Photovoltaic power generation grid-connected power limiting control method and device and computer equipment

Abstract

The application provides a photovoltaic power generation grid-connected limited power control method and device and computer equipment, and belongs to the technical field of new energy grid-connected power control. The method comprises the following steps: acquiring a load required power prediction value and a photovoltaic power station maximum output power prediction value at the next moment; judging whether the photovoltaic power station maximum output power prediction value at the next moment is greater than the sum of a power loss value generated when the photovoltaic power generation is grid-connected and the load required power prediction value at the next moment; if yes, generating an optimal duty ratio by using an MPC module which has been constructed, so that the power output by a DC / DC converter in the photovoltaic power generation under the control of a pulse modulation signal with the optimal duty ratio tracks a target power value; otherwise, adopting MPPT maximum power point tracking control on the DC / DC converter side. Through the above method, flexible power control during the grid connection of photovoltaic power generation is realized, and the method has the characteristics of fast response speed, strong real-time performance, strong adaptability to variable loads and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of new energy grid-connected power control, in particular to a photovoltaic power generation grid-connected power limiting control method, a photovoltaic power generation grid-connected power limiting control device, a computer device and a machine readable storage medium. BACKGROUND

[0002] Photovoltaic power generation has the characteristics of randomness, intermittency, periodicity and volatility, therefore, large and medium-sized photovoltaic power stations should be equipped with photovoltaic power generation grid-connected power control systems. Photovoltaic power generation system grid-connected power control refers to controlling the power delivered by the photovoltaic power station to the load in a certain way to meet the demand of the load power and track and adjust, the purpose is to make the output power of the photovoltaic power station match the load demand, realize the smooth supply of power and maximize the utilization of photovoltaic power generation.

[0003] At present, the grid-connected limited power control methods of photovoltaic power generation mainly include DC / DC converter PI control and inverter active regulation control. The DC / DC converter PI control ensures stable operation within the limited power range by accurately controlling the output voltage and current of the converter. The control process mainly includes: real-time monitoring of the output voltage and current of the DC / DC converter, calculation of the current output power, comparison of the real-time monitored output power with the set power limit value, and if the current output power exceeds the power limit value, limited power control is needed. By optimizing the control parameters of the PID algorithm, the accuracy and response speed of the limited power control can be further improved. The inverter active regulation control usually changes the phase and amplitude of the output current and voltage of the inverter to achieve the control. By controlling the conduction time of the IGBT in the inverter circuit, the size of the output voltage is changed, so as to control the amplitude of the output active power. It can be achieved by adjusting the control parameters of the inverter or using a specific control algorithm. Some advanced inverters are also equipped with active power and reactive power decoupling control function to more accurately regulate the output active power. The DC / DC converter PI control mainly has the following problems: 1) mainly suitable for linear system, and the processing of constraint condition is relatively weak; 2) poor adaptability to load change, when the load changes, the control effect of the PI controller may be affected, thereby causing the system performance to decline; 3) under the action of integral regulation, if the system error exists for a long time, the integral term may continue to grow, causing integral saturation, which may cause the system response to be excessive, and even cause the system to be unstable; 4) slow dynamic response speed, which needs to be adjusted gradually by integral action. The inverter active regulation control mainly has the following problems: 1) the power regulation of photovoltaic inverter usually involves the adjustment of multiple parameters such as voltage, frequency and switching state, which is relatively complex and slow in response; 2) photovoltaic inverter needs to have good grid adaptability to ensure stable connection with the grid and power quality, which requires photovoltaic inverter to have complex protection function and control strategy, increasing the complexity and cost of the system; 3) when photovoltaic power generation is connected to the auxiliary system of thermal power plant, there are still many problems in the output power and regulation control of the inverter, which need to be improved and optimized.

[0004] In summary, it is urgent to improve the limited power control scheme of photovoltaic power generation grid-connected. SUMMARY

[0005] Therefore, the embodiments of the present application aim to provide a photovoltaic power generation grid-connected limited power control method, a photovoltaic power generation grid-connected limited power control device, a computer device and a machine readable storage medium, to overcome one or more defects in the prior art photovoltaic power generation grid-connected limited power control scheme based on DC / DC converter PI control and inverter active regulation control.

[0006] To achieve the above object, a first aspect of the embodiment of the present application provides a photovoltaic power generation grid-connected limited power control method, which comprises:

[0007] obtaining a predicted value of power required by a load at a next moment and a predicted value of maximum output power of a photovoltaic power station;

[0008] determining whether the predicted value of maximum output power of the photovoltaic power station at the next moment is greater than a sum of a power loss value generated when the photovoltaic power generation is grid-connected and the predicted value of power required by the load at the next moment; if yes, generating an optimal duty cycle by using an MPC model that has been constructed, so that a DC / DC converter in the photovoltaic power station outputs a target power value under the control of a pulse modulation signal with the optimal duty cycle, otherwise, adopting MPPT maximum power point tracking control at the DC / DC converter side, so that the output power of the DC / DC converter tracks the predicted value of maximum output power of the photovoltaic power station at the next moment;

[0009] wherein the target power value is a power value required to be output by the DC / DC converter determined according to a scheduling instruction, and a spatial state equation of the MPC model is established based on a loop in which the photovoltaic power station where the DC / DC converter is located supplies power to the load.

[0010] In the embodiment of the present application, before determining whether the predicted value of maximum output power of the photovoltaic power station at the next moment is greater than the sum of the power loss value generated when the photovoltaic power generation is grid-connected and the predicted value of power required by the load at the next moment, the method further comprises:

[0011] composing a first time sequence from the collected characteristic variables at continuous moments;

[0012] inputting the first time sequence into a power transmission and transformation loss prediction model that has been constructed, to obtain the predicted value of power loss generated when the photovoltaic power generation is grid-connected at the next moment;

[0013] wherein the maximum collection moment of the characteristic variables is the current moment, the characteristic variables include power transmission and transformation state variables that change the power loss of the photovoltaic power generation grid-connected, and the power transmission and transformation loss prediction model is obtained by training and parameter optimization on a sequence prediction network of an initial state.

[0014] In the embodiment of the present application, the characteristic variables further include voltage amplitude fluctuation characteristics and current amplitude fluctuation characteristics of a transformer in the photovoltaic power generation grid-connected line, and the voltage amplitude fluctuation characteristics and the current amplitude fluctuation characteristics at continuous moments are obtained by using a feature extraction network that has been constructed to extract features from a second time sequence composed of collected transformer voltage amplitudes and transformer current amplitudes at continuous moments.

[0015] In the embodiment of the present application, the feature extraction network comprises a 2D convolution layer, a pooling layer and a Flatten layer connected in sequence, wherein:

[0016] a 2D convolutional layer, configured to receive a two-dimensional matrix converted from the second time sequence and perform feature extraction on the two-dimensional matrix;

[0017] a pooling layer, configured to perform a pooling operation on a feature map output after the feature extraction;

[0018] a Flatten layer, configured to convert the feature map output after the pooling operation into a one-dimensional vector.

[0019] In specific embodiments of the present application, the initial state sequence prediction network is trained and parameter-optimized, including:

[0020] a training sample is constructed using historical data of the feature variable;

[0021] the sequence prediction network is pre-trained and hyperparameter pre-optimized using the training sample;

[0022] a real sample is constructed using the real-time collected feature variable;

[0023] the sequence prediction network pre-trained and hyperparameter pre-optimized is incrementally trained and hyperparameter-optimized using the real sample;

[0024] The hyperparameter is optimized by using a particle swarm optimization algorithm.

[0025] In specific embodiments of the present application, the DC / DC converter adopts a boost circuit as a DC / DC boost circuit, and the space state equation is represented as:

[0026]

[0027] wherein x represents a state variable, A represents a state matrix, u represents an input vector, the input vector is a target voltage value output by the DC / DC converter determined according to a scheduling instruction, B represents an input matrix, represents a derivative of the state variable, y represents an output vector, the output vector is a difference between 1 and a duty cycle prediction value of a pulse modulation signal for controlling the DC / DC converter, C represents an output matrix, and D represents a direct transmission matrix. In specific embodiments of the present application, the constraint condition of the MPC model includes: the duty cycle is greater than or equal to 0 and less than or equal to 1; and the target power value is within a preset fluctuation range.

[0028] In specific embodiments of the present application, when the optimal duty cycle is generated by using the constructed MPC model, an optimization algorithm is used to optimize the prediction time domain length of the MPC model.

[0029] In specific embodiments of the present application, the optimization algorithm is used to optimize the prediction time domain length of the MPC model, including:

[0030] initialize a particle swarm;

[0031] calculate fitness values of the particles according to a fitness function;

[0032] perform a particle swarm updating step;

[0033] if the number of iterations or the fitness values of the particles of the particle swarm reach expected values, output the optimal individual in the current particle swarm as an optimal solution of the prediction time domain length, otherwise perform the particle swarm updating step again;

[0034] wherein the particle swarm updating step comprises:

[0035] determine current individual extreme values of the particles and a current global optimal solution of the particle swarm;

[0036] update the velocities and positions of the particles;

[0037] wherein the fitness function is an evaluation function representing the error degree between the output voltage of the DC / DC converter calculated by the MPC model and the target voltage value required by the DC / DC converter determined according to the scheduling instruction.

[0038] In the embodiment of the application, the power transmission and transformation state variables include ambient temperature, load and transformer use time.

[0039] The second aspect of the embodiment of the application provides a photovoltaic power generation grid-connected limited power control device, which comprises:

[0040] an acquisition module, configured to acquire a next-time load required power prediction value and a photovoltaic power station maximum output power prediction value;

[0041] a limited power control module, configured to determine whether the next-time photovoltaic power station maximum output power prediction value is greater than the sum of a power loss value generated when the photovoltaic power generation is grid-connected and the next-time load required power prediction value; if yes, generate an optimal duty ratio by using the MPC model constructed to make the DC / DC converter in the photovoltaic power station output a power tracking target power value under the control of a pulse modulation signal with the optimal duty ratio, otherwise, adopt MPPT maximum power point tracking control on the DC / DC converter side to make the output power of the DC / DC converter track the next-time photovoltaic power station maximum output power prediction value;

[0042] wherein the target power value is a power value required by the DC / DC converter to output, and the space state equation of the MPC model is established based on a loop in which the photovoltaic power station where the DC / DC converter is located supplies power to the load.

[0043] The third aspect of the embodiment of the present application provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the photovoltaic power generation grid-connected limited power control method of the first aspect of the embodiment of the present application when executing the program.

[0044] The fourth aspect of the embodiment of the present application provides a machine readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the photovoltaic power generation grid-connected limited power control method of the first aspect of the embodiment of the present application.

[0045] In the above technical solution, the real-time output power during the photovoltaic power generation grid connection is dynamically adjusted according to the prediction result of the required power of the load and the prediction result of the maximum output power of the photovoltaic power station, including two adjustment strategies, the first one is to adopt MPPT maximum power point tracking control on the DC / DC converter side, so that the output power of the DC / DC converter tracks the predicted value of the maximum output power of the photovoltaic power station at the next time, and the second one is to generate an optimal duty ratio by using the constructed MPC model, so that the output power on the DC / DC converter side tracks the required power. Based on this, the flexible adjustment of the photovoltaic power generation grid-connected power is realized, compared with the traditional limited power control, the maximum utilization of photovoltaic power generation energy is ensured, at the same time, the stable operation of the power grid and the reliability of the power supply are ensured. In addition, based on the MPC control of the DC / DC converter side, the PI control of the DC / DC converter and the active regulation control of the inverter in the traditional limited power control are abandoned, effectively overcoming the defects of weak constraint condition processing ability, poor adaptability to load changes and slow dynamic response speed of the PI control of the DC / DC converter in the traditional limited power control, and overcoming the defects of complex control process, slow response speed, complex control components, high cost and other defects of the active regulation control of the inverter in the traditional limited power control.

[0046] Other features and advantages of the embodiment of the present application will be described in detail in the following specific embodiment part. BRIEF DESCRIPTION OF DRAWINGS

[0047] The accompanying drawings are used to provide a further understanding of the embodiment of the present application, and constitute a part of the specification, and are used together with the following specific embodiment to explain the embodiment of the present application, but do not constitute a limitation on the embodiment of the present application. In the drawings:

[0048] Figure 1 The first flow chart of the photovoltaic power generation grid-connected limited power control method according to the embodiment of the present application is schematically shown;

[0049] Figure 2 The second flow chart of the photovoltaic power generation grid-connected limited power control method according to the embodiment of the present application is schematically shown;

[0050] Figure 3 Fig. 1 schematically shows a first construction process diagram of the power transmission and transformation loss prediction model in the specific application example;

[0051] Figure 4 Fig. 2 schematically shows a second construction process diagram of the power transmission and transformation loss prediction model in the specific application example;

[0052] Figure 5 Fig. 3 schematically shows a power supply loop structure diagram of a photovoltaic power station where the DC / DC converter is located in the specific application example;

[0053] Figure 6 Fig. 4 schematically shows a structure diagram of a photovoltaic power generation system connected to a 6kV thermal power plant in the specific application example;

[0054] Figure 7 Fig. 5 schematically shows a structure block diagram of the photovoltaic power generation grid-connected power limiting control device according to the embodiment of the present application;

[0055] Figure 8 Fig. 6 schematically shows a composition block diagram of the computer device according to the embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to make the objects, 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. It should be understood that the specific embodiments described herein are only used to explain and illustrate the embodiments of the present application, and are not used to limit the 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 work fall within the scope of protection of the present application.

[0057] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative positional relationship, movement condition, etc. between components in a certain specific posture (as shown in the drawings), and if the specific posture changes, the directional indications also change accordingly.

[0058] In addition, if the description of "first", "second" and the like is involved in the embodiments of the present application, the description of "first", "second" and the like is only for the purpose of description, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first", "second" can be explicitly or implicitly included at least one of the features. In addition, the technical solutions of various embodiments can be combined with each other, but it must be based on the realization of ordinary skilled in the art, when the combination of technical solutions appears contradictory or unachievable, it should be considered that the combination of technical solutions does not exist, and is not within the protection scope required by the present application.

[0059] Referring to Figure 1 The photovoltaic power generation grid-connected power limiting control method provided by the embodiments of the present application can include the following steps.

[0060] In step S100, the next time load required power prediction value and the next time photovoltaic power station maximum output power prediction value are obtained. The next time load required power prediction value can be derived from the prediction data of the load required power, and the next time photovoltaic power station maximum output power prediction value can be derived from the prediction data of the photovoltaic power station maximum output power.

[0061] In step S102, it is judged whether the next time photovoltaic power station maximum output power prediction value is greater than the sum of the power loss value generated when the photovoltaic power generation is grid-connected and the next time load required power prediction value; if yes, the optimal duty cycle is generated by using the MPC model constructed, so that the power output by the DC / DC converter in the photovoltaic power station under the control of the pulse modulation signal with the optimal duty cycle tracks the target power value. Otherwise, the MPPT maximum power point tracking control is adopted on the DC / DC converter side, so that the output power of the DC / DC converter tracks the next time photovoltaic power station maximum output power prediction value. The target power value refers to the power value required to be output by the DC / DC converter according to the scheduling instruction, i.e. the demand power value. The spatial state equation adopted by the MPC (Model Predictive Control) model is constructed based on the loop in which the photovoltaic power station where the DC / DC converter is located supplies power to the load.

[0062] It is known that the MPC model can adopt dynamic matrix control (DMC) model, model algorithm control (MAC) model and generalized predictive control (GPC) model, and known improved MPC model, etc. The spatial state equation adopted by the MPC model needs to be set in combination with specific application scenarios, and the present embodiment does not describe this part in detail.

[0063] As in the above embodiment, to realize the control of the size of the power of photovoltaic power generation grid-connected at the next moment, first, the load required power prediction value at the next moment is extracted from the prediction data of the load required power, the photovoltaic power station maximum output power prediction value at the next moment is extracted from the prediction data of the photovoltaic power station maximum output power, and then the dynamic adjustment of the photovoltaic power generation grid-connected output power is carried out in combination with the power loss value generated when the photovoltaic power is grid-connected. At the same time, in order to ensure the response speed, control cost, adaptability to variable load and the like during dynamic adjustment, when the photovoltaic power station maximum output power prediction value at the next moment is greater than the sum of the power loss value generated when the photovoltaic power is grid-connected and the load required power prediction value at the next moment, the output power control of the DC / DC converter side is carried out by using the MPC model constructed in advance, thereby realizing the flexible adjustment of the photovoltaic power generation grid-connected power, ensuring the stable operation of the power grid and the reliability of the power supply, and taking into account the real-time performance, low cost, adaptability to variable load, flexibility and scalability. Among them, the flexibility is reflected in that the dynamic adjustment of the photovoltaic power generation grid-connected output power includes two adjustment strategies. The first adjustment strategy is that when the photovoltaic power station maximum output power prediction value at the next moment is greater than the sum of the power loss value generated when the photovoltaic power is grid-connected and the load required power prediction value at the next moment, the MPC prediction tracking demand power value is used on the DC / DC converter side. The second adjustment strategy is that when the photovoltaic power station maximum output power prediction value at the next moment is less than or equal to the sum of the power loss value generated when the photovoltaic power is grid-connected and the load required power prediction value at the next moment, the MPPT maximum power point tracking control is adopted on the DC / DC converter side. The scalability is reflected in that the MPC model has strong scalability, and the control strategy of the MPC model can be adjusted according to the specific photovoltaic power generation grid-connected scene, such as adjusting the spatial state equation, the control domain length and the prediction domain length and the like.

[0064] The power loss when the photovoltaic power is grid-connected is mainly caused by the transmission and transformation lines and devices, so the power loss value generated when the photovoltaic power is grid-connected is also called the transmission and transformation loss value generated when the photovoltaic power is grid-connected.

[0065] In combination with Figure 2 As shown in the figure, in one specific embodiment of the present application, the transmission and transformation loss value is obtained by dynamic prediction, and accordingly, before judging whether the photovoltaic power station maximum output power prediction value at the next moment is greater than the sum of the power loss value generated when the photovoltaic power is grid-connected and the load required power prediction value at the next moment, the following steps can also be included:

[0066] Step S101, the power loss value generated when the photovoltaic power is grid-connected at the next moment is predicted by using the transmission and transformation prediction model constructed.

[0067] Specifically, step S101 can include:

[0068] The first time sequence is composed of the characteristic variables collected at continuous moments;

[0069] The first time sequence is input into the constructed power transmission and transformation loss prediction model to predict a power loss value generated by the photovoltaic power generation grid connection at the next time;

[0070] The maximum collection time of the characteristic variable is the current time, the characteristic variable includes a power transmission and transformation state variable changing the power transmission and transformation loss, and the power transmission and transformation loss prediction model is obtained through sequence prediction network training and parameter optimization of an initial state.

[0071] Specifically, the photovoltaic power generation grid connection line involves complex voltage boosters, voltage reducers, power transmission lines and the like, the power transmission line loss needs to consider the influence of line material, environmental temperature and heat generated by current through the wire on resistance and the like, the voltage booster and voltage reducer and the like power transformation loss needs to consider copper loss, iron loss and the like, and it can be seen that there are many influence factors leading to power transmission and transformation power loss, and the dimensions covered are wide. The power transmission and transformation state variable changing the power transmission and transformation loss described in the above embodiments should be understood as the principal components of these influence factors extracted through dimension reduction analysis or combined with expert experience. The dimension reduction analysis can use methods such as PCA principal component analysis.

[0072] In a comparative embodiment, the power transmission and transformation loss value is a preset value, and the determination process of the preset value can include the following implementation steps:

[0073] The specific photovoltaic power generation grid connection line is simplified by using, for example, the equivalent impedance method and the like, and a simplified circuit model of the photovoltaic power generation grid connection line is constructed after simplification. In the simplified circuit model, the complex voltage booster, voltage reducer and power transmission line are simplified and calculated into equivalent loss impedance by using equivalent parameters;

[0074] The power transmission and transformation loss value is calculated according to the simplified circuit model.

[0075] It is known that the equivalent impedance method regards the line resistance as a constant value when simplifying the circuit, but in fact the line resistance will change with factors such as temperature, because in most application scenarios, the ambient temperature is difficult to keep constant, thus leading to calculation result error, secondly, the equivalent impedance method mainly calculates the resistive loss in the circuit, ignoring the influence of inductive loss on the total loss of the line, in addition, the transformer power loss is affected by many factors such as voltage level, manufacturing process and material, load, ambient temperature, cooling method and loss type (iron loss and copper loss), and the equivalent for constant equivalent loss impedance will lead to calculation result error. It can be seen that in the above comparative examples, the accuracy of the transmission and transformation loss based on the preset value is limited, and the dynamic transmission and transformation loss prediction based on the transmission and transformation loss prediction model considers the real-time dynamic changes of the photovoltaic power generation grid-connected line, so that the transmission and transformation loss value of the photovoltaic power generation grid-connected power limiting control is more accurate, thereby improving the accuracy of the photovoltaic power generation grid-connected power limiting control. In addition, the dynamic transmission and transformation loss prediction based on the transmission and transformation loss prediction model enhances the flexibility and scalability of the photovoltaic power generation grid-connected power limiting control method realized by the embodiments of the application. Flexibility and scalability are reflected in: the transmission and transformation loss prediction model can be flexibly modified or extended according to different application scenarios and needs, such as introducing more feature variables to improve the prediction accuracy of the transmission and transformation loss prediction model and adapt to specific application scenarios and needs, such as modifying the dimension of the feature variable to adapt to specific application scenarios and needs.

[0076] For example, in a specific embodiment, to ensure the real-time and reliability of the transmission and transformation loss prediction, the sequence prediction network can adopt an LSTM network or the like.

[0077] As an improvement of the above various embodiments, the following embodiments propose an improved transmission and transformation loss prediction model, which introduces voltage amplitude fluctuation features and current amplitude fluctuation features of the transformer in the photovoltaic power generation grid-connected line into the feature variables as the input of the model. The voltage amplitude fluctuation features and the current amplitude fluctuation features need to be obtained through a feature extraction network. Specifically, the process of extracting the voltage amplitude fluctuation features and the current amplitude fluctuation features by the feature extraction network can include:

[0078] The second time sequence is composed of the transformer voltage amplitude and the transformer current amplitude collected at continuous time points;

[0079] The second time sequence is input into the constructed feature extraction network to extract the voltage amplitude fluctuation features and the current amplitude fluctuation features at the continuous time points.

[0080] For example, the feature extraction network can adopt a CNN network. The CNN network includes a 2D convolution layer, a pooling layer and a Flatten layer connected in sequence, wherein:

[0081] 2D convolutional layer, configured to receive the two-dimensional matrix converted from the second time sequence and perform feature extraction on the two-dimensional matrix;

[0082] Pooling layer, configured to perform a pooling operation, such as maximum value pooling, on the feature map output after the feature extraction;

[0083] Flatten layer, configured to convert the feature map output after the pooling operation into a one-dimensional vector.

[0084] Because the voltage amplitude fluctuation and the current amplitude fluctuation of the transformer will cause the change of the iron loss, the copper loss, the heat quantity of the core and the winding of the transformer winding, and further affect the power loss of the transformer, based on this, in the above improved embodiment, the voltage amplitude fluctuation feature and the current amplitude fluctuation feature of the transformer are introduced into the characteristic variable, and the prediction precision of the power transmission and transformation loss is improved.

[0085] The extraction process of the voltage amplitude fluctuation feature and the current amplitude fluctuation feature is described by taking the above CNN network as an example, and specifically as follows:

[0086] 1) The transformer voltage amplitude and the transformer current amplitude at continuous time points are collected by sensors and the like, the transformer voltage amplitude and the transformer current amplitude at continuous time points are composed into a second time sequence, and the second time sequence is converted into a two-dimensional matrix X T×V The CNN network is input, wherein the dimension T represents the time step, and the dimension V represents the transformer voltage amplitude and the transformer current amplitude.

[0087] 2) The spatial features of the two-dimensional matrix X are extracted by the 2D convolutional layer, and the convolution kernel used in the 2D convolutional layer is represented as K k×f , wherein k represents the time window size, f represents the feature window size, and the convolution operation performed at each position (t, v) in the two-dimensional matrix X can be represented as:

[0088]

[0089] In the above formula, Y i [t,v] represents the output feature map obtained by performing the convolution operation on the position (t, v) using the i-th convolution kernel, K i [m,n] represents the element in the m-th row and the n-th column in the i-th convolution kernel, and the dot product symbol in the above formula represents the element multiplication including the bias term and the ReLU activation function.

[0090] 3) In order to reduce the spatial dimension of the data, the output feature map of the 2D convolutional layer is subjected to a maximum pooling operation to obtain an output feature map Z. In the pooling window, for each position (t', v'):

[0091] Z[t', v'] = max {(t, v) e pool window(t', v')} Y[t, v];

[0092] 4) Using the Flatten layer to convert the output feature map Z into a one-dimensional vector, so as to facilitate as an input parameter of the power transmission loss prediction model.

[0093] Exemplarily, in one specific embodiment, the incremental learning method is adopted when training the sequence prediction network of the initial state, and the specific process can include:

[0094] Using the historical data of the feature variables to construct training samples;

[0095] Using the training samples to pre-train and pre-optimize the hyperparameters of the sequence prediction network;

[0096] Using the real-time collected feature variables to construct real samples;

[0097] Using the real samples to incrementally train and optimize the hyperparameters of the pre-trained and pre-optimized sequence prediction network.

[0098] As in the above embodiment, the incremental learning improves the accuracy of the power transmission loss prediction model constructed.

[0099] Exemplarily, in one specific embodiment, when training and optimizing the parameters of the initial state sequence prediction network, the particle swarm optimization algorithm (PSO algorithm) is also used for hyperparameter optimization, and the specific process can include:

[0100] Using the historical data of the feature variables to construct training samples;

[0101] Using the training samples to pre-train and pre-optimize the hyperparameters of the sequence prediction network;

[0102] Using the real-time collected feature variables to construct real samples;

[0103] Using the real samples to incrementally train and optimize the hyperparameters of the pre-trained and pre-optimized sequence prediction network.

[0104] Among them, the particle swarm optimization algorithm is used for hyperparameter optimization.

[0105] For example, the process of using the particle swarm optimization algorithm for hyperparameter optimization is as follows:

[0106] 1) Using the prediction accuracy of the sequence prediction network to design the following evaluation function:

[0107]

[0108] In the above formula, SCORE represents the evaluation function value, TP represents the true positive, TN represents the true negative, FP represents the false positive, and FN represents the false negative.

[0109] 2) Use each particle to represent a group of hyperparameters, evaluate the position of the particle swarm on the validation set using the above evaluation function, and update the speed and position of the particle swarm.

[0110] The particle velocity update formula is: The particle position update formula is: In the above two formulas, ω represents the inertia weight, c1 and c2 represent the learning factor, r1 and r2 represent random numbers, v i represents the speed of the i-th particle, p i is the position of the i-th particle, b i is the best position found so far for the i-th particle, g is the global best position found so far for the group, and H bounds represents the size limit of the hyperparameters, i.e., the boundary limit of the particle position. min max , and clip represents the clip function.

[0111] 3) Repeat the above step to obtain the global optimal solution, i.e., the optimal hyperparameters of the sequence prediction network, and apply it to the pre-training and incremental training of the sequence prediction network.

[0112] As an improvement of the above embodiments, when generating the optimal duty cycle using the constructed MPC model, an optimization algorithm is used to optimize the prediction time domain length of the MPC model. The optimization algorithm can be a genetic optimization algorithm or a particle swarm optimization algorithm.

[0113] As in the above embodiments, the prediction time domain length N in the MPC model is adjusted in real time using an optimization algorithm to fully capture the dynamic characteristics of photovoltaic power generation during grid connection, ensuring that the calculation complexity and prediction error are not increased. Preferably, the iteration process for optimizing the prediction time domain length N is synchronized with the iteration process within the MPC model to better meet the real-time update of the prediction time domain length of the MPC model.

[0114] In one specific embodiment, a particle swarm optimization algorithm is used to optimize the prediction time domain length of the MPC model, which specifically includes the following implementation steps:

[0115] Initialize the particle swarm and set the size, initial position and initial speed of the particle swarm;

[0116] Calculate the fitness value of each particle according to the fitness function;

[0117] Perform the particle swarm update step;

[0118] ​If the number of iterations or the fitness value of the particle swarm reaches the expected value, the optimal individual in the current particle swarm is output as the optimal solution of the prediction time domain length, otherwise the particle swarm updating step is executed again;

[0119] The particle swarm updating step comprises:

[0120] determining the current individual optimal solution of each particle and the current global optimal solution of the particle swarm;

[0121] updating the speed and position of each particle;

[0122] The fitness function is an evaluation function representing the error degree between the output voltage of the DC / DC converter calculated by the MPC model and the target voltage value required by the DC / DC converter determined according to the scheduling instruction.

[0123] In one specific embodiment, a genetic optimization algorithm is used to optimize the prediction time domain length of the MPC model, and the fitness function constructed is the same as the above-mentioned embodiment based on the particle swarm optimization algorithm.

[0124] The process of real-time updating the prediction time domain length is described by taking the particle swarm optimization algorithm for optimizing the prediction time domain length of the MPC model as an example, which can specifically include:

[0125] 1) An evaluation function, i.e. a fitness function, is designed based on the mean square error (MSE) between the output voltage U0' of the DC / DC converter calculated by the MPC model and the target voltage value U0 required by the DC / DC converter determined according to the scheduling instruction, and the evaluation function is expressed as:

[0126]

[0127] In the above formula, score represents the evaluation function value, and n' represents the sample quantity.

[0128] 2) Each particle represents a prediction time domain length, the position of each particle in the particle swarm is evaluated using the evaluation function designed in the above step, and the speed and position of each particle in the particle swarm are updated.

[0129] The formula for updating the particle speed is as follows: The formula for updating the particle position is as follows: In the above two formulas, ω' is the inertia weight, c1', c2' are learning factors, r1', r2' are random numbers, v i ' is the speed of the i-th particle, p i ' is the position of the i-th particle, b iThe best position found so far for the i-th particle, g' is the global best position found so far by the population, the size of the prediction horizon, i.e. the boundary constraint of the particle position N bounds = [N min , N max ].

[0130] 3) Repeat the above step to find the position of the particle with the highest fitness, i.e. the optimal prediction horizon length. Wherein, the iteration process of the particle swarm is synchronized with the iteration process of the MPC model.

[0131] In one specific application of the above embodiment, the characteristic variables only include power transmission and transformation state variables that change the power loss of photovoltaic power generation grid connection, and the power transmission and transformation state variables include environmental temperature, load, and transformer use time length, etc. The LSTM neural network is used to construct the power transmission and transformation loss prediction model, and the particle swarm optimization algorithm is introduced for hyperparameter optimization when the power transmission and transformation prediction model is constructed. Figure 3 The training and prediction process diagram of the power transmission and transformation loss prediction model at this time is shown.

[0132] In another specific application of the above embodiment, the characteristic variables also include transformer voltage amplitude fluctuation features and transformer current amplitude fluctuation features. The CNN network is used to extract the transformer voltage amplitude fluctuation features and the transformer current amplitude fluctuation features. The LSTM neural network is used to construct the power transmission and transformation loss prediction model, and the particle swarm optimization algorithm is introduced for hyperparameter optimization when the power transmission and transformation prediction model is constructed. Figure 4 The training and prediction process diagram of the power transmission and transformation loss prediction model at this time is shown.

[0133] In another specific application of the above embodiment, the photovoltaic power station adopts a Boost circuit as a DC / DC step-up circuit, and the photovoltaic power station is connected to a 6kV thermal power plant system. Figure 5 The structure diagram of the photovoltaic power station where the DC / DC converter is located to supply power to the load is shown. Figure 6 The structure diagram of the photovoltaic power generation connected to the 6kV thermal power plant system is shown. Figure 5 and Figure 6 P W represents the predicted next time power transmission and transformation loss, P C represents the predicted value of the power required by the load at the next time, P PV represents the predicted value of the maximum output power of the photovoltaic power station at the next time, CNN-PSO-LSTM model represents the power transmission and transformation loss prediction model combined with the transformer voltage amplitude fluctuation features extracted based on the CNN network, the transformer current amplitude fluctuation features extracted based on the CNN network, and the hyperparameter optimization based on the particle swarm optimization algorithm in the above embodiment. The power transmission and transformation loss prediction model is constructed based on the LSTM network.

[0134] Combination Figure 5 The photovoltaic power station in which the DC / DC converter shown in the figure supplies power to the load circuit structure diagram, the construction process of the spatial state equation of the MPC model is as follows:

[0135] 1) Establish the spatial state equation of the Boost circuit:

[0136]

[0137] Wherein, x represents the state variable, A represents the state matrix, u represents the input vector, the input vector is the target voltage value output by the DC / DC converter according to the scheduling instruction, B represents the input matrix, The derivative of the state variable is represented by, y represents the output vector, the output vector is the difference between 1 and the duty cycle prediction value of the pulse modulation signal for controlling the DC / DC converter, C represents the output matrix, and D represents the direct transmission matrix.

[0138] 2) Establish the spatial state equation of the MPC model:

[0139]

[0140] Wherein, i(t) and u(t) are state variables, the target voltage value U0 output by the DC / DC converter according to the scheduling instruction is an input, D represents the duty cycle, and D' = 1-D represents the output. In addition, in Figure 5 The circuit structure diagram shown in the figure: r L represents the resistance connected to the positive output end of photovoltaic power generation in the circuit structure, the output voltage of photovoltaic power generation is represented by Upv, and i(t) represents the current flowing through the resistance r L The resistance r L is connected with an inductor L, and the inductor L is connected with the collector of a transistor S and the positive electrode of a diode away from the one end of the resistance r L The negative electrode of the diode is connected with the first end of the resistance r C and the first end of a load R L , the second end of the resistance r C is connected with the first end of a capacitor C, and the second end of the capacitor C and the second end of the load R L are both connected to the emitter of the transistor S, the emitter of the transistor S is connected to the negative output end of photovoltaic power generation, and the voltage between the capacitor C is u(t), and the voltage between the load R L is the output voltage of the DC / DC converter.

[0141] In this specific application, the constraint conditions of the MPC model can include: U min ≤ U0 ≤ U max ; 0 ≤ D' ≤ 1, wherein U min represents the lower limit value of the fluctuation of the target voltage value U0, and Umax represents an upper limit value of fluctuation of the target voltage value U0.

[0142] Based on the above constraint conditions and the establishment of the spatial state equation, an initial MPC model is constructed, and after setting the sampling frequency, adjusting the parameters and selecting appropriate weights, a final MPC model is obtained.

[0143] Based on the pre-construction of the CNN-PSO-LSTM model and the MPC model, the specific process of the specific application of the photovoltaic power generation grid-connected power limiting control includes:

[0144] Step A1, obtaining a load required power prediction value P C and a photovoltaic power station maximum output power prediction value P PV at the next moment;

[0145] Step A2, using the CNN-PSO-LSTM model to predict the power transmission and transformation loss value P W at the next moment;

[0146] Step A3, comparing the load required power prediction value P C , the photovoltaic power station maximum output power prediction value P PV and the power transmission and transformation loss value P W at the next moment, if P W + P C ≥ P PV , then the MPPT maximum power point tracking control is adopted on the DC / DC converter side, so that the output power of the DC / DC converter tracks the photovoltaic power station maximum output power prediction value at the next moment, if P W + P C < P PV , then the MPC control based on real-time optimization in the prediction time domain length is adopted on the DC / DC converter side, so that the output power of the DC / DC converter tracks the target power value, and the target power value refers to the power value required to be output by the DC / DC converter according to the scheduling instruction.

[0147] Figure 7 The structure block diagram of the photovoltaic power generation grid-connected power limiting control device 400 according to the embodiment of the application is schematically shown. Specifically, the photovoltaic power generation grid-connected power limiting control device 400 provided by the embodiment of the application includes an acquisition module 410 and a power limiting control module 420, wherein:

[0148] The acquisition module 410 is used to acquire a load required power prediction value and a photovoltaic power station maximum output power prediction value at the next moment;

[0149] The power limiting control module 420 is configured to determine whether the maximum output power prediction value of the photovoltaic power station at the next moment is greater than the sum of the power loss value generated when the photovoltaic power generation is connected to the grid and the load power prediction value at the next moment; if yes, the optimal duty cycle is generated by using the constructed MPC model, so that the power output by the DC / DC converter in the photovoltaic power station under the control of the pulse modulation signal with the optimal duty cycle tracks the target power value; otherwise, the MPPT maximum power point tracking control is adopted at the DC / DC converter side, so that the output power of the DC / DC converter tracks the maximum output power prediction value of the photovoltaic power station at the next moment.

[0150] The target power value is a power value required to be output by the DC / DC converter determined according to the scheduling instruction, and the spatial state equation of the MPC model is established based on a loop in which the photovoltaic power station where the DC / DC converter is located supplies power to the load.

[0151] In one specific embodiment, the photovoltaic power generation grid-connected power limiting control device 400 further comprises a power transmission and transformation loss prediction module, which is configured to, before determining whether the maximum output power prediction value of the photovoltaic power station at the next moment is greater than the sum of the power loss value generated when the photovoltaic power generation is connected to the grid and the load power prediction value at the next moment, compose a first time sequence from the collected characteristic variables at continuous moments, and input the first time sequence into the constructed power transmission and transformation loss prediction model to predict the power loss value generated when the photovoltaic power generation is connected to the grid at the next moment, wherein the maximum collection moment of the characteristic variables is the current moment, the characteristic variables include power transmission and transformation state variables that change the power loss of the photovoltaic power generation connected to the grid, and the power transmission and transformation loss prediction model is obtained by training and parameter optimization on an initial state sequence prediction network.

[0152] In one specific embodiment, the characteristic variables further include voltage amplitude fluctuation characteristics and current amplitude fluctuation characteristics of a transformer in the photovoltaic power generation grid-connected line, and the voltage amplitude fluctuation characteristics and the current amplitude fluctuation characteristics at continuous moments are obtained by performing feature extraction on a second time sequence by using a constructed feature extraction network, and the second time sequence is composed of the collected transformer voltage amplitudes and transformer current amplitudes at continuous moments.

[0153] In one specific embodiment, the feature extraction network comprises a 2D convolution layer, a pooling layer and a Flatten layer connected in sequence, wherein:

[0154] The 2D convolution layer is configured to receive a two-dimensional matrix converted from the second time sequence and perform feature extraction on the two-dimensional matrix.

[0155] The pooling layer is configured to perform a pooling operation on the feature map output after the feature extraction.

[0156] Flatten layer, used to convert the feature map output after the pooling operation into a one-dimensional vector.

[0157] In one specific embodiment, the sequence prediction network of the initial state is trained and parameter optimized, including:

[0158] Using the historical data of the feature variables to construct training samples;

[0159] Using the training samples to pre-train and pre-optimize the hyperparameters of the sequence prediction network;

[0160] Using the real-time collected feature variables to construct real samples;

[0161] Using the real samples to incrementally train and optimize the hyperparameters of the pre-trained and pre-optimized sequence prediction network;

[0162] Among them, the particle swarm optimization algorithm is used to optimize the hyperparameters.

[0163] In one specific embodiment, the DC / DC converter adopts a boost circuit as the DC / DC boost circuit, and the space state equation is represented as:

[0164]

[0165] Among them, x represents the state variable, A represents the state matrix, u represents the input vector, the input vector is the target voltage value output by the DC / DC converter according to the scheduling instruction, B represents the input matrix, represents the derivative of the state variable, y represents the output vector, the output vector is the difference between 1 and the duty cycle prediction value of the pulse modulation signal for controlling the DC / DC converter, C represents the output matrix, and D represents the direct transfer matrix.

[0166] In one specific embodiment, the constraint conditions of the MPC model include: the duty cycle is greater than or equal to 0 and less than or equal to 1; the target power value is within the preset fluctuation range.

[0167] In one specific embodiment, when generating the optimal duty cycle using the MPC model that has been constructed, an optimization algorithm is used to optimize the prediction time domain length of the MPC model.

[0168] In one specific embodiment, the optimization algorithm used to optimize the prediction time domain length of the MPC model includes:

[0169] Initialize the particle swarm;

[0170] Calculate the fitness value of each particle according to the fitness function;

[0171] Perform the particle swarm update step;

[0172] If the iteration number or the fitness value of the particle in the particle swarm reaches the expected value, the optimal individual in the current particle swarm is output as the optimal solution of the prediction time domain length, otherwise the particle swarm updating step is executed again;

[0173] The particle swarm updating step comprises:

[0174] determining the current individual extremum of each particle and the current global optimal solution of the particle swarm;

[0175] updating the speed and position of each particle;

[0176] The fitness function is an evaluation function representing the error degree between the output voltage of the DC / DC converter calculated by the MPC model and the target voltage value required by the DC / DC converter determined according to the scheduling instruction.

[0177] In one specific embodiment, the power transmission and transformation state variables include ambient temperature, load and transformer usage time.

[0178] In one specific embodiment, the photovoltaic power generation grid-connected power limiting control device 400 provided by the present application can be realized in the form of a computer program, which can run on a computer device as shown in the figure. Figure 8 The memory of the computer device can store various program modules constituting the photovoltaic power generation grid-connected power limiting control device 400, such as the acquisition module 410 and the power limiting control module 420 shown in the figure. Figure 7 The computer program constituted by various program modules enables the processor to execute the steps in the photovoltaic power generation grid-connected power limiting control method of various embodiments of the present application described in the specification.

[0179] Figure 8A block diagram of a computer device according to an embodiment of the present application is shown schematically. Specifically, the computer device provided by the embodiment of the present application can be a terminal. The computer device comprises a processor A01, a network interface A02, a display screen A04, an input device A05 and a memory (not shown in the figure) connected by a system bus. Among them, the processor A01 of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises an internal memory A03 and a non-volatile storage medium A06. The non-volatile storage medium A06 stores an operating system B01 and a computer program B02. The internal memory A03 provides an environment for the running of the operating system B01 and the computer program B02 in the non-volatile storage medium A06. The network interface A02 of the computer device is used to communicate with external terminals through network connection. The computer program is executed by the processor A01 to implement a photovoltaic power generation grid-connected power limiting control method. The display screen A04 of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device A05 of the computer device can be a touch layer overlaid on the display screen, or a key, trackball or touchpad arranged on the shell of the computer device, or an external keyboard, touchpad or mouse, etc.

[0180] The embodiment of the present application also provides a machine readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the photovoltaic power generation grid-connected power limiting control method.

[0181] The embodiment of the present application also provides a computer program product, which is adapted to execute the program initialized with the following method steps when executed on a data processing device:

[0182] obtaining a predicted value of power required by a load at a next moment and a predicted value of maximum output power of the photovoltaic power station;

[0183] determining whether the predicted value of maximum output power of the photovoltaic power station at the next moment is greater than the sum of a power loss value generated when the photovoltaic power is connected to the grid and the predicted value of power required by the load at the next moment; if yes, generating an optimal duty ratio by using the MPC model constructed to make the DC / DC converter in the photovoltaic power station output a target power value under the control of a pulse modulation signal with the optimal duty ratio, otherwise, adopting MPPT maximum power point tracking control on the DC / DC converter side to make the output power of the DC / DC converter track the predicted value of maximum output power of the photovoltaic power station at the next moment;

[0184] Among them, the target power value is a power value required to be output by the DC / DC converter determined according to a scheduling instruction, and the spatial state equation of the MPC model is established based on a loop in which the DC / DC converter is located to supply power to the load.

[0185] The apparatus embodiments described above are only exemplary, in which the units as shown in the separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place, or can be distributed on a plurality of network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. Those skilled in the art can understand and implement without creative labor.

[0186] Those skilled in the art will appreciate that embodiments of the application can be provided as a method, a system, or a computer program product. Accordingly, the application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer-readable program code.

[0187] It should also be noted that the terms "comprising", "comprises", "including", "includes" or any other variation thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without further limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article or apparatus that comprises the element.

[0188] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent ones; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A photovoltaic power generation grid-connected power limiting control method, characterized in that, The method comprises: obtaining a next-time load required power prediction value and a photovoltaic power station maximum output power prediction value; determining whether the next-time photovoltaic power station maximum output power prediction value is greater than the sum of a power loss value generated when photovoltaic power generation is connected to a power grid and the next-time load required power prediction value; if yes, generating an optimal duty cycle by using an MPC model that has been constructed, so that a DC / DC converter in the photovoltaic power station outputs a target power value under the control of a pulse modulation signal with the optimal duty cycle, otherwise, adopting MPPT maximum power point tracking control on the DC / DC converter side, so that the output power of the DC / DC converter tracks the next-time photovoltaic power station maximum output power prediction value; wherein the target power value is a power value required to be output by the DC / DC converter determined according to a scheduling instruction, and a spatial state equation of the MPC model is established based on a loop in which the photovoltaic power station where the DC / DC converter is located supplies power to a load; wherein the DC / DC converter adopts a boost circuit as a DC / DC boost circuit, and the spatial state equation is expressed as: ； wherein, denotes a state variable, denotes a state matrix, denotes an input vector, the input vector being a target voltage value determined from a scheduling instruction for an output of the DC / DC converter, denotes an input matrix, denotes a derivative of the state variable, denotes an output vector, the output vector being a difference between 1 and a predicted value of a duty ratio of a pulse modulation signal for controlling the DC / DC converter, denotes an output matrix, denotes a direct transmission matrix.

2. The photovoltaic power generation grid-connected power limiting control method according to claim 1, characterized in that, Before determining whether the next-time photovoltaic power station maximum output power prediction value is greater than the sum of the power loss value generated when photovoltaic power generation is connected to the power grid and the next-time load required power prediction value, the method further comprises: composing a first time sequence from the collected characteristic variables of continuous time; inputting the first time sequence into a constructed power transmission and transformation loss prediction model to predict the next-time power loss value generated when photovoltaic power generation is connected to the power grid; wherein the maximum collection time of the characteristic variables is the current time, the characteristic variables include power transmission and transformation state variables that change photovoltaic power generation grid-connected power loss, and the power transmission and transformation loss prediction model is obtained by training and parameter optimization on an initial state sequence prediction network.

3. The grid-connected power limiting control method for photovoltaic power generation according to claim 2, characterized in that, The characteristic variables further include voltage amplitude fluctuation characteristics and current amplitude fluctuation characteristics of a transformer in a photovoltaic power generation grid-connected line, and the voltage amplitude fluctuation characteristics and the current amplitude fluctuation characteristics of continuous time are obtained by performing feature extraction on a second time sequence by using a constructed feature extraction network, the second time sequence being composed of collected transformer voltage amplitudes and transformer current amplitudes of continuous time.

4. The grid-connected power limiting control method for photovoltaic power generation according to claim 3, characterized in that, The feature extraction network comprises a 2D convolution layer, a pooling layer and a Flatten layer connected in sequence, wherein: the 2D convolution layer is configured to receive a two-dimensional matrix converted from the second time sequence and perform feature extraction on the two-dimensional matrix; the pooling layer is configured to perform a pooling operation on a feature map output after feature extraction; the Flatten layer is configured to convert the feature map output after the pooling operation into a one-dimensional vector.

5. The grid-connected power limiting control method for photovoltaic power generation according to claim 3, characterized in that, Training and parameter optimization on the initial state sequence prediction network comprises: constructing training samples by using historical data of the characteristic variables; pre-training and hyperparameter pre-optimization on the sequence prediction network by using the training samples; constructing real samples by using real-time collected characteristic variables; incremental training and hyperparameter optimization on the pre-trained and hyperparameter pre-optimized sequence prediction network by using the real samples; wherein the hyperparameters are optimized by using a particle swarm optimization algorithm.

6. The photovoltaic power generation grid-connected limited power control method according to claim 1, characterized in that, The constraint condition of the MPC model comprises: a duty cycle greater than or equal to 0 and less than or equal to 1; and a target power value within a preset fluctuation range.

7. The photovoltaic power generation grid-connected limited power control method according to claim 1, characterized by, When the optimal duty cycle is generated by using the constructed MPC model, an optimization algorithm is used to optimize the prediction time domain length of the MPC model.

8. The photovoltaic power generation grid-connected limited power control method according to claim 7, characterized in that, The optimization algorithm used to optimize the prediction time domain length of the MPC model comprises: initializing a particle swarm; calculating the fitness value of each particle according to a fitness function; performing a particle swarm updating step; if the number of iterations or the fitness value of the particles of the particle swarm reaches an expected value, outputting the optimal individual in the current particle swarm as the optimal solution of the prediction time domain length, otherwise, performing the particle swarm updating step again; wherein the particle swarm updating step comprises: determining the current individual extremum of each particle and the current global optimal solution of the particle swarm; updating the speed and position of each particle. The fitness function is an evaluation function representing the error degree between the output voltage of the DC / DC converter calculated by the MPC model and the target voltage value required by the DC / DC converter determined according to the scheduling instruction.

9. The grid-connected power limiting control method for photovoltaic power generation according to claim 2, characterized in that, The power transmission and transformation state variables comprise environmental temperature, load and transformer use time.

10. A photovoltaic power generation grid-connected power limiting control device, characterized in that, The device comprises: an acquisition module configured to acquire a load required power prediction value and a photovoltaic power station maximum output power prediction value at a next time; a power limiting control module configured to determine whether the photovoltaic power station maximum output power prediction value at the next time is greater than the sum of a power loss value generated when the photovoltaic power generation is connected to a power grid and the load required power prediction value at the next time; if yes, generate an optimal duty cycle by using the constructed MPC model, so that the DC / DC converter in the photovoltaic power station outputs power that tracks a target power value under the control of a pulse modulation signal with the optimal duty cycle, otherwise, adopt MPPT maximum power point tracking control on the DC / DC converter side, so that the output power of the DC / DC converter tracks the photovoltaic power station maximum output power prediction value at the next time; wherein the target power value is a power value required by the DC / DC converter determined according to a scheduling instruction, and a spatial state equation of the MPC model is established based on a loop in which the photovoltaic power station where the DC / DC converter is located supplies power to a load; wherein the DC / DC converter uses a boost circuit as a DC / DC boost circuit, and the spatial state equation is expressed as: ； wherein, represents a state variable, represents a state matrix, represents an input vector, the input vector being a target voltage value of the DC / DC converter output determined according to a scheduling instruction, represents an input matrix, represents a derivative of the state variable, represents an output vector, the output vector being a difference between 1 and a duty ratio prediction value of a pulse modulation signal for controlling the DC / DC converter, represents an output matrix, represents a direct transmission matrix.

11. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the photovoltaic power generation grid-connected power limiting control method of any one of claims 1 to 9 when executing the program.

12. A machine-readable storage medium having stored thereon a computer program, characterized in that The computer program implements the photovoltaic power generation grid-connected power limiting control method of any one of claims 1 to 9 when executed by the processor.

Images