A distributed photovoltaic and traction power supply collaborative voltage imbalance compensation method

Abstract

The application provides a voltage imbalance compensation method for distributed photovoltaic and traction power supply cooperation, and belongs to the technical field of distributed photovoltaic and traction power supply cooperation, and comprises the following steps: dividing a power grid into multiple sub-regions, and adopting a cooperative compensation strategy for each sub-region. Specifically, first, the compensation power required to be output by a traction power supply system compensation device SVG and a photovoltaic station in each sub-region is determined, so as to reduce the voltage imbalance degree. Then, an improved multi-agent consistency algorithm is used to optimize the compensation power output of each photovoltaic converter in each photovoltaic station. On this basis, the negative sequence current and the voltage imbalance degree of the common coupling point before and after compensation are calculated, and the constraint condition of the voltage imbalance degree is obtained. Finally, by taking the constraint condition as a reference, the specific compensation power and compensation current of each photovoltaic converter are determined by using the improved multi-agent consistency algorithm, so that the voltage imbalance degree of all common coupling points can be accurately compensated.

Description

Technical Field

[0001] The present invention belongs to the technical field of coordinated distributed photovoltaic and traction power supply technology, and in particular relates to a voltage imbalance compensation method for coordinated distributed photovoltaic and traction power supply technology. Background Art

[0002] Electrified railways have seen widespread development in recent years due to their excellent characteristics, including low carbon emissions and high transportation efficiency. As a core component of electrified railways, the traction power supply system plays a vital role in providing energy to traction loads. However, traction loads are high-power, single-phase loads with high impact. When they interact with the three-phase power system, they introduce large amounts of negative sequence current (NSC) into the grid through the point of common coupling (PCC) between the traction substation and the grid. This, in turn, causes voltage imbalance (VU) in the grid, seriously impacting the efficient operation of the grid and the traction power supply system. In some regions, for example, the voltage imbalance introduced by the traction power supply system has caused frequent disconnections of photovoltaic power stations and industrial power users along electrified railway lines, severely impacting the safe and efficient operation of the grid. Therefore, effectively controlling voltage imbalance at the PCC is crucial to ensuring the safe and efficient operation of the traction power supply system and the grid.

[0003] To improve the voltage imbalance (VU) problem caused by the traction power supply system, the traditional solution is to install compensation devices such as SVGs (Static Var Generators) in the traction power supply system to compensate for voltage imbalance. However, achieving satisfactory compensation generally requires an excessively large installed capacity, resulting in high investment costs. In recent years, with the large-scale integration of clean energy into the power system, distributed photovoltaic power generation systems have been widely integrated into the power grid along electrified railways. This opens up the possibility of synergistically compensating for VU using distributed photovoltaic systems and traction power supply systems. Some researchers have proposed using wind farms' doubly fed wind turbines and traction power supply system's SVGs to synergize VU compensation at PCC points. However, this research only addresses compensation for a single PCC point and does not consider synergistic compensation across all PCC points within the grid. Furthermore, the operating characteristics of wind power systems differ from those of photovoltaic systems, so further research is needed to investigate the use of photovoltaic systems for VU compensation.

[0004] At the same time, existing single-point compensation and regional compensation methods differ. While using all renewable energy units within a region to coordinate compensation for all PCC points within the region can effectively compensate for VU within the region as a whole, the compensation process may also cause the VU at the points where renewable energy units are connected to the grid to exceed the specified value. Therefore, it is urgent to develop a coordinated compensation method that can simultaneously address the VU compensation requirements of all PCC points within the entire grid area and ensure that the VU at the points where renewable energy units are connected meets the specified value. Summary of the Invention

[0005] In view of this, the present invention provides a voltage imbalance compensation method for coordinated distributed photovoltaic and traction power supply, which can solve the technical problem that the existing technology basically performs single-point compensation for one PCC, and does not perform multi-point overall compensation for the PCC points of all traction substations in a regional power grid.

[0006] The present invention is achieved in that:

[0007] A first aspect of the present invention provides a method for voltage imbalance compensation in a coordinated distributed photovoltaic and traction power supply system, comprising the following steps:

[0008] S10, using a grid partitioning method based on node voltage sensitivity to divide the grid into multiple sub-areas;

[0009] S20. Determine the compensation power that needs to be output by the traction power supply system compensation device SVG and the photovoltaic station in each sub-area using a coordinated compensation strategy. The coordinated compensation strategy includes obtaining the voltage imbalance of the sub-areas, calculating the average value, sorting the sub-areas, and allocating the compensation power based on the remaining capacity of the photovoltaic station and the capacity of the SVG.

[0010] S30. Using an improved multi-agent consensus algorithm to optimize the compensation power required to be output by each photovoltaic converter in each photovoltaic station; the improved multi-agent consensus algorithm calculates the ratio of the compensation power to the remaining capacity of each converter and uses graph theory to represent the connections between converters to ultimately determine the output compensation power of each converter;

[0011] S40, calculating the negative sequence reference current; calculating the negative sequence current of the common coupling point PCC before and after compensation, respectively, taking into account the voltage, current and related angle parameters of the load port and the compensation port;

[0012] S50, calculating the voltage imbalance before and after compensation at the common coupling point PCC; calculating the voltage imbalance before and after compensation based on the negative sequence current, line voltage and short-circuit capacity;

[0013] S60, calculating the unbalanced power before and after compensation to obtain the voltage imbalance constraint parameter K; by calculating the unbalanced power before and after compensation, the voltage imbalance constraint parameter K is obtained as a basis for subsequent optimization;

[0014] S70. Using the K parameter as a reference instruction and combining it with the improved multi-agent consensus algorithm, solve the compensation power and compensation current that each photovoltaic converter within each photovoltaic station needs to output; determine the compensation power output ratio of each photovoltaic converter through an iterative process, and finally calculate the specific compensation power and compensation current, which are used to achieve effective compensation for the voltage imbalance of all common coupling points in the entire power grid area through the coordinated use of distributed photovoltaic and traction power supply systems.

[0015] Specifically, step S10 is to determine the voltage sensitivity between nodes by calculating the voltage deviation factor, define the voltage sensitivity between nodes by logarithmic transformation, and finally divide the nodes with the same voltage sensitivity range into the same sub-region.

[0016] Wherein, the step S10 specifically includes:

[0017] Step 101: Propose a voltage deviation factor and calculate the voltage deviation factor according to the following formula:

[0018]

[0019] Where, ΔV + and ΔV - They are the positive and negative sequence voltage deviation factors, ΔQ + and ΔQ - represents the positive-sequence and negative-sequence reactive power at the common coupling point between the traction power supply system and the grid, respectively, and J is the power flow Jacobian matrix; Step 102, calculate the voltage sensitivity between node i and node j according to the following formula:

[0020]

[0021] Step 103: To ensure the symmetry of the VUS between node i and node j, a logarithmic transformation is used to define the VUS between the two nodes, which is represented by d as follows:

[0022] d ij =d ji =-lg|VUS ij VUS ji |;

[0023] Step 104: Divide the nodes within the same range of d into the same sub-area according to the following formula to implement grid partitioning based on voltage sensitivity:

[0024] d α≤d ij ≤d β .

[0025] The collaborative compensation strategy proposed in step S20 specifically includes:

[0026] Step 201: Obtain the voltage imbalance degree at the common coupling point between all traction substations and the power grid in each sub-area;

[0027] Step 202: Calculate the average value of the voltage unbalance at the common coupling point between all traction substations and the power grid in each sub-area;

[0028] Step 203: Sort the sub-regions in descending order based on the average value of the voltage imbalance in each sub-region;

[0029] Step 204: Starting from the sub-region with the largest average voltage imbalance, calculate the total remaining capacity of all photovoltaic stations in the sub-region;

[0030] Step 205: If the total remaining capacity of all photovoltaic stations in the sub-region is greater than the total reactive power that needs to be compensated at the common coupling points of all traction substations and the grid in the sub-region, the photovoltaic stations in the sub-region will complete the voltage imbalance compensation in the sub-region.

[0031] Step 206: If the total remaining capacity of all photovoltaic stations in the sub-region is less than the total reactive power required to be compensated at the common coupling points of all traction substations and the grid in the sub-region, the photovoltaic stations in the sub-region and the traction power supply system compensation device SVG jointly complete voltage imbalance compensation in the sub-region.

[0032] Step 207: Repeat steps 204 to 206 until voltage imbalance compensation is completed in all sub-regions.

[0033] Step 208: Sort the photovoltaic stations in each sub-region in descending order based on the remaining capacity of the photovoltaic stations, and allocate the compensation power required to be output by each photovoltaic station in this order;

[0034] Step 209: Sort the SVGs in descending order based on the remaining capacity of the traction power supply system compensation devices SVG in each sub-area, and allocate the compensation power required to be output by each SVG in this order.

[0035] Wherein, the step S30 specifically includes:

[0036] Step 301: Calculate the compensation power that each converter needs to output based on the remaining capacity of each converter, as shown below:

[0037]

[0038] Where, S uni is the compensation power allocated to the i-th converter in each PV station, S rci is the remaining capacity allocated to the i-th converter in each PV station;

[0039] Step 302: Use the basic theory of graph theory to represent the connection between each photovoltaic converter in each photovoltaic station, that is, G = {V, E, A}, where V represents the nodes of the graph, E represents the edges of the graph, and A is the adjacency matrix of the graph, which represents the connection relationship between each node, as shown below:

[0040]

[0041] Where, μ ij It is the non-negative weighted adjacent element of matrix A, which is a 0-1 binary variable.

[0042] Wherein, the step S40 specifically includes:

[0043] Step 401: Calculate the negative sequence current of the PCC before compensation. Taking phase A as an example, the calculation result is as follows:

[0044]

[0045] Step 402: Calculate the negative sequence current of the PCC after compensation. Taking phase A as an example, the calculation result is as follows:

[0046]

[0047] Where, U L and U A are the load port and phase A compensation port voltages, θ LT Represents the transformer connection angle at the load port, θ LL For U L and U A The angle between L is the load port current, m is the number of compensation ports, θ MT is the transformer connection angle of the compensation port, θ Pη is the power factor angle of the compensation port η.

[0048] Wherein, the step S50 specifically includes:

[0049] Step 501: Calculate the voltage imbalance at the PCC according to the following formula:

[0050]

[0051] Where, I -Indicates the negative sequence current value, U Line Indicates PCC line voltage, S d Indicates the short-circuit capacity of PCC;

[0052] Step 502: Calculate the voltage imbalance before compensation according to the following formula:

[0053]

[0054] Step 503: Calculate the compensated voltage imbalance according to the following formula:

[0055]

[0056] Wherein, the step S60 specifically includes:

[0057] Step 601: Calculate the unbalanced power before compensation according to the following formula:

[0058]

[0059] Step 602: Calculate the compensated unbalanced power according to the following formula:

[0060]

[0061] Step 603: Calculate the voltage imbalance constraint parameter K according to the following formula:

[0062]

[0063] Wherein, the step S70 specifically includes:

[0064] Step 701: Determine the ratio of the compensation power output of each photovoltaic inverter to its remaining capacity as follows:

[0065]

[0066] Step 702: Incorporate the result of step 701 into the following iterative process:

[0067]

[0068] Step 703: Calculate the compensation power output ratio of each photovoltaic converter as follows:

[0069]

[0070] Step 704: Calculate the compensation power of each photovoltaic converter according to the following formula:

[0071]

[0072] Step 705: Calculate the compensation current of each photovoltaic converter according to the following formula:

[0073]

[0074] The improved multi-agent consensus algorithm specifically includes:

[0075] Step 801: Change the element μ in the adjacency matrix A ij To achieve a higher convergence speed, the new assignment rules are as follows:

[0076]

[0077] Step 802: According to the assignment rule of step 801, construct an improved Laplace matrix as shown below:

[0078]

[0079] Where,

[0080] Compared with the existing technology, the beneficial effect of the voltage imbalance compensation method for coordinated distributed photovoltaic and traction power supply provided by the present invention is: it fully utilizes the distributed photovoltaic power plants along the electrified railway power grid and the SVG in the traction power supply system, adopts a systematic compensation strategy and algorithm, and realizes accurate compensation of voltage imbalance at all common coupling points in the entire power grid area.

[0081] Compared with the prior art, the solution of the present invention has the following outstanding advantages:

[0082] 1. This system comprehensively considers the voltage imbalance compensation needs of all PCCs within the grid area and adopts a coordinated compensation strategy. This strategy fully utilizes the remaining capacity of distributed photovoltaic power generation and provides compensation through the traction power supply system's SVG, effectively achieving VU compensation for all PCCs within the entire grid area. This not only improves the capacity utilization of the distributed photovoltaic station converters but also reduces the installed capacity and investment cost of the traction power supply system's compensation equipment.

[0083] 2. A grid partitioning method based on node voltage sensitivity is adopted to divide the grid into multiple sub-areas. Combined with an improved multi-agent consensus algorithm, the compensation power output between each photovoltaic converter is coordinated to ensure that the VU at the access points of all new energy units in the entire grid area meets the requirements, avoiding the problem of local over-compensation or under-compensation that may occur in single-point compensation. It solves the technical problem that the existing technology basically performs single-point compensation for one PCC and does not perform multi-point overall compensation for the PCC points of all traction substations in a regional power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Figure 1 A flow chart of the method provided by the present invention;

[0085] Figure 2 A topological diagram for implementing the method in Example 1;

[0086] Figure 3 This is the partition effect diagram of Example 1;

[0087] Figure 4 This is the collaborative compensation strategy diagram in Example 1;

[0088] Figure 5 This is the compensation effect diagram under traction load fluctuation;

[0089] Figure 6 This is a diagram of the compensation effect under photovoltaic fluctuations. DETAILED DESCRIPTION

[0090] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0091] like Figure 1 FIG. 1 is a flow chart of a method for voltage imbalance compensation for distributed photovoltaic and traction power supply coordination provided by the present invention. The method includes the following steps:

[0092] S10, using a grid partitioning method based on node voltage sensitivity to divide the grid into multiple sub-areas;

[0093] S20. Determine the compensation power that needs to be output by the traction power supply system compensation device SVG and the photovoltaic station in each sub-area using a coordinated compensation strategy. The coordinated compensation strategy includes obtaining the voltage imbalance degree of the sub-areas, calculating the average value, sorting, and allocating the compensation power based on the remaining capacity of the photovoltaic station and the capacity of the SVG.

[0094] S30. Using an improved multi-agent consensus algorithm, optimize the compensation power required to be output by each photovoltaic converter in each photovoltaic station; the improved multi-agent consensus algorithm calculates the ratio of the compensation power to the remaining capacity of each converter and uses graph theory to represent the connections between converters to ultimately determine the output compensation power of each converter;

[0095] S40, calculating the negative sequence reference current; calculating the negative sequence current of the common coupling point PCC before and after compensation, respectively, taking into account the voltage, current and related angle parameters of the load port and the compensation port;

[0096] S50, calculating the voltage imbalance before and after compensation at the common coupling point PCC; calculating the voltage imbalance before and after compensation based on the negative sequence current, line voltage and short-circuit capacity;

[0097] S60, calculating the unbalanced power before and after compensation to obtain the voltage imbalance constraint parameter K; by calculating the unbalanced power before and after compensation, the voltage imbalance constraint parameter K is obtained as a basis for subsequent optimization;

[0098] S70, using the K parameter as a reference instruction and combining it with the improved multi-agent consensus algorithm, solve the compensation power and compensation current that each photovoltaic converter within each photovoltaic station needs to output; determine the compensation power output ratio of each photovoltaic converter through an iterative process, and ultimately calculate the specific compensation power and compensation current;

[0099] S70. Using the K parameter as a reference instruction and combining it with the improved multi-agent consensus algorithm, solve the compensation power and compensation current that each photovoltaic converter within each photovoltaic station needs to output; determine the compensation power output ratio of each photovoltaic converter through an iterative process, and finally calculate the specific compensation power and compensation current, which are used to achieve effective compensation for the voltage imbalance of all common coupling points in the entire power grid area through the coordinated use of distributed photovoltaic and traction power supply systems.

[0100] The specific implementation methods of the above steps are described in detail below:

[0101] The specific implementation of step S10 is to divide the power grid into multiple sub-areas using a power grid partitioning method based on node voltage sensitivity. Specifically, the following steps are included:

[0102] Step 101: The concept of voltage deviation factor is proposed. The voltage deviation factor ΔV of the positive sequence and negative sequence is calculated according to the following formula: + and ΔV - :

[0103]

[0104] Among them, ΔQ + and ΔQ - are the positive-sequence and negative-sequence reactive powers at the common coupling point of the traction power supply system and the grid, respectively, and J is the power flow Jacobian matrix. The voltage deviation factor reflects the sensitivity of the node voltage to reactive power.

[0105] Step 102: Calculate the voltage sensitivity VUS between node i and node j according to the following formula: ij and VUS ji :

[0106]

[0107] This step aims to quantitatively describe the voltage sensitivity relationship between nodes.

[0108] Step 103: To ensure the symmetry of the VUS between node i and node j, a logarithmic transformation is used to define the VUS between the two nodes, which is represented by d as follows:

[0109] d ij =d ji =-lg|VUS ij VUS ji |;

[0110] By using logarithmic transformation, we can ensure that the voltage sensitivity between nodes in the same range of d is similar.

[0111] Step 104: Divide the nodes within the same range of d into the same sub-area according to the following formula to implement grid partitioning based on voltage sensitivity:

[0112] d α ≤d ij ≤d β ;

[0113] Among them, d α and d β is a pre-set threshold value, which can be 0.1, 0.3, etc.

[0114] In general, step S10 quantitatively describes the voltage sensitivity between nodes by calculating the voltage deviation factor, and uses logarithmic transformation to divide nodes with similar voltage sensitivity into the same sub-area, laying the foundation for the subsequent collaborative compensation strategy.

[0115] Step S20 proposes a collaborative compensation strategy, which specifically includes the following steps:

[0116] Step 201: Obtain the voltage imbalance at the common coupling points between all traction substations and the power grid in each sub-area.

[0117] Step 202: Calculate the average value of the voltage imbalance at the common coupling point between all traction substations and the power grid in each sub-area.

[0118] Step 203: sort the sub-regions in descending order according to the average value of the voltage imbalance in each sub-region.

[0119] Step 204: Starting from the sub-region with the largest average voltage imbalance, calculate the total remaining capacity of all photovoltaic stations in the sub-region.

[0120] Step 205: If the total remaining capacity of all photovoltaic stations in the sub-area is greater than the total reactive power that needs to be compensated at the common coupling points of all traction substations and the grid in the sub-area, the photovoltaic stations in the sub-area will complete the voltage imbalance compensation in the sub-area.

[0121] Step 206: If the total remaining capacity of all photovoltaic stations in the sub-area is less than the total reactive power that needs to be compensated at the common coupling point between all traction substations and the grid in the sub-area, the photovoltaic stations in the sub-area and the traction power supply system compensation device SVG jointly complete the voltage imbalance compensation in the sub-area.

[0122] Step 207: Repeat steps 204 to 206 until voltage imbalance compensation is completed in all sub-regions.

[0123] Step 208: sort the photovoltaic stations in each sub-area in descending order based on the remaining capacity of the photovoltaic stations, and allocate the compensation power required to be output by each photovoltaic station in this order.

[0124] Step 209: sort the SVGs in descending order based on the remaining capacity of the traction power supply system compensation devices SVG in each sub-area, and allocate the compensation power required to be output by each SVG in this order.

[0125] In general, the coordinated compensation strategy proposed in step S20 first evaluates the voltage imbalance of each sub-region, and then coordinates the compensation power output of each according to the remaining capacity of the photovoltaic station and SVG to achieve effective compensation for the voltage imbalance of each sub-region.

[0126] Step S30 uses an improved multi-agent consensus algorithm to optimize the compensation power that each photovoltaic converter in each photovoltaic station needs to output, specifically including the following steps:

[0127] Step 301: Based on the remaining capacity S of each converter rci Calculate the compensation power S that each converter needs to output uni , as shown below:

[0128]

[0129] This step ensures that the compensation power of each converter is proportional to its residual capacity, reflecting the principle of fairness.

[0130] Step 302: Use the basic theory of graph theory to represent the connection between each photovoltaic converter in each photovoltaic station, and construct a graph G = {V, E, A}, where V represents the nodes of the graph, E represents the edges of the graph, and A is the adjacency matrix of the graph, as shown below:

[0131]

[0132] This step establishes the topological relationship between the converters and provides a basis for subsequent consistency algorithm optimization.

[0133] In general, step S30 adopts an improved multi-agent consistency algorithm, fully considers the relationship between the converters, and reasonably allocates the compensation power according to the remaining capacity of each converter, thereby optimizing the compensation scheme inside the entire photovoltaic station.

[0134] Step S40 is aimed at calculating the negative sequence reference current, and specifically includes the following steps:

[0135] Step 401, the negative sequence current of the PCC before compensation is calculated, and the calculation result of A phase is as follows:

[0136]

[0137] Wherein, U L and U A are the voltages of the load port and the A-phase compensation port respectively, θ LT represents the transformer connection angle of the load port, θ LL is the angle between U L and U A , and I L is the current of the load port.

[0138] Step 402, the negative sequence current of the PCC after compensation is calculated, and the calculation result of A phase is as follows:

[0139]

[0140] Wherein, m is the number of compensation ports, θ MT is the transformer connection angle of the compensation port, and θ Pη is the power factor angle of the compensation port η.

[0141] By calculating the negative sequence currents of the PCC before and after compensation, the basic data for subsequent voltage unbalance degree calculation is provided.

[0142] Step S50 is aimed at calculating the voltage unbalance degree at the PCC before and after compensation, and specifically includes the following steps:

[0143] Step 501, the voltage unbalance degree at the PCC is calculated according to the following formula:

[0144]

[0145] Wherein, I - represents the negative sequence current value, U Line represents the line voltage at the PCC, and S d represents the short-circuit capacity of the PCC. The voltage unbalance degree ε U reflects the influence degree of the negative sequence current at the PCC on the line voltage.

[0146] Step 502, the voltage unbalance degree before compensation is calculated according to the following formula:

[0147]

[0148] Step 503: Calculate the compensated voltage imbalance according to the following formula:

[0149]

[0150] By comparing the voltage imbalance before and after compensation, the effectiveness of the coordinated compensation measures can be evaluated.

[0151] Step S60 is intended to calculate the unbalanced power before and after compensation, and based on this, obtain the constraint condition K parameter of the voltage imbalance degree, which specifically includes the following steps:

[0152] Step 601: Calculate the unbalanced power before compensation according to the following formula:

[0153]

[0154] Step 602: Calculate the compensated unbalanced power according to the following formula:

[0155]

[0156] Step 603: Calculate the voltage imbalance constraint parameter K according to the following formula:

[0157]

[0158] The K parameter quantitatively describes the degree of change in voltage imbalance before and after compensation, providing a basis for subsequent optimization.

[0159] Step S70 uses the improved multi-agent consensus algorithm, combined with the K parameter obtained in step S60, to solve the compensation power and compensation current that each photovoltaic converter in each photovoltaic station needs to output, specifically including the following steps:

[0160] Step 701: Determine the ratio of the compensation power output of each photovoltaic inverter to its remaining capacity as follows:

[0161]

[0162] Step 702, incorporate the result of step 701 into the following iterative process:

[0163]

[0164] This step uses a multi-agent consensus algorithm to coordinate the compensation power output ratio of each converter according to the topological relationship between the converters.

[0165] Step 703: Calculate the compensation power output ratio of each photovoltaic converter as follows:

[0166]

[0167] Step 704: Calculate the compensation power of each photovoltaic converter according to the following formula:

[0168]

[0169] Step 705: Calculate the compensation current of each photovoltaic converter according to the following formula:

[0170]

[0171] In general, step S70 uses the improved multi-agent consensus algorithm to reasonably allocate the compensation power and compensation current of each photovoltaic converter while satisfying the voltage imbalance constraint, thereby achieving effective compensation for all PCC points in the entire power grid area.

[0172] Step S80 connects the above steps in series, achieving precise compensation for voltage imbalance at all common coupling points across the entire power grid region through the coordinated utilization of distributed photovoltaic and traction power supply systems. Through grid zoning, coordinated compensation strategies, and an improved multi-agent consensus algorithm, the synergistic advantages of distributed photovoltaic and traction power supply systems are fully utilized, effectively resolving the voltage imbalance issues introduced by electrified railways. This improves the capacity utilization of distributed photovoltaic station converters and reduces the installed capacity and investment costs of compensation devices in the traction power supply system.

[0173] Specifically, the principle of the present invention is:

[0174] 1. Grid partitioning: First, the grid partitioning method based on node voltage sensitivity is used to divide the entire grid into multiple sub-areas. Specifically, by calculating the voltage deviation factor ΔV + and ΔV - , quantitatively describes the sensitivity of node voltage to reactive power. Then, according to the voltage sensitivity VUS between nodes ij and VUS ji ,The nodes with similar voltage sensitivity are divided into the same sub-area using logarithmic ,transformation.,Such grid partitioning method ensures that the voltage sensitivity of nodes in the same sub-area is relatively close, which is beneficial to the subsequent ,coordinated compensation strategy.

[0175] 2. Collaborative Compensation Strategy: Based on grid zoning, a collaborative compensation strategy is employed. First, the voltage imbalance of all PCC points within each sub-region is obtained, the average value is calculated, and the results are ranked. Then, starting with the sub-region with the highest average voltage imbalance, the remaining capacity of the PV plant within that sub-region is assessed to determine whether it is sufficient to compensate for the voltage-to-voltage (VU) load in that sub-region. If sufficient, the PV plant performs compensation; if insufficient, the PV plant and the traction power supply system's SVG jointly perform compensation. This orderly compensation approach, which fully utilizes the remaining PV capacity, effectively compensates for the voltage-to-voltage (VU) load in each sub-region.

[0176] 3. Improved Multi-Agent Consensus Algorithm: After determining the compensation division of labor within each sub-region, an improved multi-agent consensus algorithm is used to optimize the compensation power output of each PV inverter within the PV station. This algorithm first calculates the initial compensation power distribution based on the remaining capacity of each inverter. Then, using graph theory to represent the topological relationship between the inverters, iterative optimization ensures that the compensation power output ratios across all inverters are coordinated and consistent. This not only embodies the principle of fairness but also ensures the rational distribution of compensation power among the inverters.

[0177] A specific embodiment 1 of the present invention is provided below, and this embodiment 1 comprises the following steps:

[0178] Step 1: The principle of using the remaining capacity of the photovoltaic station converter to compensate for voltage imbalance is to use the remaining capacity of the photovoltaic station converter to inject reverse unbalanced power into the common coupling point between the traction substation and the power grid to offset the unbalanced power injected into the common coupling point by the traction power supply system. In a large regional power grid, when using the remaining capacity of the photovoltaic station to compensate for the voltage imbalance problem caused by the traction power supply system, there is a situation where a photovoltaic station simultaneously compensates for the voltage imbalance at the common coupling point of multiple traction substations and the power grid. Although this can make the voltage imbalance at the common coupling point of the traction substation and the power grid meet the standard, it causes the voltage imbalance at the photovoltaic station to exceed the standard. Therefore, in order to improve the compensation effect of the voltage imbalance at the common coupling point of multiple traction substations and the power grid in the entire power grid, this embodiment proposes a power grid partitioning method based on node voltage sensitivity, such as Figure 2 The figure shows a topological structure diagram of the coupling between the traction power supply system, photovoltaic station and power grid in a large regional power grid. Figure 2In the illustrated topology, the traction substation 3, the photovoltaic station 3, and the large power grid 3 are coupled and connected together through the grid node 6. In this embodiment, the power grid is divided into multiple sub-areas for zone compensation (the photovoltaic station in each sub-area only compensates the voltage imbalance at the common coupling point between the traction substation in each sub-area and the power grid), thereby ensuring that the voltage imbalance at each node of the power grid meets the standard. First, a voltage deviation factor is proposed, and the voltage deviation factor is calculated according to formulas (1) and (2):

[0179]

[0180]

[0181] Where ΔV+ and ΔV- are the positive-sequence and negative-sequence voltage deviation factors, ΔQ+ and ΔQ- represent the positive-sequence and negative-sequence reactive powers at the common coupling point between the traction power supply system and the grid, J is the power flow Jacobian matrix, and S represents the reactive voltage sensitivity of the grid.

[0182] Step 2: Calculate the voltage sensitivity between node i and node j according to equations (3) and (4):

[0183]

[0184] Step 3: To ensure the symmetry of VUS between node i and node j, the VUS between the two nodes is defined by logarithmic transformation and represented by d. The larger d is, the smaller the VUS between the two nodes is, as shown in formula (5). Furthermore, according to formula (6), the nodes with d in the same range are divided into the same sub-area to realize the grid partitioning based on voltage sensitivity, as shown in Figure 3 As shown, it ensures that the nodes in each area have the same voltage sensitivity, thereby ensuring the compensation effect of the voltage imbalance in each sub-area:

[0185] d ij =d ji =-lg|VUS ij VUS ji | (5)

[0186] d α ≤d ij ≤d β (6)

[0187] exist Figure 3 In the topological structure, the traction substation in region 1 is coupled to the grid through grid node 1, that is, the traction substation is connected to grid node 1, and the photovoltaic station is coupled to the grid through grid node 4, that is, the photovoltaic station is connected to grid node 4. The same applies to other regions. ijIt represents the partition basis calculated by the method of the present invention, has no actual meaning and unit, and is only used as a basis for partitioning the power grid. ij Identical nodes (i, j represent grid node numbers), such as node i and node j, will be assigned to the same region. The coordinate diagram indicates that nodes 1, 4, and 7 are assigned to the same region, nodes 2, 5, and 8 are assigned to the same region, and nodes 3, 6, and 9 are assigned to the same region. Therefore, during the subsequent voltage unbalance (VU) compensation process, the voltage unbalance at the common coupling point between the traction substation and the grid in each region will be compensated by the converters of the photovoltaic stations in each region and the railway's compensation devices. For example, the voltage unbalance at node 1 is compensated jointly by the compensation device of the traction substation connected to node 1 and the converters of the photovoltaic stations connected to nodes 4 and 7.

[0188] Step 4: In order to ensure the compensation effect of voltage imbalance in each sub-area after partitioning, and to make full use of the remaining capacity of the photovoltaic converter in each sub-area, a coordinated compensation strategy is proposed to accurately determine the compensation power that the traction power supply system compensation device SVG and the photovoltaic station need to output in each sub-area. Figure 4 As shown;

[0189] Step 5: After determining the compensation power required to be output by each photovoltaic station based on the collaborative compensation strategy, the improved multi-agent consensus algorithm is used to optimize the compensation power required to be output by each photovoltaic converter in each photovoltaic station. First, the compensation power required to be output by each converter is calculated based on the remaining capacity of each converter, as shown in (7):

[0190]

[0191] Step 6: Calculate the negative sequence reference current. Taking phase A as an example, the calculation results of the negative sequence current of the PCC before compensation are shown in (8), and the calculation results of the negative sequence current of the PCC after compensation are shown in (9).

[0192]

[0193] Step 7: According to the VU calculation method at the PCC shown in (10), combined with (8) and (9), the VU calculation model before and after compensation is obtained, and the VU before and after PCC compensation is calculated based on this.

[0194]

[0195] I- represents the negative sequence current value, UL represents the PCC line voltage, and Sd represents the PCC short-circuit capacity.

[0196] Step 8: Multiply (11) and (12) by 3 times 3UA to obtain the unbalanced power S-A0 and S-A1 before and after compensation, as shown in (13) and (14). Combine (13) and (14), that is, obtain the constraint condition K parameter of VU.

[0197]

[0198] Step 9: Take K parameter as a reference instruction, combine the improved multi-agent consensus algorithm, and solve the compensation power and compensation current that each photovoltaic converter needs to output inside each photovoltaic station, that is, first determine the ratio of the compensation power output of each photovoltaic inverter to its remaining capacity, as shown in (17), and put (17) into the iteration process shown in (18) to obtain the compensation power output ratio of each photovoltaic converter shown in (19), and then combine formulas (20) and (21) to calculate the compensation current size of each photovoltaic converter.

[0199]

[0200] Step 10: Based on the above steps, the precise compensation of VU of all PCC points in the entire power grid area can be completed.

[0201] In addition, regarding the improved multi-agent consensus algorithm described in the above steps, the improvement process is as follows:

[0202] First, the basic theory of graph theory is used to represent the relationship between each photovoltaic converter in each photovoltaic station, that is, G={V, E, A}, V represents the node of the graph, E represents the edge of the graph, and A is the adjacency matrix of the graph (a symmetric matrix with diagonal elements of 0), which represents the connection relationship between nodes, as shown in formulas (22) and (23). Each photovoltaic converter is regarded as an agent, which is a node of the graph, and the communication link of the information interaction between each agent is regarded as the edge E of the graph, and then all photovoltaic converters in the entire photovoltaic station together form a multi-agent system, that is, G.

[0203]

[0204] Where, μ ij is the non-negative weighted adjacency element of matrix A, which is a 0-1 binary variable.

[0205] Further, for a given topology, the consensus protocol considering no communication delay and considering communication delay can be represented by (24) and (25) respectively, that is, the relationship between input and output.

[0206]

[0207] Where, φ is the value of the input variable, ψ is the value of the output variable, τ ijrepresents the communication delay for agent i to transmit data to agent j.

[0208] (24) and (25) can be expressed in matrix form as (26) and (27).

[0209]

[0210] L=Δ-A (28)

[0211] Where L represents the Laplacian matrix of the graph.

[0212] Since the sum of all rows and columns of A is equal to 1, all eigenvalues ​​of A are less than or equal to 1. Therefore, when k increases to infinity as shown in (29), the system will converge.

[0213]

[0214] Therefore, changing the assignment rule of the elements μij in A can achieve a higher convergence speed. The new assignment rule is shown in (30):

[0215]

[0216] The improved Laplace proof is shown in (31):

[0217]

[0218] in,

[0219] To better understand and implement the present invention, the following provides Example 2 of a specific application scenario: Within a regional power grid, there is an electrified railway line, along which multiple traction substations and photovoltaic power plants are distributed. Due to the impact and single-phase nature of the traction load, a large amount of negative-sequence current is introduced into the grid through the point of common coupling (PCC) between the traction substation and the grid, causing severe voltage imbalance (VU) problems. To address this issue, the local power grid company decided to adopt the voltage imbalance compensation method proposed in this invention, which combines distributed photovoltaic and traction power supply.

[0220] First, according to step S10, the power grid in the region is divided into four sub-regions using a power grid partitioning method based on node voltage sensitivity. The specific steps are as follows:

[0221] Step 101: Calculate the voltage deviation factor ΔV of the positive sequence and negative sequence + and ΔV - The positive and negative sequence reactive power ΔQ at the common coupling point between the traction power supply system and the grid + and ΔQ -If they are 50Mvar and 30Mvar respectively, then:

[0222]

[0223] Step 102: Calculate the voltage sensitivity VUS between nodes ij and VUS ji :

[0224]

[0225] Step 103: define the voltage sensitivity d between nodes using logarithmic transformation ij :

[0226] d ij =d ji =-lg|VUS ij VUS ji |;

[0227] Divide the nodes in the same range of d into the same sub-area, assuming d α =0.1,d β =0.3, the power grid can be divided into 4 sub-areas.

[0228] Next, according to step S20, a coordinated compensation strategy is adopted to compensate for the voltage imbalance in each sub-region.

[0229] Step 201: Obtain the voltage imbalance of all PCC points in each sub-area. Assume that, through actual measurement, the voltage imbalance of the PCC points in sub-areas 1, 2, 3, and 4 is 2.8%, 3.2%, 2.6%, and 2.4%, respectively.

[0230] Step 202: Calculate the average value of the voltage imbalance of all PCC points in each sub-region. The results are as follows: Sub-region 1: 2.8% Sub-region 2: 3.2% Sub-region 3: 2.6% Sub-region 4: 2.4%

[0231] In step 203, the sub-regions are sorted according to the average voltage imbalance, and the resulting order is: sub-region 2 > sub-region 1 > sub-region 3 > sub-region 4.

[0232] Step 204: Starting from sub-region 2, where the average voltage imbalance is the largest, calculate the total remaining capacity of all PV stations within that sub-region. Assuming there are three PV stations in sub-region 2, with remaining capacities of 20 Mvar, 18 Mvar, and 15 Mvar, respectively, the total remaining capacity is 53 Mvar.

[0233] In step 205, since the total remaining capacity of all PV stations in sub-area 2 (53 Mvar) is greater than the total reactive power that needs to be compensated by all PCC points in the sub-area (assuming it is 45 Mvar), the PV stations in the sub-area will complete the voltage imbalance compensation in the sub-area.

[0234] Step 208: Sort the photovoltaic stations according to their remaining capacity: 20Mvar>18Mvar>15Mvar, and allocate compensation power in this order, i.e., 20Mvar, 18Mvar, and 15Mvar.

[0235] Step 206: For sub-area 1, the total remaining capacity of all PV stations in the sub-area (assuming 40 Mvar) is less than the total reactive power that needs to be compensated by all PCC points in the sub-area (assuming 50 Mvar). Therefore, the PV stations in the sub-area and the traction power supply system compensation device SVG jointly complete the voltage imbalance compensation in the sub-area.

[0236] Step 209: Sort the SVGs according to their remaining capacities and allocate compensation power. Assuming the remaining capacity of the SVG is 15 Mvar, allocate 10 Mvar to the SVG for compensation.

[0237] Following the above steps, the voltage imbalance compensation for sub-area 2 and sub-area 1 is completed. Similar compensation can be performed for sub-area 3 and sub-area 4.

[0238] Next, according to step S30, the improved multi-agent consensus algorithm is used to optimize the compensation power of each photovoltaic inverter within the photovoltaic station.

[0239] Step 301: Based on the remaining capacity S of each converter rci Calculate the initial compensation power S uni :

[0240]

[0241] Assuming that PV station A has three converters with residual capacities of 10 Mvar, 8 Mvar, and 6 Mvar, the initial compensation power distribution is 5 Mvar, 4 Mvar, and 3 Mvar.

[0242] Step 302: Construct a topological relationship diagram G = {V, E, A} between the converters. Assume that the connection relationship between the three converters in the photovoltaic station A is as follows: Figure 1 As shown, the adjacency matrix A is:

[0243]

[0244] Step 702: The initial compensation power allocation result is brought into the improved multi-agent consensus algorithm for iterative optimization:

[0245]

[0246] After 20 iterations, the compensation power output ratio of each converter is finally obtained.

[0247] Step 704, according to ψ * Calculate the compensation power for each converter:

[0248]

[0249] Step 705: further calculate the compensation current of each converter:

[0250]

[0251] Through the above steps, the optimization of compensation power and compensation current of each converter inside the photovoltaic station is completed.

[0252] Finally, according to steps S40-S60, the negative sequence current and voltage imbalance at the PCC before and after compensation are calculated, and the voltage imbalance constraint K parameter is obtained. Combined with the optimization results of step S70, accurate compensation of voltage imbalance at all PCC points in the entire power grid area is achieved.

[0253] Figure 5 、 Figure 6 Based on the method of the present invention Figure 3 The compensation effect of voltage imbalance at the common coupling point between the traction substation and the grid in the grid structure shown. Figure 3 It can be seen that there are three traction substations in the entire selected power grid, which are connected to the power grid through grid node 1, grid node 2, and grid node 3 respectively. That is, there are three common coupling points between the traction substation and the power grid, namely grid node 1, grid node 2, and grid node 3. The compensation target of the voltage imbalance at the common coupling point is 2% (i.e. Figure 5 , Figure 6 The compensation target points to the dotted line), that is, the voltage imbalance of grid node 1, grid node 2, and grid node 3 needs to be suppressed to below 2%.

[0254] In actual operation, the traction load in the traction power supply system and the active power output of the photovoltaic station are constantly changing. These factors have a direct impact on the compensation effect. Figure 5The compensation effect is considered under the condition of traction load fluctuation. Before the 2nd second, when no compensation is performed, the voltage imbalance at the three common coupling points exceeds the standard. After compensation is performed in the 2nd second, the voltage imbalance at the three common coupling points is limited to within 2%. Even when traction load fluctuation occurs at the 4th and 6th seconds, it can be found that the voltage imbalance at the three common coupling points can still be suppressed to below 2%. Figure 6 The compensation effect is considered when the active power output of the photovoltaic station fluctuates. Before the second second, when no compensation is performed, the voltage imbalance at the three common coupling points exceeds the standard. After compensation is performed in the second second, the voltage imbalance at the three common coupling points is limited to within 2%. Even when the active power output of the photovoltaic station fluctuates at the second and sixth seconds respectively, it can be found that the voltage imbalance at the three common coupling points can still be suppressed to below 2%.

[0255] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

Claims

1. A voltage imbalance compensation method for distributed photovoltaic and traction power supply coordination, characterized in that: The following steps are involved: S10, using a grid partitioning method based on node voltage sensitivity to divide the grid into multiple sub-areas; S20. Adopting a coordinated compensation strategy, determining the compensation power that the traction power supply system compensation device SVG and the photovoltaic station need to output in each sub-area; S30, using an improved multi-agent consensus algorithm to optimize the compensation power that each photovoltaic converter in each photovoltaic station needs to output; S40, respectively calculating the negative sequence current of the common coupling point PCC before and after compensation; S50, calculating the voltage unbalance before and after compensation based on the negative sequence current, line voltage and PCC short-circuit capacity; S60, calculating the unbalanced power before and after compensation, and obtaining a constraint condition K parameter of the voltage imbalance; S70, using the K parameter as a reference instruction and combining it with the improved multi-agent consensus algorithm, solve the compensation power and compensation current that each photovoltaic converter in each photovoltaic station needs to output; The step S30 includes: step 301, calculating the compensation power that each converter needs to output according to the remaining capacity of each converter, as shown below: Where, S uni is the compensation power allocated to the i-th converter in each PV station, S rci is the remaining capacity allocated to the i-th converter in each PV station; Step 302: Use the basic theory of graph theory to represent the connection between each photovoltaic converter in each photovoltaic station, that is, G = {V, E, A}, where V represents the nodes of the graph, E represents the edges of the graph, and A is the adjacency matrix of the graph, which represents the connection relationship between each node, as shown below: Where, μ ij is the non-negative weighted adjacent element of matrix A, a 0-1 binary variable, v i ,v j They represent the i-th and j-th nodes of the graph respectively, and the subscript n is the number of nodes; The step S70 specifically includes: Step 701: Determine the ratio of the initial compensation power of each photovoltaic inverter to its remaining capacity as follows: Step 702: Substitute the result of step 701 into the improved multi-agent consensus algorithm for iteration: Step 703: Calculate the compensation power output ratio of each photovoltaic converter as follows: Where, m is the number of compensation ports; Step 704: Calculate the compensation power of each photovoltaic converter according to the following formula: Step 705: Calculate the compensation current of each photovoltaic converter according to the following formula: Where U is the voltage at the output of the photovoltaic converter.

2. The method according to claim 1, characterized in that ,Step S10 specifically determines the voltage sensitivity between nodes by calculating the voltage deviation factor, ,and defines the voltage sensitivity between nodes using logarithmic transformation, and finally divides the nodes with the same voltage sensitivity range into the same sub-area.

3. The method according to claim 2, characterized in that The collaborative compensation strategy proposed in step S20 specifically includes: Step 201: Obtain the voltage imbalance degree at the common coupling point between all traction substations and the power grid in each sub-area; Step 202: Calculate the average value of the voltage unbalance at the common coupling point between all traction substations and the power grid in each sub-area; Step 203: Sort the sub-regions in descending order based on the average value of the voltage imbalance in each sub-region; Step 204: Starting from the sub-region with the largest average voltage imbalance, calculate the total remaining capacity of all photovoltaic stations in the sub-region; Step 205: If the total remaining capacity of all photovoltaic stations in the sub-region is greater than the total reactive power that needs to be compensated at the common coupling points of all traction substations and the grid in the sub-region, the photovoltaic stations in the sub-region will complete the voltage imbalance compensation in the sub-region. Step 206: If the total remaining capacity of all photovoltaic stations in the sub-region is less than the total reactive power required to be compensated at the common coupling points of all traction substations and the grid in the sub-region, the photovoltaic stations in the sub-region and the traction power supply system compensation device SVG jointly complete voltage imbalance compensation in the sub-region. Step 207: Repeat steps 204 to 206 until voltage imbalance compensation is completed in all sub-regions. Step 208: Sort the photovoltaic stations in each sub-region in descending order based on the remaining capacity of the photovoltaic stations, and allocate the compensation power required to be output by each photovoltaic station in this order; Step 209: Sort the SVGs in descending order based on the remaining capacity of the traction power supply system compensation devices SVG in each sub-area, and allocate the compensation power required to be output by each SVG in this order.

4. The method according to claim 3, characterized in that , the step S40 specifically includes: Step 401: Calculate the negative sequence current of the PCC before compensation. Taking phase A as an example, the calculation result is as follows: Where, U L and U A are the voltages at the load port and the compensation port of phase A, θ LT Represents the transformer connection angle at the load port, θ LL For U L and U A The angle between L is the current at the load port; Step 402: Calculate the negative sequence current of the PCC after compensation. Taking phase A as an example, the calculation result is as follows: Where U η Represents the voltage of the ηth compensation port, I η represents the current of the ηth compensation port, θ MT is the transformer connection angle of the compensation port, is the power factor angle of the compensation port η.

5. The method according to claim 4, characterized in that , the step S50 specifically includes: Step 501: Calculate the voltage imbalance at the PCC according to the following formula: Where, I - Indicates the negative sequence current value, U Line Indicates PCC line voltage, S d Indicates the short-circuit capacity of PCC; Step 502: Calculate the voltage imbalance before compensation according to the following formula: Step 503: Calculate the compensated voltage imbalance according to the following formula:

6. The method according to claim 5, characterized in that , the step S60 specifically includes: Step 601: Calculate the unbalanced power before compensation according to the following formula: Step 602: Calculate the compensated unbalanced power according to the following formula: Step 603: Calculate the voltage imbalance constraint parameter K according to the following formula:

7. The method according to claim 6, characterized in that ,The improved multi-agent consensus algorithm specifically includes: Step 801: Change the element μ in the adjacency matrix A ij To achieve a higher convergence speed, the new assignment rules are as follows: Step 802: According to the assignment rule of step 801, construct an improved Laplace matrix as shown below: Where, N i represents the set of nodes adjacent to node i, and They represent the degrees of node i and node j respectively, and ω is the weight parameter.

Images