Method for locating and sizing voltage type converter based on low-voltage alternating current and direct current system

Abstract

The invention discloses a method for locating and sizing a voltage-type converter based on a low-voltage AC / DC system, and relates to the technical field of power system planning, and the method comprises the steps: S1, constructing a mathematical model of a distributed power supply and a voltage source-type converter, and enabling the mathematical model to be determined by a photovoltaic probability model and a voltage-type converter steady-state model; s2, adopting Latin hypercube sampling to generate a multi-scene sample, and completing load flow calculation of each scene based on an alternating iteration method so as to quantify the influence of source load fluctuation on the system operation state; s3, carrying out site selection by utilizing voltage-power sensitivity analysis; and S4, constructing a multi-target model of the constant volume of the voltage type current converter, and carrying out the constant volume of the voltage type current converter based on the multi-target model. According to the invention, the problems that the site selection of the voltage type current converter depends on experience, the capacity is not matched, the consumption problem of the distributed power supply and the resulting voltage out-of-limit are solved.

Description

Technical Field

[0001] This invention relates to the field of power system planning technology, and in particular to a method for site selection and capacity determination of voltage-source converters based on low-voltage AC / DC systems. Background Technology

[0002] Against the backdrop of energy conservation, emission reduction, and environmental protection, the rapid development and integration of distributed generation, represented by distributed wind power and photovoltaics, into the distribution network is an inevitable trend. However, the randomness and volatility of distributed generation make traditional AC distribution networks prone to problems such as voltage exceeding limits and reduced economic efficiency. Expanding the network or building combiner stations can lead to reduced power quality, difficulties in peak shaving, and insufficient energy conversion efficiency. Hybrid AC / DC distribution networks have become the core direction of new low-voltage distribution networks due to their advantages: DC buses do not require phase and frequency synchronization, simplifying distributed generation access control and improving system reliability; the DC side can connect a large number of PV and energy storage devices to enhance the absorption capacity of distributed generation; voltage-type converters, as the "bridge" between AC and DC sub-networks, can actively control DC voltage and suppress disturbances to the distribution network voltage caused by high-penetration distributed generation, making them key equipment for the safe and economical operation of the system. However, AC / DC distribution networks are more complex. The location and capacity of voltage-type converters directly affect power transmission efficiency, voltage regulation effect, and operating costs. Improper location or capacity mismatch will exacerbate network losses, voltage fluctuations, and affect the absorption of distributed generation, while also increasing losses and shortening the lifespan of distribution equipment. However, existing research does not fully account for the impact of source-load fluctuations on voltage converter configuration, lacks a coordination mechanism between voltage converters and the system, focuses more on the operation control of voltage converters, lacks location and capacity analysis, and does not take into account line characteristics, voltage converter constraints and equipment tolerance, making it difficult to balance safety and economy.

[0003] CN117791751A proposes a method and system for optimizing the configuration of a low-voltage AC / DC hybrid distribution network. It involves selecting a modification structure and constructing a modification optimization model. An upper-level objective function performs preliminary optimization of the distribution network topology, while a lower-level objective function performs secondary optimization of the network configuration. The results of the secondary optimization are fed back to the preliminary optimization to update the overall objective function, and iterative calculations are performed to arrive at the optimized configuration scheme. However, this patent has significant drawbacks: the upper-level optimization model aims to minimize the capacity configuration (Cinvest), optimizing the capacity and location of voltage-type converters. The objective function is singular and does not consider the impact of photovoltaic power generation and load fluctuations. Summary of the Invention

[0004] The purpose of this invention is to propose a method for the location and capacity determination of voltage-source converters based on low-voltage AC / DC systems. This method solves the problem of distributed power source absorption and the resulting voltage over-limit through a systematic approach of fluctuation modeling, stochastic evaluation, sensitivity decision-making, and multi-objective optimization capacity determination.

[0005] A further objective of this invention is to screen the optimal access node through a comprehensive voltage-power sensitivity model and determine the capacity by combining a multi-objective model, thereby solving the problems of voltage converter site selection relying on experience and capacity mismatch, improving the system's voltage regulation and power control capabilities, and reducing equipment losses.

[0006] To achieve the above objectives, the technical method of the present invention is as follows: A method for addressing and calibrating a voltage-source converter based on a low-voltage AC / DC system includes the following steps: S1: Construct a mathematical model of distributed power sources and voltage source converters. The mathematical model is determined by the photovoltaic probability model and the voltage source converter steady-state model. S2: Latin hypercube sampling is used to generate multi-scenario samples, and power flow calculations for each scenario are completed based on the alternating iteration method to quantify the impact of source load fluctuations on the system operating status. S3: Location selection using voltage-power sensitivity analysis; S4: Construct a multi-objective model for voltage-source converter grading, and perform grading of the voltage-source converter based on the multi-objective model.

[0007] Preferably, in step S1, the photovoltaic probability model is established based on the light intensity following a Beta distribution and is used to calculate the actual active and reactive power of the photovoltaic system; the voltage-source converter steady-state model is established based on the AC-DC power transmission characteristics of the voltage-source converter and is used to distinguish the three operating states of the voltage-source converter: rectification, inversion, and zero power transmission.

[0008] Specifically, a photovoltaic (PV) probability model is constructed, with the core concept that the irradiance follows a Beta distribution. The actual value, rated value, and shape parameters of the irradiance are determined, and its probability density function is established. Based on the correlation characteristics between irradiance and PV output, the actual active and reactive power of PV are calculated under different conditions. Finally, a steady-state model of the voltage-source converter is constructed. Based on the AC-DC power transmission characteristics of the voltage-source converter, the equivalent resistance, reactance, and AC / DC side node voltages are determined. The calculation formulas for voltage relationship and phase angle difference are derived, and the correlation between DC side power and modulation ratio, efficiency, and DC side bus power is clarified. Based on the phase angle difference and voltage relationship, the three operating states of the voltage-source converter are distinguished as rectification, inversion, and zero power transmission.

[0009] As a preferred embodiment, in step S2, Latin hypercube sampling generates multiple sets of samples for photovoltaic illumination and load; the alternating iterative method decouples the system into AC subgrid and DC subgrid with the voltage-type converter as the boundary, and alternately performs power flow calculations on both sides until convergence.

[0010] Specifically, step S2 is used to quantify the impact of source load fluctuations on system power flow and obtain the power and voltage distribution of the AC and DC subgrids. Specifically, it involves: generating K sets of samples using Latin super-stereo sampling technology to address the fluctuations in photovoltaic illumination and load; decoupling the system into AC and DC subgrids with the voltage-source converter as the boundary, and initializing power flow parameters; first calculating the AC side power flow and determining convergence, then calculating the voltage-source converter power loss and over-limit parameter adjustment, and finally using the DC power injected by the voltage-source converter to calculate the DC side power flow and determine convergence. If all convergences are achieved, the current sampling is completed; otherwise, it is repeated. After K samplings, the node voltage, branch power, and expected network loss are statistically analyzed for subsequent optimization.

[0011] As a preferred option, in step S3, the multi-objective function is determined by minimizing capacity configuration, minimizing expected system network loss, minimizing voltage deviation, and maximizing the local absorption rate of distributed power sources; the constraints are determined by branch power flow constraints, voltage safety constraints, voltage-type converter operation constraints, distributed power source output constraints, and energy storage operation constraints.

[0012] As a preferred approach, the multi-objective function is normalized to a single-objective function using a linear weighting method, where the sum of the weight coefficients of each sub-objective is 1.

[0013] Specifically, step S3 establishes a multi-objective optimization model with "economy + safety" as the goal: the objective function has four core objectives, namely, minimum capacity configuration, minimum system network loss, minimum voltage deviation, and maximum local absorption rate of distributed power sources, and is normalized into a single objective function through linear weighting; the constraints include branch power flow constraints, voltage safety constraints, and distributed power source output constraints.

[0014] Preferably, in step S4, the voltage-power sensitivity model calculates the sensitivity of voltage to the active and reactive power of the voltage-source converter and combines it with weighting coefficients to obtain a comprehensive sensitivity index for evaluating the node's response capability to the voltage-source converter.

[0015] As a preferred option, the optimal access location for the voltage-source converter is selected from the set of candidate nodes. The set of candidate nodes excludes nodes directly connected to the transformer and nodes with high penetration of distributed power sources. The selection criteria are based on the comprehensive sensitivity of the nodes, ranked from largest to smallest.

[0016] As a preferred approach, in the comprehensive sensitivity model, the weighting coefficients of voltage-active power sensitivity and voltage-reactive power sensitivity are determined based on their degree of influence on system operation.

[0017] Preferably, when constructing the candidate node set, the nodes connected to the existing normal branches in the system are used as candidate access locations for the voltage-type converter.

[0018] Preferably, in step S4, the rated capacity of the voltage-source converter is determined by solving a multi-objective model that aims to minimize the capacity configuration, operation and maintenance losses, and voltage deviation of the voltage-source converter. This model is also subjected to linear weighted normalization and must meet the power and voltage operation constraints of the voltage-source converter.

[0019] Specifically, step S4 constructs a voltage-power sensitivity model for voltage-source converter access scenarios: based on the voltage and power balance equations of distribution network nodes, the active / reactive output and loss correction of the voltage-source converter are introduced to adjust the node power balance, calculate the sensitivity of voltage to the active and reactive power of the voltage-source converter, and then determine the weights to obtain the comprehensive sensitivity; subsequently, from the candidate node set excluding nodes directly connected to transformers and nodes with high penetration of distributed power sources, the optimal access location of the voltage-source converter is selected according to the comprehensive sensitivity from large to small; finally, a multi-objective model is constructed with the voltage-source converter capacity configuration, operation and maintenance losses, and voltage deviation as the objective functions. After linear weighted normalization, combined with the power / voltage constraints of the voltage-source converter, its rated capacity is determined.

[0020] The beneficial effects of this invention include: By constructing a multi-objective model that minimizes capacity configuration, network loss, and voltage deviation, and maximizes the local absorption rate of distributed power sources, and through linear weighted normalization processing, combined with constraints such as branch power flow and voltage-type converter operation, a synergistic balance between system economy and operational safety is achieved. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method of the present invention.

[0022] Figure 2 This is a flowchart detailing the location selection process for voltage-power sensitivity analysis of the voltage converter of the present invention.

[0023] Figure 3 This invention presents a collaborative planning architecture for distribution network energy storage and DC lines.

[0024] Figure 4 This is the single-phase equivalent steady-state model of the voltage-source converter of the present invention.

[0025] Figure 5 This is a system topology diagram of the present invention.

[0026] Figure 6 This is the system node voltage fluctuation diagram after optimization of Scheme 1 of the present invention.

[0027] Figure 7 This is the optimized system topology diagram of Scheme 1 of the present invention.

[0028] Figure 8 This is the system node voltage fluctuation diagram after optimization of Scheme 2 of the present invention.

[0029] Figure 9 This is the optimized system topology diagram of Scheme 2 of the present invention.

[0030] Figure 10 This is the system node voltage fluctuation diagram after optimization of Scheme 3 of the present invention.

[0031] Figure 11 This is the optimized system topology diagram of Scheme 3 of the present invention. Detailed Implementation

[0032] Example 1 See Figure 1 This is the first embodiment of the present invention, which provides a method for addressing and calibrating a voltage-source converter based on a low-voltage AC / DC system, comprising: S1: Construct mathematical models for distributed power sources and voltage source converters. The mathematical models are determined by the photovoltaic probability model and the voltage source converter steady-state model.

[0033] Specifically, a photovoltaic (PV) probability model is constructed, with the core concept that the irradiance follows a Beta distribution. The actual value, rated value, and shape parameters of the irradiance are determined, and its probability density function is established. Based on the correlation characteristics between irradiance and PV output, the actual active and reactive power of PV are calculated under different conditions. Finally, a steady-state model of the voltage-source converter is constructed. Based on the AC-DC power transmission characteristics of the voltage-source converter, the equivalent resistance, reactance, and AC / DC side node voltages are determined. The calculation formulas for voltage relationship and phase angle difference are derived, and the correlation between DC side voltage and modulation ratio, efficiency, and DC side bus voltage is clarified. Based on the phase angle difference and voltage relationship, the three operating states of the voltage-source converter are distinguished as rectification, inversion, and zero power transmission.

[0034] Furthermore, in the photovoltaic probability model, the determination of the shape parameter depends on historical illumination data. The optimal Beta distribution parameter is determined by performing a goodness-of-fit test on illumination intensity data of the target area over many years, thereby more accurately describing the random characteristics of illumination intensity.

[0035] Furthermore, the load volatility is described by establishing a load probability model based on time-series characteristics. This model comprehensively considers daily type, seasonal variation and randomness of user behavior, and uses a normal distribution or Gaussian mixture distribution to model load volatility. It is used together with the photovoltaic probability model as input for Latin hypercube sampling.

[0036] Specifically, the mathematical model of the constructed distributed power source and voltage source converter is as follows.

[0037] Irradiance is usually simulated using a Beta distribution. There is a correlation between the actual output power of photovoltaics and irradiance. The actual output power of photovoltaics can be calculated based on the irradiance value, while the rated power of photovoltaics is used as a reference value.

[0038] Specifically, voltage source converters are the main converter elements used in current flexible DC technology. Voltage source converters act as a "bridge" in AC / DC distribution networks. For ease of expression and calculation, the positive power direction is defined as the flow from the AC side to the DC side. This can be derived from... Figure 4 The topology structure visually illustrates the connection and control logic of the voltage-source converter.

[0039] Table 1 Operating Status of Voltage Source Converters As shown in Table 1, the magnitude and direction of power transmitted between the AC and DC subgrids can be controlled by controlling the phase angle and modulation ratio of the voltage-type converter, thus achieving bidirectional energy transmission with decoupled active and reactive power between the DC and AC subgrids.

[0040] S2: Latin hypercube sampling is used to generate multi-scenario samples, and power flow calculations for each scenario are completed based on the alternating iteration method to quantify the impact of source load fluctuations on the system operating status.

[0041] Specifically, step S2 is used to quantify the impact of source-load fluctuations on system power flow and obtain the power and voltage distribution of the AC and DC subgrids. Specifically, it involves: generating K sets of samples using Latin super-stereo sampling technology to address the fluctuations in photovoltaic illumination and load; decoupling the system into AC and DC subgrids with the voltage-source converter as the boundary, and initializing power flow parameters; first calculating the AC side power flow and determining convergence, then calculating the voltage-source converter power loss and over-limit parameter adjustment, and finally using the DC power injected by the voltage-source converter to calculate the DC side power flow and determine convergence. If all convergences are achieved, the current sampling is completed; otherwise, it is repeated. After K samplings, the node voltage, branch power, and expected network loss are statistically analyzed for subsequent optimization.

[0042] This invention, through the aforementioned stochastic power flow calculation, not only obtains the expected value of the system power flow, but more importantly, it obtains the probability distribution of node voltages and branch power flows. This probability distribution information provides a data foundation for subsequent risk assessment, enabling the planning scheme to be elevated from "meeting constraints" to the higher level of "risk controllable".

[0043] Specifically, in step S2, the stochastic power flow calculation of the AC / DC distribution network based on Latin hypercube sampling and alternating iteration method is as follows: S21: Latin hypercube sampling; Latin hypercube sampling is a type of Monte Carlo simulation method. It improves the sampling strategy and can achieve high sampling accuracy with a smaller sampling scale, making it highly efficient in practical applications.

[0044] The steps of Latin hyperstereosampling include: Determine the number of samples to be drawn; divide the interval from zero to one into several segments; randomly draw a value from each segment; map the drawn value to a standard normal distribution sample using the inverse function of the standard normal distribution; shuffle the sampling order.

[0045] S22: Power flow calculation for AC and DC distribution networks; given the starting node voltage and the connected end load power of the AC distribution network, feeders are used as the basic unit of calculation. The specific calculation process of forward and backward substitution is as follows: Forward power calculation: Assuming that the voltage at each node on the AC side is the rated voltage, the load power and line power loss are calculated segment by segment from the end of the line to the beginning. In the calculation process, only the power loss in each circuit element is calculated, without calculating the voltage value of each node. Finally, the current value and power loss flowing through each branch are obtained, and the power value at the beginning is obtained from this.

[0046] Backward voltage calculation: Based on the given starting node voltage value and the starting power obtained during the forward calculation process, the voltage drop is calculated segment by segment from the starting end of the line to the end, and then the voltage of each node is obtained.

[0047] When AC / DC distribution networks are calculated using the alternating solution method, any effective power flow algorithm can be used on the DC side. When the DC distribution network is an open-loop radial network or only has a weak loop network structure, the Newton-Raphson method can be used for power flow calculation.

[0048] The alternating iterative method first decouples the AC / DC distribution network into AC and DC subnetworks, and then performs power flow calculations. Its core solution is to keep the variables on one side constant and perform power flow calculations on the other side, using the voltage-source converter as a boundary. Then, the converter power flow calculation is performed, which calculates the switching power, connection point voltage, and losses of the voltage-source converter. Using the voltage-source converter as a bridge, the power of the AC / DC subnetwork is balanced through interface interaction. Finally, the power flow calculation results are obtained.

[0049] S3: Location selection using voltage-power sensitivity analysis.

[0050] As a preferred approach, before conducting sensitivity analysis, a candidate node set is first constructed. This set excludes root nodes directly connected to transformers, as these nodes have strong voltage support capabilities, and the regulation benefits of connecting voltage-source converters are limited. Nodes with already high distributed generation penetration are also excluded to avoid power backflow exacerbating voltage issues. Nodes connected to existing common branches in the system are prioritized. These nodes are key points for network reconfiguration; connecting voltage-source converters at these locations not only regulates voltage but also provides the network with flexible interconnection capabilities, enabling proactive power flow control.

[0051] Specifically, voltage-power sensitivity describes the degree to which a small change in the power of a node in a system affects the voltage of other nodes. When the power of a node changes, its voltage will also change accordingly. Therefore, analyzing the voltage of a node can determine the voltage-power sensitivity of each node in an interconnected area.

[0052] Voltage-power sensitivity is defined as the weighted sum of voltage-active power sensitivity and voltage-reactive power sensitivity, with the weighting coefficients determined based on their impact on system operation. A higher calculated sensitivity indicates that connecting a voltage-source converter to the node can more effectively regulate system power, improving the operational flexibility and stability of the distribution network.

[0053] Specifically, step S3 constructs a voltage-power sensitivity model for voltage-source converter access scenarios: based on the voltage and power balance equations of distribution network nodes, the active / reactive output and loss of the voltage-source converter are introduced to correct the node power balance, the sensitivity of voltage to the active and reactive power of the voltage-source converter is calculated, and then the weights are determined to obtain the comprehensive sensitivity; subsequently, from the candidate node set excluding nodes directly connected to transformers and nodes with high penetration of distributed power sources, the optimal access location of the voltage-source converter is selected according to the comprehensive sensitivity from large to small.

[0054] S4: Construct a multi-objective model for voltage-source converter grading, and perform grading of the voltage-source converter based on the multi-objective model.

[0055] Specifically, step S4 establishes a multi-objective optimization model with "performance indicators + safety" as the goal, and normalizes it into a single objective function through the linear weighting method; the constraints include branch power flow constraints, voltage safety constraints, distributed power output constraints, and energy storage operation constraints.

[0056] The objective functions include minimizing capacity configuration, minimizing expected system network losses, minimizing voltage deviation, and maximizing the local absorption rate of distributed generation. Capacity configuration includes the rated capacity of the converter and its corresponding operating losses. Expected system network losses cover the losses on the AC side, DC side, and voltage-type converters. Voltage deviation is measured by the sum of squares of the deviations between the node voltage and the rated voltage. The local absorption rate of distributed generation is calculated as the ratio of the total output of distributed generation to the power flowing to the main grid.

[0057] Multi-objective normalization processing: First, the network loss of AC / DC distribution network is converted into power loss to achieve the unification of the dimensions of capacity configuration and network loss expectation; then, the sub-objective function is normalized by linear weighting to be converted into a single objective function, and the sum of the weight coefficients of each sub-objective is 1.

[0058] Branch power flow constraints are based on the DistFlow power flow model to ensure power balance between AC and DC branches; voltage safety constraints limit node voltages to within permissible ranges; voltage-type converter operation constraints include reactive power control, active power balance, and apparent power limitation; distributed generation output constraints limit the active power output range of photovoltaics; energy storage operation constraints include charging and discharging power and capacity limitations.

[0059] Example 2 See Figures 2 to 11 As a second embodiment of the present invention, the present invention adopts Figure 2 The calculation flowchart shown is used for calculation, and the simulation analysis is based on... Figure 5 The system topology diagram shown verifies the effectiveness and necessity of the proposed model.

[0060] A low-voltage power distribution system was used to verify the effectiveness of the proposed voltage-source converter location and capacity optimization method. The system structure and connection diagram are shown below. Figure 5 .

[0061] Specifically, the system includes 32 normally closed branches and 5 normally open branches, with a rated voltage of 415V and a reference power S. B =10MW, with a total load active power and reactive power of 715kW and 500kvar respectively. The system connects four photovoltaic generators at nodes 7, 10, 24 and 27 with rated capacities of 70kVA, 70kVA, 100kVA and 40kVA respectively. Two energy storage units with a rated capacity of 60kWh are connected at nodes 10 and 24. All photovoltaic generators and energy storage units operate at unity power factor.

[0062] S1: Construct a mathematical model of distributed power sources and voltage source converters.

[0063] Based on 10 years of solar irradiance data provided by the local meteorological station, the shape parameters α=2.1 and β=4.3 of the Beta distribution were determined using the maximum likelihood estimation method. The KS test verified that this distribution fits the actual solar irradiance distribution well at the 5% significance level. The conversion efficiency between photovoltaic output and solar irradiance was set at 16.5%, taking into account the effects of temperature degradation and inverter efficiency.

[0064] A Gaussian mixture distribution is used to describe load fluctuations, with different distribution parameters for weekdays and weekends. Cluster analysis is used to divide the load curves into typical daily patterns, with each pattern corresponding to a set of Gaussian distribution parameters, thereby more accurately simulating the temporal characteristics and randomness of the load.

[0065] The steady-state model of the voltage-source converter adopts Figure 4The single-phase equivalent model shown has an equivalent resistance R_c = 0.01 pu and an equivalent reactance X_c = 0.05 pu. The modulation ratio M operates in the range of 0.8 to 1.0, the phase angle difference δ ranges from -30° to +30°, and the efficiency μ = 98%. These parameters are determined based on the technical specifications of actual commercial voltage converter equipment.

[0066] S2: Latin hypercube sampling is used to generate multi-scenario samples, and power flow calculations for each scenario are completed based on the alternating iteration method to quantify the impact of source load fluctuations on the system operating status.

[0067] 2000 daily operating scenarios were generated using Latin hypercube sampling. The sampling process ensured that the probability distribution of each input variable (PV output and load at each node) was fully sampled, and the correlation between variables was considered through Cholesky decomposition.

[0068] First, the power injected into the DC side by the voltage-source converter is fixed, and the power flow of the AC subnet is calculated using the forward-backward substitution method. Then, based on the voltage and power of the AC side interface nodes, the power loss of the voltage-source converter is calculated and its power injected into the DC side is updated. Finally, the power flow of the DC subnet is solved using the Newton-Raphson method. This process is repeated iteratively until both sides converge.

[0069] The stochastic power flow calculation results show that when the system is not connected to a voltage-source converter, multiple nodes have voltage over-limit problems, as shown in Table 2.

[0070] Table 2 Number of node crossings and maximum deviation voltage As shown in Table 2, voltage exceedances occurred at 10 nodes, with the number of exceedances ranging from 2 to 4. Nodes at the end of the feeder (such as 8, 10, 16, and 24) experienced low voltage due to line voltage drop, with a maximum deviation as low as 0.9398 pu; while nodes near the photovoltaic access point (such as 15, 20, 21, 29, 30, and 31) experienced high voltage due to power backfeeding, with a maximum deviation as high as 1.6615 pu.

[0071] S3: Location selection using voltage-power sensitivity analysis.

[0072] Specifically, voltage-power sensitivity analysis was used to perform calculations over a 24-hour period. The first calculation showed that node 8-16 had the highest improved sensitivity value and should be the primary connection node for the voltage-source converter. After selecting node 8-16 as the primary connection node, based on... Figure 1The voltage source converter optimization configuration process calculates the second improved sensitivity value of the candidate voltage source converter connection nodes when the next voltage source converter is connected. Based on the calculation results, the optimal grid connection point for the second voltage source converter is determined to be nodes 10-21. This process is repeated to calculate the third improved sensitivity value, determining the optimal grid connection point for the third voltage source converter to be nodes 29-31. Due to the high investment cost of voltage source converters, this embodiment sets the maximum number of voltage source converter connection points to three.

[0073] S4: Construct a multi-objective model for voltage-source converter grading, and perform grading of the voltage-source converter based on the multi-objective model.

[0074] Specifically, the relevant parameters for the voltage-source converter are as follows: the unit selectable installation capacity of the voltage-source converter is 10kVA; the unit capacity investment cost of the voltage-source converter is 1000 yuan / kVA; the converter loss coefficient, the annual operation and maintenance cost coefficient of the voltage-source converter, and the annual power supply loss cost coefficient of the distribution system are 0.02, 0.01, and 0.08, respectively. The electricity price is 0.5 yuan per kilowatt-hour, and the annual operating time is 4200 hours. In the multi-objective normalization process, the sum of the weight coefficients of each sub-objective is 1; the voltage deviation penalty coefficient and the distributed power generation absorption penalty coefficient are 0.1 and 0.05, respectively.

[0075] The site selection and capacity determination results for the voltage-source converter are as follows: Table 3 Comparison of various indicators for different voltage-source converter access schemes Analysis of Table 3 shows that Scheme 3 is the optimal voltage-source converter access scheme: 100kVA voltage-source converters are connected at nodes 8-16, and 20kVA voltage-source converters are connected at nodes 9-20. Its average voltage deviation is 0.020pu, significantly improved compared to Scheme 1 and close to Scheme 2; the curtailment rate is 0.98%, lower than both Schemes 1 and 2, indicating more efficient photovoltaic absorption. The annual comprehensive cost is 25,363 yuan. Although slightly higher than Scheme 2, it offers superior voltage regulation accuracy and curtailment rate optimization, resulting in the best overall benefits and meeting the photovoltaic absorption and voltage stability requirements of low-voltage systems.

[0076] Specifically, before connecting the voltage-type converter, power flow calculations showed that nodes 8, 10, 15, 16, 20, 21, 24, 29, 30, and 31 experienced one or more voltage over-limit situations within 24 hours. Figure 6-11 The voltage variation curves of system nodes at different time periods were optimized for different schemes. After optimizing the configuration of the voltage-source converter using this method, no voltage over-limit occurred at any node throughout the day after it was connected to the voltage-source converter.

[0077] Analysis of the simulation results shows that the proposed method for optimizing the location and capacity of voltage-source converters in low-voltage AC / DC systems improves the system's operational stability and economy. It also verifies the beneficial effects of the invention: by constructing a multi-objective model that minimizes capacity configuration, network loss, voltage deviation, and the local absorption rate of distributed power sources, and through linear weighted normalization, combined with constraints such as branch power flow and voltage-source converter operation, a synergistic balance between system technical economy and operational safety is achieved.

Claims

1. A method for site selection and capacity determination of voltage source converters based on low voltage AC / DC systems, characterized in that, The method comprises the following steps: S1: constructing a mathematical model of the distributed power supply and the voltage source converter, the mathematical model being determined by a photovoltaic probability model and a voltage source converter steady-state model; S2: generating multiple scenario samples by using Latin hypercube sampling, and completing power flow calculation of each scenario based on an alternating iteration method to quantify the influence of source and load fluctuation on system operation state; S3: using voltage-power sensitivity analysis for site selection; S4: constructing a multi-objective model of voltage source converter capacity determination, and determining the capacity of the voltage source converter based on the multi-objective model.

2. The method for locating and sizing voltage source converters based on low voltage AC / DC systems according to claim 1, characterized in that, In step S1, the photovoltaic probability model is established based on Beta distribution of illumination intensity, and is used to calculate actual active power and reactive power of the photovoltaic power supply; the voltage source converter steady-state model is established based on AC-DC power transmission characteristics of the voltage source converter, and is used to distinguish three operating states of the voltage source converter, i.e. rectification, inversion and power zero transmission.

3. The method for siting and sizing voltage source converters based on low voltage AC / DC systems according to claim 1, characterized in that, In step S2, the Latin hypercube sampling generates multiple groups of samples for photovoltaic illumination and load under typical scenarios; the alternating iteration method decouples the system into an AC sub-network and a DC sub-network by taking the voltage source converter as a boundary, and alternately performs power flow calculation on both sides until convergence.

4. The method for locating and sizing voltage source converters based on low voltage AC / DC systems according to claim 1, characterized in that, In step S3, a multi-objective function is determined by minimum annual investment cost, minimum system loss, minimum voltage deviation and maximum local consumption rate of the distributed power supply; constraint conditions are determined by branch power flow constraint, voltage safety constraint, voltage source converter operation constraint, distributed power supply output constraint and energy storage operation constraint.

5. The method for voltage source converter siting and sizing in a low voltage AC / DC system according to claim 4, wherein, The multi-objective function is normalized into a single objective function by linear weighting method, wherein the sum of weight coefficients of each sub-objective is 1.

6. The method for voltage source converter siting and sizing in a low voltage AC / DC system according to claim 1, wherein, In step S4, a voltage-power sensitivity model is obtained by calculating sensitivity of voltage to active power and reactive power of the voltage source converter, and combining the weight coefficients to obtain a comprehensive sensitivity index for evaluating response capability of nodes to voltage source converter access.

7. The method for locating and sizing voltage source converters based on low voltage AC / DC systems according to claim 1 or 6, characterized in that, The optimal access location of the voltage source converter is selected from a set of candidate nodes, the set of candidate nodes excludes nodes directly connected with transformers and distributed power supply high penetration rate nodes, and the selection is based on ordering of the comprehensive sensitivity of the nodes from large to small.

8. The method of claim 1 or 6, wherein, In the comprehensive sensitivity model, the weight coefficients of voltage-active power sensitivity and voltage-reactive power sensitivity are determined according to the influence degree on system operation.

9. The method for voltage source converter siting and sizing in a low voltage AC / DC system according to claim 7, wherein, When the set of candidate nodes is constructed, nodes connected with original on-load branches in the system are taken as candidate access locations of the voltage source converter.

10. The method for siting and sizing voltage source converters based on low voltage AC / DC systems according to claim 1, wherein, In step S4, the rated capacity of the voltage source converter is determined by solving a multi-objective model with the minimum voltage deviation as the target, the model is also normalized by linear weighting, and needs to meet power and voltage operation constraints of the voltage source converter.

Images