9.

Images