Converter adaptive hybrid synchronous control method based on steady-state margin prediction

Abstract

The invention discloses a converter adaptive hybrid synchronous control method based on steady-state margin prediction, and belongs to the technical field of new energy grid connection, and the method comprises the steps: constructing a network construction type VSC converter adaptive hybrid synchronous control dq impedance model; a steady-state margin prediction model based on DNN is constructed; s2, collecting training data, and training the steady-state margin prediction model in the step S2 by using the collected training data to obtain a trained steady-state margin prediction model; based on the steady-state margin prediction model in the step S3, solving the optimal solution of the adaptive coefficient ka by using an interior point method under the condition of giving the steady-state margin; and based on the steady-state margin prediction model in the step S3 and in combination with the optimization algorithm in the step S4, obtaining an optimal solution of the adaptive coefficient ka for parameter adjustment, and realizing adaptive small-interference stable control of the converter.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy grid connection, and in particular to a method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction. Background Art

[0002] As the interface between renewable energy and the power grid, the stability of the converter is directly related to the reliability of the power grid operation. In the current renewable energy grid-connected control technology, the converter control strategy can be roughly divided into two types: grid-following control and grid-building control. The main difference between the two is the different ways of achieving grid voltage synchronization. Among them, the grid-following converter tracks the grid information through the phase-locked link; the grid-building control realizes the synchronization function through the power synchronization link that simulates the characteristics of the synchronous generator. The difference in synchronization methods will lead to significant differences in the oscillation instability characteristics of the two control strategies. Specifically, the grid-following inverter will become unstable under weak grid conditions because the phase-locked loop cannot normally track the grid information; the grid-building inverter is not easy to synchronize under strong grid conditions due to the lack of damping; in addition, the stability performance of the two also shows dual characteristics in terms of renewable energy output. Therefore, under the condition of rapid and large fluctuations in renewable energy and power load, it is difficult to maintain the stability of the grid-converter system using a single control strategy, and the power system urgently needs a new converter control strategy.

[0003] Adaptive hybrid synchronous control is a control strategy that integrates a phase-locked loop into the power synchronization link. It has the characteristics of both grid-following inverters and grid-building inverters, and can be adjusted by adjusting the adaptive coefficient k a To adjust the output ratio. However, the existing adaptive hybrid synchronous control only adjusts parameters according to the short-circuit ratio, while ignoring the stability impact caused by the fluctuation of the new energy itself; on the other hand, the existing short-circuit ratio parameter adjustment method only gives the linear change of the adaptive coefficient with the short-circuit ratio, while the actual system steady-state margin has nonlinear characteristics with the short-circuit ratio, and is related to the composition of each control link of the grid-type VSC converter. Therefore, the original parameter adjustment method cannot adapt well to the actual steady-state margin change relationship, and is not universal. In addition, this parameter adjustment method can only achieve a qualitative improvement in the stability of the system with small disturbances, and cannot accurately and quantitatively control the steady-state margin. In fact, despite adjusting the adaptive coefficient k a It will improve the small disturbance stability of the grid-converter system, but it will also reduce the proportional coefficient of the power loop and thus reduce the active power dynamic tracking performance of the converter. A higher adaptive coefficient will also cause monotonic instability in the grid-type converter under weak grid conditions.

[0004] Therefore, how to provide an adjustable adaptive coefficient tuning method that has universality and high accuracy is a technical problem that needs to be urgently solved by technical personnel in this field. Summary of the invention

[0005] To this end, the present invention provides a converter adaptive hybrid synchronous control method based on steady-state margin prediction to solve the problems in the prior art.

[0006] In order to achieve the above object, the present invention provides the following technical solutions: The method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction comprises the following steps: Step S1: constructing a dq impedance model for adaptive hybrid synchronous control of a grid-connected VSC converter; Step S2: constructing a steady-state margin prediction model based on deep neural networks (DNN); Step S3: Collect training data, and calculate the steady-state margin of the converter under the working condition corresponding to the training data in combination with the dq impedance model established in step S1, which is used as the output label of the steady-state margin prediction model in step S2. The steady-state margin prediction model in step S2 is trained using the collected training data to obtain a trained steady-state margin prediction model; Step S4: Based on the steady-state margin prediction model in step S3, the adaptive coefficient is solved using the interior point method under the condition of a given steady-state margin. k a The optimal solution of Step S5: Based on the steady-state margin prediction model in step S3, the short-circuit ratio and active power reference values ​​of the real-time input grid are used to generate the dynamic prediction result of the steady-state margin of the converter, and the adaptive coefficient is obtained by combining the optimization method of step S4. k a The optimal solution is used to adjust the parameters and realize adaptive small disturbance stable control of the converter.

[0007] Furthermore, in step S1, the method for constructing a dq impedance model for adaptive hybrid synchronous control of a grid-connected VSC converter specifically includes the following steps: Step S101: constructing a physical model of a converter for adaptive hybrid synchronous control; Step S102: Based on the dynamic process of the phase-locked loop and the power synchronization link, a small signal dq impedance modeling method is used according to the physical model to construct small signal equations of the hybrid synchronization link, the reactive power-voltage droop control link, the voltage control link, the current control link, the coordinate transformation link and the physical circuit; Step S103: The small signal equations established in step S102 are simultaneously eliminated to obtain the converter dq impedance transfer function.

[0008] Furthermore, in step S2, a method for constructing a DNN steady-state margin prediction model specifically includes the following steps: Step S201: Analyze the key factors affecting the small disturbance stability of the power grid-converter system, select the key features affecting the small disturbance stability of the power grid-converter system, and use feature engineering to construct the input label X of the steady-state margin prediction model. I ; Step S202: considering the number of input labels and output labels, determining the DNN structure of the steady-state margin prediction model, and selecting a loss function L to optimize the weights and biases of the steady-state margin prediction model in step S3; Step S203: Use the steady-state margin solved in step S1 as the output label Y I , combined with the input label X in S201 I Construct sample data set D I ={X I , Y I}, used for regression training of the steady-state margin prediction model in step S3.

[0009] Furthermore, the input label X of the steady-state margin prediction model in step S201 is I It is expressed as: ; in, k aI is the converter adaptation coefficient in the sample data, P refI is the reference value of the converter operating condition characteristic active power in the sample data, SCR I is the grid side short-circuit ratio in the sample data.

[0010] Furthermore, the DNN structure of the steady-state margin prediction model in step S202 is a four-layer structure, specifically including an input layer, an output layer and two hidden layers, wherein the activation function type of the hidden layer neurons is selected as Sigmoid function, and the activation function of the output layer neurons is a linear function. The output of each layer of neurons is expressed by the following formula: ; ; Among them, w and b represent weight and bias respectively; n Represents the number of layers where the neuron is located, and stipulates that the input layer is the 0th layer; p represent n -1st floor p neurons; q represent n Layer q neurons;k represent n -1 layer corresponds to the number of neurons; m represent n The layer corresponds to the number of neurons; h q Represents n The hidden layer q The output of a neuron; w p,q Represents n -1st floor p neurons and n Layer q The weights between neurons; x p yes n -1st floor p The output of a neuron; b q yes n -1 layer bias node and n layer q The bias between neurons; f (X I ) represents the output of the output layer neurons.

[0011] Furthermore, the loss function L in step S202 is expressed as: ; Among them, MSE is the mean square error, N is the number of samples, I is the Ith sample data, and Y I is the steady-state margin solved in step S1, f (X I ) represents the output of the output layer neurons.

[0012] Furthermore, the output label Y in step S203 I It is expressed as: ; Among them, GM I is the steady-state margin in the sample data.

[0013] Furthermore, in step S4, the adaptive coefficient k is solved a The optimal solution method specifically includes the following steps: Step S401: Determine the steady-state margin reference value GM of the grid-converter system ref , design optimization objective function F(k a ) is the steady-state margin prediction model prediction value and the steady-state margin reference value GM in step S2 ref The square difference is calculated as: ; in, f (X I ) represents the output of the output layer neuron, specifically the predicted value of the steady-state margin prediction model in step S2, X I is the corresponding input label; Step S402: Setting the inverter adaptive coefficient k a The optimal initial value k a0 , and set upper and lower thresholds, which correspond to the inequality constraints of the optimal problem; Step S403: Set the number of iterations and the objective function threshold, and use the interior point method to solve the converter adaptive coefficient k a The optimal solution under this condition.

[0014] Furthermore, the training data in step S3 includes the short-circuit ratio of the grid side, the reference value of active power, the dq-axis component of the PCC point voltage, the dq-axis component of the PCC point current, the dq-axis component of the converter arm voltage and the dq-axis component of the converter arm current.

[0015] The present invention has the following advantages: 1. The grid-type VSC converter adaptive hybrid synchronous control dq impedance model constructed by the present invention takes into account the dynamic influence of the phase-locked loop and the power synchronization link on the control angle θ, derives the transformation equations of the electrical quantities in the system coordinate system and the control coordinate system, obtains the accurate full-order impedance transfer function expression, and improves the calculation accuracy of the steady-state margin.

[0016] 2. The DNN-based steady-state margin prediction model constructed in the present invention takes into account the impact of changes in the output of new energy stations on the steady-state margin on the basis of considering changes in the short-circuit ratio. The trained steady-state margin prediction model can reflect the nonlinear characteristics between the system steady-state margin and multiple operating conditions of the converter, and can realize fast and accurate steady-state margin estimation under a wide range of converter operating conditions.

[0017] 3. The adaptive hybrid synchronous control method constructed by the present invention can solve the adaptive coefficient k based on the existing steady-state margin prediction model. a The optimal solution method can achieve quantitative control of the steady-state margin, and can improve small disturbance stability or active power dynamic tracking performance according to system needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0019] The structures, proportions, sizes, etc. illustrated in this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with the technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantial technical significance. Any structural modification, change in proportion or adjustment of size shall still fall within the scope of the technical contents disclosed in the present invention without affecting the effects and purposes that can be achieved by the present invention.

[0020] Figure 1 It is a flow chart of the adaptive hybrid synchronous control method of the grid-type VSC converter based on DNN steady-state margin prediction of the present invention; Figure 2 The topology structure of the LC grid-connected converter and the block diagram of the adaptive hybrid synchronous control implemented in the present invention; Figure 3 A full-order small signal model diagram of the adaptive hybrid synchronous control of the grid-connected VSC converter constructed by the present invention; Figure 4 A structural diagram of the DNN-based steady-state margin prediction model constructed by the present invention; Figure 5 A schematic diagram of the training process of the DNN-based steady-state margin prediction model constructed in the present invention; Figure 6 A schematic diagram of the optimization parameter adjustment based on the DNN steady-state margin prediction model constructed in the present invention. DETAILED DESCRIPTION

[0021] The following is a description of the implementation of the present invention by specific embodiments. People familiar with the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] like Figure 1 As shown, the adaptive hybrid synchronization control method of the grid-type VSC converter based on DNN steady-state margin prediction includes the following steps: Step S1: constructing a grid-type VSC converter adaptive hybrid synchronous control dq impedance model, the purpose of which is to calculate the steady-state margin of the grid-converter system by combining grid-side impedance information and using the generalized Nyquist theorem under different converter conditions; Step S2: constructing a steady-state margin prediction model based on deep neural networks (DNN) to achieve steady-state margin prediction of the converter under multiple working conditions; Step S3: Collect training data, and calculate the steady-state margin of the converter under the working condition corresponding to the training data in combination with the dq impedance model established in step S1, which is used as the output label of the steady-state margin prediction model in step S2. The steady-state margin prediction model in step S2 is trained using the collected training data to obtain a trained steady-state margin prediction model; Step S4: Based on the steady-state margin prediction model in step S3, the adaptive coefficient is solved using the interior point method under the condition of a given steady-state margin. k a The optimal solution of Step S5: Based on the steady-state margin prediction model in step S3, the short-circuit ratio and active power reference values ​​of the real-time input grid are used to generate the dynamic prediction result of the steady-state margin of the converter, and the adaptive coefficient is obtained by combining the optimization algorithm in step S4. k a The optimal solution is used to adjust the parameters and realize adaptive small disturbance stable control of the converter.

[0023] like Figure 2 The LC grid-connected converter topology and adaptive hybrid synchronous control block diagram are shown in the figure. In the topology, the converter is connected to the inverter through a filter inductor. L f and filter capacitors C f Connected to the grid, which is powered by an infinite power voltage source V g and equivalent inductance L g The adaptive hybrid synchronous control structure consists of four parts: hybrid synchronous loop, reactive power-voltage droop control loop, voltage control loop, and current control loop. Figure 2 It can be seen that the hybrid synchronization loop is composed of a phase-locked loop and a power synchronization link. The converter phase angle signal is obtained by summing the power feedback signal at the PCC point and the q-axis voltage feedback signal: ; In the formula, k a is the adaptive coefficient; k p and k PLL They are respectively the active power-frequency of the power synchronization link controller ( P - f ) proportional coefficient and proportional coefficient of the phase-locked loop controller; k ip is the integral control coefficient of the power synchronization link controller; v oq is the PCC voltage measurement value q Axis component; ω 0 is the rated angular frequency of the power grid; s represents the variable in the system coordinate system; P*is the reference value of the converter active power; P LPF is the measured value of active power after being processed by a first-order low-pass filter, and its calculation formula is: ; In the formula, P is the measured value of the converter active power; ω LPF is the cutoff frequency of a first-order low-pass filter.

[0024] The reactive power-voltage droop control loop uses droop control to generate PCC points d Shaft voltage reference value v od * : ; In the formula, k q is the droop coefficient of the reactive-voltage link controller; Q* It is the reactive power of the converter after being processed by the first-order low-pass filter LPF Q LPF Reference value of v 0 is the preset amplitude of the PCC point voltage.

[0025] In the voltage control loop, the q-axis voltage reference v oq * Generally, it is set to 0. The voltage reference value is processed by the voltage control loop and the current control loop to generate the bridge arm voltage reference value. v cd * and v cq * : ; ; Where: i cd *、 i cq *、 i cd , i cq The bridge arm currents are d Axis components and q Reference and measured values ​​of axis components; v od , v oq They are the PCC voltage measurements. d Axis components and q Axis component; kpv and k iv are the proportional coefficient and integral coefficient of the voltage loop; k pi and k ii are the proportional and integral coefficients of the current loop.

[0026] Compared with the general hybrid synchronization control structure, this control strategy introduces an adaptive coefficient in the hybrid synchronization link. k a To adjust the output ratio of the phase-locked loop link and the power synchronization link, so as to achieve the purpose of influencing the steady-state margin of the system, and the total output value of the hybrid synchronization link is 1.

[0027] In step S1, the method for constructing a dq impedance model of an adaptive hybrid synchronous control of a grid-connected VSC converter specifically includes the following steps: Step S101: constructing a physical model of a converter for adaptive hybrid synchronous control; Step S102: Based on the dynamic process of the phase-locked loop and the power synchronization link, a small signal dq impedance modeling method is used according to the physical model to construct small signal equations of the hybrid synchronization link, the reactive power-voltage droop control link, the voltage control link, the current control link, the coordinate transformation link and the physical circuit; Step S103: The small signal equations established in step S102 are simultaneously eliminated to obtain the converter dq impedance transfer function.

[0028] like Figure 3 The full-order small signal model of the grid-type VSC converter adaptive hybrid synchronous control constructed in step S1 shown in the figure, according to the impedance definition, the impedance transfer function can be expressed as: ; Among them, Z vsc ( s ) represents the full-order small signal model of adaptive hybrid synchronous control of grid-type VSC converter. The specific expressions of the transfer function matrices corresponding to the model are as follows: ; Among them, Z g , Z f , Y c Derived from the physical circuit small signal equation, they represent the grid equivalent inductance impedance matrix, filter inductance impedance matrix and filter capacitor admittance matrix respectively; G vo-vo , G vo-io , G io-vo , G io-io , G ic-vo , G ic-io , G vc-vo , Gvc-io It is derived from the small signal equations of the hybrid synchronous control link and the coordinate transformation link, and represents the PCC point voltage transformation voltage component matrix, PCC point voltage transformation current component matrix, grid side current transformation voltage component matrix, grid side current transformation current component matrix, bridge arm current transformation voltage component matrix, bridge arm current transformation current component matrix, bridge arm voltage transformation voltage component matrix, bridge arm voltage transformation current component matrix; G cv , G dev Derived from the small signal equation of the voltage control link, they represent the voltage loop PI control matrix and the filter capacitor coupling matrix respectively; G ci , G dei Derived from the small signal equation of the current control link, they represent the current loop PI control matrix and the filter inductor coupling matrix respectively; G QI , G QV Derived from the small signal equation of reactive power-voltage droop control link, they represent reactive power droop current component matrix and reactive power droop voltage component matrix respectively; I od , I oq They are the steady-state values ​​of the bridge arm current d Axis components and q Axis component; I cd , I cq They are the steady-state values ​​of the PCC point current. d Axis components and q Axis component; V cd , V cq They are the steady-state values ​​of the bridge arm voltage d Axis components and q Axis component; V od , V oq They are the steady-state values ​​of the voltage at the PCC point d Axis components and q axis component; in addition, s represents the variable in the system coordinate system, and the symbol "Δ" represents the small signal component. i o is the preset amplitude of the grid-side current, i od , i oq The grid-side current is d Axis components and q Axis component.

[0029] Then, the grid-side impedance model can be expressed as: ; Moreover, according to the generalized Nyquist stability criterion, the stability of the grid-converter system can be predicted by the characteristic locus. The characteristic root locus is the characteristic value locus of the impedance ratio parameterized as a frequency function and can be expressed as: ; In the formula, I_2 represents the 2×2 identity matrix, λ It's Z vsc ( s ) and Z g ( s ) and det is the matrix determinant.

[0030] Moreover, the steady-state margin of the system can be expressed by the amplitude margin on the Nyquist diagram, and its specific expression is: ; Where GM is the steady-state margin of the grid-converter system, ω c is the crossing frequency corresponding to the characteristic root locus, and j is an imaginary unit.

[0031] In step S2, a method for constructing a DNN steady-state margin prediction model is constructed, which specifically includes the following steps: Step S201: Analyze the key factors affecting the small disturbance stability of the power grid-converter system, select the key features affecting the small disturbance stability of the power grid-converter system, and use feature engineering to construct the input label X of the steady-state margin prediction model. I ; Step S202: considering the number of input labels and output labels, determining the DNN structure of the steady-state margin prediction model, and selecting a suitable loss function L for optimizing the weights and biases of the steady-state margin prediction model in step S3; Step S203: Use the steady-state margin solved in step S1 as the output label Y I , combined with the input label X in S201 I Construct sample data set D I ={X I , Y I}, used for regression training of the steady-state margin prediction model in step S3.

[0032] In order to predict the steady-state margin, it is necessary to screen the input labels that affect the system steady-state margin. These features mainly include the output of new energy, the short-circuit ratio of the grid side ( SCR ) and the adaptive coefficient ( k a ) three indicators. Considering that new energy units usually only provide active power to the grid, the output of new energy can be calculated by the reference value of active power P refInstead, input the sample set X of labels I It can be expressed as: ; in, k aI is the converter adaptive coefficient in the sample data, P refI is the reference value of the converter operating condition characteristic active power in the sample data, SCR I is the grid side short-circuit ratio in the sample data.

[0033] The structure of the steady-state margin prediction model based on DNN constructed in step S2 is as follows: Figure 4 The DNN model is a high-order nonlinear model with three inputs and one output. It has a four-layer structure, consisting of an input layer X1-X3, an output layer Y1, and two hidden layers M1-M 11 and N1-N 11 The number of neurons in each layer is 3-11-11-1. The activation function type of the hidden layer neurons is Sigmoid function, while the activation function of the output layer neurons is a linear function. The output of each layer of neurons can be expressed by the following formula: ; ; Among them, w and b represent weight and bias respectively; n Represents the number of layers where the neuron is located, and stipulates that the input layer is the 0th layer; p represent n -1st floor p neurons; q represent n Layer q neurons; k represent n -1 layer corresponds to the number of neurons; m represent n The layer corresponds to the number of neurons; h q Represents n The hidden layer q The output of a neuron; w p,q Represents n -1st floor p neurons and n Layer q The weights between neurons; x p yes n -1st floor p The output of a neuron; b q yesn -1 layer bias node and n layer q The bias between neurons; f (X I ) represents the output of the output layer neurons.

[0034] The steady-state margin prediction model is trained using the back-propagation algorithm and the Levenberg-Marquardt method, and the mean square error (MSE) is used as the loss function for model training. L , the specific expression is: ; In the formula, N represents the number of samples, I represents the Ith sample data, and Y I The steady-state margin calculated by the dq impedance model constructed by S1 can be expressed as: ; Among them, GM I is the steady-state margin in the sample data.

[0035] In step S3, the training data includes the short-circuit ratio of the grid side, the reference value of active power, the dq-axis component of the PCC point voltage, the dq-axis component of the PCC point current, the dq-axis component of the converter bridge arm voltage, and the dq-axis component of the converter bridge arm current. The process of collecting training data can be described as follows: Using MATLAB / SINMULINK software to build Figure 2 The simulation model shown in the figure is combined with the MATLAB script to call MATLAB / SINMULINK for batch simulation. Set the active power reference value P ref By 0pu-1pu, short circuit ratio SCR From 2-20, adaptive coefficient k a The value changes evenly from 0 to 0.9, and a total of 3200 sets of data are collected, and the training set, test set, and validation set are divided into training set, test set, and validation set according to the ratio of 3:1:1. The output label Y of the steady-state margin prediction model I During training, the steady-state value of each operating point and the grid impedance calculation are imported into the dq impedance model constructed in step S1. The trained steady-state margin prediction model can be used for online prediction of the steady-state margin.

[0036] The stability margin model training process is as follows: Figure 5 As shown, the signal data set [P ref , SCR , k a ], combining the dq impedance model of step S1 and the grid impedance to obtain the stability margin data set [P ref , SCR , ka ,GM], and after normalization preprocessing, the training data set [P ref , SCR , k a ] → GM train And the test data set [P ref , SCR , k a ] → GM test , train and test the neural network according to the training data set and the test data set, adjust the number of model layers and nodes, and obtain the mean square error MSE that reflects the accuracy and prediction ability of the model. If the MSE is less than the set threshold MSE limit , then continue to process the neural network further, otherwise, retrain the neural network.

[0037] In step S4, the adaptive coefficients are solved k a The optimal solution process is as follows Figure 6 As shown, according to the above training results, the neural network is further processed, the neural network is denormalized, the steady-state margin is evaluated, and the steady-state margin reference value GM is set ref and optimal initial value k a0 , the interior point method is used for optimization. When the iteration process reaches the threshold of the objective function or the corresponding number of iterations, the iteration will be stopped and the adaptation coefficient will be obtained. k a The optimal solution, otherwise continue to adjust the adaptation coefficient k a Iterate the process, input the characteristic measurement signal and k a Take the value and perform normalization preprocessing on the neural network.

[0038] In step S4, the adaptive coefficient is solved k a The optimal solution method specifically includes the following steps: Step S401: Determine the steady-state margin reference value GM of the grid-converter system ref , design optimization objective function F( k a ) is the steady-state margin prediction model prediction value and the steady-state margin reference value GM in step S2 ref The square difference is calculated as: ; Among them, f(X I ) represents the predicted value of the steady-state margin prediction model in step S2, X I is the corresponding input label; Step S402: Setting the inverter adaptive coefficient k a Optimal initial value of k a0 , and set upper and lower thresholds, which correspond to the inequality constraints of the optimal problem; Step S403: Set the number of iterations and the objective function threshold, and use the interior point method to solve the converter adaptive coefficient k a The optimal solution under this condition.

[0039] Since the DNN model is essentially a high-order nonlinear function, it is quite difficult to reversely deduce the input label from the output label, and it is necessary to use an iterative method of numerical optimization to solve it. First, define the objective function F ( k a ), the objective function is the predicted value f(X I ) and steady-state margin reference value GM ref In addition, in order to keep the converter from losing control of active power, the adaptive coefficient k a Need to be constrained. The two together constitute the constraints of the optimization problem, which can be specifically expressed as: ; According to the above constraints, the objective function and inequality constraints are transformed into Lagrangian functions, and the interior point method is used for optimization. The specific expression is: ; In the formula, μ It is a barrier parameter that controls the strength of the constraint and gradually decreases as the iteration proceeds so that the solution gradually approaches the corresponding constraint boundary. The update rule of this parameter is: ; In the formula, β is the reduction factor used to gradually reduce the barrier parameter μ , r is a discrete index.

[0040] Optimization direction Δ of the interior point method α It can be determined by the gradient of the Lagrangian function and the Hessian matrix, and its specific expression is: ; Then, the adaptive coefficients are updated by the Newton step size k a , the specific expression is: ; In the formula,α It is the iteration step size, which is automatically solved by the backtracking line search algorithm during the iteration process to ensure that the objective function value continues to decrease until the constraint conditions are met.

[0041] Furthermore, when the iteration process reaches the target function threshold or the corresponding number of iterations, the iteration will stop. k a As variables, and the range of inequality constraints is small, the control system can quickly calculate the values ​​of the gradient and Hessian matrix, which can meet the needs of online computing of the system.

[0042] The steady-state margin prediction model trained in step S5 can calculate the adaptive coefficient based on the input characteristic data of the system k a The optimal solution is adjusted online and the prediction results of the system's steady-state margin are generated in real time.

[0043] Although the present invention has been described in detail above by general description and specific embodiments, it is obvious to those skilled in the art that some modifications or improvements can be made to the present invention. Therefore, these modifications or improvements made without departing from the spirit of the present invention all belong to the scope of protection claimed by the present invention.

Claims

1. A converter adaptive hybrid synchronous control method based on steady-state margin prediction, characterized in that: The following steps are involved: Step S1: constructing a dq impedance model for adaptive hybrid synchronous control of a grid-connected VSC converter; Step S2: construct a steady-state margin prediction model based on DNN; Step S3: Collect training data, and calculate the steady-state margin of the converter under the working condition corresponding to the training data in combination with the dq impedance model established in step S1, which is used as the output label of the steady-state margin prediction model in step S2. The steady-state margin prediction model in step S2 is trained using the collected training data to obtain a trained steady-state margin prediction model; Step S4: Based on the steady-state margin prediction model in step S3, the adaptive coefficient is solved using the interior point method under the condition of a given steady-state margin. k a The optimal solution of Step S5: Based on the steady-state margin prediction model in step S3, the short-circuit ratio and active power reference values ​​of the real-time input grid are used to generate the dynamic prediction result of the steady-state margin of the converter, and the adaptive coefficient is obtained by combining the optimization method of step S4. k a The optimal solution is used to adjust the parameters and realize adaptive small disturbance stable control of the converter.

2. The method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction according to claim 1, characterized in that: In step S1, the method for constructing a dq impedance model of an adaptive hybrid synchronous control of a grid-type VSC converter specifically comprises the following steps: Step S101: constructing a physical model of a converter for adaptive hybrid synchronous control; Step S102: Based on the dynamic process of the phase-locked loop and the power synchronization link, a small signal dq impedance modeling method is used according to the physical model to construct small signal equations of the hybrid synchronization link, the reactive power-voltage droop control link, the voltage control link, the current control link, the coordinate transformation link and the physical circuit; Step S103: The small signal equations established in step S102 are simultaneously eliminated to obtain the converter dq impedance transfer function.

3. The method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction according to claim 1, characterized in that: In step S2, a method for constructing a DNN steady-state margin prediction model specifically includes the following steps: Step S201: Analyze the key factors affecting the small disturbance stability of the power grid-converter system, select the key features affecting the small disturbance stability of the power grid-converter system, and use feature engineering to construct the input label X of the steady-state margin prediction model. I ; Step S202: considering the number of input labels and output labels, determining the DNN structure of the steady-state margin prediction model, and selecting a loss function L to optimize the weights and biases of the steady-state margin prediction model in step S3; Step S203: Use the steady-state margin solved in step S1 as the output label Y I , combined with the input label X in S201 I Construct sample data set D I ={X I , Y I }, used for regression training of the steady-state margin prediction model in step S3.

4. The method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction according to claim 3, characterized in that: The input label X of the steady-state margin prediction model in step S201 is I It is expressed as: ； in, k aI is the converter adaptation coefficient in the sample data, P refI is the reference value of the converter operating condition characteristic active power in the sample data, SCR I is the grid side short-circuit ratio in the sample data.

5. The method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction according to claim 3, characterized in that: The DNN structure of the steady-state margin prediction model in step S202 is a four-layer structure, specifically including an input layer, an output layer and two hidden layers, wherein the activation function type of the hidden layer neurons is selected as Sigmoid function, and the activation function of the output layer neurons is a linear function. The output of each layer of neurons is expressed by the following formula: ； ； Among them, w and b represent weight and bias respectively; n Represents the number of layers where the neuron is located, and stipulates that the input layer is the 0th layer; p represent n -1st floor p neurons; q represent n Layer q neurons; k represent n -1 layer corresponds to the number of neurons; m represent n The layer corresponds to the number of neurons; h q Represents n The hidden layer q The output of a neuron; w p,q Represents n -1st floor p neurons and n Layer q The weights between neurons; x p yes n -1st floor p The output of a neuron; b q yes n -1 layer bias node and n layer q The bias between neurons; f (X I ) represents the output of the output layer neurons.

6. The method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction according to claim 3, characterized in that: The loss function L in step S202 is expressed as: ； Among them, MSE is the mean square error, N is the number of samples, I is the Ith sample data, and Y I is the steady-state margin solved in step S1, f (X I ) represents the output of the output layer neurons.

7. The method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction according to claim 3, characterized in that: The output label Y in step S203 I It is expressed as: ； Among them, GM I is the steady-state margin in the sample data.

8. The method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction according to claim 1, characterized in that: In step S4, the adaptive coefficient is solved k a The optimal solution method specifically includes the following steps: Step S401: Determine the steady-state margin reference value GM of the grid-converter system ref , design optimization objective function F( k a ) is the steady-state margin prediction model prediction value and the steady-state margin reference value GM in step S2 ref The square difference is calculated as: ； in, f (X I ) represents the output of the output layer neuron, specifically the predicted value of the steady-state margin prediction model in step S2, X I is the corresponding input label; Step S402: Setting the inverter adaptive coefficient k a Optimal initial value of k a0 , and set upper and lower thresholds, which correspond to the inequality constraints of the optimal problem; Step S403: Set the number of iterations and the objective function threshold, and use the interior point method to solve the converter adaptive coefficient k a The optimal solution under this condition.

9. The method for adaptive hybrid synchronous control of a converter based on steady-state margin prediction according to claim 1, characterized in that: The training data in step S3 includes the short-circuit ratio of the grid side, the reference value of active power, the dq-axis component of the PCC point voltage, the dq-axis component of the PCC point current, the dq-axis component of the converter arm voltage and the dq-axis component of the converter arm current.

Images