Large-scale photovoltaic access area voltage multi-time scale optimal trend prediction control method and system

Abstract

The invention discloses a voltage multi-time-scale optimal prediction control method and system for a large-scale photovoltaic access station area, and relates to the technical field of power systems, and the method comprises the steps: recognizing a regulation and control moment in a typical scene based on the voltage deviation of a grid-connected point, constructing a long-time-scale prediction control model driven by voltage sensitivity, and carrying out the prediction control of the long-time-scale prediction control model; generating an adjustable resource parameter adjusting quantity; constructing a short-time scale prediction control model driven by high-frequency sampling characteristic quantity, executing time sequence voltage prediction and error function optimization, and outputting a staged adjustment strategy; based on the voltage fluctuation evaluation index and the control method selection function, determining the time scale of the control strategy; according to the method, a multi-time-scale prediction control model is constructed, and a voltage fluctuation evaluation index and a control method selection function are introduced, so that dynamic switching of transformer area voltage control strategies is realized, the responsiveness and the adaptive ability of voltage regulation and control in a complex fluctuation scene are improved, and the voltage stability under photovoltaic high-proportion access is enhanced.

Description

Technical Field

[0001] This invention relates to the field of power system technology, specifically to a method and system for predictive control of voltage trends at multiple time scales in large-scale photovoltaic grid-connected distribution areas. Background Technology

[0002] Large-scale distributed photovoltaic (PV) grid connection has severely impacted the local absorption of power in distribution substations. The inability of loads to absorb PV power has led to frequent power flow surpluses and significant voltage increases at grid connection points in distribution substations. Simultaneously, the large-scale connection of flexible loads such as electric vehicles to distribution substations has exacerbated voltage quality problems due to their intermittent and unpredictable nature, adversely affecting transformer equipment, line losses, and personal safety, threatening the safe and reliable operation of power supply in distribution substations. Therefore, in the current scenario of numerous new power sources and loads coexisting, the need for comprehensive management of power quality in distribution substations is extremely urgent. With the further increase in the scale of future grid connection, ensuring the safe, economical, and reliable operation of the power system has become a pressing practical challenge that needs to be addressed.

[0003] Currently, most voltage quality trend optimization and iterative prediction models for distribution transformer areas start from voltage quality and fault impact. They assess the voltage quality stability of the system by establishing risk indicators and use relevant analysis methods to screen power supply quality status assessment indicators. Based on the correlation between the indicators and stability indices, fuzzy sets of indicators for different stability levels are determined, thereby achieving an intuitive assessment of the transient stability of the power system after a fault. However, the risk assessment factors have a high degree of subjectivity and fail to achieve advance control of voltage quality, resulting in difficulty in rapid response to complex scenarios or sudden changes in scenarios, and poor predictive control performance.

[0004] Existing power quality trend optimization and iterative prediction models, on the one hand, establish risk indicators based on fault characteristics and impacts, but do not consider or utilize customer-side power supply service data, and give little consideration to the temporal sequence of power quality indicators, usually only analyzing and determining their warning thresholds. Therefore, it is difficult to achieve early prevention and control of power quality risks. On the other hand, in the research on dynamic response control of regional adjustable resources, existing evaluation models pay little attention to the performance characteristics of new sources and loads and their impact mechanisms on the dynamic response capability of the system's adjustable resources, and cannot quantify the flexibility of the distribution system under massive new source and load disturbances. Summary of the Invention

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

[0006] Therefore, the technical problem solved by this invention is that existing methods for optimizing and predicting voltage quality trends in distribution areas lack advance control over power supply quality risks, have low model response speed, cannot adapt to voltage disturbances caused by photovoltaic output and load fluctuations, and have problems with how to achieve dynamic switching of multi-timescale control models under different voltage fluctuation scenarios.

[0007] To address the aforementioned technical problems, this invention provides the following technical solution: a multi-timescale optimal predictive control method for voltage in large-scale photovoltaic (PV) grid-connected distribution areas, comprising: identifying control moments under typical scenarios based on grid connection point voltage deviation; constructing a long-timescale predictive control model driven by voltage sensitivity; generating adjustable resource parameter adjustment quantities; constructing a short-timescale predictive control model driven by high-frequency sampling feature quantities; performing time-series voltage prediction and error function optimization; and outputting a phased adjustment strategy; determining the timescale of the control strategy based on voltage fluctuation evaluation indicators and a control method selection function; the timescale of the control strategy includes: constructing voltage fluctuation evaluation indicators based on the degree of PV power output fluctuation and load power randomness within the distribution area; and dynamically outputting corresponding control model switching identifiers using a control method selection function.

[0008] As a preferred embodiment of the large-scale photovoltaic access area voltage multi-timescale optimal prediction and control method described in this invention, the control time includes: using the time before the sampling point where the grid connection point voltage deviation is greater than a preset threshold under typical scenarios as the control benchmark, and using the time set for adjusting the adjustable resource parameters of the area to construct the control time set.

[0009] As a preferred embodiment of the large-scale photovoltaic access area voltage multi-timescale optimal prediction and control method described in this invention, the voltage sensitivity includes calculating the response rate to changes in grid connection point voltage based on the active and reactive power of the photovoltaic inverter, the transmission power of the flexible interconnect, the active and reactive power of the energy storage, and the inductance and capacitance of the LCL filter.

[0010] As a preferred embodiment of the large-scale photovoltaic access area voltage multi-timescale optimal prediction and control method described in this invention, the long-time-scale prediction and control model includes: based on the control time set, adjusting the parameters of the adjustable resources corresponding to the voltage sensitivity ranking results in sequence, and combining the adjustment capability constraints of the adjustable resources to generate a set of adjustment quantities that meet the voltage control target.

[0011] As a preferred embodiment of the large-scale photovoltaic access area voltage multi-timescale optimal prediction and control method described in this invention, the short-timescale prediction and control model includes: extracting feature quantities and control quantities in each sampling period, constructing a predicted voltage expression, generating the voltage prediction value for the next period, and constructing an error function based on the squared difference between the voltage prediction value and the measured voltage value.

[0012] As a preferred embodiment of the large-scale photovoltaic access area voltage multi-timescale optimal prediction and control method described in this invention, the phased adjustment strategy includes setting an initial adjustment percentage as the starting value, and releasing the remaining adjustment amount sequentially according to a set period and increment ratio.

[0013] As a preferred embodiment of the large-scale photovoltaic access area voltage multi-timescale optimal prediction and control method described in this invention, the time scale of the control strategy includes: constructing a voltage fluctuation evaluation index based on the degree of photovoltaic power output fluctuation and the randomness of load power; using a control method selection function to output the time scale switching identifier of the control strategy; and dynamically selecting between a long-timescale prediction and control model and a short-timescale prediction and control model.

[0014] Another objective of this invention is to provide a large-scale photovoltaic access area voltage multi-timescale optimal predictive control system. This system can dynamically switch between long-timescale and short-timescale predictive control models based on voltage fluctuation evaluation indicators and control method selection functions. This solves the problems of current voltage quality control technologies that rely on a single control time domain and cannot adapt to different fluctuation scenarios, including response lag and control mismatch.

[0015] As a preferred embodiment of the large-scale photovoltaic grid connection voltage multi-timescale optimization predictive control system of the present invention, it includes: a multi-timescale predictive control model construction module, a phased adjustment execution module, and a control strategy timescale switching module; the multi-timescale predictive control model construction module is used to extract the control time set based on the grid connection point voltage deviation, calculate and sort the voltage sensitivity of photovoltaic inverters, flexible interconnection, energy storage, and LCL filters, and construct a long-timescale predictive control model and a short-timescale predictive control model to complete parameter adjustment amount generation, voltage prediction, and error optimization, respectively; the phased adjustment execution module is used to perform phased release of the output adjustment amount, and form a phased adjustment strategy based on the initial adjustment amount percentage and period increment setting; the control strategy timescale switching module is used to construct a voltage fluctuation evaluation index based on the degree of photovoltaic output fluctuation and the randomness of load power, and determine the applicable timescale switching indicator of the control through the control method selection function, and dynamically switch between the long-timescale predictive control model and the short-timescale predictive control model.

[0016] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement a method for multi-timescale optimal predictive control of voltage in large-scale photovoltaic access areas.

[0017] A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of a method for multi-timescale optimal predictive control of voltage in large-scale photovoltaic access areas are implemented.

[0018] The beneficial effects of this invention are as follows: The large-scale photovoltaic access distribution area voltage multi-timescale optimal predictive control method provided by this invention, based on the long-timescale predictive control model constructed from the adjustable resources of the distribution area, achieves forward response and stable guidance of voltage trends for typical voltage deviation scenarios, significantly improving the accuracy and stability of voltage regulation under typical solar radiation and load patterns; the short-timescale predictive control model constructed based on high-frequency sampling characteristic quantities achieves millisecond-level iterative control for scenarios with rapid voltage fluctuations, reducing voltage fluctuation rate and enhancing dynamic adaptability to high-frequency disturbance scenarios; based on voltage fluctuation evaluation index and control method selection function, the method improves the judgment accuracy, switching timeliness and overall control closed-loop integrity of the control strategy under variable operating environments. This invention achieves higher collaborative performance and scenario coverage in terms of the accuracy of distribution area voltage regulation, the adaptability of the adjustment model and the timeliness of the control strategy. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is an overall flowchart of a method for predicting and controlling the voltage of a large-scale photovoltaic access area over multiple time scales, as provided in Embodiment 1 of the present invention.

[0021] Figure 2 This is a voltage quality trend optimization multi-timescale predictive control model diagram for a large-scale photovoltaic access area voltage multi-timescale optimal predictive control method provided in Embodiment 1 of the present invention.

[0022] Figure 3 This is a schematic diagram of the voltage quality trend optimization multi-timescale predictive control principle of a large-scale photovoltaic access area voltage multi-timescale optimal predictive control method provided in Embodiment 1 of the present invention.

[0023] Figure 4 The diagram below shows the logical framework of the voltage quality trend optimization multi-timescale predictive control model for a large-scale photovoltaic access area voltage multi-timescale optimal predictive control method provided in Embodiment 1 of the present invention. Detailed Implementation

[0024] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0025] Example 1, referring to Figures 1-4 As an embodiment of the present invention, a method for predictive control of voltage trends over multiple time scales in large-scale photovoltaic grid-connected distribution areas is provided, comprising:

[0026] S1: Based on the voltage deviation at the grid connection point, identify the control timing in typical scenarios, construct a long-term predictive control model driven by voltage sensitivity, and generate adjustable resource parameter adjustment amounts.

[0027] Furthermore, such as Figures 2-3 The control time includes the time before the sampling point where the voltage deviation at the grid connection point is greater than the preset threshold under typical scenarios, which is used as the control benchmark, and the set of times for adjusting the adjustable resource parameters of the transformer area is constructed.

[0028] Furthermore, in typical scenarios within a distribution area, photovoltaic output and user energy consumption exhibit certain patterns, and the corresponding voltage at the photovoltaic grid connection point in the distribution area also shows a pattern of change. Specifically, the moment preceding the sampling point where the voltage deviation exceeds a preset threshold by 7% in a typical scenario is selected as the control moment, expressed as:

[0029]

[0030] Among them, U D U represents the voltage deviation at a PCC (point of convergence between multiple power entities, such as distributed generation and the grid), i.e., the relative value of the voltage deviation after regulation, measured as a percentage of the voltage adjustment range, and should meet the restriction of ≤7%. PCC Indicates the voltage at the photovoltaic grid connection point in the affected area, U N It represents the nominal voltage at the grid connection point and is used to normalize the voltage deviation.

[0031] It should be noted that voltage sensitivity includes calculating the response rate to changes in grid connection point voltage based on the active and reactive power of the photovoltaic inverter, the transmission power of the flexible interconnect, the active and reactive power of the energy storage, and the inductance and capacitance of the LCL filter.

[0032] It should be noted that selecting the photovoltaic grid connection point in the distribution area as the key node, measuring the voltage waveform at the grid connection point using an oscilloscope, and changing the parameters of adjustable resources in the distribution area, such as the power of the photovoltaic inverter, the power of energy storage, and the power transmission of the flexible interconnection device, and measuring the grid connection point voltage before and after the change, combined with obtaining the voltage sensitivity parameters of each adjustable resource in the distribution area, is a preferred scheme, expressed as follows:

[0033]

[0034] Where, η F U represents the voltage sensitivity of the photovoltaic grid connection point corresponding to variable F. PCC The voltage at the photovoltaic grid connection point represents the voltage at which the photovoltaic grid connection point is adjusted, and F represents the adjustable resource variable of the transformer area that is regulated by changing the voltage at the photovoltaic grid connection point.

[0035] The voltage sensitivity set {η} is composed of the active and reactive power parameters of the photovoltaic inverter, the transmission power parameters of the flexible interconnection device, the energy storage power parameters, the inductance and capacitance parameters of the filter circuit, and the grid connection point voltage sensitivity corresponding to the adjustable resource parameters of the distribution area, and is expressed as:

[0036]

[0037] in, The grid connection point voltage sensitivity of the active and reactive power parameters of the photovoltaic inverter; The grid connection point voltage sensitivity for the transmission power parameters of flexible interconnect devices; η C The grid connection point voltage sensitivity corresponds to the inductance and capacitance parameters of the LCL filter; The grid connection point voltage sensitivity for active and reactive power parameters of energy storage.

[0038] It should also be noted that the constraint expressions for the active power and reactive power parameters of the photovoltaic inverter are as follows:

[0039]

[0040] Among them, P pv Q represents the active power regulation of a photovoltaic inverter. pv Let α represent the reactive power regulation of the photovoltaic inverter, α represent the photovoltaic active power regulation coefficient (used to prevent power fluctuations caused by overvoltage due to high photovoltaic penetration), and S represent the rated capacity of the photovoltaic inverter (used to constrain the upper limit of photovoltaic grid-connected power). The expression for the transmission power parameter constraint of the flexible interconnection device is as follows:

[0041]

[0042] Among them, P co Q represents the active power regulation of a flexible interconnect device.co S represents the active power regulation of a flexible interconnect device. sud Indicates the total available power capacity of the flexible interconnected transformer area; S Laud The total capacity required by the load in the flexible interconnected distribution area, and the constraint expressions for the inductance and capacitance parameters of the LCL filter, are expressed as follows:

[0043]

[0044] Where, η LCL The values ​​represent the adjustment amounts of the inductor and capacitor parameters of the LCL filter, where L represents the inductor parameter and C represents the capacitor parameter. The constraint expressions for the active and reactive power parameters of the energy storage are expressed as follows:

[0045]

[0046] Where SOC represents the energy storage capacity, SOC min State of Charge (SOC) represents the minimum energy storage capacity. max S represents the maximum energy storage capacity. sto Initial energy storage capacity, P sto Q represents the active power regulation of energy storage. sto S represents the reactive power regulation of energy storage. N This indicates the rated capacity of the energy storage.

[0047] It should also be noted that the long-term predictive control model includes, based on the control time set, adjusting the parameters of the adjustable resources corresponding to the voltage sensitivity ranking results in sequence, and combining the adjustment capability constraints of the adjustable resources to generate a set of adjustment quantities that meet the voltage control target.

[0048] It should also be noted that a preferred embodiment of the long-term predictive control model is expressed as follows:

[0049]

[0050] Where ΔU represents the voltage value to be regulated by predictive control, serving as the target adjustment amount for voltage deviation at the grid connection point; n represents the total number of adjustable resources participating in voltage regulation; i represents the index of the current number of adjustable resources; and ΔF... i Let ηi represent the adjustment amount of the adjustable resource parameter corresponding to the i-th voltage sensitivity, used as the control input. η1 represents the maximum value among all voltage sensitivities, i.e., the sensitivity of the adjustable resource with the strongest response capability. i U represents the voltage sensitivity of the adjustable resources of the i-th transformer area to the grid connection point voltage, reflecting the degree of response to voltage changes caused by a unit adjustment. DU represents the voltage deviation at a PCC (point of convergence between multiple power entities, such as distributed generation and the grid), i.e., the relative value of the voltage deviation after regulation, measured as a percentage of the voltage adjustment range, and should meet the restriction of ≤7%. N It represents the nominal voltage at the grid connection point and is used to normalize the voltage deviation.

[0051] It should also be noted that by identifying control times under typical scenarios based on grid connection point voltage deviation, a control time set is constructed. Combined with the voltage sensitivity set composed of photovoltaic inverters, flexible interconnection devices, energy storage systems, and LCL filter parameters, the long-term predictive control model is driven by sensitivity sorting to generate parameter adjustment quantities that meet the adjustable resource capacity constraints of the distribution area. This achieves coordinated optimization of voltage control time and control path, effectively distinguishing it from the single control strategy of existing methods that cannot simultaneously take into account control timing and response priority.

[0052] S2: Construct a short-time-scale predictive control model driven by high-frequency sampling features, perform time-series voltage prediction and error function optimization, and output a staged adjustment strategy.

[0053] Furthermore, such as Figures 2-3 The short-time-scale predictive control model includes extracting the feature quantities and control quantities within each sampling period, constructing the predicted voltage expression, generating the predicted voltage value for the next period, and constructing an error function based on the squared difference between the predicted voltage value and the measured voltage value.

[0054] Furthermore, a preferred scheme for constructing the predicted voltage expression is expressed as follows:

[0055]

[0056] Where u(t) represents the predicted grid-connected voltage value in the current sampling period, x(t) represents the system characteristic quantity extracted in the current sampling period, including photovoltaic output, load power, and historical voltage changes, Δu(t) represents the voltage regulation amount generated by predictive control in the current sampling period, and x(t+1) represents the characteristic quantity state in the next sampling period. As a preferred scheme of the predicted voltage data in the next sampling period, which is the evolution result of the system state, it is expressed as:

[0057] {u}={u(t+1∣t),u(t+2∣t+1),...u(t+H∣t-1+H)}

[0058] Where u(t+H|t-1+H) represents the predicted grid connection point voltage at time t+H, which is predicted from the measured value at time t-1+H; H represents the total number of cycles in the prediction control time window, which is used to define the time span of voltage prediction; and t represents the index of the current sampling cycle, which corresponds to the actual time of the system state and characteristic quantity.

[0059] A preferred scheme for constructing the error function is expressed as:

[0060]

[0061] Where u(t) represents the predicted grid connection voltage value in the current sampling period, u(t|t-1) represents the predicted grid connection voltage value at time t predicted from the measured value at time t-1, a represents the starting time index of the prediction evaluation, k represents the time window width of the prediction error evaluation, and (u(t)-u(t|t-1)) 2 This represents the squared error between the predicted and measured values, used to eliminate the directional influence of the sign and ensure the convergence of the error function.

[0062] It should be noted that the phased adjustment strategy includes setting an initial adjustment percentage as the starting value, and releasing the remaining adjustment amount sequentially according to the set period and increment ratio.

[0063] It should also be noted that a preferred scheme for the short-timescale predictive control model is expressed as follows:

[0064]

[0065] Where Δu(t) represents the voltage regulation generated by predictive control in the current sampling period, u(t+1|t)-u(t) represents the predicted grid connection point voltage at time t based on the measurement value at time t+1, n represents the total number of adjustable resources participating in voltage regulation, i represents the index of the current number of adjustable resources, and ΔF i η represents the adjustment amount of the adjustable resource parameter of the substation corresponding to the i-th voltage sensitivity at time t, and η1 represents the maximum value among all voltage sensitivities, that is, the sensitivity of the adjustment resource with the strongest response capability. i U represents the voltage sensitivity of the adjustable resources of the i-th transformer area to the grid connection point voltage, reflecting the degree of response to voltage changes caused by a unit adjustment. D U represents the voltage deviation at a PCC (point of convergence between multiple power entities, such as distributed generation and the grid), i.e., the relative value of the voltage deviation after regulation, measured as a percentage of the voltage adjustment range, and should meet the restriction of ≤7%. N It represents the nominal voltage at the grid connection point and is used to normalize the voltage deviation.

[0066] It should also be noted that by constructing a short-time-scale predictive control model driven by high-frequency sampling features, a voltage prediction expression is constructed based on the features and control quantities extracted from the sampling period. The output staged adjustment strategy is optimized by combining the error function between the predicted and measured values, so as to realize the dynamic release of voltage regulation on a periodic basis. This forms a complementary control mechanism with the long-time-scale model, which is effectively different from the single time-domain prediction method in the existing technology that lacks high-speed feedback and cannot adapt to the scenario of rapid voltage fluctuation.

[0067] S3: Determine the time scale of the control strategy based on the voltage fluctuation evaluation index and the control method selection function.

[0068] Furthermore, such as Figure 4 The time scale of the control strategy includes constructing a voltage fluctuation assessment index based on the degree of photovoltaic power output fluctuation and the randomness of load power, using a control method selection function to output the time scale switching identifier of the control strategy, and dynamically selecting between a long-time-scale predictive control model and a short-time-scale predictive control model.

[0069] It should be noted that PC represents a voltage fluctuation assessment index constructed based on the degree of photovoltaic power output fluctuation and the randomness of load power, expressed as:

[0070]

[0071] Among them, U PCC T represents the voltage at the photovoltaic grid connection point. reg P represents a sampling period, serving as a timescale reference for calculating voltage fluctuation assessment. L Q represents the active power currently consumed by the user. L P represents the reactive power currently consumed by the user. PV Q represents the active power generated by photovoltaics. PV ε represents the reactive power generated by photovoltaics. d The voltage fluctuation coefficient is used to comprehensively correct for the fluctuations in photovoltaic power output and the randomness of user energy consumption. F(PC) represents the corresponding control method selection function, expressed as:

[0072]

[0073] When F(PC) is 1, considering that the voltage fluctuations in the current distribution area are relatively frequent and complex, it is necessary to switch to a short-time-scale trend optimization predictive control method for regulation; when F(PC) is 0, considering that the voltage fluctuations in the current distribution area are within a certain range, the trend optimization can be completed based on the regulation time obtained from typical scenarios, and a long-time-scale trend optimization predictive control method is adopted for regulation.

[0074] It should also be noted that by constructing a voltage fluctuation evaluation index PC based on the degree of photovoltaic power output fluctuation and the randomness of load power, and by using the control method selection function F(PC) to output the time scale switching identifier of the control strategy, dynamic selection between long-time-scale predictive control model and short-time-scale predictive control model is realized, and an adaptive control strategy switching mechanism driven by voltage disturbance characteristics is established, which is effectively different from the single control method in the existing technology that has a fixed control time domain and lacks the ability to judge voltage fluctuation.

[0075] Example 2 is an embodiment of the present invention, which provides a large-scale photovoltaic access area voltage multi-timescale optimization prediction and control system, including a multi-timescale prediction and control model construction module, a staged adjustment execution module, and a control strategy timescale switching module.

[0076] The multi-timescale predictive control model construction module is used to extract the control time set based on the grid connection point voltage deviation, calculate and sort the voltage sensitivity of photovoltaic inverters, flexible interconnection, energy storage and LCL filtering, and construct long-timescale predictive control models and short-timescale predictive control models to complete parameter adjustment generation, voltage prediction and error optimization, respectively.

[0077] The phased adjustment execution module is used to perform phased release of the output adjustment amount, forming a phased adjustment strategy based on the initial adjustment amount percentage and the periodic increment setting.

[0078] The control strategy time scale switching module is used to construct voltage fluctuation evaluation index based on the degree of photovoltaic power output fluctuation and load power randomness. It determines the applicable time scale switching indicator through the control method selection function and dynamically switches between the long time scale predictive control model and the short time scale predictive control model.

Claims

1. A large-scale photovoltaic access substation voltage multi-time scale optimization prediction control method, characterized in that, The method comprises the following steps: Based on the voltage deviation of the grid-connected point, the regulation time in the typical scenario is identified, a long-time scale prediction control model driven by voltage sensitivity is constructed, and the adjustment amount of the adjustable resource parameter is generated; A short-time scale prediction control model driven by high-frequency sampling feature quantity is constructed, time series voltage prediction and error function optimization are performed, and a phased adjustment strategy is output; Based on the voltage fluctuation evaluation index and the control method selection function, the time scale of the control strategy is determined; The time scale of the control strategy includes constructing a voltage fluctuation evaluation index according to the photovoltaic output fluctuation degree and the load power randomness in the area, and dynamically outputting the corresponding control model switching identifier by using the control method selection function.

2. The large scale photovoltaic access substation voltage multi-time scale optimization prediction control method of claim 1, wherein: The regulation time includes: The time set of the adjustable resource parameter adjustment in the area is constructed by taking the time point before the sampling point where the voltage deviation of the grid-connected point in the typical scenario is greater than the preset threshold as the regulation reference, and the regulation time set is constructed.

3. The large scale photovoltaic access substation voltage multi-time scale optimization prediction control method of claim 2, wherein: The voltage sensitivity includes: According to the active and reactive power of the photovoltaic inverter, the transmission power of the flexible interconnection, the active and reactive power of the energy storage, and the inductance and capacitance of the LCL filter, the response rate to the voltage change of the grid-connected point is calculated respectively.

4. The large scale photovoltaic access substation voltage multi-time scale optimization prediction control method of claim 3, wherein: The long-time scale prediction control model includes: Based on the regulation time set, the adjustable resources corresponding to the sorting results of the voltage sensitivity are adjusted in sequence, and the adjustment amount set satisfying the voltage regulation target is generated by combining the adjustment capacity constraint conditions of the adjustable resources.

5. The large scale photovoltaic access substation voltage multi-time scale optimization prediction control method of claim 4, wherein: The short-time scale prediction control model includes: The feature quantity and control quantity in each sampling period are extracted, a prediction voltage expression is constructed, a next period voltage prediction value is generated, and an error function is constructed based on the square difference between the voltage prediction value and the voltage measured value.

6. The large scale photovoltaic access substation voltage multi-time scale optimization prediction control method of claim 5, wherein: The phased adjustment strategy includes: The initial adjustment amount percentage is set as the starting value, and the remaining adjustment amount is released in sequence according to the set period and the increment ratio.

7. The large scale photovoltaic access substation voltage multi-time scale optimization prediction control method of claim 6, wherein: The time scale of the control strategy includes: Based on the photovoltaic output fluctuation degree and the load power randomness, a voltage fluctuation evaluation index is constructed, a time scale switching identifier of the control strategy is output by using a control method selection function, and dynamic selection is performed between the long-time scale prediction control model and the short-time scale prediction control model.

8. A large-scale photovoltaic access substation voltage multi-time scale optimal prediction control system, adopting the large-scale photovoltaic access substation voltage multi-time scale optimal prediction control method according to any one of claims 1 to 7, characterized in that: It comprises a multi-time scale prediction control model construction module, a phased adjustment execution module, and a control strategy time scale switching module; The multi-time scale prediction control model construction module is used to extract the regulation time set based on the voltage deviation of the grid-connected point, calculate and sort the voltage sensitivity of the photovoltaic inverter, flexible interconnection, energy storage and LCL filter, construct the long-time scale prediction control model and the short-time scale prediction control model, and respectively complete the parameter adjustment amount generation, voltage prediction and error optimization; The phased adjustment execution module is used to execute the phased release of the output adjustment amount, and form a phased adjustment strategy according to the initial adjustment amount percentage and the period increment setting; The control strategy time scale switching module is used to construct a voltage fluctuation evaluation index according to the photovoltaic output fluctuation degree and the load power randomness, determine the time scale switching identifier suitable for regulation by using a control method selection function, and dynamically switch the long-time scale prediction control model and the short-time scale prediction control model. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor implements the steps of the large-scale photovoltaic access substation voltage multi-time scale optimal predictive control method of any one of claims 1 to 7 when the computer program is executed.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the large-scale photovoltaic access substation voltage multi-time scale optimal predictive control method of any one of claims 1 to 7.

Images