Fault location method for distribution network with distributed power based on AOA

Abstract

The application discloses a kind of based on AOA's fault location method of distribution network with distributed power supply, belong to power system fault diagnosis technical field.The method of the application comprises: first, distribution network topology is discretized into feeder section, actual fault current array is generated using the fault current information collected by feeder terminal unit, and switch expectation function is established to associate feeder section fault state with expected fault current array;Second, construct evaluation function based on the difference degree between actual fault current array and expected fault current array;Finally, improved archimedes optimization algorithm is used for optimization solution.The method of the application effectively solves the problem that the direction of power flow is variable due to the access of distributed power supply to distribution network, and the adaptability of traditional methods is insufficient, and realizes fast and accurate fault section positioning in single-point and multi-point fault scenarios.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power system fault diagnosis, and particularly relates to a fault positioning method for a distribution network containing distributed power sources based on AOA. BACKGROUND

[0002] Accurate positioning of distribution network faults is a key foundation for ensuring power supply reliability and rapid recovery. Under the background of energy transformation, renewable energy sources such as solar energy have been massively connected to distribution networks in the form of distributed power sources due to their flexibility and economic advantages, resulting in increasingly complex system topologies. The randomness and intermittency of distributed power output have significantly reduced the accuracy of traditional fault positioning methods. Therefore, fault positioning for distribution networks containing distributed power sources has become a research hotspot in the field of distribution networks.

[0003] When there is an interphase short-circuit fault in a distribution network, the fault characteristics are obvious, and accurate fault information can be collected and extracted using feeder terminal unit (FTU) devices. Through a positioning algorithm that is fast and has high fault tolerance, the fault section can be positioned, which is the main research method at present. The main process is that the fault current information detected by the FTU is uploaded to the supervisory control and data acquisition (SCADA) system, and the fault section is calculated through intelligent algorithms. At present, distribution network fault positioning methods are diversified, but the large-scale connection of distributed power sources (DG) has led to complex system topologies and variable power flow directions, posing serious challenges to traditional positioning techniques. The mainstream intelligent optimization algorithm realizes positioning by constructing a fault state evaluation function, but the convergence efficiency and fault tolerance performance vary due to differences in algorithm architecture, and there is an urgent need to improve the positioning accuracy and speed through core algorithm improvement. Existing optimization algorithms (such as genetic algorithms and particle swarm algorithms) have a certain degree of fault tolerance, but they often face problems such as the need for a large number of populations, multiple iterations, or being easily trapped in local optimal solutions.

[0004] Therefore, there is an urgent need in the prior art for a new technical solution to solve this problem. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a fault positioning method for a distribution network containing distributed power sources based on AOA, which solves the problem of poor positioning accuracy and speed of the existing fault positioning method for a distribution network containing distributed power sources.

[0006] The technical solution adopted by the present application is to provide a fault positioning method for a distribution network containing distributed power sources based on AOA, comprising the following steps:

[0007] Step 1: fault information collection and discretization coding: the power distribution network is abstracted as a plurality of feeder sections segmented by switch nodes, and a binary coded individual position vector is generated with the fault state of each feeder section as an element; an actual fault current array is generated according to the actual fault current information reported by the feeder terminal unit (FTU) installed on the power distribution network; an expected fault current array is generated according to the individual position vector and a preset switch expectation function;

[0008] Step 2: evaluation function design: a fitness function is constructed for calculating the difference between the actual fault current array and the expected fault current array;

[0009] Step 3: Archimedes optimization algorithm for fault location: the fitness function is optimized and solved by using an improved Archimedes optimization algorithm, including initializing the population, iteratively updating the density, volume, acceleration and position vector of the individual, and converting the continuous position vector into binary coding through discrete mapping; when the iteration termination condition is met, the globally optimal individual position vector is output, and the binary coding thereof corresponds to the located fault feeder section.

[0010] The coding rule of the actual fault current array in step 1 is that, for any switch node:

[0011] If no fault current is detected or cannot be detected, it is coded as 0;

[0012] If the fault current is detected and its direction is the same as the specified positive direction, it is coded as 1;

[0013] If the fault current is detected and its direction is opposite to the specified positive direction, it is coded as -1;

[0014] Wherein, the specified positive direction is: when there is no distributed power supply in the feeder section, the direction from the system power supply to the load; when there is a distributed power supply in the feeder section, the direction from the system power supply to the distributed power supply.

[0015] The specific expression of the switch expectation function in step 1 is:

[0016] ;

[0017] In the formula: n1 is the total number of feeder sections upstream of the i-th switch; n2 is the total number of feeder sections downstream of the i-th switch; n DG is the total number of downstream distributed power supplies; K j is the j-th distributed power supply switch downstream; L j is the fault state of the j-th feeder section in the upstream or downstream section.

[0018] The specific expression of the fitness function in step 2 is:

[0019] ;

[0020] wherein: W is a very large number; N is the total number of switches in the distribution network with distributed power supply; S i is the actual fault current array information of the i-th switch node detected by the FTU device; is the expected fault current state array information of each switch node; n is the total number of feeder sections in the distribution network with distributed power supply; is the preset weight coefficient; is the sum of the number of all fault feeder sections.

[0021] The discretization process of the improved Archimedean optimization algorithm in step 3 is to map the continuous position vector into discrete values of 0 or 1 by using a sigmoid function.

[0022] The optimization solution in step 3 balances the global exploration and local exploitation capabilities of the algorithm by setting the transition operator TF and the density factor d, and the calculation formula of the transition operator TF and the density factor d is:

[0023] ;

[0024] wherein, t represents the current iteration number, T max represents the maximum iteration number.

[0025] The process of iteratively updating the individual position in step 3 includes:

[0026] If TF≤0.5, global exploration is performed, and the update formula of the individual acceleration acc and position x is:

[0027] ;

[0028] ;

[0029] ;

[0030] If TF>0.5, local development is performed, and the update formula of the individual acceleration acc and position x is:

[0031] ;

[0032] ;

[0033] wherein, represents the i thThe acceleration of the individual, den, vol, and acc represent the density, volume, and acceleration of the individual, the subscript mr represents a random individual, the subscript best represents the current optimal individual, C1 and C2 are constants, rand is a random number in [0, 1], F is a direction factor, and T is a constant.

[0034] The direction factor F is calculated as follows:

[0035] ;

[0036] Wherein, , is a constant.

[0037] A computer readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the steps of the AOA-based fault location method for a distribution network containing distributed power sources.

[0038] The above design scheme can bring the following beneficial effects:

[0039] The standard continuous domain Archimedes optimization algorithm (AOA) is improved by discretization (such as using Sigmoid mapping strategy), which is adapted to the binary coding characteristics of the fault state of the distribution network, so that fault section positioning can be realized in single-point and multi-point fault scenarios. Compared with traditional algorithms (such as genetic algorithm and particle swarm algorithm), the AOA algorithm can quickly converge to the optimal solution, effectively improving the solution speed of fault location, thereby effectively shortening the overall time length of fault handling and improving the efficiency of power supply recovery. And fully considering the variability of the topology structure and power flow direction of the distribution network after the access of distributed power sources, by clearly defining the positive direction of the fault current and constructing the corresponding switch expectation function, the method can flexibly adapt to complex scenarios with different numbers and different positions of distributed power sources. Whether the distribution network is without DG, partially accesses DG or fully accesses DG, the method shows good positioning performance. The AOA algorithm successfully migrates the advantages of the AOA algorithm in engineering optimization problems to the field of power system fault diagnosis, providing a new solution to the problem of fault location of active distribution networks. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1 It is the overall flowchart of the AOA-based fault location method for a distribution network containing distributed power sources.

[0041] Figure 2 It is the IEEE-33 node power system diagram of the AOA-based fault location method for a distribution network containing distributed power sources. DETAILED DESCRIPTION

[0042] The application is further illustrated in conjunction with the accompanying drawings and specific embodiments:

[0043] The AOA-based fault location method for distribution network with distributed power supply includes the following steps:

[0044] Step 1: Fault information collection and discretization coding.

[0045] Step 2: Design of evaluation function.

[0046] Step 3: Archimedes optimization algorithm for fault location.

[0047] The method for fault information collection and discretization coding in step 1 is as follows:

[0048] The sectional switch, tie switch, circuit breaker and other elements in the distribution network are taken as nodes, and the distribution network is divided into several feeder sections by multiple nodes. The fault operating state (0 for normal and 1 for fault) of each feeder section is taken as the position vector of the individual in the Archimedes optimization algorithm, and the expected fault current array of each node is generated according to the position vector. When a fault occurs in a feeder section of the distribution network, the actual fault current state of each node is detected by the FTU device, and the actual fault current array is generated. The difference between the expected fault current array corresponding to each individual and the fault current array actually reported by the FTU is used to construct the fitness function, which is used as the iteration basis of the Archimedes optimization algorithm. In the iteration process of the algorithm, the position of the individual (i.e. the state of the feeder section) is updated by simulating the interaction of density, volume and acceleration of the object in the fluid. After density adaptive update, position vector discrete mapping and acceleration guidance, if the termination condition is met, the global optimal individual is output, and its position vector represents the fault state of each feeder section of the current distribution network, thereby realizing the positioning of the fault feeder section.

[0049] The standard Archimedes optimization algorithm is only applicable to continuous domain problems, while the distribution network fault location problem is based on discrete domain, so the Archimedes optimization algorithm is discretized, and the population individuals are binary coded. Each dimension of the position component of each individual corresponds to the fault state of a feeder section: 1 for fault and 0 for no fault. Therefore, the position vector of an individual can be represented as L1, L2, …, Ln, where Ln takes the value of 0 or 1.

[0050] When distributed generation (DG) sources are connected to the distribution network, the network topology, power flow distribution, and power flow direction of the entire distribution network will change. To adapt to this change, the positive direction of the node fault current detected by the FTU device is specified as follows: when there is no DG source in the feeder segment, the direction from the system power source to the load is the positive direction; when there is a DG source in the feeder segment, the direction from the system power source to the DG source is the positive direction. According to this specification, there are three cases for encoding the node fault current, as shown in Table 1.

[0051] Table 1. Coding of node fault currents

[0052]

[0053] The expected fault state of a switching node is closely related to the fault states of its upstream and downstream feeder segments, and the state of the distributed power supply switches in the downstream segment. A switching expectation function is established. For a given switch in a feeder segment, the feeder segment from the system power source to that switch is defined as the upstream segment, and the feeder segment from that switch to the distributed power supply or load is defined as the downstream segment. Then, the expected fault current state function for each switching node is... It can be represented as follows:

[0054] ;

[0055] In the formula: n1 is the total number of feeder segments upstream of the i-th switch; n2 is the total number of feeder segments downstream of the i-th switch; n DG K represents the total number of downstream distributed power sources. j For the j-th distributed power switch downstream; L j This represents the fault state of the j-th feeder segment in the upstream or downstream section. The desired fault current array is then generated from this.

[0056] The method for designing the evaluation function in step 2 is as follows:

[0057] When a fault occurs in a feeder section of a distribution network containing distributed generation, its fitness is represented by the difference between the actual fault current array detected by the FTU equipment and the expected function array of the switch. Only when the actual fault current information and the expected function value are most similar can the difference between the two be minimized to obtain the optimal solution. The expansion rate function can be expressed as follows:

[0058] ;

[0059] In the formula: W is a maximum number (taken as 10). 2 To ensure the evaluation function remains positive, the problem is transformed into a maximum value optimization problem; N represents the total number of switches in a distribution network containing distributed generation sources; S i This refers to the actual fault current array information of the i-th switching node detected by the FTU device. is the expected fault current state array information of each switch node, n is the total number of feeder sections in the distribution network with distributed power supply; is the weight coefficient, and 0.5 is taken herein to prevent misjudgment; is the sum of the number of all fault feeder sections.

[0060] The method for fault location in step 3 is achieved by using the Archimedes optimization algorithm.

[0061] The Archimedes optimization algorithm is inspired by the Archimedes law, simulates the process that an object immersed in a liquid constantly adjusts its own properties through collision or the buoyancy it receives, and finally reaches a balanced state, and proposes a kind of optimization algorithm based on population. In the initialization stage, the algorithm randomly generates an initial population of a certain size, evaluates the initial population, selects the optimal individual, and uses it to guide the population. Next, the transfer operator and the density factor are set to control and balance the global exploration and local development of the algorithm. The operation of constantly updating the individual position and the optimal solution is repeated until the termination condition is met.

[0062] In the initialization stage, the AOA initializes the population position as shown in the following formula:

[0063] ;

[0064] In the formula, [lb i , ub i ] represents the search space.

[0065] The density (den), volume (vol) and acceleration (acc) of each individual are initialized as shown in the following formula:

[0066] ;

[0067] In this stage, the AOA will evaluate the population, select the current optimal individual (x best ), the density (den best ) of the optimal individual, the volume (vol best ) of the optimal individual and the acceleration (acc best ) of the optimal individual. And with the density and volume of the optimal individual, update the volume and density of other individuals:

[0068] ;

[0069] Next, set the transfer operator TF and the density factor d:

[0070] ;

[0071] Where, t represents the current iteration number, T maxrepresents the maximum number of iterations.

[0072] Global exploration phase (collision between objects):

[0073] If TF≤ 0.5, AOA performs global exploration, and the acceleration of the individual is updated as follows:

[0074] ;

[0075] wherein, represents the acceleration of the i th th individual in the t+1 generation, and are the density and volume of the random individual in the current iteration.

[0076] AOA normalizes the acceleration of the individual according to the following formula and is used for the update of the position of the individual:

[0077] ;

[0078] wherein, represents the normalized acceleration of the i th th individual in the t+1 generation; u and l are used to adjust the range of normalization.

[0079] In this phase, the position of the individual is updated as follows:

[0080] ;

[0081] wherein, C1 is a fixed constant, is a random number.

[0082] Local development phase (no collision between objects):

[0083] If TF> 0.5, AOA performs local development, and the acceleration of the individual is updated according to the following formula, and the acceleration is normalized:

[0084] ;

[0085] In this phase, the position of the individual is updated by the following formula:

[0086] ;

[0087] wherein, C2 is a fixed constant; , and ; F is a direction factor, which is used to determine the position update direction of the iteration, and is defined as:

[0088] ;

[0089] wherein, , is a constant.

[0090] The flowchart of the application is shown in the figure Figure 1 The simulation experiment is based on the Matlab platform, and the mathematical model of the IEEE-33 node distribution network structure containing DG is built to verify the effectiveness of the method, and the system structure is shown in the figure Figure 2 The AOA algorithm is used, and in the case of distributed power supply access to the distribution network, multiple single-point and multi-point fault comparison simulation experiments are carried out.

[0091] (1) Single-point fault

[0092] For the single-point fault positioning simulation of the multi-distributed power supply distribution network model, [K1, K2, K3] = [1, 1, 1], that is, the distributed power supply is connected to the distribution network, if the fault occurs in section L28, the FTU device reports the information as [1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, -1, 0, 0, 0, 0], and the output positioning result is [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0], and the positioning result is L28.

[0093] (2) Multi-point fault

[0094] The simulation of multi-point fault positioning is carried out, [K1, K2, K3] = [1, 1, 1], that is, the distributed power supply is connected to the distribution network, if the faults occur in sections L10, L28 and L32, the FTU device reports the information as [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, 1, -1, 1, 1, 1, 0], and the output positioning result is [0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0], and the positioning result is L10, L28 and L32.

[0095] The embodiments of the application are not limited by the above examples, and any changes, modifications, substitutions, combinations and simplifications made without departing from the spirit and principles of the application shall be equivalent replacement methods, and are included in the protection scope of the application.

Claims

1. A fault location method for distribution network with distributed generation based on AOA, characterized in that: The method comprises the following steps: Step 1: fault information acquisition and discretization coding: abstracting the power distribution network into multiple feeder sections segmented by switch nodes, generating a binary-coded individual position vector with the fault state of each feeder section as an element; generating an actual fault current array according to the actual fault current information reported by the feeder terminal unit (FTU) installed on the power distribution network; generating an expected fault current array according to the individual position vector and a preset switch expectation function; Step 2: evaluation function design: constructing a fitness function for calculating the difference between the actual fault current array and the expected fault current array; Step 3: Archimedes optimization algorithm for fault location: using an improved Archimedes optimization algorithm to optimize and solve the fitness function, including initializing the population, iteratively updating the density, volume, acceleration and position vector of the individual, and converting the continuous position vector into a binary code through discrete mapping; when the iteration termination condition is met, output the globally optimal individual position vector, and the binary code thereof corresponds to the located fault feeder section; The optimization solving balances the global exploration and local development capabilities of the algorithm by setting a transition operator (TF) and a density factor (d), and the calculation formula of the transition operator (TF) and the density factor (d) is: ； where t denotes the current iteration number, T max denotes the maximum number of iterations. The process of iteratively updating the individual position includes: If TF≤0.5, perform global exploration, and the update formula of the individual acceleration (acc) and position (x) is: ； ； ； If TF>0.5, perform local development, and the update formula of the individual acceleration (acc) and position (x) is: ； ； wherein, denotes the i th th individual in the t+1 generation, den, vol, acc represent the density, volume and acceleration of the individual, the subscript mr represents the random individual, the subscript best represents the current optimal individual, C1, C2 are constants, rand is a random number in [0,1], F is a direction factor, and T is a constant.

2. The AOA based fault location method for distribution network with distributed generation sources as claimed in claim 1 wherein: The coding rule of the actual fault current array in step 1 is that, for any switch node: If no fault current is detected or cannot be detected, it is coded as 0; If the fault current is detected and its direction is the same as the specified positive direction, it is coded as 1; If the fault current is detected and its direction is opposite to the specified positive direction, it is coded as -1; Wherein, the specified positive direction is: when there is no distributed power supply in the feeder section, the direction from the system power supply to the load; when there is a distributed power supply in the feeder section, the direction from the system power supply to the distributed power supply.

3. The method for fault location in AOA based distribution network with distributed generation as claimed in claim 1 wherein: The specific expression of the switch expectation function in step 1 is: ； where: n1 is the total number of feeder sections upstream of the ith switch; n2 is the total number of feeder sections downstream of the ith switch; n DG is the total number of downstream distributed power sources; K j is the jth distributed power source switch downstream; L j is the fault status of the jth feeder section in either the upstream or downstream section.

4. The method for fault location in AOA based distribution network with distributed generation as claimed in claim 1, wherein: The specific expression of the fitness function in step 2 is: ； Wherein: W is a very large number; N is the total number of switches in the distribution network with distributed power supply; S i is the actual fault current array information of the i-th switch node detected by the FTU device; is the expected fault current state array information of each switch node; n is the total number of feeder sections in the distribution network with distributed power supply; is the preset weight coefficient; is the sum of the number of all fault feeder sections.

5. The method for fault location in AOA based distribution network with distributed generation as claimed in claim 1, wherein: The discretization process of the improved Archimedes optimization algorithm in step 3 is to map the continuous position vector into a discrete value of 0 or 1 using a sigmoid function.

6. The method for fault location in AOA based distribution network with distributed generation as claimed in claim 1, wherein: The calculation method of the direction factor F is: ； wherein , is a constant.

7. A computer readable storage medium having stored thereon a computer program, characterized in that: The program is executed by the processor to realize the steps of the method in any one of claims 1 to 6.

Images