Risk assessment and optimization method based on abnormal operation state of low-voltage distribution network

Abstract

The application discloses a risk assessment and optimization method based on abnormal operation state of low-voltage power distribution network, relates to the field of low-voltage power distribution network risk assessment and optimization, and comprises a low-voltage power distribution network risk assessment model, wherein the low-voltage power distribution network risk assessment model comprises a mathematical model unit, a deterministic three-phase power flow improved calculation unit, a source and load probability distribution unit and a Monte Carlo probability power flow calculation unit, abnormal operation state assessment parameters of the low-voltage power distribution network are defined based on the low-voltage power distribution network risk assessment model, risk weight coefficients are configured to the abnormal operation state assessment parameters of the low-voltage power distribution network to obtain comprehensive risk, and the comprehensive risk P sys The genetic algorithm is used to evaluate and optimize the risk of the low-voltage power distribution network, the above method is adopted to effectively reduce the risk of abnormal operation states, such as overvoltage and three-phase imbalance exceeding limit, of the low-voltage power distribution system in a region with high penetration rate of distributed new energy, and the probability statistical data of the power flow distribution of each phase and the neutral line are quickly and accurately calculated.

Description

Technical Field

[0001] This invention relates to the field of risk assessment and optimization of low-voltage distribution networks, and in particular to risk assessment and optimization methods based on abnormal operating conditions of low-voltage distribution networks. Background Technology

[0002] Low-voltage distribution networks typically operate radially with a single power flow direction. However, in recent years, with the rapid development of the national economy and the tremendous changes in the power industry, distributed power sources such as wind and solar power are increasingly being connected, making the operating environment of low-voltage distribution networks increasingly complex.

[0003] On the power supply side, distributed power sources, represented by photovoltaics, are extensively connected to low-voltage distribution networks; on the load side, single-phase high-power loads, represented by electric vehicles, are also extensively connected to low-voltage distribution networks. On the one hand, some of these new power sources are connected to the low-voltage distribution network in a single-phase manner, increasing the power difference between phases and exacerbating the imbalance between the three phases, leading to increased system losses and shortened equipment lifespan. Simultaneously, the low-voltage distribution system uses a three-phase four-wire system (ABC three-phase + N-line), and the power flow analysis method differs significantly from traditional three-phase symmetrical power flow calculations, requiring new power flow calculation methods. On the other hand, these new power sources have strong uncertainties, making the network power flow of the low-voltage distribution network more complex and increasing the risk of backflow, voltage exceeding limits, or overload. However, constrained by both safety and economic factors, traditional 400V low-voltage distribution networks still mainly adopt a radial structure. Their network structure is simple, relying on a single power source, lacking load transfer capabilities, and has low reliability. Currently, user power outages caused by low-voltage distribution networks account for more than 40%. The addition of new power sources will further exacerbate the risks of system backflow and voltage / current exceeding limits.

[0004] Against this backdrop, low-voltage distribution networks face a more complex operating environment and exhibit more complex operating characteristics. To further improve the technical foundation for power supply reliability and operating efficiency of low-voltage distribution networks, it is necessary to quantify the risks of abnormal operating states of low-voltage distribution networks and propose targeted risk power linkage control strategies.

[0005] To address these issues, there is an urgent need for risk assessment and optimization methods based on the abnormal operating conditions of low-voltage distribution networks. Summary of the Invention

[0006] To address the aforementioned issues and fill the gaps in research on risk assessment and optimization control of low-voltage power distribution systems, this application proposes a risk assessment and optimization method based on abnormal operating conditions of the low-voltage power distribution network for low-voltage power distribution systems with distributed renewable energy access, including the following steps: S1. Construct a low-voltage distribution network risk assessment model, which includes a mathematical model unit, a deterministic three-phase power flow improvement calculation unit, a source-load probability distribution unit, and a Monte Carlo probabilistic power flow calculation unit. S2. Risk assessment results are obtained by conducting a risk assessment of abnormal operating conditions of low-voltage distribution networks based on a low-voltage distribution network risk assessment model. S3. Based on the risk assessment results, the risks of the low-voltage distribution network are optimized using a genetic algorithm.

[0007] Preferably, the mathematical model units in S1 include line units, load units, distributed photovoltaic units, and transformer units: Line Unit: The line unit uses a simplified П-type equivalent circuit to describe the line, and obtains the resistance and reactance parameters based on the line model; Load Unit: The load unit sets the initial active power of each load node. reactive power Set the rated voltage of the load node. , To calculate the complex power at the k-th iteration, a hybrid load equation is constructed for constant power P, constant impedance Z, and constant current I. Distributed photovoltaic unit: The distributed photovoltaic unit defines distributed photovoltaic as PQ node, and photovoltaic output is defined as negative constant power load; Transformer Unit: The transformer unit adopts a single-phase transformer. A type of equivalent model describes the transformer, setting parameters such as the total resistance R of the transformer's high and low voltage windings. T The total reactance X of the high and low voltage windings of the transformer T The conductance G of the transformer T and the transformer's susceptance B T .

[0008] Preferably, the calculation steps of the deterministic three-phase power flow improvement calculation unit in S1 include: Step 1: Combine the resistance and reactance parameters with the topological relationship of the line to obtain the n-dimensional node admittance matrix; Extend the n-dimensional node admittance matrix to a 4n-dimensional node admittance matrix; The 4n-dimensional nodal admittance matrix is ​​reduced in dimension to obtain a 3n×3n 3n-dimensional nodal admittance matrix. ; Step 2: Classify the nodes in the low-voltage power distribution system to obtain balancing nodes and PQ nodes. The balancing nodes are the access points of the low-voltage power distribution system, and the remaining nodes are PQ nodes. For the slack node, the three phases A, B, C, and the neutral line on the primary side of the equivalent circuit of the distribution transformer are taken as the slack node, and their per-unit voltage values ​​are given as follows: , , , And it remains unchanged during the iteration process; For node PQ, both its active and reactive power are known. Set the initial voltage values ​​for the three phases (A, B, C) and neutral line of each node to be... , , , And update it before the next iteration begins; Step 3: Calculate the current imbalance based on the power flow equations of the low-voltage distribution network. ,test If the absolute value of the value meets the convergence accuracy, the calculation ends and the voltage data of each node is output; otherwise, the calculation continues. Step 4: Solve the fundamental correction equations based on the Jacobian matrix to obtain the correction amounts for the real and imaginary parts of the voltages at each PQ node. , ; Step 5: Combine the corrections for the real and imaginary parts of the voltages at each PQ node. , Correct the voltage of each PQ node, and calculate the neutral line voltage and current of each node by using the relationship between the three-phase currents of ABC and the neutral line current. Then execute step 3 to perform the next iteration calculation. The expression for the relationship between the three-phase currents (A, B, and C) and the neutral current is: ,in, , , , These are the current vectors flowing through phase lines ABC and the N line, respectively.

[0009] The expression for the neutral line voltage at a node is: ; in, , , , These are the voltage vectors of phase lines ABC and the N line, respectively.

[0010] The preferred expression for the basic correction equation is: ; Where J is the Jacobian matrix, and its elements are 6×6 matrices. and For current correction vector The real and imaginary parts; and For node voltage correction vector The real and imaginary parts.

[0011] Preferably, the construction steps of the source load probability distribution unit include: The load probability density function is constructed based on the random variables of active power and reactive power. The expression of the load probability density function is as follows: ; in, and This represents the mean and variance of the active power at the initial stage of load node iteration at a certain moment; and This represents the mean and variance of reactive power at the initial stage of load node iteration at a certain moment. Let active power be a random variable. Let reactive power be a random variable. Pi is the mathematical constant of a circle.

[0012] The source-load probability distribution unit is obtained by constructing the probability density function of the active power output of the photovoltaic grid-connected system based on the load probability density function. The probability density function of the active power output of a grid-connected photovoltaic system is expressed as follows: ; ; Among them, P max S represents the maximum active power output of the photovoltaic equipment, and S represents the total area of ​​the photovoltaic panels. The combined energy conversion efficiency of the photovoltaic panel and inverter. and Let be the shape parameter of the Beta distribution. The maximum light intensity for the next hour is determined by the mean and variance of the light intensity for the next hour. P G It is a random variable, that is, the independent variable of the probability density function, which characterizes the probability distribution of photovoltaic power output.

[0013] Preferably, the calculation steps of the Monte Carlo probabilistic power flow calculation unit include: a. Latin hypercube sampling is performed on the probability distribution of photovoltaic power sources and loads calculated in the source-load probability distribution unit. Each time a scene is generated, Latin hypercube sampling is performed on each device in a fixed order to obtain the power of each device, thereby obtaining a large-scale random scene set. b. Call the deterministic three-phase power flow improvement calculation unit to solve the system state under each scenario; c. Use the error bounds method to determine the number of target scenarios required to be generated by the deterministic three-phase power flow improvement computing unit. And by dynamically controlling the number of scenes generated in the Monte Carlo simulation; d. In probabilistic power flow calculations, node voltage statistics and error analysis are performed after every 10% of the minimum scenario calculations are completed. If the error... If the convergence condition is met, the probabilistic power flow calculation is completed; otherwise, 10% of the target scene numbers are generated for calculation until the convergence condition is met.

[0014] e. Statistically analyze the distribution of electrical quantities at each node across all scenarios and represent it using a frequency distribution histogram; f. Use kernel density estimation to process the frequency distribution histogram of each node voltage, and then obtain the probability density function of each node voltage.

[0015] Preferably, the abnormal operation status assessment parameters for medium and low voltage distribution networks in S2 include the voltage over-limit probability. Three-phase imbalance degree exceeding the limit probability Branch current over-limit probability ; Comprehensive Risk P sys The expression is: ; in, These are the voltage over-limit probabilities. Three-phase imbalance degree exceeding the limit probability Branch current over-limit probability Weighting coefficients; Preferably, in S3, based on comprehensive risk P sys The specific content of optimizing the risks of low-voltage distribution networks using genetic algorithms is as follows: S301: Set the objective function and constraint function, and encode each device in combination with the power encoding method of each device. For energy storage devices, the output size is directly encoded. For photovoltaic power sources, the output size is limited to regulate the photovoltaic power source. Set the output coefficient k of the photovoltaic power source, which has a value range of 0-1 and can be directly encoded according to the conversion between decimal and binary. S302. Concatenate the codes of each device to obtain the chromosome in the genetic algorithm, that is, the individual in the population; S303: Solve the individual population using a genetic algorithm to optimize the risk of the low-voltage distribution network.

[0016] In summary, the risk assessment and optimization method for abnormal operating conditions of low-voltage distribution networks, based on this invention, has the following advantages compared to traditional technologies: 1. This invention addresses the unique structure of a low-voltage distribution network with three phases ABC and a neutral line. It proposes to reduce the matrix dimension to 3n×3n by utilizing the relationship between the three phase currents ABC and the neutral line current, while retaining the parameters of the three phases ABC and eliminating the independent dimension of the neutral line. This simplifies the calculation while ensuring the accuracy of power flow calculation. 2. This invention employs a multi-objective optimization genetic algorithm, which balances the goals of reducing overall risk and minimizing photovoltaic curtailment, while also improving power supply reliability, reducing the risk of power outages, optimizing operational efficiency, supporting a high proportion of new source load access, and contributing to the construction of new power systems.

[0017] The technical method of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the generation of the node admittance matrix for low-voltage distribution network power flow calculation according to the present invention. Figure 2 This is the coding method for the energy storage device of the present invention; Figure 3 This is a flowchart illustrating the steps of the risk assessment and optimization method for abnormal operating conditions of low-voltage distribution networks according to the present invention. Detailed Implementation

[0019] The technical method of the present invention will be further described below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of this application.

[0020] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the scope of this application and its application or use.

[0021] Techniques, systems, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, they should be considered part of the instruction manual.

[0022] In all the examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0023] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0024] Example 1 This invention provides a risk assessment and optimization method based on abnormal operating conditions of low-voltage distribution networks, such as... Figure 3 As shown, considering the risk assessment of a three-phase four-wire low-voltage distribution network with three-phase + N lines and new source loads, matrix dimensionality reduction is used in power flow calculation. The parameters of the three phases (A, B, and C) are retained, and the independent dimension of the neutral line is eliminated, simplifying the calculation while ensuring the accuracy of power flow calculation. A genetic algorithm is used to optimize the probability characteristic parameter of the low-voltage distribution network. The specific details are as follows: S1. Construct a low-voltage distribution network risk assessment model. The low-voltage distribution network risk assessment model includes a mathematical model unit, a deterministic three-phase power flow improvement calculation unit, a source-load probability distribution unit, and a Monte Carlo probabilistic power flow calculation unit. The mathematical model unit receives actual physical parameter inputs, including transformer parameters, line lengths, and cable parameters, and then constructs the node admittance matrix of the entire network. The deterministic three-phase power flow improvement calculation unit, as the basis of the Monte Carlo probabilistic power flow calculation unit, provides power flow calculation method support for a single scenario generated by Monte Carlo sampling. The source-load probability distribution unit determines the probability density function of Monte Carlo sampling by analyzing the probabilistic characteristics of the output or absorption power parameters of the power source and load. The Monte Carlo probabilistic power flow calculation unit integrates the node admittance matrix generated by the mathematical model unit, the improved three-phase power flow algorithm, and the sampling data of each source load to calculate the probability distribution of the system power flow.

[0025] Furthermore, the mathematical model units in S1 include line units, load units, distributed photovoltaic units, and transformer units: Line Unit: The line unit uses a simplified П-type equivalent circuit to describe the line, assuming that the three-phase line parameters are symmetrical, the line parameters of the three phase lines are different from those of the neutral line, and the resistance and reactance parameters are obtained based on the line model; Load Unit: The load unit sets the initial active power of each load node. reactive power Set the rated voltage of the load node. , To calculate the complex power at the k-th iteration, a hybrid load equation is constructed for constant power P, constant impedance Z, and constant current I. The expression for the hybrid load equation is as follows: ; Where j is the imaginary unit, The defined ordinal parameter has no practical meaning. The proportion of each type of load to the total load, among which When =0, it represents a constant power load. =1 represents a constant impedance load. =2 represents a constant current load; V k This represents the voltage value of the node during the k-th iteration.

[0026] Distributed Photovoltaic Unit: The distributed photovoltaic unit defines distributed photovoltaic as a PQ node, and photovoltaic output is defined as a negative constant power load. Therefore, the expression for the complex power injected into each phase by the photovoltaic power source at the k-th iteration is: ; in, Inject complex power into the photovoltaic power source of each phase in the k-th iteration. Injecting active power into the initial photovoltaic power source, Inject reactive power into the initial photovoltaic power source.

[0027] Transformer Unit: The transformer unit uses a single-phase transformer Γ-type equivalent model to describe the transformer, and sets the parameter R. T X T G T and B T The expression for solving the parameters is: ; Among them, R T Total resistance (Ω) of transformer high and low voltage windings, X T Total reactance (Ω) of transformer high and low voltage windings, G T Transformer conductance (S), B T Transformer susceptance (S), U N U is the rated voltage of the transformer. k % represents the percentage of the transformer's short-circuit voltage, P k The transformer's short-circuit loss, P0 is the transformer's no-load loss, I0% is the transformer's no-load current percentage, S N This refers to the rated capacity of the transformer.

[0028] Furthermore, the calculation steps for the deterministic three-phase power flow improvement calculation unit in S1 include: Step 1: Obtain the n-dimensional nodal admittance matrix by combining resistance, reactance parameters, and line topology; extend the n-dimensional nodal admittance matrix to a 4n-dimensional nodal admittance matrix, as follows: Figure 1 As shown; Understandably, for a n For a node-based low-voltage distribution network, the three-phase +N line voltage vector of all nodes in the system is defined. With the three-phase +N line current vector Then, the relationship between the bus voltage and current, which is governed by Kirchhoff's current and voltage laws, is as follows: ,in, For a 4 n ×4 n A complex matrix.

[0029] After obtaining the admittance matrix of the low-voltage distribution network nodes, the relationship between bus voltage and current can be expressed as a... n Nodal current injection equations for a three-phase +N low-voltage distribution network: ; Each element is in matrix form, where U kRepresents a node k The set of voltages at the +N line node of the three phases. I k Represents a node k The collection of injected currents at the upper three-phase +N line nodes. This represents the self-admittance or mutual admittance matrix of each node.

[0030] For nodes k Three-phase +N line current and voltage: ; ; For matrix Y k Inverse the matrix to obtain the nodal impedance matrix. Z k， Right now: ; Then the nodes are obtained k The matrix expressions for the three-phase +N line current and voltage are as follows: ; In power distribution lines, the parameters of the three-phase lines can be considered symmetrical, that is: ; In the formula, Z SP Z represents the self-impedance of a three-phase line. MP Z represents the mutual impedance of a three-phase line. MG Z represents the mutual impedance of a three-phase line to ground. NN This represents the self-impedance of the neutral line; Meanwhile, the expression for the relationship between the three-phase currents (A, B, and C) and the neutral current is: ; Node k Expanding the matrix expressions for the three-phase +N line current and voltage by rows, and combining them with the expressions for the three-phase current and neutral line current in equation ABC, we have: .

[0031] After elementary row operations on the matrix, the parameters of the three phase lines can be decoupled from the parameters of the neutral line: .

[0032] Thus, if the potential of the neutral line is used as the reference potential, the line parameters of the neutral line can be reduced to the three phase lines, eliminating the row and column corresponding to the neutral line: .

[0033] To simplify the writing, the following rules are given: ; The matrix described above, reduced to three dimensions, can be written as: .

[0034] For matrix Inverting the matrix yields the dimension-reduced node admittance matrix: .

[0035] Therefore, the node injection current equations for the entire low-voltage power distribution system after dimension reduction can be obtained as follows: .

[0036] The 4n-dimensional nodal admittance matrix is ​​reduced in dimension to obtain a 3n×3n 3n-dimensional nodal admittance matrix. .

[0037] Step 2: Classify the nodes in the low-voltage power distribution system to obtain balancing nodes and PQ nodes. Since there are generally no large-capacity devices that maintain voltage amplitude except for the system access point, the balancing node is the low-voltage power distribution system access point, and the remaining nodes are PQ nodes.

[0038] For the slack node, the three phases A, B, C, and the neutral line on the primary side of the equivalent circuit of the distribution transformer are taken as the slack node, and their per-unit voltage values ​​are given as follows: , , , And it remains unchanged during the iteration process.

[0039] For node PQ, both its active and reactive power are known. Set the initial voltage values ​​for the three phases (A, B, C) and neutral line of each node to be... , , , And it is updated before the next iteration begins.

[0040] node neutral line voltage during iteration Use the results calculated in the previous iteration.

[0041] Calculate the current imbalance at each node, expand the injection current equation row by row to obtain the first row-by-row expression of the injection current equation: .

[0042] In the formula, The sub-blocks belonging to the line self-admittance or mutual admittance in the line node admittance matrix are third-order square matrices. and This is a three-dimensional column vector of the voltage and current corresponding to this sub-block.

[0043] Calculate the node injection current, which is derived from the power injected into each node and the power absorbed by the load: .

[0044] Substituting these equations into the block-wise expression of the injection current equation, we obtain the equilibrium equation for the nodal injection current: .

[0045] Expanding the equilibrium equations for the injected current at nodes composed of sub-blocks further row-wise yields the second-order row-wise block expression for the injected current equations: .

[0046] In the formula, x , y For the row and column indices of the sub-blocks of the node admittance matrix, The correction factor for the node injection current is used; therefore, the second-order row-block expression of the injection current equation is also the correction equation for power flow calculation. This equation is a complex equation. By decomposing its real and imaginary parts according to the decomposition expression method, the initial correction equation of the system power flow equation is obtained. The expression of the initial correction equation of the system power flow equation is as follows: .

[0047] The decomposition expression is: .

[0048] The injection current equation is expressed in a row-block format as the current balance equation for each node. Fully expanding this expression yields 3n equations, among which... and For current correction vector The real and imaginary parts; and The matrix obtained by extracting the real and imaginary parts of each element in the nodal admittance matrix; and For node voltage vectors The real and imaginary parts.

[0049] Step 3: Calculate the current imbalance based on the power flow equations of the low-voltage distribution network. ,test If the absolute value of the value meets the convergence accuracy, the calculation ends and the voltage data of each node is output; otherwise, the calculation continues.

[0050] By linearizing the initial modified equations of the system's power flow equations and arranging their first derivatives with respect to their respective variables in a certain manner, we can obtain the expression for the basic modified equations of the system's power flow equations as follows: .

[0051] inJ For Jacobian matrices, and For current correction vector The real and imaginary parts; and For node voltage correction vector The real and imaginary parts.

[0052] Step 4: Solve the fundamental correction equations based on the Jacobian matrix to obtain the correction amounts for the real and imaginary parts of the voltages at each PQ node. , .

[0053] Step 5: Combine the corrections for the real and imaginary parts of the voltages at each PQ node. , Correct the voltages at each PQ node: , , This is the corrected node voltage.

[0054] Then, by calculating the neutral line voltage and current at each node through the relationship between the three-phase currents (A, B, and C) and the neutral line current, step 3 is executed to perform the next iteration calculation.

[0055] The power flow equations of the system are iteratively solved based on the given initial voltage values ​​until the iterations converge, thus obtaining the solution for the deterministic power flow calculation.

[0056] Furthermore, the calculation process for each element in the Jacobian matrix includes: .

[0057] If i = j, then add the following correction term to the above formula: .

[0058] in: .

[0059] .

[0060] in and The matrix obtained by extracting the real and imaginary parts of each element in the nodal admittance matrix. and For node voltage vectors The real and imaginary parts.

[0061] Furthermore, the construction steps of the source load probability distribution unit include: The probability distribution of the load is reflected in the mathematical model as the initial input active and reactive power. Therefore, a load probability density function is constructed based on the random variables of active and reactive power. The expression of the load probability density function is as follows: .

[0062] in, and This represents the mean and variance of the active power at the initial stage of load node iteration at a certain moment; and This represents the mean and variance of reactive power at the initial stage of load node iteration at a certain moment. Let active power be a random variable. Let reactive power be a random variable. Pi is the mathematical constant of a circle.

[0063] Constructing a probability distribution model for distributed photovoltaic power sources: The source-load probability distribution unit is obtained by constructing the probability density function of the active power output of the photovoltaic grid-connected system based on the load probability density function.

[0064] The probability density function of the active power output of a grid-connected photovoltaic system is expressed as follows: .

[0065] .

[0066] Among them, P max S represents the maximum active power output of the photovoltaic equipment, and S represents the total area of ​​the photovoltaic panels. The combined energy conversion efficiency of the photovoltaic panel and inverter. and Let be the shape parameter of the Beta distribution. This represents the maximum light intensity within the next hour, and its value is determined by the mean and variance of the light intensity within the next hour. P G It is a random variable, that is, the independent variable of the probability density function.

[0067] Furthermore, the calculation steps of the Monte Carlo probabilistic power flow calculation unit include: a. Latin hypercube sampling is performed on the probability distribution of photovoltaic power sources and loads calculated in the source-load probability distribution unit. Each time a scene is generated, Latin hypercube sampling is performed on each device in a fixed order to obtain the power of each device, thereby obtaining a large-scale random scene set.

[0068] b. Call the deterministic three-phase power flow improvement calculation unit to solve the system state under each scenario.

[0069] c. Use the error bounds method to determine the number of target scenarios required to be generated by the deterministic three-phase power flow improvement computing unit. And by dynamically controlling the number of scenes generated in the Monte Carlo simulation.

[0070] d. In probabilistic power flow calculations, node voltage statistics and error analysis are performed after every 10% of the minimum scenario calculations are completed. If the error... If the convergence condition is met, the probabilistic power flow calculation is completed; otherwise, 10% of the target scene numbers are generated for calculation until the convergence condition is met.

[0071] e. Statistically analyze the distribution of electrical quantities at each node across all scenarios and represent it using a frequency distribution histogram.

[0072] f. Use kernel density estimation to process the frequency distribution histogram of each node voltage, and then obtain the probability density function of each node voltage.

[0073] Furthermore, the number of target scenarios The expression is: .

[0074] in, This indicates the accuracy of probabilistic power flow calculations, i.e., the error bound. This represents the overall standard deviation of the load and photovoltaic power in the system. For confidence level, This represents the Z-value corresponding to the confidence level.

[0075] The expression for error analysis is: .

[0076] in, This represents the total number of scenarios for which deterministic power flow calculations have been completed. Indicates to The standard deviation of node voltage was obtained by statistical analysis of various scenarios.

[0077] The S2 medium and low voltage distribution network abnormal operation status assessment parameters include voltage over-limit probability. Three-phase imbalance degree exceeding the limit probability Branch current over-limit probability .

[0078] Comprehensive Risk P sys The expression is: .

[0079] in, These are the voltage over-limit probabilities. Three-phase imbalance degree exceeding the limit probability Branch current over-limit probability Weighting coefficients; .

[0080] The probability density function representing voltage is calculated by the improved deterministic three-phase power flow calculation unit, V. min and V max To set the upper and lower limits for normal operation of node voltage.

[0081] .

[0082] The probability density function representing the three-phase voltage imbalance is calculated by the deterministic three-phase power flow improvement calculation unit.

[0083] .

[0084] The probability density function representing the branch current is calculated by the improved deterministic three-phase power flow calculation unit. d The Greek letters that follow are all integral variables, and their dimensions are consistent with the upper and lower limits of integration.

[0085] S3, Based on comprehensive risk P sys The risks of low-voltage distribution networks are optimized using genetic algorithms. The specific content is as follows: S301: Define the objective function and constraint functions, and encode each device according to its power encoding method. For energy storage devices, their output power is directly encoded, such as... Figure 2 As shown, the sign bit indicates the power absorbed or emitted by the energy storage device, with 0 representing absorbed power and 1 representing emitted power; the value bit indicates the magnitude of its absorbed or emitted power, which can be directly encoded based on the conversion between decimal and binary. For photovoltaic power sources, the output power is controlled by limiting its output magnitude. The output coefficient k of the photovoltaic power source is set, with a value range of 0-1, which can be directly encoded based on the conversion between decimal and binary.

[0086] The objective function includes: The first optimization objective function for the comprehensive operational risk of the low-voltage power distribution system, obtained through probabilistic power flow calculation, is expressed as follows: .

[0087] The second optimization objective function for the photovoltaic power output in the low-voltage power distribution system is expressed as follows: .

[0088] in, This indicates the theoretical maximum output of the photovoltaic power source. This indicates the maximum output of the photovoltaic power source after limitations are imposed. This represents the total amount of curtailed solar power from all photovoltaic sources within the system.

[0089] The constraint function is the fitness function of an individual in the genetic algorithm, and the expression of the fitness function is: .

[0090] in, This represents the total amount of light wasted by the system after normalization. This indicates the overall operational risk of the system.

[0091] S302. Concatenate the codes from each device to obtain the chromosomes in the genetic algorithm, i.e., the individuals in the population. .

[0092] S303: Solve the individual in the population according to the genetic algorithm, and then optimize the risk of the low voltage distribution network. S3031: A set of binary codes is randomly generated by the computer.

[0093] S3032: Decode each binary code to obtain the corresponding source-storage-output strategy and calculate its fitness.

[0094] S3033: Check if any individual in the source-storage-output strategy group has fitness that meets the requirements. If not, continue to perform the genetic algorithm optimization operation; if it does, output the source-storage-output strategy that meets the requirements and stop the calculation.

[0095] S3034: Perform copying, crossover, and mutation operations on the binary code group to obtain the next generation of binary code group, and return to S3032.

[0096] S3035: If an individual with satisfactory fitness is still not obtained after a sufficient number of iterations, output the source-load output strategy with the lowest probability of running risk during all iterations.

[0097] This invention constructs a three-phase + N-line model, employs probabilistic power flow and Latin hypercube sampling to quantify the risks of voltage exceeding limits, three-phase imbalance, and branch current exceeding limits. It clarifies the risk assessment mode and optimizes the number of scenarios through the error bounds method. Combined with kernel density estimation, it covers the entire source-load state, avoids excessive scenario generation leading to wasted computing resources, improves risk assessment efficiency, and uses optimization algorithms to reduce the risk of abnormal operation and improve the reliability and stability of low-voltage power distribution system operation.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical methods of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical methods of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical methods to deviate from the spirit and scope of the technical methods of the present invention.

Claims

1. A risk assessment and optimization method based on abnormal operating conditions of low-voltage distribution networks, characterized in that, Includes the following steps: S1. Construct a low-voltage distribution network risk assessment model, which includes a mathematical model unit, a deterministic three-phase power flow improvement calculation unit, a source-load probability distribution unit, and a Monte Carlo probabilistic power flow calculation unit. S2. Based on the low-voltage distribution network risk assessment model, define the assessment parameters for abnormal operation status of the low-voltage distribution network, and configure risk weight coefficients for these parameters to obtain the comprehensive risk P. sys ; S3, Based on comprehensive risk P sys Risk assessment and optimization of low-voltage distribution networks using genetic algorithms; The mathematical model units in S1 include line units, load units, distributed photovoltaic units, and transformer units: Line Unit: The line unit uses a simplified П-type equivalent circuit to describe the line, and obtains the resistance and reactance parameters based on the line model; Load Unit: The load unit sets the initial active power of each load node. reactive power Set the rated voltage of the load node. , To calculate the complex power at the k-th iteration, a hybrid load equation is constructed for constant power P, constant impedance Z, and constant current I. Distributed photovoltaic unit: The distributed photovoltaic unit defines distributed photovoltaic as PQ node, and photovoltaic output is defined as negative constant power load; Transformer Unit: The transformer unit adopts a single-phase transformer. A type of equivalent model describes the transformer, setting parameters such as the total resistance R of the transformer's high and low voltage windings. T The total reactance X of the high and low voltage windings of the transformer T The conductance G of the transformer T and the transformer's susceptance B T ; The calculation steps of the deterministic three-phase power flow improvement calculation unit in S1 include: Step 1: Combine the resistance and reactance parameters with the topological relationship of the line to obtain the n-dimensional node admittance matrix; Extend the n-dimensional node admittance matrix to a 4n-dimensional node admittance matrix; The 4n-dimensional nodal admittance matrix is ​​reduced in dimension to obtain a 3n×3n 3n-dimensional nodal admittance matrix. ; Step 2: Classify the nodes in the low-voltage power distribution system to obtain balancing nodes and PQ nodes. The balancing nodes are the access points of the low-voltage power distribution system, and the remaining nodes are PQ nodes. For the slack node, the three phases A, B, C, and the neutral line on the primary side of the equivalent circuit of the distribution transformer are taken as the slack node, and their per-unit voltage values ​​are given as follows: , , , And it remains unchanged during the iteration process; For node PQ, both its active and reactive power are known. Set the initial voltage values ​​for the three phases (A, B, C) and neutral line of each node to be... , , , And update it before the next iteration begins; Step 3: Calculate the current imbalance based on the power flow equations of the low-voltage distribution network. ,test If the absolute value of the value meets the convergence accuracy, the calculation ends and the voltage data of each node is output; otherwise, the calculation continues. Step 4: Solve the fundamental correction equations based on the Jacobian matrix to obtain the correction amounts for the real and imaginary parts of the voltages at each PQ node. , ; Step 5: Combine the corrections for the real and imaginary parts of the voltages at each PQ node. , The voltages of each PQ node are corrected, and the neutral line voltage and current of each node are calculated by using the relationship between the three-phase currents of ABC and the neutral line current. Then, step 3 is executed to perform the next iteration calculation.

2. The risk assessment and optimization method based on abnormal operating conditions of low-voltage distribution networks according to claim 1, characterized in that, The expression for the basic correction equation is: ； Where J is the Jacobian matrix, and its elements are 6×6 matrices. and For current correction vector The real and imaginary parts; and For node voltage correction vector The real and imaginary parts.

3. The risk assessment and optimization method based on abnormal operating conditions of low-voltage distribution networks according to claim 2, characterized in that, The construction steps of the source load probability distribution unit include: The load probability density function is constructed based on the random variables of active power and reactive power. The expression of the load probability density function is as follows: ； in, and This represents the mean and variance of the active power at the initial stage of load node iteration at a certain moment; and This represents the mean and variance of reactive power at the initial stage of load node iteration at a certain moment. Let active power be a random variable. Let reactive power be a random variable. Pi; The source-load probability distribution unit is obtained by constructing the probability density function of the active power output of the photovoltaic grid-connected system based on the load probability density function. The probability density function of the active power output of a grid-connected photovoltaic system is expressed as follows: ； ； in, P max S represents the maximum active power output of the photovoltaic equipment, and S represents the total area of ​​the photovoltaic panels. The combined energy conversion efficiency of the photovoltaic panel and inverter. and Let be the shape parameter of the Beta distribution. The maximum light intensity in the next hour. P G It is a random variable, that is, the independent variable of the probability density function.

4. The risk assessment and optimization method based on abnormal operating conditions of low-voltage distribution networks according to claim 3, characterized in that, The calculation steps of the Monte Carlo probabilistic power flow calculation unit include: a. Latin hypercube sampling is performed on the probability distribution of photovoltaic power sources and loads calculated in the source-load probability distribution unit. Each time a scene is generated, Latin hypercube sampling is performed on each device in a fixed order to obtain the power of each device, thereby obtaining a large-scale random scene set. b. Call the deterministic three-phase power flow improvement calculation unit to solve the system state under each scenario; c. Use the error bounds method to determine the number of target scenarios required to be generated by the deterministic three-phase power flow improvement computing unit. ; d. In probabilistic power flow calculations, node voltage statistics and error analysis are performed after every 10% of the minimum scenario calculations are completed. If the error... If the convergence condition is met, the probabilistic power flow calculation is completed; otherwise, 10% of the target scene numbers are generated for calculation until the convergence condition is met. e. Statistically analyze the distribution of electrical quantities at each node across all scenarios and represent it using a frequency distribution histogram; f. Use kernel density estimation to process the frequency distribution histogram of each node voltage, and then obtain the probability density function of each node voltage.

5. The risk assessment and optimization method based on abnormal operating conditions of low-voltage distribution networks according to claim 4, characterized in that, The S2 medium and low voltage distribution network abnormal operation status assessment parameters include voltage over-limit probability. Three-phase imbalance degree exceeding the limit probability Branch current over-limit probability ; Comprehensive Risk P sys The expression is: ； in, These are the voltage over-limit probabilities. Three-phase imbalance degree exceeding the limit probability Branch current over-limit probability Weighting coefficients; ； The probability density function representing voltage is calculated by the improved deterministic three-phase power flow calculation unit, V. min and V max To set the upper and lower limits for normal operation of node voltage; ； The probability density function representing the three-phase voltage imbalance is calculated by the deterministic three-phase power flow improvement calculation unit; ； The probability density function representing the branch current is calculated by the improved deterministic three-phase power flow calculation unit.

6. The risk assessment and optimization method based on abnormal operating conditions of low-voltage distribution networks according to claim 5, characterized in that, S3 is based on comprehensive risk P sys The specific content of optimizing the risks of low-voltage distribution networks using genetic algorithms is as follows: S301: Set the objective function and constraint function, and encode each device in combination with the power encoding method of each device. For energy storage devices, the output size is directly encoded. For photovoltaic power sources, the output size is limited to regulate the photovoltaic power source. Set the output coefficient k of the photovoltaic power source, which has a value range of 0-1, and encode it directly based on the conversion between decimal and binary. S302. Concatenate the codes of each device to obtain the chromosome in the genetic algorithm, that is, the individual in the population; S303: Solve the individual population using a genetic algorithm to optimize the risk of the low-voltage distribution network.

7. The risk assessment and optimization method based on abnormal operating conditions of low-voltage distribution networks according to claim 6, characterized in that, The objective function includes: The first optimization objective function for the comprehensive operational risk of the low-voltage power distribution system, obtained through probabilistic power flow calculation, is expressed as follows: ； The second optimization objective function for the photovoltaic power output in the low-voltage power distribution system is expressed as follows: ； in, This indicates the theoretical maximum output of the photovoltaic power source. This indicates the maximum output of the photovoltaic power source after limitations are imposed. This represents the total amount of curtailed solar power from all photovoltaic sources within the system. The constraint function is the fitness function of an individual in the genetic algorithm, and the expression of the fitness function is: ； in, This represents the total amount of light wasted by the system after normalization. This indicates the overall operational risk of the system. This is the individual fitness function in the genetic algorithm; The expression is: 。

Images