Power grid dispatching method and system based on photovoltaic power generation power prediction

Abstract

The invention discloses a power grid dispatching method and system based on photovoltaic power generation power prediction. The method comprises the following steps: establishing a power grid double-layer control model comprising a main power grid, a photovoltaic power generation module, a standby diesel generator and an energy storage system; defining constraint conditions of the power grid double-layer control model; performing power prediction on each photovoltaic power generation module by using the photovoltaic power generation power prediction model; constructing a target function of an upper-layer controller by taking improvement of the photovoltaic power generation utilization rate and reduction of power fluctuation as targets; under the constraint condition, an optimal power flow curve is optimized and determined through an upper-layer controller; and according to the photovoltaic power generation power prediction value, constructing a lower-layer control objective function, and determining an optimal power grid dispatching scheme through a lower-layer controller. The photovoltaic power generation utilization rate is improved, the power fluctuation is reduced, the scheduling cost is reduced, the overall efficiency and stability of the system are improved, falling into a local optimal solution is avoided, the overall efficiency of the system is maintained on a long-time scale, and clean energy is utilized to the maximum extent.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a power grid dispatching method and system based on photovoltaic power generation prediction. Background Art

[0002] With the continuous development of photovoltaic, wind and hydropower clean energy, more and more clean energy is connected to the power grid system. In order to reduce the fluctuations of clean energy caused by weather, the clean energy system is often equipped with corresponding energy storage systems and backup diesel generators. The energy storage system can store electricity when there is excess photovoltaic power generation, and release electricity when there is insufficient photovoltaic power generation, smoothing the power output and stabilizing the operation of the power grid. The backup diesel generator can start quickly when the main power grid is out of power or photovoltaic power generation is insufficient, providing a reliable emergency power supply to ensure the continuous power supply of critical loads. The photovoltaic power generation module provides clean energy, the energy storage system stores excess electricity, and the diesel generator serves as a backup. The three complement each other to achieve efficient use of renewable energy.

[0003] The current power grid dispatching often uses model predictive control (MPC) control technology. MPC can predict system behavior at multiple future moments and ensure the optimal performance of the system at future moments by optimizing the control strategy within the entire prediction period.

[0004] However, the current MPC control technology is prone to falling into local optimal solutions. Although the local optimal scheduling plan can be obtained within each time step, it cannot guarantee to remain at the global optimal solution. In particular, there will be deviations on a long time scale. During long-term operation, it may gradually deviate from the global optimal path. The accumulated small deviations lead to a decline in the overall efficiency and benefits of the system, and clean energy cannot be maximized, affecting the stability of the power grid. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is: the existing cloud service platform calculation method has large power consumption, load rate rewards and penalties, high cost, and the optimization problem of how to distribute task requests to each host to achieve load balancing.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a grid dispatching method based on photovoltaic power generation power prediction, comprising:

[0008] Establish a two-layer control model of the power grid including the main power grid, photovoltaic power generation modules, backup diesel generators and energy storage systems;

[0009] Define the constraints of the two-layer control model of the power grid;

[0010] Use the photovoltaic power generation prediction model to predict the power of each photovoltaic power generation module;

[0011] To improve the utilization rate of photovoltaic power generation and reduce power fluctuations, the objective function of the upper-level controller is constructed;

[0012] Under the constraints, the best power flow curve is determined by optimizing the upper controller;

[0013] According to the predicted value of photovoltaic power generation, the lower-level control objective function is constructed, and the optimal grid dispatching plan is determined through the lower-level controller.

[0014] As a preferred solution of the grid dispatching method based on photovoltaic power generation prediction described in the present invention, the grid two-layer control model includes an upper controller for determining the optimal power flow curve and a lower controller for determining the optimal grid dispatching plan.

[0015] As a preferred solution of the power grid dispatching method based on photovoltaic power generation prediction described in the present invention, wherein: using a photovoltaic power generation prediction model to predict the power of each photovoltaic power generation module includes: the photovoltaic power generation prediction model includes a weather prediction unit and a power prediction unit;

[0016] The weather prediction unit includes a first convolutional layer, a first gating unit, and a first attention layer connected in series, and the power prediction unit includes a second convolutional layer, a second gating unit, and a second attention layer connected in series;

[0017] The first and second attention layers are both connected to the fully connected network, and the fully connected layer is connected to the prediction layer;

[0018] Obtain historical weather data and historical photovoltaic power generation data. The weather features in the historical weather data are extracted through the first convolutional layer in the weather prediction unit, and the power features in the historical photovoltaic power generation data are extracted through the second convolutional layer in the power prediction unit;

[0019] The weather hidden feature in the weather feature is extracted by the first gating unit in the weather prediction unit, and the power hidden feature in the power feature is extracted by the second gating unit in the power prediction unit.

[0020] Attention weights are added to the weather hidden features at each moment through the first attention layer in the weather prediction unit, and attention weights are added to the power hidden features at each moment through the second attention layer in the power prediction unit.

[0021] The weather hidden features and power hidden features after adding attention weights are concatenated through a fully connected network. Based on the concatenated fusion features, power prediction is performed through the prediction layer to obtain a preliminary predicted power.

[0022] Based on the time periodicity and the nearest neighbor similarity, the preliminary predicted power is corrected to determine the final predicted power. According to the correlation predicted power, the preliminary predicted power is corrected to determine the final predicted power.

[0023] As a preferred solution of the power grid dispatching method based on photovoltaic power generation prediction described in the present invention, wherein: using the photovoltaic power generation prediction model to predict the power of each photovoltaic power generation module includes, the photovoltaic power generation prediction model also includes, with the goal of improving the utilization rate of photovoltaic power generation and reducing power fluctuations, constructing the objective function of the upper controller includes calculating the correlation coefficient between each adjacent day in the year before the prediction day:

[0024]

[0025] S ij =f(w ij ,w 00 )

[0026] V ij =g(w ij )

[0027] Among them, C i j represents the correlation coefficient between the forecast day and the jth adjacent day in the previous i years, T i j represents the time correlation coefficient of the jth adjacent day in the previous i years, k1 represents the year correlation coefficient, k2 represents the date correlation coefficient, S i j represents the weather correlation coefficient of the jth adjacent day in the previous i year, w i j represents the weather type of the jth consecutive day in the previous i year, w 00 Represents the weather type forecast value of the forecast day. The weather type forecast value of the forecast day is calculated based on the weather hidden features after adding the attention weight. V i j represents the randomness coefficient of the jth consecutive day in the previous i years, V 00 represents the randomness coefficient of the forecast day;

[0028] The correlation prediction power is calculated based on the correlation coefficient between the predicted day and each adjacent day in previous years:

[0029]

[0030] Among them, P c Represents the correlation prediction power of the prediction day, P ij represents the power generation of the jth consecutive day in the previous i years, n represents the total number of years, and m represents the total number of consecutive days;

[0031] The specific method of correcting the preliminary predicted power according to the correlation predicted power is as follows:

[0032]

[0033] Among them, P t represents the final predicted power at time t, represents the preliminary predicted power at time t, P ct represents the correlation prediction power at time t in the prediction day, and λ represents the correction coefficient.

[0034] As a preferred solution of the power grid dispatching method based on photovoltaic power generation prediction described in the present invention, wherein: with the goal of improving photovoltaic power generation utilization and reducing power fluctuations, the objective function of the upper controller is constructed, and the upper control objective function is specifically:

[0035]

[0036] Among them, J H represents the upper control objective function, min represents the minimum value, represents the power flow at time t, represents the average power flow, T represents the total observation time, λ1 represents the weight coefficient of the power flow term, and λ2 represents the weight coefficient of the power flow fluctuation term. Among them, those skilled in the art can set the size of the weight coefficient λ1 of the power flow term and the weight coefficient λ2 of the power flow fluctuation term according to actual conditions;

[0037] The power flow is specifically:

[0038] P PCC =P + -P -

[0039] Among them, P P CC represents power flow, P + Represents the active power in the upstream direction of the node, P - Indicates the active power in the downstream direction of the node;

[0040] The smaller the power flow indicated by the power flow item, the higher the degree of self-sufficiency within the microgrid, and the greater the utilization rate of photovoltaic power generation.

[0041] As a preferred solution of the power grid dispatching method based on photovoltaic power generation prediction described in the present invention, wherein: under the constraint condition, optimizing and determining the optimal power flow curve through the upper-level controller includes, through the upper-level controller, under the constraint condition, taking the upper-level control objective function as the target, using a particle swarm optimization algorithm to determine the optimal power flow curve;

[0042] Initialize the population, which contains multiple particles, each particle represents a feasible power flow curve

[0043] line, calculate the fitness of each particle, determine the individual optimal position and the global optimal position, and the formula is expressed as:

[0044]

[0045] in, represents the velocity of the ith particle at the t+lth iteration, represents the velocity of the i-th particle at the t-th iteration, ω represents the inertia weight factor, c1 represents the individual learning factor, c2 represents the global learning factor, r1 and r2 both represent random numbers between 0 and l, represents the individual optimal position of the i-th particle at the t-th iteration, represents the global optimal position of the i-th particle at the t-th iteration, represents the position of the i-th particle at the t-th iteration, represents the position of the i-th particle at the t+l-th iteration;

[0046] Determine whether the maximum number of iterations has been reached. If so, output the power flow curve represented by the particle with the highest fitness value as the optimal power flow curve. Otherwise, return to continue iterating.

[0047] The replacement probability is:

[0048]

[0049] Among them, P represents the replacement probability, e represents the natural constant, δ() represents the fitness function, and C t represents the temperature at the tth iteration;

[0050] Update the temperature:

[0051] T i+1 =εT i

[0052] Where ε represents the temperature drop coefficient, T i +1 indicates the temperature at the i+1th iteration, T i represents the temperature at the i-th iteration;

[0053] When the temperature is high, it is easier to accept a poor solution, thus increasing the chance of jumping out of the local optimum. As the number of iterations increases, the temperature gradually decreases, causing the algorithm to perform more global searches in the early stages and gradually focus on local searches in the later stages, improving the optimization effect.

[0054] The specific temperature reduction coefficient is:

[0055]

[0056] Among them, A1 represents the cooling oscillation factor of the first half cycle, B1 represents the cooling floating factor of the first half cycle, A2 represents the cooling oscillation factor of the second half cycle, and B2 represents the cooling floating factor of the second half cycle.

[0057] As a preferred solution of the grid dispatching method based on photovoltaic power generation prediction described in the present invention, wherein: according to the photovoltaic power generation prediction value, a lower-level control objective function is constructed, and the optimal grid dispatching scheme is determined by the lower-level controller, and the lower-level control objective function is specifically:

[0058]

[0059] Among them, J L represents the lower layer control objective function, min represents the minimum value, represents the power flow at time t, represents the reference power flow, T represents the total observation time, C t represents the dispatch cost at time t, μ1 represents the weight coefficient of the power flow deviation term, and μ2 represents the weight coefficient of the dispatch cost term;

[0060] By flexibly adjusting the weight coefficients to adapt to different operating requirements and strategies, it ensures that the lower-level control can achieve the overall optimization goal of the upper-level control and improve the overall performance of the system;

[0061] The specific scheduling cost is:

[0062]

[0063] Among them, C t represents the scheduling cost at time t, represents the main grid electricity price at time t, N represents the amount of electricity purchased by the main power grid at time t. G Indicates the number of standby diesel generators, represents the unit power generation cost of the nth standby diesel generator, represents the power generation of the nth standby diesel generator at time t, N E SS represents the number of energy storage systems, represents the unit energy storage cost of the nth energy storage system, represents the storage power of the nth energy storage system at time t, N P V represents the number of photovoltaic power generation modules, represents the unit power generation cost of the nth photovoltaic power generation module, P n t P V represents the predicted power generation value of the nth photovoltaic power generation module at time t.

[0064] A power grid dispatching system based on photovoltaic power generation power prediction, wherein:

[0065] The main power grid provides basic power supply and transmission functions;

[0066] Photovoltaic power generation modules, which use photovoltaic technology to generate electricity;

[0067] Backup diesel generators to provide emergency power supply when the main power grid or photovoltaic power generation is insufficient;

[0068] Energy storage systems store electricity when there is excess power and release it when there is a power shortage, thus smoothing out power output;

[0069] The upper controller is responsible for determining the best power flow curve, optimizing the utilization of photovoltaic power generation and reducing power fluctuations;

[0070] The lower-level controller optimizes the grid dispatching scheme based on the predicted value of photovoltaic power generation and the optimal power flow curve, reduces the deviation between the actual power flow and the reference power flow, and reduces the dispatching cost;

[0071] Processor, which executes grid dispatch algorithms and control strategies;

[0072] The memory stores relevant computer-readable instructions and data for executing the scheduling method.

[0073] A computer device comprises: a memory and a processor; the memory stores a computer program, wherein the processor implements the steps of any one of the methods of the present invention when executing the computer program.

[0074] A computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of any one of the methods of the present invention.

[0075] Beneficial effects of the invention: The power grid dispatching method based on photovoltaic power generation prediction provided by the invention improves the utilization rate of photovoltaic power generation, reduces power fluctuations, reduces dispatching costs, and improves the overall efficiency and stability of the system by constructing a double-layer control model, avoids falling into local optimal solutions, maintains the overall efficiency of the system on a long time scale, maximizes the use of clean energy, promotes the development of renewable energy, and ultimately achieves efficient, economical and sustainable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0077] Figure 1 An overall flow chart of a power grid dispatching method based on photovoltaic power generation prediction provided by the first embodiment of the present invention;

[0078] Figure 2 A schematic structural diagram of a power grid dispatching method based on photovoltaic power generation prediction provided in the first embodiment of the present invention. DETAILED DESCRIPTION

[0079] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0080] Example 1, reference Figure 1 , is an embodiment of the present invention, and provides a power grid dispatching method based on photovoltaic power generation power prediction, comprising:

[0081] S1: Establish a two-layer control model of the power grid including the main power grid, photovoltaic power generation modules, backup diesel generators and energy storage systems.

[0082] The power grid two-layer control model includes an upper-layer controller for determining the optimal power flow curve and a lower-layer controller for determining the optimal power grid dispatching plan.

[0083] It should be noted that the energy storage system can store electricity when there is excess photovoltaic power generation, and release electricity when there is insufficient photovoltaic power generation, smoothing power output and stabilizing grid operation. The backup diesel generator can start quickly when the main grid is out of power or photovoltaic power generation is insufficient, providing a reliable emergency power supply and ensuring continuous power supply to critical loads. The photovoltaic power generation module provides clean energy, the energy storage system stores excess electricity, and the diesel generator serves as a backup. The three complement each other to achieve efficient use of renewable energy.

[0084] S2: Define the constraints of the two-layer control model of the power grid.

[0085] In a possible implementation, the constraints specifically include:

[0086] Power balance constraints:

[0087]

[0088] in, N represents the amount of electricity purchased by the main power grid at time t. G Indicates the number of standby diesel generators, represents the power generation of the nth standby diesel generator at time t, N P V represents the number of photovoltaic power generation modules, represents the power generation of the nth photovoltaic power generation module at time t, N E SS represents the number of energy storage systems, P n t DCh,ESS represents the discharge power of the nth energy storage system at time t, P n t Ch,ESS represents the nth energy storage, R represents the load, I t represents the current at time t, represents the load power at time t. It should be noted that the power balance constraint can ensure that at each time step t, the total power generation and purchased power of the power grid can meet the total load demand. Here, the power purchase of the main power grid, the power generation of the backup diesel generator, the photovoltaic power generation, the charging and discharging power of the energy storage system, and the power flow and load current losses are considered.

[0089] Standby diesel generator power constraints:

[0090] P G ≤P G,Nom

[0091] Among them, P G Indicates the power generation of the standby diesel generator, P G,Nom Indicates backup diesel generator

[0092] The nominal power of the machine.

[0093] It should be noted that the power constraint of the standby diesel generator can ensure that the power generated by the standby diesel generator does not exceed its nominal power, prevent the generator from overloading, protect the equipment and extend its life. Energy storage system charging power constraint:

[0094] bs + ≤β·bs max

[0095] Among them, bs + Indicates the charging power of the energy storage system, bs max Represents the maximum charging power of the energy storage system, and β represents the charging and discharging efficiency. It should be noted that the energy storage system charging power constraint can limit the charging power of the energy storage system to not exceed a part of its maximum charging power, ensuring that the charging process is within the range allowed by the equipment. Energy storage system discharge power constraint:

[0096] bs - ≤(1-β)·bs max

[0097] Among them, bs - Indicates the discharge power of the energy storage system, bsmax Represents the maximum charging power of the energy storage system, and β represents the charging and discharging efficiency. It should be noted that the energy storage system discharge power constraint can limit the energy storage system's discharge power to no more than a portion of its maximum charging power, ensuring that the charging process is within the range allowed by the device. Energy storage system storage capacity constraint:

[0098]

[0099] Among them, C ESS represents the storage capacity of the energy storage system, Indicates the maximum storage capacity of the energy storage system, Indicates the minimum storage capacity of the energy storage system. It should be noted that the energy storage system storage capacity constraint can ensure that the storage capacity of the energy storage system is between the maximum and minimum storage capacities, preventing the energy storage system from overcharging or overdischarging, thereby protecting the battery life and performance. Charge and discharge capacity balance constraint:

[0100] C Ch =C DCh

[0101] in, Indicates the charging capacity, It should be noted that the charge-discharge capacity balance constraint can ensure that the charge capacity is equal to the discharge capacity and maintain the energy balance of the energy storage system.

[0102] |E G +E PV -E Load -E Loss |=E ESS

[0103] Among them, E G Indicates the energy of the standby diesel generator, E P V represents the energy of photovoltaic power generation module, E L oad

[0104] Energy balance constraint: represents the energy consumed by the load, Indicates the energy lost by power abandonment, E E SS represents the energy of the energy storage system. It should be noted that the energy balance constraint can ensure the energy balance between the backup diesel generator energy, photovoltaic power generation module energy, load consumption energy, abandoned power loss energy and energy storage system energy, ensuring the energy conservation of the system. Energy storage system storage power constraint:

[0105] P ESS =βP Ch +(1-β)P DCh

[0106] Among them, P ESS represents the storage power of the energy storage system, P C h represents the charging power, P D Ch represents the discharge power, and β represents the charge and discharge efficiency. It should be noted that the storage power constraint of the energy storage system can ensure that the energy conversion process of the energy storage system conforms to the actual situation.

[0107] Maximum and minimum storage power constraints of energy storage system:

[0108]

[0109] in, represents the maximum storage power of the energy storage system, Indicates the minimum storage power of the energy storage system.

[0110] It should be noted that the maximum and minimum storage power constraints of the energy storage system can ensure that the storage power of the energy storage system is between the maximum and minimum storage powers, preventing the energy storage system from being overcharged or over-discharged.

[0111] Energy storage system energy update constraints:

[0112]

[0113] Among them, E t +1 ESS represents the stored energy of the energy storage system at time t+1, represents the stored energy of the energy storage system at time t, It represents the storage power of the energy storage system at time u, and du represents the differentiation of u.

[0114] It should be noted that the energy update constraint of the energy storage system can determine the change in energy of the energy storage system over time by integrating the storage power of the energy storage system, thereby ensuring the accuracy of the energy change process.

[0115] Energy balance constraints of energy storage system within time interval:

[0116]

[0117] in, represents the stored energy of the energy storage system in the first time interval, Represents the stored energy of the energy storage system in the kth time interval.

[0118] It should be noted that the energy balance constraint within the time interval of the energy storage system can ensure the energy consistency and balance of the energy storage system within different time intervals, and ensure the stable operation of the system in different time periods.

[0119] Node power flow constraints:

[0120]

[0121] in, Represents the active power in the upstream direction of the node, P - Represents the active power in the downstream direction of the node, V N om represents the nominal voltage, I max Indicates the maximum current that the node can withstand.

[0122] S3: Use the photovoltaic power generation prediction model to predict the power of each photovoltaic power generation module.

[0123] It should be noted that the node power flow constraint can ensure that the sum of the upstream and downstream active powers of the node does not exceed the nominal power, protecting the grid equipment and ensuring its safe operation.

[0124] Optionally, the photovoltaic power generation power prediction model includes a weather prediction unit and a power prediction unit. The weather prediction unit includes a first convolutional layer, a first gating unit, and a first attention layer connected in series, and the power prediction unit includes a second convolutional layer, a second gating unit, and a second attention layer connected in series, the first attention layer and the second attention layer are both connected to a fully connected network, and the fully connected layer is connected to the prediction layer.

[0125] Optionally, historical weather data and historical photovoltaic power generation data are obtained. The weather features in the historical weather data are extracted by the first convolutional layer in the weather prediction unit, and the power features in the historical photovoltaic power generation data are extracted by the second convolutional layer in the power prediction unit. The weather hidden features in the weather features are extracted by the first gating unit in the weather prediction unit, and the power hidden features in the power features are extracted by the second gating unit in the power prediction unit. Attention weights are added to the weather hidden features at each moment by the first attention layer in the weather prediction unit, and attention weights are added to the power hidden features at each moment by the second attention layer in the power prediction unit. The weather hidden features and the power hidden features after adding the attention weights are spliced ​​through a fully connected network. According to the fusion features after splicing, power prediction is performed through the prediction layer to obtain a preliminary predicted power. Based on time periodicity and neighbor similarity, the preliminary predicted power is corrected to determine the final predicted power. According to the correlation predicted power, the preliminary predicted power is corrected to determine the final predicted power.

[0126] In a possible implementation, convolution features in historical weather data and power features in historical photovoltaic power generation data are extracted respectively by convolution kernels:

[0127]

[0128] in, represents the output of the current convolutional layer, represents the output of the previous convolutional layer, Represents the convolution kernel weight of the current convolution layer, b′ i Represents the bias term of the current convolutional layer, M j Represents the selected input feature map, and σ() represents the activation function. Pooling is performed on the extracted convolutional features:

[0129]

[0130] Among them, down represents the downsampling function, represents the output of the current pooling layer, Represents the output of the previous pooling layer. The features after pooling are summarized to obtain weather features and power features respectively.

[0131] In a possible implementation, extracting weather hidden features in weather features through a first gating unit in a weather prediction unit, and extracting power hidden features in power features through a second gating unit in a power prediction unit, specifically includes:

[0132] z t =σ(W uz x t +W hz hw t-1 +b z )

[0133] r t =σ(W ur x t +W hr hw t-1 +b r )

[0134] c t =σ(W uc x t +W hc (r t ⊙hw t-1 )+b c )

[0135] jw t =(1-z t )⊙hw t-1 +z t ⊙c t

[0136] Among them, z t represents the output vector of the update gate at time t, σ() represents the activation function, and w u z represents the weight matrix between the input layer and the update gate, x t represents the weather characteristics or power characteristics at time t, W hz represents the self-connection weight matrix of the update gate between time t and time t-1, hw t-1 represents the hidden state vector at time tl, b z represents the bias term of the update gate, r t represents the output vector of the reset gate at time t, represents the weight matrix between the input layer and the reset gate, represents the self-connection weight matrix between the reset gate at time t and time tl, b r represents the bias term of the reset gate, c t represents the output vector of the candidate layer at time t, W uc Represents the weight matrix between the input layer and the candidate layer, W h c represents the self-connection weight matrix of the candidate layer between time t and time t-1, b c represents the bias term of the candidate layer, ⊙ represents the element-wise product operation, hw t Represents the hidden state vector at time t. Among them, the activation function is such as ReLU, tanh, sigmoid, etc., and the appropriate activation function can be selected according to the specific needs of the problem.

[0137] In a possible implementation, adding attention weights to weather hidden features at each moment through the first attention layer, and adding attention weights to power hidden features at each moment through the second attention layer in the power prediction unit, specifically includes:

[0138] e t =σ(W h hw t +b h )

[0139]

[0140] hw′ t =α t hw t

[0141] Among them, e t represents the attention score at time t, σ() represents the activation function, represents the attention weight matrix, hw t represents the hidden state vector at time t, b h represents the bias term of the attention layer, α t represents the attention weight at time t, exp() represents an exponential function with a natural constant as the base, T represents the total length of the data, and hw′ trepresents the weather hidden feature or power hidden feature after adding the attention weight. In a possible implementation, the weather hidden feature and the power hidden feature after adding the attention weight are concatenated through a fully connected network:

[0142] H t =(hw′ t ,hp′ t )

[0143] Among them, H t represents the fusion features after splicing, hw t Represents the weather hidden features after adding attention weights, hp′ t Indicates the power hidden feature after adding the attention weight. In a possible implementation, according to the fusion feature after splicing, power prediction is performed through the prediction layer to obtain the preliminary predicted power:

[0144]

[0145] in, represents the initial prediction power at time t, σ() represents the activation function, W H represents the prediction weight matrix, H t represents the fusion features after splicing, b H In one possible implementation, the correlation coefficient between each adjacent day in the year before the forecast day is calculated:

[0146] Among them, C i j represents the correlation coefficient between the forecast day and the jth adjacent day in the previous i years, T i j represents the number of relationships:

[0147]

[0148] V ij =g(w ij )

[0149] The time correlation coefficient of the jth adjacent day in the previous i years, k1 represents the year correlation coefficient, k2 represents the date correlation coefficient, S i represents the weather correlation coefficient of the jth adjacent day in the previous i years, f() represents the weather correlation function, and w i j represents the weather type of the jth consecutive day in the previous i year, w 00 The weather type prediction value of the forecast day is calculated by adding the attention weight to the weather hidden features in 20 rows. i j represents the randomness coefficient of the jth consecutive day in the previous i years, V 00Represents the randomness coefficient of the forecast day, g() random function. Calculate the correlation forecast power based on the correlation coefficient between the forecast day and each adjacent day in the previous year:

[0150]

[0151] Among them, P c Represents the correlation prediction power of the prediction day, P i j represents the power generation of the jth consecutive day in the previous i years, n represents the total number of years, and m represents the total number of consecutive days. In a possible implementation, the specific method of correcting the preliminary predicted power according to the correlation predicted power is:

[0152]

[0153] Among them, P t represents the final predicted power at time t, P t represents the preliminary predicted power at time t, P ct represents the correlation prediction power at time t in the prediction day, and λ represents the correction coefficient.

[0154] S4: Construct the objective function of the upper-level controller with the goal of improving the utilization rate of photovoltaic power generation and reducing power fluctuations.

[0155] In a possible implementation manner, the upper layer control objective function is specifically:

[0156]

[0157] Among them, J H represents the upper control objective function, min represents the minimum value, represents the power flow at time t, Represents the average power flow, T represents the total observation time, λ1 represents the weight coefficient of the power flow term, and λ2 represents the weight coefficient of the power flow fluctuation term. Wherein. Those skilled in the art can set the size of the weight coefficient λ1 of the power flow term and the weight coefficient λ2 of the power flow fluctuation term according to actual conditions, and the present invention does not limit it. It should be noted that the smaller the power flow indicated by the power flow term, the higher the degree of self-sufficiency within the microgrid, and the greater the utilization rate of photovoltaic power generation. In the present invention, by constructing the upper-level control objective function, a balance can be found between optimizing the use of power flow in the power grid and reducing fluctuations, thereby improving the utilization rate of photovoltaic power generation and reducing operating costs. At the same time, the reliability and stability of the system can be improved. These benefits combined will help to achieve efficient, economical and sustainable operation of the power grid.

[0158] Optionally, the power flow is specifically:

[0159] P PCC =P+ -P -

[0160] Among them, P P CC represents power flow, P + Represents the active power in the upstream direction of the node, P - Indicates the active power in the downstream direction of the node.

[0161] S5: Under the constraints, the optimal power flow curve is determined through the upper-level controller optimization.

[0162] Through the upper-level controller, the optimal power flow curve is determined using the particle swarm optimization algorithm with the upper-level control objective function as the target under the constraints.

[0163] Among them, Particle Swarm Optimization (PSO) is an optimization algorithm based on swarm intelligence. Inspired by the foraging behavior of bird flocks, the algorithm finds the optimal solution to the problem through information sharing and collaboration among individuals.

[0164] Specifically, the inverse of the above-layer control objective function is used as the fitness function of the particle swarm optimization algorithm.

[0165] Initialize the population, which contains multiple particles, each particle represents a feasible power flow curve.

[0166] Calculate the fitness of each particle.

[0167] Determine the individual optimal position and the global optimal position.

[0168]

[0169] in, represents the velocity of the ith particle at the t+lth iteration, represents the velocity of the i-th particle at the t-th iteration, ω represents the inertia weight factor, c1 represents the individual learning factor, c2 represents the global learning factor, r1 and r2 both represent random numbers between 0 and l, represents the individual optimal position of the i-th particle at the t-th iteration, represents the global optimal position of the i-th particle at the t-th iteration, represents the position of the i-th particle at the t-th iteration, represents the position of the i-th particle at the t+1th iteration.

[0170] By calculating the speed and position of the particle, combined with the guidance of the individual optimal position and the global optimal position, it is possible to effectively balance the global search and the local search, and quickly approach the optimal solution. Recalculate the fitness of each particle. When the fitness of the current particle is better than the individual optimal solution of the particle itself. Update the individual optimal solution of the particle itself. When the fitness of the current particle is worse than the individual optimal solution of the particle itself, update the individual optimal solution of the particle itself with a certain replacement probability. When the fitness of the current particle is better than the global optimal, update the global optimal solution.

[0171] It should be noted that when the fitness of the current particle is better than the individual optimal solution of the particle itself, the individual optimal solution is directly updated. This adaptive update mechanism can ensure that the particle swarm promptly follows the better solution found in the search process and improve the convergence speed of the algorithm. When the fitness of the current particle is worse than the individual optimal solution of the particle itself, the individual optimal solution is updated with a certain replacement probability. This mechanism introduces a certain degree of randomness to prevent the individual optimal solution from falling into the local optimum, thereby enhancing the global exploration capability.

[0172] Determine whether the maximum number of iterations has been reached. If so, output the power flow curve represented by the particle with the highest fitness value as the optimal power flow curve. Otherwise, return to continue iterating.

[0173] The replacement probability is:

[0174]

[0175] Among them, P represents the replacement probability, e represents the natural constant, δ() represents the fitness function, and C t Represents the temperature at the tth iteration. The temperature is updated according to the following formula:

[0176] T i+1 =εT i

[0177] Where ε represents the temperature drop coefficient, T i +1 indicates the temperature at the i+lth iteration, T i represents the temperature at the i-th iteration.

[0178] It should be noted that when the temperature is higher, it is easier to accept worse solutions, thereby increasing the chance of escaping the local optimum. As the number of iterations increases, the temperature gradually decreases, allowing the algorithm to perform more global searches in the early stages and gradually focus on local searches in the later stages, thereby improving the optimization effect.

[0179] The specific temperature reduction coefficient is:

[0180]

[0181] Among them, A1 represents the cooling oscillation factor of the first half cycle, B1 represents the cooling floating factor of the first half cycle, A2 represents the cooling oscillation factor of the second half cycle, and B2 represents the cooling floating factor of the second half cycle.

[0182] It should be noted that by dividing the cooling process into two cycles, the temperature changes in different stages can be finely controlled to adapt to different optimization requirements and search strategies. In each cycle, the temperature change not only considers exponential decay, but also combines sinusoidal oscillation and floating factors to make the temperature change smoother and more flexible. At the same time, the sinusoidal oscillation characteristics introduce a certain temperature fluctuation, avoiding premature convergence, increasing the diversity of the solution space, and reducing the risk of falling into the local optimum.

[0183] In the present invention, by using the particle swarm optimization algorithm in the upper layer control, its advantages such as strong global search capability, fast convergence speed, simple implementation, and adaptability to dynamic environments can be fully utilized to effectively improve the efficiency and quality of power grid dispatch optimization. This method can ensure that in a complex power grid operation environment, the optimal power flow solution that meets multiple constraints is found, thereby improving the overall performance and stability of the system.

[0184] S6: According to the predicted value of photovoltaic power generation, the lower-level control objective function is constructed, and the optimal grid dispatching plan is determined through the lower-level controller.

[0185] According to the predicted power generation value of each photovoltaic power generation module, the optimal power flow curve is used as the reference power flow, and the lower control objective function of the lower controller is constructed with the goal of reducing the deviation between the actual power flow and the reference power flow and reducing the scheduling cost.

[0186] In a possible implementation, the lower layer control objective function is specifically:

[0187]

[0188] Among them, J L represents the lower layer control objective function, min represents the minimum value, represents the power flow at time t, represents the reference power flow, T represents the total observation time, C t represents the dispatch cost at time t, μ1 represents the weight coefficient of the power flow deviation term, and μ2 represents the weight coefficient of the dispatch cost term.

[0189] Among them, those skilled in the art can set the weight coefficient μ1 of the power flow deviation term and the weight coefficient μ2 of the scheduling cost term according to actual conditions, and the present invention does not limit this.

[0190] In the present invention, by constructing the lower-level control objective function, a balance can be found between minimizing the power flow deviation and the dispatching cost, thereby improving the operating efficiency, stability and economy of the system. At the same time, by flexibly adjusting the weight coefficients to adapt to different operating requirements and strategies, it is ensured that the lower-level control can achieve the overall optimization goal of the upper-level control and improve the overall performance of the system.

[0191] Optionally, the scheduling cost is specifically:

[0192]

[0193] Among them, C t represents the scheduling cost at time t, represents the main grid electricity price at time t, N represents the amount of electricity purchased by the main power grid at time t. G Indicates the number of standby diesel generators, represents the unit power generation cost of the nth standby diesel generator, represents the power generation of the nth standby diesel generator at time t, N ESS represents the number of energy storage systems, represents the unit energy storage cost of the nth energy storage system, represents the storage power of the nth energy storage system at time t, N PV Indicates the number of photovoltaic power generation modules, represents the unit power generation cost of the nth photovoltaic power generation module, Represents the predicted power generation value of the nth photovoltaic power generation module at time t.

[0194] In the present invention, the actual cost of different power resources in the power grid can be fully considered, the dispatching decision can be optimized, and the economic benefits and operating efficiency of the system can be improved. At the same time, the cost parameters can be flexibly adjusted to adapt to different market environments and operating conditions, enhance the robustness and stability of the system, promote the priority use of renewable energy, and achieve efficient, economical and sustainable operation of the power grid.

[0195] Substitute the predicted power generation value of each photovoltaic power generation module predicted by the photovoltaic power generation prediction model into the lower-level controller, under the constraints of the constraints, and take the lower-level control objective function as the goal to determine the optimal grid scheduling plan.

[0196] It should be noted that the predicted power values ​​of each photovoltaic power generation module obtained by substituting them into the photovoltaic power generation power prediction model can reduce the uncertainty brought by photovoltaic power generation. Combined with the regulation capability of the energy storage system, it can maintain the stability of the power flow of the power grid, enhance the reliability of the system, and enable the scheduling plan to more accurately reflect the actual situation.

[0197] In one possible implementation, the predicted power generation value of each photovoltaic power generation module predicted by the photovoltaic power generation prediction model is substituted, and the optimal grid scheduling plan is determined by the lower-level controller under the constraints of the constraints and taking the lower-level control objective function as the target using the particle swarm optimization algorithm.

[0198] It should be noted that how to use the particle swarm optimization algorithm to determine the optimal power grid dispatching plan can refer to the relevant technical solutions for determining the optimal power flow curve using the particle swarm optimization algorithm mentioned above to avoid repetition.

[0199] The present invention further provides a power grid dispatching system 20 based on photovoltaic power generation prediction, which is applied to the above-mentioned power grid dispatching method based on photovoltaic power generation prediction, and comprises:

[0200] Processor 201.

[0201] The memory 202 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 201, a power grid dispatching method based on photovoltaic power generation prediction as in the method embodiment is implemented.

[0202] The power grid dispatching system 20 based on photovoltaic power generation prediction provided by the present invention can execute the above-mentioned power grid dispatching method based on photovoltaic power generation prediction and achieve the same or similar technical effects. To avoid repetition, the present invention will not go into details.

[0203] On the other hand, this embodiment also provides a power grid dispatching system based on photovoltaic power generation prediction, which includes:

[0204] The main power grid provides basic power supply and transmission functions.

[0205] Photovoltaic power generation module, using photovoltaic technology to generate electricity.

[0206] Backup diesel generators provide emergency power supply when the main grid or photovoltaic power generation is insufficient.

[0207] Energy storage systems store electricity when there is excess electricity and release it when there is a shortage of electricity, thus smoothing out the power output.

[0208] The upper controller is responsible for determining the optimal power flow curve, optimizing photovoltaic power generation utilization and reducing power fluctuations.

[0209] The lower-level controller optimizes the grid dispatching scheme according to the predicted value of photovoltaic power generation and the optimal power flow curve, reduces the deviation between the actual power flow and the reference power flow, and reduces the dispatching cost.

[0210] Processor, executes grid dispatch algorithms and control strategies.

[0211] The memory stores relevant computer-readable instructions and data for executing the scheduling method.

[0212] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0213] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0214] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0215] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0216] Example 2

[0217] The test environment includes a power grid that includes a main power grid, photovoltaic power generation modules, backup diesel generators and energy storage systems. The test scenario is set up in an industrial park, which is equipped with 5 photovoltaic power generation modules, each with a rated power of 50kW; 2 backup diesel generators, each with a rated power of 100kW; and 1 energy storage system with a capacity of 200kWh.

[0218] A two-layer control model of the power grid was constructed. The upper control model mainly includes the main power grid, photovoltaic power generation modules, backup diesel generators and energy storage systems, and is responsible for determining the optimal power flow curve. The lower control model is further refined on the basis of the upper control model to determine the optimal power grid dispatching plan.

[0219] The constraints of the two-layer control model of the power grid are defined, including power balance constraints, backup diesel generator power constraints, and energy storage system charge and discharge power constraints. These constraints ensure the stable operation of the power grid under various operating conditions. For example, the power balance constraint ensures that the total power generation of the power grid is equal to the total load demand at each time step, the backup diesel generator power constraint ensures that the generator power does not exceed its rated power, and the energy storage system charge and discharge power constraint limits the charge and discharge power of the energy storage system to within the allowed range of the equipment. A photovoltaic power generation prediction model is constructed. The model uses historical weather data and photovoltaic power generation data, extracts features through convolutional neural networks, and combines time series analysis methods to predict photovoltaic power generation in the next 24 hours. The input of the prediction model includes parameters such as temperature, humidity, and solar radiation, and the output is the power prediction value of each photovoltaic power generation module.

[0220] After obtaining the predicted value of photovoltaic power generation, the objective function of the upper-level controller is constructed with the goal of improving photovoltaic power generation utilization and reducing power fluctuations. This objective function optimizes the operation of the power grid by minimizing power flow fluctuations and maximizing photovoltaic power generation utilization.

[0221] The upper controller optimizes and determines the best power flow curve under the constraints. The particle swarm optimization algorithm is used to solve the problem and initialize the particle swarm. Each particle represents a feasible power flow curve. Through iterative calculation, the power flow curve represented by the particle with the highest fitness is finally determined.

[0222] According to the predicted value of photovoltaic power generation, the lower control objective function is constructed. The objective function optimizes the dispatching scheme of the power grid by minimizing the deviation between the actual power flow and the reference power flow, as well as the dispatching cost. Finally, the optimal power grid dispatching scheme is determined by the lower controller under the constraint conditions, and the same particle swarm optimization algorithm is used to solve it, ensuring that the dispatching scheme can achieve efficient and stable operation of the system under the premise of meeting various constraints. The experimental data are shown in Table 1.

[0223] Table 1 Experimental data table

[0224]

[0225] From the data in the table, it can be seen that the present invention effectively improves the utilization rate of photovoltaic power generation by constructing a double-layer control model and using a photovoltaic power generation power prediction model. Specifically, the photovoltaic power generation power is close to the rated power of 50kW, and the photovoltaic power generation power fluctuation range of each test object is between 45-50kW, with a small fluctuation range. In addition, the power of the standby diesel generator is basically stable between 90-97kW, and the charging and discharging power of the energy storage system is between 20-30kW, which meets the set constraints.

[0226] In terms of dispatching costs, by optimizing the dispatching scheme, the dispatching costs of all test objects were lower than 0.36 yuan / kWh, with the lowest being 0.31 yuan / kWh, showing the significant advantages of the present invention in reducing dispatching costs. In terms of power fluctuations, the power fluctuations of all test objects were controlled within 5%, showing a high system stability, with scores above 9.3 points, and the highest reaching 9.8 points.

[0227] By comparing with the existing technologies, the present invention shows obvious innovation and advantages in photovoltaic power generation prediction and grid dispatch optimization. The existing technologies often find it difficult to optimize both power fluctuation and dispatch cost at the same time, while the present invention achieves dual optimization of power fluctuation and dispatch cost through a double-layer control model and a particle swarm optimization algorithm. In addition, the photovoltaic power generation prediction model is used to make the dispatch plan more accurate, further improving the stability and economy of the system.

[0228] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A power grid dispatching method based on photovoltaic power generation power prediction, characterized in that: include: Establish a two-layer control model of the power grid including the main power grid, photovoltaic power generation modules, backup diesel generators and energy storage systems; Define the constraints of the two-layer control model of the power grid; Use the photovoltaic power generation prediction model to predict the power of each photovoltaic power generation module; To improve the utilization rate of photovoltaic power generation and reduce power fluctuations, the objective function of the upper-level controller is constructed; Under the constraints, the best power flow curve is determined by optimizing the upper controller; According to the predicted value of photovoltaic power generation, the lower-level control objective function is constructed, and the optimal grid dispatching plan is determined through the lower-level controller.

2. The grid dispatching method based on photovoltaic power generation prediction according to claim 1, characterized in that: The power grid double-layer control model includes an upper-layer controller for determining an optimal power flow curve and a lower-layer controller for determining an optimal power grid dispatching solution.

3. The grid dispatching method based on photovoltaic power generation prediction according to claim 2, characterized in that: Using a photovoltaic power generation prediction model to predict the power of each photovoltaic power generation module includes: the photovoltaic power generation prediction model includes a weather prediction unit and a power prediction unit; The weather prediction unit includes a first convolutional layer, a first gating unit, and a first attention layer connected in series, and the power prediction unit includes a second convolutional layer, a second gating unit, and a second attention layer connected in series; The first and second attention layers are both connected to the fully connected network, and the fully connected layer is connected to the prediction layer; Obtain historical weather data and historical photovoltaic power generation data. The weather features in the historical weather data are extracted through the first convolutional layer in the weather prediction unit, and the power features in the historical photovoltaic power generation data are extracted through the second convolutional layer in the power prediction unit; The weather hidden feature in the weather feature is extracted by the first gating unit in the weather prediction unit, and the power hidden feature in the power feature is extracted by the second gating unit in the power prediction unit. Attention weights are added to the weather hidden features at each moment through the first attention layer in the weather prediction unit, and attention weights are added to the power hidden features at each moment through the second attention layer in the power prediction unit. The weather hidden features and power hidden features after adding attention weights are concatenated through a fully connected network. Based on the concatenated fusion features, power prediction is performed through the prediction layer to obtain a preliminary predicted power. Based on the time periodicity and the nearest neighbor similarity, the preliminary predicted power is corrected to determine the final predicted power. According to the correlation predicted power, the preliminary predicted power is corrected to determine the final predicted power.

4. The grid dispatching method based on photovoltaic power generation prediction according to claim 3, characterized in that: The photovoltaic power generation prediction model is used to predict the power of each photovoltaic power generation module. The photovoltaic power generation prediction model also includes, with the goal of improving the utilization rate of photovoltaic power generation and reducing power fluctuations, constructing the objective function of the upper controller, including calculating the correlation coefficient between each adjacent day in the year before the prediction day: Among them, C i j represents the correlation coefficient between the predicted day and the jth adjacent day in the previous i years, T i j represents the time correlation coefficient of the jth adjacent day in the previous i years, k1 represents the year correlation coefficient, k2 represents the date correlation coefficient, S i j represents the weather correlation coefficient of the jth adjacent day in the previous i year, w i j represents the weather type of the jth consecutive day in the previous i year, w 00 Represents the weather type forecast value of the forecast day. The weather type forecast value of the forecast day is calculated based on the weather hidden features after adding the attention weight. V i j represents the randomness coefficient of the jth consecutive day in the previous i years, V 00 represents the randomness coefficient of the forecast day; The correlation prediction power is calculated based on the correlation coefficient between the predicted day and each adjacent day in previous years: Among them, P c Represents the correlation prediction power of the prediction day, P ij represents the power generation of the jth consecutive day in the previous i years, n represents the total number of years, and m represents the total number of consecutive days; The specific method of correcting the preliminary predicted power according to the correlation predicted power is as follows: Among them, P t represents the final predicted power at time t, represents the preliminary predicted power at time t, P ct represents the correlation prediction power at time t in the prediction day, and λ represents the correction coefficient.

5. The grid dispatching method based on photovoltaic power generation prediction according to claim 4, characterized in that: With the goal of improving the utilization rate of photovoltaic power generation and reducing power fluctuations, the objective function of the upper-level controller is constructed, and the upper-level control objective function is specifically: Among them, J H represents the upper control objective function, min represents the minimum value, represents the power flow at time t, represents the average power flow, T represents the total observation time, λ1 represents the weight coefficient of the power flow term, and λ2 represents the weight coefficient of the power flow fluctuation term. Among them, those skilled in the art can set the size of the weight coefficient λ1 of the power flow term and the weight coefficient λ2 of the power flow fluctuation term according to actual conditions; The power flow is specifically: P PCC =P + -P - Among them, P P CC represents power flow, P + Represents the active power in the upstream direction of the node, P - Indicates the active power in the downstream direction of the node; The smaller the power flow indicated by the power flow item, the higher the degree of self-sufficiency within the microgrid, and the greater the utilization rate of photovoltaic power generation.

6. The grid dispatching method based on photovoltaic power generation prediction according to claim 5, characterized in that: Under the constraint conditions, optimizing and determining the best power flow curve through the upper-level controller includes, through the upper-level controller, under the constraint conditions, taking the upper-level control objective function as a target, using a particle swarm optimization algorithm to determine the best power flow curve; Initialize the population, which contains multiple particles. Each particle represents a feasible power flow curve. Calculate the fitness of each particle, determine the individual optimal position and the global optimal position, and the formula is expressed as: in, represents the velocity of the ith particle at the t+lth iteration, represents the velocity of the ith particle at the tth iteration, ω represents the inertia weight factor, c1 represents the individual learning factor, c2 represents the global learning factor, r1 and r2 both represent random numbers between 0 and l, represents the individual optimal position of the i-th particle at the t-th iteration, represents the global optimal position of the i-th particle at the t-th iteration, represents the position of the i-th particle at the t-th iteration, represents the position of the i-th particle at the t+l-th iteration; Determine whether the maximum number of iterations has been reached. If so, output the power flow curve represented by the particle with the highest fitness value as the optimal power flow curve. Otherwise, return to continue iterating. The replacement probability is: Among them, P represents the replacement probability, e represents the natural constant, δ() represents the fitness function, and C t represents the temperature at the tth iteration; Update the temperature: T i+1 =εT i Where ε represents the temperature drop coefficient, T i +1 indicates the temperature at the i+1th iteration, T i represents the temperature at the i-th iteration; When the temperature is high, it is easier to accept a poor solution, thus increasing the chance of jumping out of the local optimum. As the number of iterations increases, the temperature gradually decreases, causing the algorithm to perform more global searches in the early stages and gradually focus on local searches in the later stages, improving the optimization effect. The specific temperature reduction coefficient is: Among them, A1 represents the cooling oscillation factor of the first half cycle, B1 represents the cooling floating factor of the first half cycle, A2 represents the cooling oscillation factor of the second half cycle, and B2 represents the cooling floating factor of the second half cycle.

7. The grid dispatching method based on photovoltaic power generation prediction according to claim 6, characterized in that: According to the predicted value of photovoltaic power generation, the lower-level control objective function is constructed, and the optimal grid dispatching scheme is determined by the lower-level controller. The specific lower-level control objective function is: Among them, J L represents the lower layer control objective function, min represents the minimum value, represents the power flow at time t, P ′ t ef represents the reference power flow, T represents the total observation time, C t represents the dispatch cost at time t, μ1 represents the weight coefficient of the power flow deviation term, and μ2 represents the weight coefficient of the dispatch cost term; By flexibly adjusting the weight coefficients to adapt to different operating requirements and strategies, it ensures that the lower-level control can achieve the overall optimization goal of the upper-level control and improve the overall performance of the system; The specific scheduling cost is: Among them, C t represents the scheduling cost at time t, represents the main grid electricity price at time t, N represents the amount of electricity purchased by the main power grid at time t. G Indicates the number of standby diesel generators, represents the unit power generation cost of the nth standby diesel generator, represents the power generation of the nth standby diesel generator at time t, N E SS represents the number of energy storage systems, represents the unit energy storage cost of the nth energy storage system, represents the storage power of the nth energy storage system at time t, N P V represents the number of photovoltaic power generation modules, represents the unit power generation cost of the nth photovoltaic power generation module, P n t P V represents the predicted power generation value of the nth photovoltaic power generation module at time t.

8. A power grid dispatching system based on photovoltaic power generation prediction using the method according to any one of claims 1 to 7, characterized in that: The main power grid provides basic power supply and transmission functions; Photovoltaic power generation modules, which use photovoltaic technology to generate electricity; Backup diesel generators to provide emergency power supply when the main power grid or photovoltaic power generation is insufficient; Energy storage systems store electricity when there is excess power and release it when there is a power shortage, thus smoothing the power output; The upper controller is responsible for determining the best power flow curve, optimizing the utilization of photovoltaic power generation and reducing power fluctuations; The lower-level controller optimizes the grid dispatching scheme based on the predicted value of photovoltaic power generation and the optimal power flow curve, reduces the deviation between the actual power flow and the reference power flow, and reduces the dispatching cost; Processor, which executes grid dispatch algorithms and control strategies; The memory stores relevant computer-readable instructions and data for executing the scheduling method.

9. A computer device comprising: Memory and processor; The memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of the power grid dispatching method based on photovoltaic power generation prediction are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of a power grid dispatching method based on photovoltaic power generation prediction are implemented.

Images