A power system reactive voltage control optimization method, system, medium and device

Abstract

The application discloses a power system reactive power and voltage control optimization method, system, medium and equipment. The method comprises the following steps: establishing a voltage-reactive power multivariate linear regression model based on a power system flow equation; obtaining training data; solving the regression model by using the training data to obtain a coefficient matrix W; verifying the accuracy of the regression model determined by the coefficient matrix W; selecting a static reactive power compensation device as a reactive power regulating device, embedding the voltage-reactive power multivariate linear regression model into optimization, taking the minimum sum of voltage deviation of each node of the power system as the target, and establishing a reactive power-voltage control optimization model; solving the reactive power-voltage control optimization model to obtain reactive power scheduling strategies of each node at different times, and realizing reactive power and voltage control optimization. The application adopts a data-driven method, takes the minimum node voltage deviation during voltage control of the power system as the target, and establishes a voltage-reactive power multivariate linear regression model and a reactive power scheduling optimization model based on sensitivity.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power system control, and particularly relates to a power system reactive voltage control optimization method, system, medium and equipment. BACKGROUND

[0002] The voltage level is one of the standards for measuring the power quality of a power system, and excessively high or low voltage will have adverse effects on power equipment and even users. Ensuring that the bus voltage value is within the specified range and close to the rated voltage value is one of the basic tasks of power system operation scheduling. When the voltage of the system deviates from the allowed value, power system voltage control must be performed to maintain the voltage of the power system within the allowed range.

[0003] Power system voltage control is achieved by taking measures to reasonably arrange the reactive power injection of each bus of the power system, changing the reactive power flow distribution of the power system, and ensuring that the bus voltage is within the required range. In order to implement the energy low-carbon transformation strategy, the penetration rate of renewable energy in the power system is increasing, and the proposal of the "double carbon" target and the proposal of "building a new power system dominated by new energy" make it inevitable for renewable energy to be connected to the power system in large scale. Traditional voltage control methods are based on power flow calculation and rely on the physical model of power system power facilities. However, the large-scale connection of renewable energy significantly increases the uncertainty, randomness and complexity of power system operation, making it challenging to establish an accurate system model. Traditional voltage control based on physical models is difficult to apply to new power systems dominated by new energy. Moreover, the traditional voltage regulation method through switching of capacitor banks has discrete values, can only achieve step difference regulation, and has a large impact on the system when switched. The number of switching operations is limited, and the voltage regulation effect is not ideal. SUMMARY

[0004] The main purpose of the present application is to overcome the shortcomings and deficiencies of the prior art, and to provide a power system reactive voltage control optimization method, system, medium and equipment. The present application adopts a data-driven method, takes the minimum node voltage deviation during power system voltage control as the target, and establishes a voltage-reactive multi-linear regression model based on sensitivity and a reactive power dispatch optimization model.

[0005] In order to achieve the above purpose, the technical scheme adopted by the present application is as follows:

[0006] In one aspect of the present application, a power system reactive voltage control optimization method is provided, comprising the following steps:

[0007] A linear regression model of node voltage and system node injected reactive power, i.e. a voltage-reactive multi-linear regression model, is established based on the power system power flow equation;

[0008] Record the reactive power load of each node, the reactive power output of the generator and the voltage value of each node under different operation scenarios to form training data;

[0009] Solve the voltage-reactive power multivariate linear regression model by using the training data to obtain a coefficient matrix W, Ψ is a voltage-reactive power sensitivity matrix, is a constant vector.

[0010] Perform power flow calculation to verify the accuracy of the voltage-reactive power multivariate linear regression model determined by the obtained coefficient matrix W.

[0011] Select a static reactive power compensation device as a reactive power regulation device, embed the voltage-reactive power multivariate linear regression model into optimization with the minimum sum of voltage deviation squares of each node in the power system as the target to establish a reactive power-voltage control optimization model.

[0012] Solve the established reactive power-voltage control optimization model to obtain the reactive power scheduling strategy of each node at different times to realize reactive power-voltage control optimization.

[0013] As a preferred technical solution, the voltage-reactive power multivariate linear regression model is specifically:

[0014]

[0015] Wherein, V(t) is the voltage vector of the power system node at time t; Q(t) is the reactive power injection vector of the system node at time t; Ψ(t) is the voltage-reactive power sensitivity matrix at time t; is a constant vector at time t; and ε is an error vector.

[0016] As a preferred technical solution, the recording of the reactive power load of each node, the reactive power output of the generator and the voltage value of each node under different operation scenarios to form training data is specifically:

[0017] Randomly generate load by using a power flow calculation tool, generate an operation scenario according to the load, obtain the node voltage value under different operation scenarios by power flow calculation, record the reactive power load of each node, the reactive power output of the generator and the voltage value of each node under different operation scenarios to form training data.

[0018] As a preferred technical solution, the sensitivity matrix Ψ and the constant vector in the proposed voltage-reactive power multivariate linear regression model are determined by a linear regression method based on least squares. Specifically,

[0019] For a single node voltage data y and each node reactive power injection data X, the parameter ω of the corresponding row of the coefficient matrix W is obtained by a linear regression method based on least squares, and the parameters obtained by multiple node regression together constitute the coefficient matrix W.

[0020] determining a target function of the least square based linear regression method, when X T When X is a full rank matrix or a positive definite matrix, the parameters in the voltage-reactive power multivariate linear regression model are solved directly by using the first order condition, that is, ω=(X T X) -1 X T y;When X T X does not satisfy the full rank matrix or the positive definite matrix condition, a penalty term is added, and the gradient descent method is used to calculate the parameters in the voltage-reactive power multivariate linear regression model.

[0021] As a preferred technical solution, the target function Y(ω) of the least square based linear regression method is:

[0022]

[0023] Wherein, y ω , y are predicted values and true values respectively, and X is a sample matrix of independent variables.

[0024] As a preferred technical solution, the power flow calculation is performed to verify the accuracy of the voltage-reactive power multivariate linear regression model determined by the coefficient matrix W, and specifically:

[0025] The accuracy of the voltage-reactive power multivariate linear regression model is verified by comparing the voltage calculated by using the voltage-reactive power multivariate linear regression model with the voltage calculated by using the classic Newton-Raphson method power flow calculation, and comparing the voltage calculated by using the linear power flow with the voltage calculated by using the classic Newton-Raphson method power flow calculation.

[0026] As a preferred technical solution, the reactive power-voltage control optimization model is:

[0027]

[0028] Wherein, v before (t) is a voltage vector before optimization, Ψ(t) is a voltage-reactive power sensitivity matrix at t, Δq(t) is a reactive power scheduling vector, v ref (t) is a node voltage reference vector, Δ q (t)、 are a reactive power scheduling lower limit and upper limit vector respectively, v (t)、 are a node voltage lower limit and upper limit vector respectively.

[0029] Another aspect of the present application provides a data-driven based power system reactive power voltage control optimization system, comprising:

[0030] A power flow module is configured to establish a voltage-reactive power multi-linear regression model of node voltage and system node reactive power injection according to power system power flow equations.

[0031] A data module is configured to record training data of node reactive power load, generator reactive power output and node voltage value under different operation scenarios.

[0032] A calculation module is configured to obtain a coefficient matrix W by solving the voltage-reactive power multi-linear regression model established by the power flow module according to the training data of the data module. Ψ is a voltage-reactive power sensitivity matrix, is a constant vector.

[0033] A verification module is configured to perform power flow calculation and verify the accuracy of the voltage-reactive power multi-linear regression model determined by the coefficient matrix W obtained by the calculation module.

[0034] A selection module is configured to select a static reactive power compensation device as a reactive power regulation device, and establish a reactive power-voltage control optimization model with the minimum sum of power system node voltage deviation squares as the target.

[0035] An optimization module is configured to solve the reactive power-voltage control optimization model established by the selection module to obtain node reactive power scheduling strategies at different time and realize reactive power-voltage control optimization.

[0036] Another aspect of the present application provides a computer readable storage medium storing one or more programs, the one or more programs including instructions that when executed by a computing device cause the computing device to perform any of the power system reactive power-voltage control optimization methods according to the above.

[0037] Another aspect of the present application provides a computing device comprising:

[0038] one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for performing any of the power system reactive power-voltage control optimization methods according to the above.

[0039] Compared with the prior art, the present application has the following advantages and beneficial effects:

[0040] (1)Adopting a data-driven method, unknown parameters in a voltage-reactive power multivariate linear model are estimated by regression using historical operation data of a power system, compared with a traditional method relying on a physical model of the power system, the method proposed in the application is based on a multivariate linear model and does not rely on a physical model, so that the method can better cope with the diversity of power system load and the randomness and uncertainty of distributed renewable energy, and is more suitable for a new type of power system with distributed renewable energy being connected continuously;

[0041] (2) When reactive power flow of the power system changes, the traditional Newton-Raphson method is used to calculate the voltage value of each node of the system, although the result is more accurate, but the Newton-Raphson method takes a long time to calculate the power flow; the voltage-reactive power multivariate linear regression model based on sensitivity is used to calculate the voltage value of each node, the calculation speed is fast, and the relative error of the calculation result is small, which can well agree with the Newton-Raphson method; at the same time, compared with linear power flow calculation, the linear model based on sensitivity has faster calculation speed and more accurate result;

[0042] (3) The static var compensator (SVC) is selected as the reactive power regulating device, the SVC is a device connected in parallel with the system and supplying or absorbing reactive power from the system, which is composed of capacitors and various reactance elements. The traditional voltage regulating method using switched capacitors has discrete reactive power regulating capacity, can only realize step difference regulation, and the number of switching operations of the capacitors is limited; the SVC is used for voltage regulation, which has fast response speed, can realize continuous two-way reactive power regulation, has small impact on the system during regulation, and has no limitation on the number of operations.

[0043] The technical solutions of the application will be further described in detail below with the aid of the drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 It is a flowchart of the reactive power-voltage control optimization based on data driving;

[0045] Figure 2 It is a flowchart of the voltage control algorithm;

[0046] Figure 3 It is a verification diagram of the multivariate linear model based on sensitivity, which is a relative error diagram of node voltage calculation by the linear model based on sensitivity and the Newton-Raphson method under a certain scenario;

[0047] Figure 4 It is a relative error diagram of node voltage calculation by the linear power flow method and the Newton-Raphson method under a certain scenario;

[0048] Figure 5 It is a comparison of the objective function values before and after optimization at each time;

[0049] Figure 6 The voltage value comparison of each node before and after optimization at a certain time;

[0050] Figure 7 The reactive power scheduling value of each node at a certain time. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0052] In the description of the present application, it should be understood that the terms "include" and "contain" indicate the existence of described features, whole, steps, operations, elements and / or components, but do not exclude the existence or addition of one or more other features, whole, steps, operations, elements, components and / or sets thereof.

[0053] It should also be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification and the appended claims of the present application, unless otherwise clear from the context, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0054] It should be further understood that the term "and / or" used in the specification and the appended claims of the present application means any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0055] Various structural diagrams according to the disclosed embodiments of the present application are shown in the drawings. These diagrams are not drawn to scale, in which certain details are exaggerated for the purpose of clear expression, and certain details can be omitted. The shapes of various regions, layers and their relative size and positional relationship shown in the drawings are only exemplary, and in actuality, there can be deviations due to manufacturing tolerances or technical limitations, and a person skilled in the art can additionally design regions / layers with different shapes, sizes and relative positions according to actual needs.

[0056] EMBODIMENTS

[0057] As Figure 1As shown, the data-driven-based reactive power and voltage control optimization method of the power system of the application comprises six parts, which are establishing a multiple linear regression model, generating training data, solving the voltage-reactive multiple linear regression model, verifying the accuracy of the sensitivity-based multiple linear regression model, establishing a voltage-reactive control optimization model, and solving the voltage-reactive control optimization model. The specific steps are as follows:

[0058] S1, based on the power system flow equation, a linear regression model of node voltage and system node injected reactive power is established;

[0059] The relationship between voltage and reactive power is:

[0060] v(t)=f(q(t)) (1)

[0061] Wherein, f is a multi-input, multi-output nonlinear function, and q(t) is a bus reactive power injection vector.

[0062] At time t, when the function f is available, the voltage regulation is carried out by using the relationship. However, this relationship is difficult to adapt to large-scale power systems containing many random components. In order to realize a real-time applicable method, for the power system with strong coupling of voltage and reactive power, f is approximated as a linear affine function:

[0063]

[0064] Wherein, Ψ represents the voltage-reactive sensitivity matrix, represents a constant vector.

[0065] The relationship between the change amount of the power injected by the bus of the power system and the change amount of the voltage is represented as:

[0066]

[0067] Wherein, Δp and Δq are the change vectors of the active power and the reactive power injected by the bus respectively; Δv and Δθ are the change vectors of the voltage amplitude and the phase angle respectively; the partial derivative matrices J pθ 、J pv 、J qθ 、J qv of different power to different state quantities together constitute the Jacobian matrix J.

[0068] Under normal circumstances, the value of Δv is related to Δp, Δq and Δθ. In the transmission network, the line reactance value is much larger than the resistance, which makes the Jacobian matrix J pv 、J qθThe value of J is very small. Therefore, in the power grid, the change of the bus active power injection is mainly related to the voltage phase angle, the change of the bus reactive power injection is mainly related to the voltage amplitude, the influence of the voltage phase angle and the active power on the voltage amplitude is ignored, and the relationship between the voltage amplitude change and the bus reactive power injection is represented as:

[0069] Δv=J qv -1 Δq (4)

[0070] Comparing formula (2) and formula (4), the voltage-reactive power sensitivity matrix Ψ can be approximately estimated according to J qv . Therefore, for a power system with k nodes, the relationship between the voltage and the reactive power is approximately represented as:

[0071]

[0072] wherein V(t) is a system node voltage vector at time t; Q(t) is a system node reactive power injection vector at time t; Ψ(t) is a voltage-reactive power sensitivity matrix at time t; is a constant vector at time t; when the system topology is unchanged, the sensitivity matrix and the constant vector at different times are the same; ε is an error vector, representing the error value between the predicted value and the actual output.

[0073] Formula (5) is a typical multiple linear regression model, and a linear regression method based on least squares is used to solve the voltage-reactive power sensitivity matrix Ψ and the constant vector

[0074] S2, generating training data;

[0075] A large number of loads are randomly generated by using a power flow calculation tool (such as matpower), a large number of operation scenarios are generated, node voltage values under different operation scenarios are obtained through power flow calculation, and the reactive power load and the generator reactive power output (i.e. the node reactive power injection value) of each node under different operation scenarios and the node voltage values are recorded to form the training data, which is used to determine the unknown parameters in the voltage-reactive power multiple linear regression model established in step S1;

[0076] S3, using the training data obtained in step S2, a linear regression method based on least squares is used to solve the multiple linear regression model established in step S1, the unknown parameters in the sensitivity-based voltage-reactive power multiple linear regression model are determined, and a coefficient matrix W is obtained: for a single node voltage data y and each node reactive power injection data X, the parameters ω obtained by the linear regression method based on least squares are the corresponding row of the matrix W, and the parameters obtained by multiple node regression together constitute the coefficient matrix W,

[0077] The least square method is a method for solving linear regression model, the principle is to make the error square sum of all observations and corresponding regression estimates minimum, that is, the square sum of each element in ε is minimum, the partial derivative of the objective function is solved, and the first order condition is used to solve the parameters. For a single bus voltage and system each node reactive power injection multivariate linear regression model, unknown parameters are ω = [ω0ω1…ω k ], ω0 is the constant value in the regression model, and is the value of corresponding node in formula (5) ω i (i = 1, …, k) is the voltage reactive sensitivity value, and is the value of corresponding node in formula (5).

[0078] The objective function of the linear regression method based on least square is:

[0079]

[0080] Where, y ω , y are the predicted value and the true value, and X is the sample matrix of independent variable.

[0081] When X T X is a full rank matrix or a positive definite matrix, the normal equation method is used to directly solve the gradient as follows:

[0082]

[0083] Let Then:

[0084] X T Xω-X T y = 0 (8)

[0085] When |X T X|≠0, we have:

[0086] ω = (X T X) -1 X T y (9)

[0087] When X T X does not satisfy the full rank matrix or positive definite matrix condition, the gradient descent method is used for solution; In order to prevent overfitting, a penalty term is added:

[0088] ω = (X T X + λI) -1 X T y (10)

[0089] S4, use the traditional method to carry out power flow calculation, verify the accuracy of the voltage-reactive multivariate linear regression model determined by the coefficient matrix W obtained in step S3; Figure 1 The data-driven reactive power-voltage control optimization flowchart is based on the multivariate linear model.

[0090] The accuracy of the voltage-reactive power multiple linear regression model is verified by comparing the voltage calculated by the sensitivity-based multiple linear model with the voltage calculated by the classic Newton-Raphson power flow calculation. Meanwhile, the voltage calculated by the linear power flow is compared with the voltage calculated by the classic Newton-Raphson power flow calculation. The relative error of the two is compared to verify that the accuracy of the sensitivity-based multiple linear regression model is higher than that of the linear power flow calculation.

[0091] The sensitivity-based multiple linear regression model is used to calculate the system bus voltage, which avoids complex power flow calculation, but the error between the predicted value and the true value increases compared with accurate power flow calculation. Based on the example in the power flow calculation software matpower, the power flow is calculated by changing the reactive power load of each node and using the Newton-Raphson method to obtain the voltage value of each node and the reactive power injection value of each generator node. The reactive power injection value of each node is calculated, and then the voltage value of each node is calculated by using the established sensitivity-based voltage-reactive power multiple linear regression model, and compared with the more accurate Newton-Raphson calculation result, the relative error is calculated, as shown in Figure 3 The relative error calculation result under a certain operating scenario is shown in FIG. 5; at the same time, the relative error of the voltage value of each node calculated by the linear power flow and the voltage value of each node calculated by the Newton-Raphson power flow is calculated, as shown in Figure 4 The relative error calculation result under the same scenario is shown in FIG. 6. Through analysis and comparison, the voltage calculated by the sensitivity-based multiple linear regression model can well agree with the result of the Newton-Raphson power flow calculation, and the relative error is small; compared with the linear power flow calculation result, the result calculated by the sensitivity-based multiple linear regression model is more accurate.

[0092] S5, a static var compensator (SVC) is selected as a reactive power regulating device, a voltage-reactive power multiple linear regression model established in S1 is embedded into optimization (the accuracy of the voltage-reactive power multiple linear regression model is verified in S4, so the voltage-reactive power multiple linear regression model can be directly embedded into the optimization model in S5) to establish a reactive power-voltage control optimization model;

[0093] The SVC is a device composed of capacitors and various reactance elements connected in parallel with the system and supplying or absorbing reactive power from the system, which is the most commonly used Flexible AC Transmission Systems (FACTS) device for reactive power compensation and voltage control, and can realize bidirectional continuous regulation of reactive power.

[0094] A voltage and reactive power control optimization model is established by taking the minimum of the sum of squares of voltage deviation of each node as the target, and by simultaneously constraining the reactive power scheduling value and the node voltage value at each time point:

[0095]

[0096] wherein v after is a node voltage vector after reactive power scheduling; v ref is a node reference voltage vector; Δq, Δ q , are reactive power scheduling, lower limit and upper limit vectors of reactive power scheduling respectively; v , are lower limit and upper limit vectors of voltage respectively. The constraint condition indicates that the reactive power scheduling value and the node voltage value at each time point should be within the specified range.

[0097] According to the voltage-reactive power multi-element linear regression model, there are:

[0098]

[0099] wherein v before is a node voltage vector before reactive power scheduling, and Ψ, is known, and the q value at time t is based on the predicted value of load and generator output.

[0100] Therefore, the reactive power-voltage control optimization model is expressed as:

[0101]

[0102] S6, based on the MATLAB calling CPLEX solver, the reactive power-voltage control optimization model established in step S5 is solved, and the reactive power scheduling strategy of each node at different time points is obtained, so as to realize the optimization of reactive power and voltage control.

[0103] The model shown in formula (13) is a quadratic programming model, and the voltage and reactive power optimization model is solved based on the MATLAB calling CPLEX solver, so as to obtain the reactive power scheduling strategy of each node at different time points.

[0104] Please refer to Figure 2 , Figure 2 The flow chart of the data-driven based power system reactive power and voltage control optimization method proposed in the application comprises data input before optimization, determination of coefficient matrix W, parameters and solution of the optimization model.

[0105] In still another embodiment of the present application, a data-driven power system reactive voltage control optimization system is provided, which can be used to implement the data-driven power system reactive voltage control optimization method described above. Specifically, the data-driven power system reactive voltage control optimization system comprises a power flow module, a data module, a calculation module, a verification module, a selection module, and an optimization module.

[0106] The power flow module is configured to establish a linear regression model of node voltage and system node injected reactive power according to power system power flow equations.

[0107] The data module is configured to record training data composed of node reactive load, generator reactive output, and node voltage values under different operating scenarios.

[0108] The calculation module is configured to solve the voltage-reactive power multivariate linear regression model established by the power flow module using a least squares-based linear regression method according to the training data of the data module to obtain a coefficient matrix W.

[0109] The verification module is configured to perform power flow calculation to verify the accuracy of the voltage-reactive power multivariate linear regression model determined by the coefficient matrix W obtained by the calculation module.

[0110] The selection module is configured to select a static reactive power compensation device as a reactive power regulating device, and to establish a reactive power-voltage control optimization model with the objective of minimizing the sum of squares of power system node voltage deviations.

[0111] The optimization module is configured to call a CPLEX solver based on MATLAB to solve the reactive power-voltage control optimization model established by the selection module to obtain node reactive power scheduling strategies at different times, thereby achieving reactive voltage control optimization.

[0112] In still another embodiment of the present application, a terminal device is provided, which comprises a processor and a memory, the memory being configured to store a computer program, the computer program comprising program instructions, and the processor being configured to execute the program instructions stored in the computer storage medium. The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, and are suitable for implementing one or more instructions, and are particularly suitable for loading and executing one or more instructions to implement a corresponding method flow or a corresponding function; the processor in the embodiments of the present application can be used for the operation of the power system reactive power and voltage control optimization method based on data driving, comprising:

[0113] establishing a linear regression model of node voltage and system node injected reactive power based on power system flow equation;

[0114] recording the training data composed of the reactive power load of each node, the generator reactive power output and the voltage value of each node under different operating scenarios;

[0115] using the training data, solving the multivariate linear regression model by using the least square based linear regression method to obtain the coefficient matrix W;

[0116] performing flow calculation to verify the accuracy of the voltage-reactive power multivariate linear regression model determined by the coefficient matrix W;

[0117] selecting a static reactive power compensation device as a reactive power regulating device, and establishing a reactive power-voltage control optimization model with the minimum sum of power system node voltage deviation squares as the target;

[0118] solving the reactive power-voltage control optimization model by using the CPLEX solver based on MATLAB to obtain the reactive power scheduling strategy of each node at different time, and realizing the reactive power and voltage control optimization.

[0119] In another embodiment of the present application, the present application also provides a storage medium, specifically a computer readable storage medium (Memory), which is a memory device in a terminal device, used for storing programs and data. It can be understood that the computer readable storage medium herein can include a built-in storage medium in the terminal device, and of course can also include an expansion storage medium supported by the terminal device. The computer readable storage medium provides a storage space, which stores an operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer readable storage medium herein can be a high-speed RAM memory, or a non-volatile memory such as at least one disk memory.

[0120] The one or more instructions stored in the computer readable storage medium can be loaded and executed by the processor to realize the corresponding steps of the data-driven power system reactive voltage control optimization method in the above embodiments; the one or more instructions in the computer readable storage medium are loaded and executed by the processor to perform the following steps:

[0121] A linear regression model of node voltage and system node injected reactive power is established based on power system flow equations;

[0122] Record the training data composed of the reactive power load of each node, the reactive power output of the generator, and the voltage value of each node under different operating scenarios;

[0123] Using the training data, a multivariate linear regression model is solved by using a least squares-based linear regression method to obtain a coefficient matrix W;

[0124] Perform power flow calculation to verify the accuracy of the voltage-reactive multivariate linear regression model determined by the coefficient matrix W;

[0125] Selecting a static reactive power compensation device as a reactive power regulating device, and establishing a reactive-voltage control optimization model with the minimum sum of voltage deviation squares of each node in the power system as the target;

[0126] Solving the reactive-voltage control optimization model based on MATLAB calling CPLEX solver to obtain the reactive power scheduling strategy of each node at different times, and realizing the reactive voltage control optimization.

[0127] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0128] Based on the case30 system in the power flow calculation software MATPOWER, the reactive power and load of each node at 24 time points in a day are given, and the reactive power voltage control optimization method is used for optimization, and the objective function before and after optimization at each time point is obtained as shown in the table. Figure 5 The table shows that the objective function value before optimization is f before, and the objective function value after optimization at each time point is f after. The objective function after optimization is obviously reduced. Figure 6 The table shows that the objective function value before optimization is f before, and the objective function value after optimization at each time point is f after. The objective function after optimization is obviously reduced. Figure 7 The table shows that the objective function value before optimization is f before, and the objective function value after optimization at each time point is f after. The objective function after optimization is obviously reduced.

[0129] In summary, the power system reactive power voltage control optimization method and system based on data driving has the following advantages:

[0130] First, the data-driven method is used to estimate the unknown parameters in the voltage-reactive power multivariate linear model by using the historical operation data of the power system. Compared with the traditional method relying on the physical model of the power system, the method proposed in the present application is based on the multivariate linear model and does not rely on the physical model, which can better cope with the diversity of power system load, the randomness and uncertainty of distributed renewable energy, and is more suitable for the new type of power system with distributed renewable energy access.

[0131] Second, when the reactive power flow of the power system changes, the traditional Newton-Raphson method is used to calculate the voltage values of each node in the system. Although the result is more accurate, the Newton-Raphson method takes a long time to calculate the power flow. The voltage-reactive power multi-linear regression model based on sensitivity is used to calculate the voltage values of each node, which is fast in calculation and has a relatively small error in the calculation result, and can be well matched with the Newton-Raphson method. At the same time, compared with linear power flow calculation, the linear model based on sensitivity is faster in calculation and more accurate in result.

[0132] Third, the static var compensator (SVC) is selected as the reactive power regulation device. The SVC is a device composed of capacitors and various reactance elements connected in parallel with the system and supplying or absorbing reactive power from the system. The traditional reactive power regulation capacity using switched capacitors for voltage regulation is a discrete value, which can only achieve step regulation, and the number of switching operations of the capacitors is limited. The SVC for voltage regulation not only has fast response speed, but also can realize continuous two-way regulation of reactive power, and has small impact on the system during regulation and no limit on the number of operations.

[0133] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0134] The present application is described with reference to flowcharts and / or block diagrams according to the method, device (system), and computer program product of the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The functions specified in one flow or multiple flows and / or blocks Figure 1 The functions specified in one flow or multiple flows and / or blocks

[0135] These computer program instructions can also be stored in a computer readable storage medium that can guide the computer or other programmable data processing devices to work in a specific way, so that the instructions stored in the computer readable storage medium produce a manufactured product including instruction devices that implement the functions specified in the flowcharts and / or block diagrams. Figure 1one or more processes and / or blocks Figure 1 the function specified in the one or more blocks.

[0136] These computer program instructions can also be loaded into computer or other programmable data processing devices, so that a series of operational steps are performed on the computer or other programmable data processing devices to generate a computer-implemented process, so that the instructions executed on the computer or other programmable data processing devices provide a process for implementing the flow Figure 1 one or more processes and / or blocks Figure 1 the function specified in the one or more blocks.

[0137] The above is only to illustrate the technical idea of the present application, and cannot limit the protection scope of the present application. Any modification made according to the technical idea of the present application on the basis of the technical scheme falls within the protection scope of the claims of the present application.

Claims

1. A method for optimizing reactive voltage control in a power system, characterized by, The method comprises the following steps: a linear regression model of node voltage and reactive power injection of each node of the system is established based on power system flow equation, namely, a voltage-reactive power multivariate linear regression model; voltage values of each node under different operation scenarios are recorded to form training data; Solving a voltage-reactive power multivariate linear regression model by using the training data to obtain a coefficient matrix W, Ψ is a voltage-reactive power sensitivity matrix, is a constant vector; the sensitivity matrix Ψ and the constant vector in the voltage-reactive power multivariate linear regression model are determined by a least square-based linear regression method Specifically: for single node voltage data y and reactive power injection data X of each node, parameters ω of corresponding rows of coefficient matrix W are obtained through a linear regression method based on least squares, and parameters obtained through regression of multiple nodes are combined to form coefficient matrix W; Determine the objective function of the least squares based linear regression method, when X T When X is a full rank matrix or a positive definite matrix, directly use the first order condition to solve the parameters in the voltage-reactive power multivariate linear regression model, that is, ω = (X T X) -1 X T y; When X T X does not satisfy the full rank matrix or positive definite matrix condition, add a penalty term, and use the gradient descent method to calculate the parameters in the voltage-reactive power multivariate linear regression model; a power flow calculation is performed to verify accuracy of the voltage-reactive power multivariate linear regression model determined by the obtained coefficient matrix W; a static reactive power compensation device is selected as a reactive power regulation device, a voltage-reactive power multivariate linear regression model is embedded into optimization, and a reactive power-voltage control optimization model is established with the minimum sum of voltage deviation squares of each node of the power system as a target; the reactive power-voltage control optimization model is solved to obtain reactive power scheduling strategies of each node at different time, and reactive power-voltage control optimization is realized.

2. The power system reactive voltage control optimization method of claim 1, wherein, The voltage-reactive power multivariate linear regression model is specifically: Wherein, V(t) is the voltage vector of the power system node at time t; Q(t) is the reactive power vector injected by the system node at time t; Ψ(t) is the voltage-reactive sensitivity matrix at time t; is the constant vector at time t; and ε is the error vector.

3. The power system reactive voltage control optimization method of claim 1, wherein, The voltage values of each node under different operation scenarios are recorded to form training data, specifically: random loads are generated by using a power flow calculation tool, operation scenarios are generated according to the loads, node voltage values under different operation scenarios are obtained through power flow calculation, and reactive power loads of each node, reactive power outputs of generators and voltage values of each node under different operation scenarios are recorded to form training data.

4. The power system reactive voltage control optimization method of claim 1, wherein, The objective function Y(ω) of the linear regression method based on least squares is: where y ω , y are the predicted and true values, respectively, and X is the matrix of samples of the independent variables.

5. The power system reactive voltage control optimization method of claim 1, wherein, The power flow calculation is performed to verify accuracy of the voltage-reactive power multivariate linear regression model determined by the obtained coefficient matrix W, specifically: the accuracy of the voltage-reactive power multivariate linear regression model is verified by comparing the voltage calculated by using the voltage-reactive power multivariate linear regression model with the voltage calculated by using a classic Newton-Raphson method power flow calculation, and by comparing the voltage calculated by using linear power flow calculation with the voltage calculated by using the classic Newton-Raphson method power flow calculation.

6. The power system reactive voltage control optimization method of claim 1, wherein, The reactive power-voltage control optimization model is: where v before (t) is the voltage vector before optimization, Ψ(t) is the voltage-reactive power sensitivity matrix at time t, Δq(t) is the reactive power dispatch vector, v ref (t) is the node voltage reference vector, Δ q (t) is the voltage vector after optimization, Δq(t) is the reactive power dispatch vector, and v are the lower and upper reactive power dispatch vectors, respectively, v (t) is the node voltage reference vector, and Δ are the lower and upper node voltage vectors, respectively.

7. A data-driven based power system reactive voltage control optimization system, characterized in that, The method is applied to any one of the power system reactive power-voltage control optimization methods in claims 1 to 6, and comprises: a power flow module for establishing a voltage-reactive power multivariate linear regression model of node voltage and reactive power injection of each node of the system based on power system flow equation; a data module for recording voltage values of each node under different operation scenarios to form training data; The computing module is configured to obtain a coefficient matrix W by using a voltage-reactive power multi-element linear regression model established by the power flow module according to the training data of the data module, Ψ is a voltage-reactive power sensitivity matrix, is a constant vector; a verification module for performing power flow calculation to verify accuracy of the voltage-reactive power multivariate linear regression model determined by the coefficient matrix W obtained by the calculation module; a selection module for selecting a static reactive power compensation device as a reactive power regulation device, establishing a reactive power-voltage control optimization model with the minimum sum of voltage deviation squares of each node of the power system as a target; an optimization module for solving the reactive power-voltage control optimization model established by the selection module to obtain reactive power scheduling strategies of each node at different time, and realizing reactive power-voltage control optimization.

8. A computer-readable storage medium storing one or more programs, the one or more programs comprising instructions that when executed by a computer cause the computer to perform a method of any of claims 1-7. The one or more programs, when executed by a computing device, cause the computing device to perform any of the power system reactive voltage control optimization methods as claimed in claims 1-6.

9. A computing device, comprising: comprise: one or more processors, memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing any of the power system reactive voltage control optimization methods as claimed in claims 1-6.

Images