A distributed reactive power and voltage fast optimization control method

Abstract

The application provides a kind of distributed reactive power voltage fast optimization control method, it is related to power system optimization control technical field, the system data of initial time of period is obtained first in the application, call reactive power voltage optimization model in period to process, obtain period optimal operating point;Then based on the optimal operating point, measure the active power of new energy in each time in period, calculate new energy short-time fluctuation rate evaluation index and amplitude evaluation index;According to fluctuation rate evaluation index, judge whether to trigger new energy reactive power voltage optimization control;When triggering, according to amplitude evaluation index, select small fluctuation under distributed reactive power voltage fast optimization control based on linear sensitivity or large fluctuation under distributed reactive power voltage fast optimization control based on nonlinear sensitivity, realize the efficient dynamic tracking of system optimal operating point.The application can effectively improve the operation safety and economy of high proportion new energy power system.

Description

Technical Field

[0001] This invention relates to the field of power system optimization control technology, specifically to a distributed reactive power and voltage fast optimization control method. Background Technology

[0002] Building a new power system with renewable energy as the mainstay is an important way to achieve the current "dual carbon" goals. With the large-scale grid connection of renewable energy sources such as wind power and photovoltaics, the operating characteristics of the power system have undergone profound changes. The output of renewable energy has significant volatility and uncertainty. Its short-term and large-scale power fluctuations cause frequent changes in the power flow distribution of the system, making it difficult to maintain the optimal operating point of the system. This not only leads to increased active power losses and decreased operating economy, but also increases the risk of voltage exceeding limits, seriously threatening the safe and stable operation of the system.

[0003] In existing technologies, centralized reactive power optimization methods are typically used to address reactive power and voltage optimization control problems. These methods, based on global information, achieve system-level reactive power scheduling and voltage regulation by solving nonlinear power flow models, possessing the theoretical advantage of converging to the global optimum. However, centralized optimization methods are highly dependent on communication networks and central computing nodes. The required data acquisition, model solving, and control command issuance cycles are typically on the order of hours or minutes (e.g., 15 minutes or 5 minutes), making it difficult to meet the real-time and rapid control requirements of large-scale power systems under short-term fluctuations in renewable energy sources. On the other hand, while the fastest-responding primary voltage control in the power system can achieve rapid reactive power response within seconds, its control range is limited to a local area, lacking overall coordination of voltage deviation and economic distribution of reactive power across the entire network, making it difficult to achieve global optimization goals. Furthermore, traditional reactive power and voltage regulation equipment, such as capacitor banks and on-load tap-changing transformers, are limited by mechanical lifespan and response delays, making it difficult to efficiently match the regulation speed requirements of short-term fluctuations in renewable energy sources.

[0004] In recent years, distributed optimization control methods have gradually gained attention due to their advantages in reducing communication dependence, improving control response speed, and fully utilizing local adjustment resources. However, existing distributed control strategies are mostly based on fixed time periods or local voltage deviations for adjustment, failing to fully account for the impact of renewable energy fluctuation rates and amplitudes on system operation, and lacking refined identification of fluctuation characteristics and adaptive control mechanisms. Furthermore, in voltage sensitivity modeling, traditional methods often employ linear approximations, which are applicable to small fluctuations, but exhibit significant linear sensitivity errors in scenarios with large renewable energy fluctuations, making it difficult to accurately reflect the nonlinear coupling relationship between voltage and power, thus affecting the optimization control effect.

[0005] In view of the above, this application is hereby submitted. Summary of the Invention

[0006] This invention provides a distributed reactive power voltage fast optimization control method, which can at least partially improve the above-mentioned problems.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] A distributed reactive power voltage fast optimization control method, comprising:

[0009] Obtain system data at the initial moment of the time period, call the preset reactive voltage optimization model for the time period to process the system data at the initial moment of the time period, and obtain the optimal operating point for the time period;

[0010] Based on the optimal operating point of the time period, the active power of new energy at each moment within the time period is measured, and the short-term fluctuation rate evaluation index and amplitude evaluation index of new energy are calculated based on the active power of new energy.

[0011] Based on the fluctuation rate assessment index, determine whether to trigger the new energy reactive voltage optimization control and generate the judgment result;

[0012] When the judgment result is triggered, distributed reactive power voltage fast optimization control is performed based on the amplitude evaluation index and the preset distributed reactive power voltage optimization model. The distributed reactive power voltage fast optimization control includes distributed reactive power voltage fast optimization control based on linear sensitivity under small fluctuations in new energy sources and distributed reactive power voltage fast optimization control based on nonlinear sensitivity under large fluctuations in new energy sources.

[0013] In summary, this method first acquires system data at the initial moment of a time period and processes it using a reactive power and voltage optimization model for that period to obtain the optimal operating point. Second, based on this optimal operating point, it measures the active power of renewable energy at each moment within the time period and calculates the short-term fluctuation rate and amplitude evaluation indices for renewable energy. Then, it determines whether to trigger renewable energy reactive power and voltage optimization control based on the fluctuation rate evaluation indices. Finally, when triggered, it selects between distributed reactive power and voltage rapid optimization control based on linear sensitivity under small fluctuations or distributed reactive power and voltage rapid optimization control based on nonlinear sensitivity under large fluctuations, based on the amplitude evaluation indices, to achieve efficient dynamic tracking of the system's optimal operating point. This invention, through a dual evaluation mechanism of fluctuation rate and amplitude, achieves adaptive adjustment of the optimization cycle and precise control under different fluctuation scenarios, effectively improving the operational safety and economy of high-proportion renewable energy power systems.

[0014] Compared with existing technologies, this method has the following advantages: 1. In high-proportion renewable energy power systems, when the power of widely distributed renewable energy sources fluctuates, the distributed reactive power and voltage optimization model proposed in this paper can achieve precise optimization and adjustment of reactive power at each renewable energy grid connection point. Compared with traditional independent voltage regulation methods for renewable energy grid connection, this method can effectively reduce the disturbance impact of renewable energy active power fluctuations and reactive power regulation behavior on the overall system voltage stability and objective function, significantly improve the system voltage operation safety and economy, and ultimately achieve precise tracking of the system's optimal operating point. 2. The nonlinear mapping algorithm derived in this paper can effectively reveal the nonlinear coupling mechanism between voltage fluctuations and renewable energy power fluctuations. Compared with traditional analysis methods based on linear sensitivity, this algorithm significantly reduces voltage calculation errors under scenarios of large renewable energy power fluctuations; the algorithm is essentially an efficient approximate power flow calculation method, which can provide a fast and reliable calculation path for real-time acquisition of system voltage status under renewable energy fluctuation conditions, providing support for subsequent optimization control. 3. A distributed reactive power and voltage optimization control strategy based on dual evaluation of renewable energy fluctuation rate and amplitude is proposed: The optimization cycle is dynamically adjusted according to the renewable energy power fluctuation rate, which effectively reduces the redundant consumption of system computing resources and communication costs; Based on the renewable energy power fluctuation amplitude, reactive power and voltage optimization control strategies based on linear sensitivity and nonlinear sensitivity are constructed for two typical scenarios of small fluctuation and large fluctuation, respectively, which significantly improves the ability to track the optimal operating point of the system under different renewable energy fluctuation scenarios. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the distributed reactive power voltage rapid optimization control method provided in this embodiment of the invention.

[0016] Figure 2 This is a schematic diagram illustrating the main characteristics of short-term fluctuations in new energy provided in this embodiment of the invention.

[0017] Figure 3 This is a flowchart of the event triggering mechanism provided in the embodiments of the present invention.

[0018] Figure 4 This is a schematic diagram of the reactive power adjustment of new energy sources under different optimization models provided in the embodiments of the present invention.

[0019] Figure 5 This is a schematic diagram of active power network loss for different optimization models provided in the embodiments of the present invention.

[0020] Figure 6 These are schematic diagrams of system voltages for different optimization models provided in embodiments of the present invention.

[0021] Figure 7 This is a power output curve of new energy provided in an embodiment of the present invention.

[0022] Figure 8 This is a schematic diagram of the short-term fluctuation rate assessment of new energy provided in an embodiment of the present invention.

[0023] Figure 9 This is a schematic diagram of the short-term fluctuation amplitude assessment of new energy provided in an embodiment of the present invention.

[0024] Figure 10 This is a power output curve of new energy provided in an embodiment of the present invention.

[0025] Figure 11 These are reactive power curves of new energy sources obtained through different optimization methods provided in the embodiments of the present invention.

[0026] Figure 12 This is a system voltage surface diagram of the conventional optimization method provided in the embodiments of the present invention.

[0027] Figure 13 This is a system voltage surface diagram of the distributed reactive voltage fast optimization control strategy based on linear sensitivity provided in the embodiments of the present invention.

[0028] Figure 14 This is a system voltage surface diagram of a distributed reactive voltage fast optimization control strategy based on nonlinear sensitivity provided in an embodiment of the present invention.

[0029] Figure 15 This is a statistical chart showing the calculation time of the optimal power flow method and the distributed reactive voltage fast optimization control strategy based on linear sensitivity provided in the embodiments of the present invention.

[0030] Figure 16 This is a statistical chart showing the calculation time of a distributed reactive voltage fast optimization control strategy based on nonlinear sensitivity, provided in an embodiment of the present invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0032] refer to Figure 1 As shown, the first embodiment of the present invention discloses a distributed reactive power voltage fast optimization control method, which can be executed by a distributed reactive power voltage fast optimization control device (hereinafter referred to as the control device), specifically, by one or more processors within the control device, to implement the following method:

[0033] S1, Obtain system data at the initial moment of the time period, call the preset reactive voltage optimization model for the time period to process the system data at the initial moment of the time period, and obtain the optimal operating point for the time period;

[0034] The steps for constructing the reactive power voltage optimization model for the specified time period are as follows: Using the minimum active power network loss as the objective function, construct the active power network loss objective function: ,in, For the system node set, Let be the voltage amplitude at node i. Let i be the grounding conductance. For the system branch set, Let J be the voltage magnitude at node j. Let be the voltage phase angle difference between node i and node j. The conductance of branch l;

[0035] With minimizing voltage deviation as the objective function, the voltage deviation objective function is constructed as follows: , Let be the reference voltage of node i; and then perform a weighted summation of the active power loss objective function and the voltage deviation objective function to obtain the overall objective function. , These are the weighting coefficients for the active power network loss objective function. These are the weighting coefficients for the voltage deviation objective function; in practical applications, if operational economy is a greater priority, they can be increased. The value of can be increased if voltage safety is a greater concern. The possible values ​​of ;

[0036] Establish equality constraints, which are power balance equations. ,in, Let be the active power of the new energy source at node i. Let be the fluctuation of the active power of new energy sources at node i. Let be the active power of the generator at node i. Let be the active power of the load at node i. This indicates that node j is connected to node i. Let the mutual conductance of nodes i and j be the nodal admittance matrix. Let the mutual amperage of nodes i and j in the nodal admittance matrix be denoted as . Let i be the reactive power of the new energy source. Let i be the SVC reactive power. Let i be the reactive power of the generator. Let be the reactive power of the load at node i;

[0037] Establish inequality constraints, including branch transmission power constraints, voltage amplitude constraints, voltage phase angle constraints, generator reactive power constraints, new energy reactive power constraints, and SVC reactive power constraints;

[0038] The allowable range for discrete variable constraints established based on discrete variables is as follows: , , Among them, discrete variables include the transformer turns ratio and the number of capacitor banks in operation. This is the lower limit of the transformer turns ratio for branch l. Let l be the transformer turns ratio. This represents the upper limit of the transformer turns ratio for branch l. This represents the lower limit of the reactive power of the capacitor bank at node i. Let be the reactive power of the capacitor bank at node i. This represents the upper limit of the reactive power of the capacitor bank at node i. This represents the number of capacitor banks in operation. Let be the reactive power capacity of a single capacitor bank at node i;

[0039] Based on the overall objective function, equality constraints, inequality constraints, and discrete variable constraints, the basic model for reactive power voltage optimization is obtained by programming in the pre-set Matlab software and calling the Cplex function to solve the problem.

[0040] Based on the reactive power and voltage optimization model, the fluctuation of active power of new energy sources, the fluctuation of voltage amplitude, and the fluctuation of voltage phase angle are all set to 0 to obtain the reactive power and voltage optimization model for the time period.

[0041] The specific limitation of the inequality constraint is: the transmission power of a branch does not exceed its maximum transmission capacity. , The active power transmitted by branch l. The reactive power transmitted by branch l, The maximum transmission capacity of branch l;

[0042] The voltage amplitude and voltage phase angle at the node should be within the following ranges: , , , This represents the lower limit of the voltage amplitude at node i. Let be the fluctuation value of the voltage amplitude at node i. This represents the upper limit of the voltage amplitude at node i. This is the lower limit of the voltage phase angle at node i. Let i be the voltage phase angle at node i. Let be the fluctuation value of the voltage phase angle at node i. Let be the upper limit of the voltage phase angle at node i. Let represent the functional relationship between the voltage amplitude at node i and the active power fluctuation of each new energy source. Let be the functional relationship between the voltage phase angle of node i and the active power fluctuation of each renewable energy source, where m is the number of renewable energy nodes. This refers to the fluctuation in active power of new energy sources at the first node. This refers to the fluctuation in active power of new energy sources at the second node. Let be the fluctuation of active power of new energy sources at the m-th node;

[0043] The reactive power of generators, reactive power of new energy sources, and reactive power of SVC should be within the following ranges: , , , This represents the lower limit of the reactive power of the generator at node i. Let i be the reactive power of the generator. Let be the upper limit of the reactive power of the generator at node i. This represents the lower limit of the reactive power of new energy sources at node i. Let i be the reactive power of the new energy source. Let i be the upper limit of the reactive power of new energy sources. This is the lower limit of the SVC reactive power at node i. Let i be the SVC reactive power. is the upper limit of the SVC reactive power of node i.

[0044] Specifically, in this embodiment, firstly, system data at the initial moment of the time period is acquired, and a preset reactive power and voltage optimization model for the time period is called to process the system data at the initial moment of the time period to obtain the optimal operating point for the time period. The construction process of this reactive power and voltage optimization model is as follows: With the goal of minimizing the combined active power loss and voltage deviation, a basic reactive power and voltage optimization model based on AC power flow is constructed. This model belongs to a nonlinear programming problem, using the fluctuation of active power from new energy sources as random variables, voltage amplitude and phase angle as state variables, and generator reactive power, new energy reactive power, and the output of static var compensators (SVC, for example) as control variables.

[0045] Specifically, the objective function of the reactive power voltage optimization basic model is established. Considering that reducing active power network loss can improve the economic efficiency of system operation, the objective function is to minimize active power network loss; considering that fluctuations in new energy sources can cause voltage fluctuations or even exceed limits, the objective function is to minimize voltage deviation. The established model is a multi-objective optimization problem. By weighted summing of each objective, the overall objective function of the reactive power voltage optimization basic model can be obtained. Subsequently, equality constraints, inequality constraints, and discrete variable constraints of the reactive power voltage optimization basic model are established sequentially. Finally, the Cplex function is called in Matlab to solve the reactive power voltage optimization basic model. This method is implemented in the improved IEEE-39 node system. With the combined goal of minimizing active power network loss and voltage deviation, a reactive power voltage optimization basic model based on AC power flow is constructed. This model belongs to a nonlinear programming problem and is denoted as Model I.

[0046] Since the fundamental model for reactive power-voltage optimization based on AC power flow is a nonlinear stochastic programming problem, the solution process is complex and difficult to efficiently meet the real-time requirements of large-scale system optimization under short-term fluctuations in renewable energy sources. To address this, a time decoupling mechanism is proposed to decouple the above model into a time-period reactive power-voltage optimization model (e.g., 5 min) and a short-term linear reactive power-voltage optimization control model (e.g., 5 s).

[0047] The objective function and constraints of the reactive power and voltage optimization model are the same as those of the basic reactive power and voltage optimization model, but the impact of short-term fluctuations in renewable energy is not considered (i.e., the fluctuations in renewable energy active power, voltage amplitude, and voltage phase angle are all zero). This model is a deterministic nonlinear programming model, and its time scale is T. k The time period is referred to as 5 minutes. That is, a time decoupling mechanism is used to decouple Model I into a reactive voltage optimization model for a time period (5 minutes) and a linear reactive voltage optimization control model for a short time period (5 seconds). The linear reactive voltage optimization control model for the short time period is denoted as Model II.

[0048] S2, based on the optimal operating point of the time period, measures the active power of new energy at each moment within the time period, and calculates the short-term fluctuation rate evaluation index and amplitude evaluation index of new energy based on the active power of new energy.

[0049] Specifically, step S2 further includes: defining a short-term fluctuation rate assessment index for a single renewable energy source as the degree of deviation in the system's operating state caused by changes in the active power of the renewable energy source at adjacent times, the formula of which is: , For a moment The volatility rate assessment index is a comprehensive index. For ease of analysis, 1 is used as the base value, and the volatility rate assessment index is converted to a per-unit value. For node j at time... The active power of new energy sources below For node j at time... The active power of new energy sources below , All are coefficients. For the slight increase in network loss, This is an average calculation; when a volatility rate assessment reference value is selected. Then, by comparing and evaluating indicators Compared with reference value The magnitude of the fluctuation rate can be used to determine whether the system is in a scenario with a fast or slow fluctuation rate at that moment.

[0050] Calculating time using voltage sensitivity Voltage fluctuation at grid connection point caused by new energy fluctuations , For node j at time... Fluctuation of active power of new energy sources Let $\frac{ ... For node j in time period The new energy benchmark active power;

[0051] Calculate the actual voltage fluctuation value at the grid connection point The amplitude evaluation index is defined as the calculation error of the sensitivity analysis method. , For a moment The amplitude evaluation index has the same dimensions as voltage. In practice, nonlinear sensitivity can be used to approximate the actual voltage fluctuation value to improve calculation speed. When a fluctuation amplitude evaluation reference value is selected... Then, by comparing and evaluating indicators and reference value The magnitude of the fluctuation can be used to determine whether the system is experiencing large or small fluctuations at that moment.

[0052] In this embodiment, short-term fluctuation rate and amplitude evaluation indicators for new energy sources are established. Based on the evaluation of the new energy fluctuation rate, an event triggering mechanism for dynamic adjustment of reactive power optimization cycle is constructed. Based on the fluctuation amplitude evaluation, a distributed reactive power and voltage optimization control strategy based on linear and nonlinear sensitivity is constructed to achieve efficient dynamic tracking of the optimal operating point of the system under different new energy fluctuation amplitudes, thereby improving the operational safety and economy of high-proportion new energy power systems.

[0053] Taking a single renewable energy node as the object, if the combined change in voltage and active power network loss caused by fluctuations in the active power of renewable energy at adjacent times is small, then the fluctuation rate is considered slow. Figure 2 In (a) time t n Similarly, at time t m The fluctuation rate is fast. The fluctuation amplitude is used to describe the magnitude of the deviation of the renewable energy active power output from the benchmark active power. If the fluctuation in renewable energy active power output at a certain moment does not cause a significant error in the sensitivity analysis method, the fluctuation amplitude is considered small. Figure 2 (b) at time t n Sensitivity analysis, mathematically speaking, is a linear mapping algorithm. Based on linearization at the operating point, it uses sensitivity to analyze voltage changes caused by variations in node power. Similarly, at time t... m The fluctuation range is large.

[0054] Specifically, to quantify the comprehensive impact of short-term fluctuations in renewable energy on voltage and operational economy, and to correspond with the event triggering mechanism, a single renewable energy short-term fluctuation rate assessment index is defined as the degree of deviation in system operating state (active power network loss and voltage) caused by changes in renewable energy active power at adjacent times. Secondly, to correspond with the proposed distributed reactive power and voltage rapid optimization control strategy, the magnitude of renewable energy short-term fluctuations is evaluated using whether renewable energy fluctuations cause significant errors in the sensitivity analysis method. First, voltage sensitivity is used to calculate the grid connection point voltage fluctuation value caused by renewable energy fluctuations. Then, the actual voltage fluctuation value at the grid connection point is calculated, and the assessment index for the magnitude of renewable energy short-term fluctuations is defined as the calculation error of the sensitivity analysis method.

[0055] In simple terms, this step constructs short-term fluctuation rate and amplitude assessment indicators for a single renewable energy source, proposes a time-triggered mechanism based on fluctuation rate assessment, and designs distributed reactive power and voltage optimization control strategies based on linear and nonlinear sensitivity for small and large fluctuations of the renewable energy source, respectively. Specifically, it assumes that nodes {4, 15, 32, 33, 36, 37} of the IEEE-39 node system are connected to renewable energy sources, and their active power output is as follows: Figure 7 As shown, the initial active power is consistent with the data at the original location of the IEEE-39 node standard test system, which are {500, 320, 800, 630, 700, 940} MW respectively.

[0056] Fluctuation rate analysis: The short-time fluctuation rate of new energy refers to the degree of deviation in the system's operating state caused by changes in the active power of new energy at adjacent moments. Its evaluation index is... The impact of renewable energy fluctuations on active power grid losses and voltage is comprehensively considered. Taking renewable energy nodes {4, 33} as an example, their fluctuation rates are evaluated, with reference values... Take 0.5 × 10⁻⁴ pu. From Figure 7 It can be seen that the evaluation index of the new energy fluctuation rate at node 4 In timing simulations, the overall value is less than the reference value. Therefore, this example considers the renewable energy fluctuation rate at node 4 to be slow. Evaluation index for renewable energy fluctuation rate at node 33. Overall greater than the reference value This example assumes that the renewable energy fluctuation rate at node 33 is fast. There are two reasons for the discrepancies in the renewable energy fluctuation rate assessment results: firstly, the magnitude of the renewable energy active power fluctuation varies; secondly, the sensitivity or incremental rate of system voltage and active power losses to changes in renewable energy active power at different locations differs.

[0057] Fluctuation amplitude analysis uses the criterion of whether short-term fluctuations in renewable energy sources cause significant errors in sensitivity analysis to assess the magnitude of fluctuations. Taking renewable energy access nodes {36, 37} as an example, assuming the active power fluctuation of renewable energy gradually increases from -80MW to +80MW in 1MW intervals, sensitivity analysis is used to calculate the voltage fluctuation value of the renewable energy node and compare it with the actual voltage fluctuation value. Figure 9 It can be seen that as the fluctuation of active power from new energy sources increases, the calculation error of the sensitivity analysis method gradually increases. When the reference value... Using 2×10⁻⁴ PU, a graph shows that the straight lines of the fluctuation amplitude assessment index and the reference value intersect at the red dot. The horizontal axis corresponding to the red dot is the dividing point for distinguishing the magnitude of the new energy fluctuations. Within the dividing point range, the sensitivity analysis method has a small calculation error; outside the range, the sensitivity analysis method has a large calculation error. Based on this, this example considers the condition where the overall fluctuation of the new energy active power within 5 minutes is less than 40MW as a small fluctuation scenario, and the condition where the overall fluctuation of the new energy active power within 5 minutes is greater than 40MW as a large fluctuation scenario.

[0058] S3, determine whether to trigger the new energy reactive voltage optimization control based on the fluctuation rate evaluation index, and generate the judgment result;

[0059] Specifically, step S3 further includes: constructing a judgment index based on the volatility rate assessment index. To assess the reference time of the last execution of renewable energy reactive power voltage optimization control. At the time The degree to which fluctuations in the active power of new energy sources cause deviations in the system's operating state. The fluctuation rate evaluation index at time t. Trigger threshold for optimizing the reactive voltage control event triggering mechanism for new energy sources; It is an evaluation index for the degree of deviation in system operating state caused by fluctuations in a single new energy source, taking into account the comprehensive impact on voltage and active power network losses;

[0060] When the judgment indicator is determined to be greater than or equal to the trigger threshold, the judgment result is triggered, and distributed reactive power voltage rapid optimization control is performed to quickly adjust the reactive power of new energy sources, while simultaneously setting the reference time... ;

[0061] When the judgment indicator is determined to be less than the trigger threshold, the reactive power of the new energy source remains unchanged, and the process proceeds to the next moment.

[0062] In this embodiment, during time period T k The process of the new energy reactive power voltage optimization control event triggering mechanism is as follows: Figure 3 As shown in the figure, N tThis represents the number of moments contained in a time period (if the time period is 5 minutes and each moment is 5 seconds, then N is...). t (60). Taking a single new energy node (corresponding to the grid connection node of a new energy power station) as the object, its controller measures its own active power locally at each moment, and constructs a judgment index based on the fluctuation rate assessment.

[0063] If the judgment indicator is greater than or equal to the trigger threshold, the new energy node controller executes new energy reactive power voltage optimization control, quickly adjusts the new energy reactive power, and simultaneously sets the reference time... If the judgment indicator is less than the threshold, the reactive power of the new energy source remains unchanged, and the process proceeds to the next time step. The new energy reactive power voltage optimization control event triggering mechanism can adaptively adjust the optimization cycle based on the fluctuation of individual new energy sources, thereby reducing the number of reactive power adjustments and control costs.

[0064] Choosing an appropriate threshold is key to designing an event triggering mechanism; let the coefficient... , and All are 1, by setting a coefficient To distinguish the thresholds at different new energy nodes. As shown in Table 1, the coefficients... The values ​​are assigned based on the degree of impact of renewable energy fluctuations on the system's operating status, with the coefficient corresponding to renewable energy nodes having a greater impact. The values ​​are also large, such as node 37, which corresponds to the coefficient of new energy nodes with a small impact. The values ​​are also small, such as node 15. According to the statistics in Table 2, in the 2-hour time-series simulation (1440 time points), the number of reactive power optimization control executions for nodes {4, 15, 32, 33, 36, 37} are {125, 197, 332, 263, 338, 414} times, respectively, which are reductions of {91.32%, 86.32%, 76.95%, 81.74%, 76.53%, 71.25%}. The average number of reactive power optimization control executions is reduced by 80.69%, meaning that the renewable energy node controller executes renewable energy reactive voltage optimization control once every 25 seconds on average. Therefore, the event triggering mechanism can effectively reduce the number of renewable energy reactive voltage optimization control operations and the cost of renewable energy reactive power control.

[0065] Table 1. Statistics on the Implementation of Reactive Power Optimization Control for New Energy Sources:

[0066]

[0067] S4, when the judgment result is triggered, distributed reactive power voltage fast optimization control is performed according to the amplitude evaluation index and the preset distributed reactive power voltage optimization model. The distributed reactive power voltage fast optimization control includes distributed reactive power voltage fast optimization control based on linear sensitivity under small fluctuations in new energy and distributed reactive power voltage fast optimization control based on nonlinear sensitivity under large fluctuations in new energy.

[0068] The construction steps of the distributed reactive power-voltage optimization model are as follows: The active power network loss objective function of the basic reactive power-voltage optimization model is used as the original objective function, and linearization is performed to obtain the objective function of the short-time linear reactive power-voltage optimization control model. , For the set of new energy nodes in the system, Let be the incremental rate of change of active power at node j in the original objective function. For the short-term fluctuation of new energy at node j at time t, Let be the incremental rate of change of reactive power at node j (node ​​j is a new energy node) of the original objective function. Let be the reactive power adjustment of new energy sources at node j at time t. For the SVC node set in the system, Let $k$ be the incremental rate of the reactive power change at node $k$ (node ​​$k$ is an SVC node). Let be the SVC reactive power regulation of node k at time t;

[0069] Linearizing the equality constraints of the reactive voltage optimization basic model, the node voltage fluctuation values ​​are obtained as follows: , Let be the fluctuation value of the voltage amplitude at node i at time t. Let be the voltage phase angle fluctuation value of node i at time t. Let $\frac{j}{j}$ be the amount by which a unit change in active power from a renewable energy source causes a change in the voltage amplitude at node $i$. Let $\frac{j}{j}$ be the amount by which a unit change in reactive power at node $i$ causes a change in the voltage amplitude at node $i$. Let $\frac{j}{j}$ be the change in voltage phase angle at node $i$ caused by a unit change in active power from the renewable energy source. Let $\frac{j}{j}$ be the change in voltage phase angle at node $i$ caused by a unit change in reactive power from the renewable energy source. Let be the sensitivity of the voltage amplitude at node i to the change in reactive power of the SVC at node k. The sensitivity of the voltage phase angle at node i to the change in reactive power of the SVC at node k;

[0070] The inequality constraints for constructing a short-time linear reactive power-voltage optimization control model are given, where the voltage exceedance that should not occur at any node i is: , The voltage amplitude obtained by optimizing the reactive voltage at node i during the time period. The voltage phase angle obtained by optimizing the reactive voltage for node i during the time period; the reactive power of new energy sources and the reactive power of SVC should be within the following range: , The reactive power of new energy obtained by solving the reactive voltage optimization model for node j during the time period is... This represents the lower limit of the reactive power at node j during the specified time period. Let be the upper limit of the reactive power of node j during the time period. The SVC reactive power obtained by solving the reactive voltage optimization model for node k during the time period is... This represents the lower limit of the reactive power of the SVC during the time period at node k. This represents the upper limit of the reactive power of the SVC during the time period at node k;

[0071] Based on the objective function The linearized equality constraints and the inequality constraints of the short-time linear reactive power-voltage optimization control model are used to obtain the short-time linear reactive power-voltage optimization control model.

[0072] The objective function of the short-time linear reactive power-voltage optimization control model is represented by the sum of M sub-objective functions in partitioned form, as shown in the formula: , The sum of M sub-objective functions. For the first sub-objective function, This is the second sub-objective function. Let M be the Mth sub-objective function, and M be the number of sub-objective functions. This refers to the fluctuation in active power of new energy sources. For reactive power regulation of new energy sources and SVC, Let be the fluctuation of active power from new energy sources in the first sub-objective function. The second sub-objective function represents the fluctuation in active power from new energy sources. Let be the fluctuation of active power from new energy sources in the Mth sub-objective function. The reactive power regulation of new energy sources and SVC is the first sub-objective function. The reactive power regulation of new energy sources and SVC is the second sub-objective function. For the reactive power regulation of new energy sources and SVC in the Mth sub-objective function;

[0073] The node voltage fluctuation values ​​of the short-time linear reactive power-voltage optimization control model are represented by partitions, resulting in a block matrix form. , This represents the voltage fluctuation value corresponding to the first sub-objective function. This represents the voltage fluctuation value corresponding to the second sub-objective function. This represents the voltage fluctuation value corresponding to the Mth sub-objective function. The sensitivity block matrix of voltage fluctuation in the Mth partition to active power fluctuation of new energy in the Mth partition is the sensitivity of voltage fluctuation value of the Mth sub-objective function to reactive power regulation of new energy and SVC in the Mth sub-objective function.

[0074] Substituting the block matrix form into the voltage limit violation formula of the short-time linear reactive power-voltage optimization control model, and expressing it as multiple sub-constraint forms, yields the system voltage inequality constraint set. , This represents the voltage fluctuation value. This is the first voltage inequality constraint form. This is the second voltage inequality constraint form. Let M be the form of the voltage inequality constraint, where each partition has a corresponding form of voltage inequality constraint.

[0075] The ranges of renewable energy reactive power and SVC reactive power in the short-time linear reactive power-voltage optimization control model are expressed as M sub-constraints, resulting in the system reactive power inequality constraint set. , This is the first constraint form of the reactive power inequality. This is the second reactive power inequality constraint form. Let M be the reactive power inequality constraint form, where each partition has a corresponding reactive power inequality constraint form;

[0076] Based on the objective function after partitioning, the block matrix form, the system voltage inequality constraint set, and the system reactive power inequality constraint set, a spatial decoupling mechanism is obtained, and a distributed reactive power and voltage optimization model is constructed based on the spatial decoupling mechanism.

[0077] Specifically, step S4 further includes the following steps for rapid optimization control of distributed reactive power and voltage based on linear sensitivity under small fluctuations in renewable energy: at a single renewable energy grid-connected node, a distributed reactive power and voltage optimization model constructed based on a spatial decoupling mechanism is used for rapid optimization control of distributed reactive power and voltage, and a voltage reference variable is introduced. Its value is the maximum voltage obtained by solving the reactive power and voltage optimization model for the time period (taking the voltage close to the upper limit as an example), so that the system voltage after the fluctuation of the active power of new energy and the adjustment of reactive power at this node is not greater than the voltage reference variable.

[0078] The objective function of the distributed reactive power voltage optimization model under small fluctuations in new energy sources is: The constraints of the distributed reactive power voltage optimization model under small fluctuations in new energy sources are: the node voltages satisfy: The permissible range for reactive power of new energy sources is: .

[0079] In this embodiment, a short-time linear reactive power-voltage optimization control model is first constructed. This short-time linear reactive power-voltage optimization control model is a linear approximation of the basic reactive power-voltage optimization model. The increment of the original objective function is used as the new objective function, voltage fluctuation is used as the state variable, and the reactive power adjustment of new energy sources and static var compensators (SVCs, for example) are used as control variables. For ease of analysis, this model assumes that short-time fluctuations in new energy sources will not cause the line transmission power to exceed its limit, i.e., the influence of the maximum transmission capacity formula is ignored. Simultaneously, generator nodes are considered as PV nodes, autonomously maintaining a constant voltage; therefore, the model does not consider controlling the reactive power of generators, i.e., the influence of the generator reactive power range formula is ignored. Next, the objective function of the basic reactive power-voltage optimization model is linearized, and the power balance equation of the basic reactive power-voltage optimization model is linearized, constructing its inequality constraints. Using a time decoupling mechanism, the basic reactive power-voltage optimization model based on AC power flow can be decomposed into a time-period (e.g., 5-minute) reactive power-voltage optimization model and a short-time (e.g., 5-second) linear reactive power-voltage optimization control model. The latter's objective function and constraints are both linear functions, thus it belongs to the linear programming problem and has a low solution complexity. However, its state variable is the voltage fluctuation value of the entire system, and the control variables are the new energy and SVC reactive power regulation of the entire system. This model still belongs to the global optimization model.

[0080] In this embodiment, we assume that the renewable energy access nodes {35, 36, 37, 38} in the IEEE-39 node system have active power of {650, 560, 540, 830} MW, respectively. The fluctuation ratio of renewable energy active power gradually increases from 0 to 5% in intervals of 0.1%. Different optimization models are used to determine the optimal operating state of the system under the above conditions. Model I uses the reactive power of renewable energy at nodes {35, 36, 37, 38} as the control variable, while Model II uses the reactive power adjustment of renewable energy at those nodes as the control variable. All other boundary conditions are the same: voltage upper and lower limits are 1.1 pu and 0.9 pu, the base capacity is 100 MVA, and the coefficients are... and Set the values ​​to 1 and 0 respectively. Write a program in Matlab and directly call the fmincon function to solve the above optimization model.

[0081] From the perspective of control variables, when the fluctuation of renewable energy is small (e.g., less than 2%), the reactive power regulation of renewable energy in Model II is approximately equal to that in Model I. The difference between the two increases slightly as the proportion of renewable energy fluctuation increases. This is because Model II does not consider system nonlinearity factors, such as... Figure 4 As shown. Overall, the optimization results of Model II and Model I are basically consistent within the allowable error range, indicating that it is feasible to use a linear reactive power-voltage optimization control model to approximate the basic reactive power-voltage optimization model based on AC power flow under short-term fluctuations in renewable energy. From the objective function perspective, when the proportion of renewable energy fluctuations is small (e.g., less than 2%), the objective function value of Model II is approximately equal to that of Model I. As the proportion of renewable energy fluctuations increases, the objective function value of Model I increases non-linearly, while the objective function value of Model II increases linearly, and the difference between the two gradually increases, as shown... Figure 5 As shown. From the perspective of state variables, under different new energy fluctuation ratios, the system voltage optimized by Model II is basically consistent with the system voltage of Method I, as shown. Figure 6 As shown, this further verifies the feasibility of the linear reactive power-voltage optimization control model. The solution time of the two optimization models under different renewable energy fluctuation ratios was statistically analyzed to assess model complexity. Table 2 shows that Model I is a nonlinear programming problem with an average solution time of 4.73 seconds, while Model II is a linear programming problem with a solution time in the millisecond range. Therefore, the linear reactive power-voltage optimization control model can effectively reduce model complexity while ensuring the accuracy of the optimization results.

[0082] Table 2. Average solution time for different optimization models:

[0083]

[0084] In this embodiment, the short-time linear reactive power-voltage optimization control model still solves for control variables uniformly from the perspective of the entire network, which is highly dependent on global communication capabilities. Therefore, a spatial decoupling mechanism considering the voltage sensitivity of the entire system is proposed (by constructing voltage reference variables, spatial decoupling of voltage constraints and objective functions is achieved, thereby constructing a distributed reactive power-voltage optimization model at each renewable energy grid-connected node to reduce model size and complexity), to further reduce model size and complexity. It is assumed that the power system with renewable energy grid connection can be divided into M partitions (or divided according to renewable energy grid-connected nodes), with the entire system having random variables (renewable energy active power fluctuations) and system control variables (reactive power regulation of renewable energy and SVC).

[0085] Specifically, the objective function of the linear reactive power-voltage optimization control model is represented as the sum of M sub-objective functions by partitioning, thus achieving spatial decoupling of the objective function. Next, spatial decoupling of the voltage constraint conditions is performed. Traditional partitioning methods divide the power grid into different control regions and assume that the effect of the partitioned control variables is completely limited to their own region, with a sensitivity of 0 to state variables in other regions, i.e., S... ij =0 However, in real power systems, it is impossible to find such a completely decoupled partition. As the interconnection of power grids deepens, the coupling between regions tends to increase, and this idealized assumption is difficult to meet the requirements of modern power system development. Therefore, a spatial decoupling mechanism considering system voltage sensitivity is proposed. This mechanism restricts the control variables to the partition, while the state variable is the voltage fluctuation value of the entire system, i.e., S. ij Not equal to 0 This ensures that the solution results of the partitioned optimization model still possess a certain degree of globality. Substituting the block matrix form into the voltage limit-crossing formula of the short-time linear reactive power-voltage optimization control model, the state variable in the sub-constraints is the system voltage fluctuation value, considering the superposition effect of control variables from different partitions. Subsequently, the reactive power constraint space decoupling is further performed.

[0086] As the above analysis shows, the objective function and reactive power constraints can be directly decoupled into M partitions. However, voltage is affected by the coupling of control variables in each partition. Therefore, handling the voltage constraints is key to achieving spatial decoupling. Sufficient conditions for effective decoupling are given for the system voltage inequality constraint set: the system voltage of a single partition optimization model is within the allowable range; simultaneously, considering the superposition effect of all partition optimization models, the system voltage does not exceed the limit. The proposed spatial decoupling mechanism has the following advantages: firstly, it fully utilizes the regional characteristics of reactive power and voltage, effectively reducing the dimension of the optimization problem; secondly, because this mechanism considers the coupling effect of different partition control variables on system voltage, compared with the completely decoupled traditional partitioning method, its optimization results still retain a certain degree of global optimality and are more in line with the application needs of actual power grids.

[0087] Specifically, in this embodiment, a distributed reactive power voltage optimization control strategy is designed based on linear voltage sensitivity and spatial decoupling mechanism to achieve reactive power optimization adjustment at each renewable energy node with small fluctuations. A voltage reference variable is introduced to ensure that the system voltage after reactive power adjustment and fluctuations in renewable energy at that node does not exceed the voltage reference variable. According to the principle of linear superposition, under the combined action of the distributed reactive power optimization control models of all renewable energy grid-connected nodes, the system voltage will not exceed the voltage reference variable, thus satisfying the requirement that the voltage does not exceed the upper limit, and thereby achieving voltage constraint decoupling. Similarly, for operating conditions where the voltage is close to the lower limit, the voltage reference variable takes the minimum value of the system reference voltage; while for operating conditions with medium voltage levels, the combined action of the distributed reactive power optimization control models of each renewable energy grid-connected node usually does not lead to voltage exceeding the limit, and the voltage constraint can be directly decoupled.

[0088] Spatial decoupling mechanisms can significantly reduce the solution scale of optimization problems, while also facilitating local optimization and regulation of reactive power from renewable energy sources and reducing dependence on centralized framework computing and communication capabilities. Based on the spatial decoupling mechanism, a distributed reactive power optimization control model is constructed at the renewable energy grid-connected node. Its state variable is the voltage fluctuation value of the entire system, and the control variable is the reactive power regulation amount of the renewable energy source.

[0089] Specifically, in this embodiment, the steps of the distributed reactive power and voltage fast optimization control based on nonlinear sensitivity under large fluctuations in renewable energy are as follows: Based on the distributed reactive power and voltage fast optimization control based on linear sensitivity under small fluctuations in renewable energy, a nonlinear mapping algorithm based on Taylor series inversion is used. Specifically, at any optimal operating point in any time period, the AC power flow equation is expanded to a higher order (taking the third order as an example) to obtain the nonlinear equation of the node injected power increment with respect to the voltage increment. , The increase in active power injected into the node. The reactive power increment injected into the node For Jacobian matrices, It is 2 factorial. This represents the voltage magnitude increment at the node. This represents the voltage phase angle increment at the node. For transpose, For Hessian matrix, It is the factorial of 3. The third-order partial derivative matrix of the AC power flow equation, It is a higher-order term;

[0090] The nonlinear mapping algorithm is derived by analogy with the inversion of Taylor series, and the nonlinear equations are... By analogy with Taylor series, we obtain the analogy formula: , Let be the dependent variable, representing the increment of active power and reactive power injected at the node, where a, b, and c are coefficients. , where are the independent variables, representing the voltage magnitude increment and voltage phase angle increment at the node;

[0091] Suppose the inverse function of the Taylor series is , , , All are coefficients; substituting the inverse function into the analogy formula, we obtain about The expression: And by making the coefficients of each order equal on both sides of the equation, we obtain the coefficients of each order and the expression for the inverse function: , ;

[0092] By drawing an analogy between the inverse function expression and the power system, and combining it with practical engineering analysis, an analytical expression for the third-order nonlinear mapping algorithm of voltage increment with respect to node injected power increment is obtained. S is the sensitivity matrix, which is obtained by inverting the Jacobian matrix;

[0093] A second-order nonlinear mapping algorithm is used for fast optimization control of distributed reactive power and voltage. The second-order nonlinear sensitivity matrix... The formula is: .

[0094] In this embodiment, due to the significant calculation error in nonlinear voltage sensitivity under large fluctuations, a fast optimization control strategy for distributed reactive power voltage considering nonlinear sensitivity under large fluctuations in new energy sources is proposed, based on the above. First, the nonlinear mapping algorithm is derived, then the nonlinear voltage sensitivity is calculated. Finally, the nonlinear voltage sensitivity is substituted into the formula of the distributed reactive power voltage optimization model for small fluctuations, constructing a distributed reactive power voltage optimization control model based on nonlinear sensitivity to improve the reactive power voltage optimization control effect under large fluctuations in new energy sources.

[0095] To overcome the shortcomings of the linear sensitivity analysis method, a nonlinear mapping algorithm based on Taylor series inversion is proposed to improve the accuracy of voltage calculation under large fluctuations in new energy sources and to reveal the nonlinear relationship between new energy output fluctuations and voltage changes. To obtain the nonlinear equation of voltage increment with respect to node injected power increment, equation (30) must be inverted, but this process involves the calculation of higher-order partial derivatives, which is difficult to obtain directly. Therefore, this section derives the nonlinear mapping algorithm by analogy with Taylor series inversion. In addition, the analytical expression of the third-order nonlinear mapping algorithm reveals the nonlinear relationship between voltage increment and node injected power increment. When the node injected power increment is known (such as the active power fluctuation of new energy sources), the node voltage increment can be directly calculated using this equation, and then the system voltage can be obtained. Therefore, the proposed nonlinear mapping algorithm is an approximate power flow calculation method. When the second-order and higher-order expansion terms are ignored, the analytical expression of the third-order nonlinear mapping algorithm degenerates into the linear mapping algorithm. In comparison, the nonlinear mapping algorithm can effectively improve the accuracy of voltage calculation, but its calculation time is also longer because it involves the calculation of higher-order partial derivatives. To balance computational time and accuracy, this section employs a second-order nonlinear mapping algorithm to implement distributed reactive power tracking optimization control, and provides the second-order nonlinear sensitivity matrix. It should be noted that the sensitivity matrix S and Hessian matrix H in the formula only need to be calculated once at the initial moment of the time period and can be applied to all moments throughout the entire time period.

[0096] It should be noted that the formula structure of the distributed reactive voltage fast optimization model under large-amplitude fluctuations is similar to that under small-amplitude fluctuations, but the sensitivity coefficient in the model is a nonlinear voltage sensitivity. This model has high accuracy in large-amplitude fluctuation scenarios.

[0097] Specifically, in this embodiment, the effectiveness of the distributed reactive power and voltage fast optimization control strategy based on fluctuation amplitude assessment is further verified through time-series simulation of the IEEE-39 node system. The simulation duration is set to 1 hour, including 12 time periods and 720 time points. Figure 10 As shown, the dashed line represents the active power of new energy measured during the time period, and the solid line represents the active power of new energy measured at any moment. The distributed reactive power and voltage fast optimization control strategies based on linear sensitivity and nonlinear sensitivity are used to optimize the control of the system.

[0098] The reactive power curves of new energy sources with different optimization methods are as follows: Figure 11As shown, the optimal power flow method can obtain the optimal reactive power of renewable energy sources through global optimization at each time step, but the system faces significant computational and communication challenges. Traditional optimization methods keep the reactive power of renewable energy sources constant throughout the time period, and their values ​​deviate significantly from the optimal reactive power. Under small fluctuations, the reactive power of renewable energy sources obtained by the distributed reactive voltage and reactive power fast optimization control strategy based on nonlinear sensitivity and the distributed reactive voltage and reactive power fast optimization control strategy based on linear sensitivity are basically consistent. However, under large fluctuations, the deviation between the two is significant. Among them, the reactive power curve of renewable energy sources obtained by the distributed reactive voltage and reactive power fast optimization control strategy based on nonlinear sensitivity is closer to the optimal reactive power curve. Taking the renewable energy at node 36 as an example, at time 337 (i.e., 28 minutes and 5 seconds), the reactive power of the renewable energy obtained by the optimal power flow method is 43.15 MVar, the reactive power of the renewable energy obtained by the traditional optimization method is 61.62 MVar, the reactive power of the renewable energy obtained by the distributed reactive power and voltage fast optimization control strategy based on linear sensitivity is 22.98 MVar, and the reactive power of the renewable energy obtained by the distributed reactive power and voltage fast optimization control strategy based on nonlinear sensitivity is 36.15 MVar, which is the smallest difference from the optimal reactive power. Therefore, the distributed reactive power and voltage fast optimization control strategy based on nonlinear sensitivity can effectively improve the ability to track the optimal reactive power in scenarios with large fluctuations.

[0099] Table 3 shows the average active power loss of different optimization methods. The active power loss of the distributed reactive power voltage fast optimization control strategy based on linear and nonlinear sensitivity is slightly higher than the optimal loss value, but significantly lower than that of the traditional optimization method. Therefore, the proposed strategy can effectively reduce active power loss and ensure the economic efficiency of system operation.

[0100] Table 3 Average active power network loss for different optimization methods:

[0101]

[0102] Traditional optimization methods have an optimization cycle of minutes (5 minutes), which is insufficient to effectively address the adverse effects of short-term fluctuations in new energy sources on voltage within 5 minutes. Therefore, in time-series simulations, the system exhibits varying degrees of voltage exceedance, such as... Figure 12 As shown, nodes {28, 29} reached a maximum over-limit amplitude of 0.015 pu during the 6th time period (25min-30min). The distributed reactive power and voltage fast optimization control strategy based on linear sensitivity improves voltage safety by optimizing and adjusting the reactive power of new energy sources locally at each moment. However, because it ignores the nonlinear factors of the system, slight voltage over-limit phenomena still occur in scenarios with large fluctuations in new energy sources, such as... Figure 13 As shown in the figure, the distributed reactive power voltage fast optimization control strategy based on nonlinear sensitivity did not exhibit voltage limit exceedance during the entire time-series simulation. Figure 14 As shown, this demonstrates that the method can improve the voltage optimization control effect in scenarios with large fluctuations.

[0103] like Figure 15 As shown, the optimal power flow method takes approximately 2.56 seconds to perform one optimization, with a maximum time approaching 10 seconds. Considering the time required for collecting information and issuing commands across the entire network, the optimal power flow method cannot meet the requirement of 5-second timescale optimization control. The distributed reactive power and voltage fast optimization control strategy based on linear sensitivity takes less than 10ms to perform one optimization, significantly shortening the optimization timescale while reducing dependence on network-wide communication and improving system operating efficiency. Statistical analysis of the computation time of the distributed reactive power and voltage fast optimization control strategy based on nonlinear sensitivity is presented, such as... Figure 16 As shown, the maximum time for executing one optimization is approximately 70ms, and the average time is 36.1ms. The average computation time of the distributed reactive power and voltage fast optimization control strategy based on nonlinear sensitivity is approximately 30.1 times that of the distributed reactive power and voltage fast optimization control strategy based on linear sensitivity. This is because the former considers the nonlinear factors of the system, making its calculation process more complex. In summary, the distributed reactive power and voltage fast optimization control strategy based on nonlinear sensitivity can effectively improve the ability to track the optimal operating point under scenarios of large fluctuations in renewable energy, but its computation time is greater than that of the distributed reactive power and voltage fast optimization control strategy based on linear sensitivity.

[0104] This embodiment demonstrates the feasibility of the proposed method. The proposed strategy, through time decoupling mechanism, spatial decoupling mechanism, construction of distributed reactive power voltage optimization model, evaluation of new energy fluctuation rate and amplitude, and design of distributed reactive power voltage fast optimization control strategy, can effectively eliminate voltage over-limit under different fluctuation amplitudes of new energy, reduce the active power network loss objective function, and improve voltage safety level and operating efficiency.

[0105] In summary, this method first constructs a basic reactive power-voltage optimization model based on AC power flow, aiming to minimize both active power network loss and voltage deviation. Second, it proposes a time decoupling mechanism to construct a reactive power-voltage optimization model for time periods (e.g., 5 minutes) and a linear reactive power-voltage optimization control model for short periods (e.g., 5 seconds). Third, it proposes a spatial decoupling mechanism that considers the voltage sensitivity of the entire system. By constructing a voltage reference variable, it achieves spatial decoupling between voltage constraints and the objective function, thereby constructing a distributed reactive power-voltage optimization model at each renewable energy grid-connected node to reduce model size and complexity. Finally, it establishes evaluation indices for the short-term fluctuation rate and amplitude of renewable energy, constructs an event-triggered mechanism for dynamic adjustment of the reactive power optimization cycle based on the renewable energy fluctuation rate, and constructs a distributed reactive power-voltage rapid optimization control strategy based on linear and nonlinear sensitivity based on the fluctuation amplitude index. This enables efficient dynamic tracking of the optimal operating point of the system under different renewable energy fluctuation amplitudes, improving the operational safety and economy of high-proportion renewable energy power systems.

[0106] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A distributed reactive power and voltage rapid optimization control method, characterized in that, include: Obtain system data at the initial moment of the time period, call the preset reactive voltage optimization model for the time period to process the system data at the initial moment of the time period, and obtain the optimal operating point for the time period; Based on the optimal operating point for a given time period, the active power of renewable energy at each moment within that time period is measured. Then, based on the active power of renewable energy, short-term fluctuation rate and amplitude evaluation indicators are calculated, specifically: The short-term fluctuation rate assessment index for a single renewable energy source is defined as the degree of deviation in the system's operating state caused by changes in the active power of the renewable energy source at adjacent times. Its formula is: , For a moment The volatility rate assessment index under the following conditions For node j at time... The active power of new energy sources below For node j at time... The active power of new energy sources below , All are coefficients. For the slight increase in network loss, For averaging, For the system node set, The change in voltage amplitude at node i caused by a unit change in active power generated by the new energy source at node j; Calculating time using voltage sensitivity Voltage fluctuation at grid connection point caused by new energy fluctuations , For node j at time... Fluctuation of active power of new energy sources Let $\frac{ ... For node j in time period The new energy benchmark active power; Calculate the actual voltage fluctuation value at the grid connection point The amplitude evaluation index is defined as the calculation error of the sensitivity analysis method. , For a moment The following magnitude assessment indicators; Based on the fluctuation rate assessment index, it is determined whether to trigger the reactive power voltage optimization control of new energy sources, and a judgment result is generated, specifically as follows: Based on the volatility rate assessment index, construct a judgment index. To assess the reference time of the last execution of renewable energy reactive power voltage optimization control. At the time The degree to which fluctuations in the active power of new energy sources cause deviations in the system's operating state. The fluctuation rate evaluation index at time t. The trigger threshold for optimizing the reactive power voltage control event triggering mechanism for new energy sources. This refers to the set of new energy nodes in the system. When the judgment indicator is determined to be greater than or equal to the trigger threshold, the judgment result is triggered, and distributed reactive power voltage rapid optimization control is performed to quickly adjust the reactive power of new energy sources, while simultaneously setting the reference time... ; When the judgment indicator is determined to be less than the trigger threshold, the reactive power of the new energy source remains unchanged, and the process proceeds to the next time step. When the judgment result is triggered, distributed reactive power voltage fast optimization control is performed based on the amplitude evaluation index and the preset distributed reactive power voltage optimization model. The distributed reactive power voltage fast optimization control includes distributed reactive power voltage fast optimization control based on linear sensitivity under small fluctuations in new energy sources and distributed reactive power voltage fast optimization control based on nonlinear sensitivity under large fluctuations in new energy sources.

2. The distributed reactive power and voltage rapid optimization control method according to claim 1, characterized in that, The steps for constructing the time-period reactive voltage optimization model are as follows: With minimizing active power network loss as the objective function, the following objective function is constructed: ,in, For the system node set, Let be the voltage amplitude at node i. Let i be the grounding conductance. For the system branch set, Let J be the voltage magnitude at node j. Let be the voltage phase angle difference between node i and node j. The conductance of branch l; With minimizing voltage deviation as the objective function, the voltage deviation objective function is constructed as follows: , Let be the reference voltage of node i; and then perform a weighted summation of the active power loss objective function and the voltage deviation objective function to obtain the overall objective function. , These are the weighting coefficients for the active power network loss objective function. These are the weighting coefficients for the voltage deviation objective function; Establish equality constraints, which are the power balance equations. ,in, Let be the active power of the new energy source at node i. Let be the fluctuation of the active power of new energy sources at node i. Let be the active power of the generator at node i. Let be the active power of the load at node i. This indicates that node j is connected to node i. Let the mutual conductance of nodes i and j be the nodal admittance matrix. Let the mutual amperage of nodes i and j in the nodal admittance matrix be denoted as . Let i be the reactive power of the new energy source. Let i be the SVC reactive power. Let i be the reactive power of the generator. Let be the reactive power of the load at node i; Establish inequality constraints, including branch transmission power constraints, voltage amplitude constraints, voltage phase angle constraints, generator reactive power constraints, new energy reactive power constraints, and SVC reactive power constraints; The allowable range for discrete variable constraints established based on discrete variables is as follows: , , Among them, discrete variables include the transformer turns ratio and the number of capacitor banks in operation. This is the lower limit of the transformer turns ratio for branch l. Let l be the transformer turns ratio. This represents the upper limit of the transformer turns ratio for branch l. This represents the lower limit of the reactive power of the capacitor bank at node i. Let be the reactive power of the capacitor bank at node i. This represents the upper limit of the reactive power of the capacitor bank at node i. This represents the number of capacitor banks in operation. Let be the reactive power capacity of a single capacitor bank at node i; Based on the overall objective function, equality constraints, inequality constraints, and discrete variable constraints, the basic model for reactive power voltage optimization is obtained by programming in the pre-set Matlab software and calling the Cplex function to solve the problem. Based on the reactive power and voltage optimization model, the fluctuation of active power of new energy sources, the fluctuation of voltage amplitude, and the fluctuation of voltage phase angle are all set to 0 to obtain the reactive power and voltage optimization model for the time period.

3. The distributed reactive power and voltage rapid optimization control method according to claim 2, characterized in that, The specific restrictions of the inequality constraints are as follows: Branch transmission power shall not exceed its maximum transmission capacity: , The active power transmitted by branch l. The reactive power transmitted by branch l, The maximum transmission capacity of branch l; The voltage amplitude and voltage phase angle at the node should be within the following ranges: , , , This represents the lower limit of the voltage amplitude at node i. Let be the fluctuation value of the voltage amplitude at node i. This represents the upper limit of the voltage amplitude at node i. This is the lower limit of the voltage phase angle at node i. Let i be the voltage phase angle at node i. Let be the fluctuation value of the voltage phase angle at node i. Let be the upper limit of the voltage phase angle at node i. Let represent the functional relationship between the voltage amplitude at node i and the active power fluctuation of each new energy source. Let be the functional relationship between the voltage phase angle of node i and the active power fluctuation of each renewable energy source, where m is the number of renewable energy nodes. This refers to the fluctuation in active power of new energy sources at the first node. This refers to the fluctuation in active power of new energy sources at the second node. Let be the fluctuation of active power of new energy sources at the m-th node; The reactive power of generators, reactive power of new energy sources, and reactive power of SVC should be within the following ranges: , , , This represents the lower limit of the reactive power of the generator at node i. Let i be the reactive power of the generator. Let be the upper limit of the reactive power of the generator at node i. This represents the lower limit of the reactive power of new energy sources at node i. Let i be the reactive power of the new energy source. Let i be the upper limit of the reactive power of new energy sources. This is the lower limit of the SVC reactive power at node i. Let i be the SVC reactive power. is the upper limit of the SVC reactive power of node i.

4. The distributed reactive power and voltage rapid optimization control method according to claim 3, characterized in that, The construction steps of the distributed reactive power voltage optimization model are as follows: The active power network loss objective function of the reactive power-voltage optimization basic model is used as the original objective function and linearized to obtain the objective function of the short-time linear reactive power-voltage optimization control model. , For the set of new energy nodes in the system, Let be the incremental rate of change of active power at node j in the original objective function. For the short-term fluctuation of new energy at node j at time t, Let be the incremental rate of change of reactive power at node j in the original objective function. Let be the reactive power adjustment of new energy sources at node j at time t. For the SVC node set in the system, Let be the incremental rate of change of reactive power at node k in the original objective function. Let be the SVC reactive power regulation of node k at time t; Linearizing the equality constraints of the reactive voltage optimization basic model, the node voltage fluctuation values ​​are obtained as follows: , Let be the fluctuation value of the voltage amplitude at node i at time t. Let be the voltage phase angle fluctuation value of node i at time t. Let $\frac{j}{j}$ be the amount by which a unit change in active power from a renewable energy source causes a change in the voltage amplitude at node $i$. Let $\frac{j}{j}$ be the amount by which a unit change in reactive power at node $i$ causes a change in the voltage amplitude at node $i$. Let $\frac{j}{j}$ be the change in voltage phase angle at node $i$ caused by a unit change in active power from the renewable energy source. Let $\frac{j}{j}$ be the change in voltage phase angle at node $i$ caused by a unit change in reactive power from the renewable energy source. Let be the sensitivity of the voltage amplitude at node i to the change in reactive power of the SVC at node k. The sensitivity of the voltage phase angle at node i to the change in reactive power of the SVC at node k; The inequality constraints for constructing a short-time linear reactive power-voltage optimization control model are given, where the voltage exceedance that should not occur at any node i is: , The voltage amplitude obtained by optimizing the reactive voltage at node i during the time period. The voltage phase angle obtained by optimizing the reactive voltage for node i during the time period; the reactive power of new energy sources and the reactive power of SVC should be within the following range: , The reactive power of new energy obtained by solving the reactive voltage optimization model for node j during the time period is... This represents the lower limit of the reactive power at node j during the specified time period. Let be the upper limit of reactive power at node j during the specified time period. The SVC reactive power obtained by solving the reactive voltage optimization model for node k during the time period is... This represents the lower limit of the reactive power of the SVC during the time period at node k. The upper limit of the reactive power of the SVC during the time period at node k; Based on the objective function The linearized equality constraints and the inequality constraints of the short-time linear reactive power-voltage optimization control model are used to obtain the short-time linear reactive power-voltage optimization control model.

5. The distributed reactive power and voltage rapid optimization control method according to claim 4, characterized in that, Also includes: The objective function of the short-time linear reactive power-voltage optimization control model is represented by the sum of M sub-objective functions in partitioned form, as shown in the formula: , The sum of M sub-objective functions. For the first sub-objective function, This is the second sub-objective function. Let M be the Mth sub-objective function, and M be the number of sub-objective functions. This refers to the fluctuation in the active power of new energy sources in the entire system. This refers to the reactive power regulation of the entire system's new energy sources and SVC. Let be the fluctuation of active power from new energy sources in the first sub-objective function. The second sub-objective function represents the fluctuation in active power from new energy sources. Let be the fluctuation of active power from new energy sources in the Mth sub-objective function. The reactive power regulation of new energy sources and SVC is the first sub-objective function. The reactive power regulation of new energy sources and SVC is the second sub-objective function. For the reactive power regulation of new energy sources and SVC in the Mth sub-objective function; The node voltage fluctuation values ​​of the short-time linear reactive power-voltage optimization control model are represented by partitions, resulting in a block matrix form. , This represents the voltage fluctuation value corresponding to the first sub-objective function. This represents the voltage fluctuation value corresponding to the second sub-objective function. This represents the voltage fluctuation value corresponding to the Mth sub-objective function. Let be the sensitivity block matrix of voltage fluctuation in the Mth partition to the active power fluctuation of new energy sources in the Mth partition; Substituting the block matrix form into the voltage limit violation formula of the short-time linear reactive power-voltage optimization control model, and expressing it as multiple sub-constraint forms, yields the system voltage inequality constraint set. , This represents the voltage fluctuation value of the entire system. This is the first voltage inequality constraint form. This is the second voltage inequality constraint form. Let M be the form of the voltage inequality constraint, where each partition has a corresponding form of voltage inequality constraint. The ranges of renewable energy reactive power and SVC reactive power in the short-time linear reactive power-voltage optimization control model are expressed as M sub-constraints, resulting in the system reactive power inequality constraint set. , This is the first constraint form of the reactive power inequality. This is the second reactive power inequality constraint form. Let M be the reactive power inequality constraint form, where each partition has a corresponding reactive power inequality constraint form; Based on the objective function after partitioning, the block matrix form, the system voltage inequality constraint set, and the system reactive power inequality constraint set, a spatial decoupling mechanism is obtained, and a distributed reactive power and voltage optimization model is constructed based on the spatial decoupling mechanism.

6. The distributed reactive power and voltage rapid optimization control method according to claim 5, characterized in that, The steps for rapid optimization control of distributed reactive power and voltage based on linear sensitivity under small fluctuations in new energy sources are as follows: At a single renewable energy grid-connected node, a distributed reactive power and voltage optimization model based on a spatial decoupling mechanism is used for rapid optimization control of distributed reactive power and voltage, and a voltage reference variable is introduced. This ensures that the system voltage after the fluctuation of active power from new energy sources and the adjustment of reactive power at this node is not greater than the voltage reference variable. The objective function of the distributed reactive power voltage optimization model under small fluctuations in new energy sources is: The constraints of the distributed reactive power voltage optimization model under small fluctuations in new energy sources are: the node voltages satisfy: The permissible range for reactive power of new energy sources is: .

7. The distributed reactive power and voltage rapid optimization control method according to claim 6, characterized in that, The steps for rapid optimization control of distributed reactive power and voltage based on nonlinear sensitivity under conditions of significant fluctuations in renewable energy are as follows: Based on the distributed reactive power and voltage fast optimization control with linear sensitivity under small fluctuations in renewable energy, a nonlinear mapping algorithm based on Taylor series inversion is used. In this algorithm, at any optimal operating point in a given time period, the AC power flow equations are expanded to a higher order to obtain a nonlinear equation for the node injection power increment with respect to the voltage increment. , The increase in active power injected into the node. The reactive power increment injected into the node For Jacobian matrices, It is 2 factorial. This represents the voltage magnitude increment at the node. This represents the voltage phase angle increment at the node. For transpose, For Hessian matrix, It is the factorial of 3. The third-order partial derivative matrix of the AC power flow equation, It is a higher-order term; The nonlinear mapping algorithm is derived by analogy with the inversion of Taylor series, and the nonlinear equations are... By analogy with Taylor series, we obtain the analogy formula: , Let be the dependent variable, representing the increment of active power and reactive power injected at the node, where a, b, and c are coefficients. , where are the independent variables, representing the voltage magnitude increment and voltage phase angle increment at the node; Suppose the inverse function of the Taylor series is , , , All are coefficients; substituting the inverse function into the analogy formula, we obtain about The expression: And by making the coefficients of each order equal on both sides of the equation, we obtain the coefficients of each order and the expression for the inverse function: , ; By drawing an analogy between the inverse function expression and the power system, and combining it with practical engineering analysis, an analytical expression for the third-order nonlinear mapping algorithm of voltage increment with respect to node injected power increment is obtained. S is the sensitivity matrix, which is obtained by inverting the Jacobian matrix; A second-order nonlinear mapping algorithm is used for fast optimization control of distributed reactive power and voltage. The second-order nonlinear sensitivity matrix... The formula is: .

Images