CN-122000850-A - Ship medium-voltage direct-current power system fault reconstruction method and system

CN122000850ACN 122000850 ACN122000850 ACN 122000850ACN-122000850-A

Abstract

The invention discloses a fault reconstruction method and a fault reconstruction system for a ship medium-voltage direct current power system, wherein the method comprises the steps of constructing a node-branch diagram based on a topological structure of the power system, and constructing an expansion association matrix based on the node-branch diagram by utilizing a graph theory; detecting connectivity of each node and each branch in the node-branch diagram in real time, updating detection results to an extended incidence matrix, determining a fault area of a power system based on abrupt elements in the extended incidence matrix, isolating the fault area, constructing a recovery power supply path set based on the extended incidence matrix by taking the fault branch as a starting point and the branch with connectivity as an end point, constructing a multi-objective fault reconstruction model by taking maximized load recovery and generator operation efficiency and minimum switching times as objective functions and meeting the operation constraint condition of the power system, solving the recovery power supply path set by combining the multi-objective fault reconstruction model by utilizing an improved SABPSO algorithm to obtain a reconstruction scheme, and executing the reconstruction scheme in the power system.

Inventors

ZHAO ZHIGAO
ZHANG KAIJIE
LU NING
Peng Dengchao
CHEN DIQIU
ZHANG XINGLONG
Yi Ruijia
GUO YI

Assignees

上海船舶运输科学研究所有限公司
上海海事大学

Dates

Publication Date: 20260508
Application Date: 20260122

Claims (10)

1. The fault reconstruction method of the ship medium-voltage direct-current power system is characterized by comprising the following steps of: The method comprises the steps of establishing an expansion incidence matrix, namely, based on the topological structure of a ship medium voltage direct current power system, taking a generator, a main switchboard and a port and starboard contact bus as main nodes, taking a regional switchboard as sub-nodes, taking a load as a load node, taking a circuit breaker as a branch, constructing a node-branch diagram together, constructing an expansion incidence matrix based on the node-branch diagram by utilizing graph theory, wherein the node-branch diagram reflects the running state of the ship medium voltage direct current power system in real time, and elements of the expansion incidence matrix represent the connectivity of each node and each branch in the node-branch diagram; If the elements in the extended incidence matrix are suddenly changed, determining fault area information of the medium-voltage direct-current power system of the ship based on the suddenly changed elements and corresponding values thereof, wherein the fault area information comprises a fault branch and a power failure node; constructing an alternative recovery power supply path set for the power-losing node based on the extended incidence matrix, taking the fault branch as a starting point and taking a branch with connectivity as an ending point; Dividing the load into a first-stage load, a second-stage load and a third-stage load according to the importance of the load, and discretizing each stage of load into a plurality of expansion loads, wherein each of the first-stage expansion load and the second-stage expansion load comprises a normal power supply path and a standby power supply path, and each of the third-stage expansion loads only comprises a normal power supply path; constructing a multi-objective fault reconstruction model by taking the maximization of the load recovery degree, the minimization of the switching operation times and the maximization of the generator operation efficiency as targets, taking the constraint conditions that only one power supply path of the primary expansion load and the secondary expansion load is conducted, the generator capacity is larger than the total capacity of the connection load and the line node capacity is smaller than the maximum threshold of the line capacity as constraints; An optimal restoration path solving step, namely carrying out optimal path solving on the alternative restoration power supply path set by utilizing an improved SABPSO algorithm obtained by combining a simulated annealing algorithm with a discrete binary particle swarm algorithm, wherein the improved SABPSO algorithm reserves the discrete optimizing characteristic of the discrete binary particle swarm algorithm and the global optimizing characteristic of the simulated annealing algorithm, adjusts the particle searching step length through self-adaptive inertia weight and optimizes the global optimal solution screening by combining a roulette selection mechanism; and the fault reconstruction scheme executing step is that a switching instruction is generated based on the fault reconstruction scheme and is executed in the ship medium voltage direct current power system.
2. The method of claim 1, wherein in the step of establishing the extended incidence matrix, a main switchboard included in the main node is divided into a main switchboard node and a secondary switchboard node, and a regional switchboard included in the secondary node is a load switchboard node; The method comprises the steps of traversing main nodes and sub nodes in sequence by a tree search algorithm in real time according to the node-branch diagram, reversely verifying load nodes, detecting connectivity of each node and each branch, wherein the node-branch diagram is used for traversing adjacent main distribution nodes of each generator main node by a breadth-first search algorithm in real time, traversing the main distribution nodes, sub distribution nodes, load distribution nodes and corresponding branches layer by adopting a depth-first search algorithm from each generator main node, establishing node-branch connectivity relation until traversing to the load nodes or the last stage of the sub distribution nodes, and then backtracking to the generator main nodes layer by layer from the searched load nodes, and combining the node-branch connectivity relation to obtain connectivity information of each node and each branch.
3. The method according to claim 1, wherein in the fault checking and information extracting step, whether the medium voltage dc power system of the ship is faulty is determined by whether an element in the extended correlation matrix is suddenly changed; if the elements in the extended incidence matrix are mutated, determining a fault branch based on the mutated elements and the corresponding values thereof; for each generator main node, determining the state of the generator main node based on the values of all elements in the corresponding column of the generator main node in the extended incidence matrix, wherein if the values of all elements are 0, the generator main node loses power; for each load node, determining all elements of a row corresponding to the load node in the extended incidence matrix, screening cross elements of each generator node corresponding to the elements, and if the value of all the cross elements is 0, powering off the load node; And determining the row and the column corresponding to each panel node in the expansion incidence matrix aiming at each panel node formed by the main panel contained in the main node and the regional panel contained in the sub node, if the values of all elements in the row are 0 and the values of the crossing elements in the column corresponding to all the generator main nodes are 0, the panel node loses power, otherwise, the panel node is normal.
4. The method according to claim 1, wherein in the multi-objective fault reconstruction model building step, the multi-objective function is as follows: , , , , Wherein maxf denotes the maximum of the multiple objective function, And max Indicating that the degree of load recovery is maximized, With min Indicating that the number of switching operations is minimized, And max Indicating that the generator operating efficiency is maximized, The weight coefficient is represented by a number of weight coefficients, The priority coefficient is represented as such, Representing the power state of the i-th stage extension load, Representing the power state of the j-th secondary expansion load, Representing the power state of the z-th three-stage extension load, Representing the load of the i-th stage extension load, Representing the load of the j-th secondary expansion load, Represents the load of the z-th three-stage expansion load, h represents the number of the first-stage expansion loads, k represents the number of the second-stage expansion loads, m represents the number of the third-stage expansion loads, Indicating the i-th three-stage extended load switch state, The j-th stage expansion load or the second stage expansion load path switching state is represented, l represents the total number of the stage expansion load and the second stage expansion load, Indicating the i-th generator operating efficiency, and n indicating the total number of generators.
5. The method according to claim 4, wherein in the multi-objective fault reconstruction model building step, a constraint that only one power supply path is turned on by the primary expansion load and the secondary expansion load is as follows: , Wherein, the Representing the normal power supply paths of the primary expansion load and the secondary expansion load, A standby power supply path representing a primary expansion load and a secondary expansion load; Constraints for generator capacity being greater than the total capacity of the connected load are as follows: , Wherein, the Representing the capacity of the i-th generator, Represents the j-th load connected with the i-th generator, and m represents the total number of loads connected with the i-th generator; the constraint that the line node capacity is less than the line capacity maximum threshold is as follows: , Wherein, the Representing the capacity of the line node, Representing the line capacity maximum threshold.
6. The method of claim 1, wherein in the optimal restoration path solving step, the step of performing optimal path solving on the set of alternative restoration supply paths using a modified SABPSO algorithm includes: s1, configuring particle population quantity, maximum iteration times, particle speed range, initial temperature, cooling rate and termination conditions; s2, initializing the speed of each particle in the speed range of the particle, taking each stage of expansion load as the dimension of each particle, and initializing the position of each particle in the alternative recovery power supply path set range; s3, respectively calculating initial fitness of each particle by utilizing the multi-objective function, taking the initial fitness of each particle as an individual optimal fitness, and selecting the individual optimal fitness with highest probability from all individual optimal fitness by utilizing a roulette selection algorithm as a global optimal fitness; S4, respectively calculating the current fitness of each particle at the current temperature by utilizing the multi-objective function; s5, updating the speed of each dimension in each particle according to the individual optimal fitness, the global optimal fitness and the self-adaptive inertia weight corresponding to the particle, so as to adjust the particle searching step length; S6, checking whether each particle meets the constraint condition, and eliminating particles which do not meet the constraint condition; S7, comparing the current fitness corresponding to each particle with the individual optimal fitness according to Metropolis criterion, judging whether to select the current fitness as a new individual optimal fitness, and finally determining the current individual optimal fitness of each particle; S8, gradually cooling the current temperature according to the cooling rate by taking the configured initial temperature as a starting point; And S9, judging whether the configured termination condition is met, if not, returning to the step S4, and if so, outputting particles corresponding to the current global optimal fitness as a fault reconstruction scheme.
7. The method according to claim 6, wherein in the optimal restoration path solving step, the speed of each dimension in the particle is updated based on the individual optimal fitness, the global optimal fitness and the adaptive inertia weight corresponding to the particle, and the formula is as follows: , Wherein, the Representing the speed of the ith particle in the jth dimension for the t +1 th iteration, Representing the speed of the ith particle in the jth dimension for the t-th iteration, Representing the adaptive inertial weights of the vehicle, Representing the individual learning factors of the particles, Representing the global learning factor(s), Representing the random number between 0,1, Representing the individual optimal fitness of the ith particle in the jth iteration in the jth dimension, Representing the global optimum fitness of the ith particle in the jth dimension for the t-th iteration, Representing the position of the ith particle in the jth dimension for the t-th iteration; updating the position of each dimension in the particle based on the speed of each dimension after the particle is updated, and if the extension load type corresponding to the dimension is the primary extension load or the secondary extension load, updating the position of the dimension by using the following formula: , Wherein, the Representing the j-th dimension updated position of the i-th particle, A minimum speed threshold value is indicated and, Representing a maximum speed threshold, sigmoid representing a probability function; if the extension load type corresponding to the dimension is the three-level extension load, updating the position of the dimension by using the following formula: , Wherein rand represents a random number between [0-1 ].
8. The method of claim 6, wherein in the optimal recovery path solving step, a linear adjustment strategy is used to determine the adaptive inertial weight, and the formula is as follows: , Wherein, the Representing the adaptive inertial weight of the ith iteration, i representing the number of iterations, ger representing the maximum number of iterations, Representing the adaptive inertial weight maximum threshold, Representing an adaptive inertial weight minimum threshold.
9. The method of claim 6, wherein in the optimal restoration path solving step, the metapolis criterion is as shown in the following formula: , Where p represents the probability that the current fitness of the particle is the new individual optimal fitness of the particle, Indicating the current fitness of the particle, Representing the current individual optimal fitness of the particle, Representing the temperature of the ith iteration; if the current fitness corresponding to the particle is smaller than the individual optimal fitness, taking the current fitness as a new individual optimal fitness of the particle; If the current fitness corresponding to the particle is greater than or equal to the individual optimal fitness, then The formula calculates the probability that the current fitness is the new individual optimal fitness of the particle; And finally randomly generating a number of 0-1, taking the current fitness as a new individual optimal fitness of the particle if the generated number is less than or equal to P, and not changing the individual optimal fitness corresponding to the particle if the generated number is more than P.
10. The fault reconstruction system of the ship medium-voltage direct-current power system is characterized by comprising an expansion incidence matrix building module, a fault checking and information extracting module, a recovery power supply path building module, a multi-target fault reconstruction model building module, an optimal recovery path solving module and a fault reconstruction scheme executing module which are connected in sequence, The system comprises an expansion incidence matrix building module, a node-branch diagram, a tree search algorithm, a node-branch diagram and a node-branch diagram, wherein the topology structure of a ship medium voltage direct current power system is based on a generator, a main switchboard and a port-starboard contact bus as main nodes, a regional switchboard as sub-nodes, loads as load nodes and circuit breakers as branches, the node-branch diagram is used for constructing node-branch diagrams together, the expansion incidence matrix is constructed based on the node-branch diagram by utilizing a graph theory, elements of the expansion incidence matrix represent the connectivity of each node and each branch in the node-branch diagram in real time, and the node-branch diagram is used for traversing the main nodes and the sub-nodes in real time and reversely verifying the load nodes by adopting the tree search algorithm, detecting the connectivity of each node and each branch and updating the detection result to the expansion incidence matrix in real time; The fault detection and information extraction module is used for determining fault area information of the ship medium-voltage direct-current power system based on the abrupt elements and corresponding values if the elements in the extended incidence matrix are abrupt, wherein the fault area information comprises a fault branch and a power failure node; The recovery power supply path construction module is used for constructing an alternative recovery power supply path set for the power-losing node based on the expansion incidence matrix, taking the fault branch as a starting point and taking a branch with connectivity as an ending point; The multi-objective fault reconstruction model building module is used for dividing the load into a first-stage load, a second-stage load and a third-stage load according to the importance of the load, and discretizing each stage of load into a plurality of expansion loads, wherein each of the first-stage expansion load and the second-stage expansion load comprises a normal power supply path and a standby power supply path, each of the third-stage expansion load only comprises a normal power supply path; The optimal restoration path solving module is used for carrying out optimal path solving on the alternative restoration power supply path set by utilizing an improved SABPSO algorithm obtained by combining a simulated annealing algorithm with a discrete binary particle swarm algorithm, wherein the improved SABPSO algorithm reserves the discrete optimizing characteristic of the discrete binary particle swarm algorithm and the global optimizing characteristic of the simulated annealing algorithm, adjusts particle searching step length through self-adaptive inertia weight and optimizes global optimal solution screening by combining a roulette selection mechanism; The fault reconstruction scheme execution module is used for generating a switching instruction based on the fault reconstruction scheme and executing the switching instruction in the ship medium-voltage direct-current power system.

Description

Ship medium-voltage direct-current power system fault reconstruction method and system Technical Field The invention relates to the technical field of fault reconstruction of a ship medium-voltage direct-current power system, in particular to a fault reconstruction method and system of a ship medium-voltage direct-current power system. Background The ship power system is evolving from a traditional medium-voltage alternating current system to a medium-voltage direct current (MVDC) system, and the MVDC system has become the core development direction of modern ship power systems such as military drive ships, large scientific research ships and the like by virtue of the core technical advantages of high energy transmission efficiency, excellent space utilization rate, low total life cycle cost and the like. However, in the actual operation process, once an electrical fault (such as short circuit, circuit breaking and the like) occurs in the MVDC system, the system vitality and the operation reliability are directly threatened, so that not only is the shutdown and the power interruption of key equipment possibly caused, but also the overall breakdown of the system is caused in serious cases, and a great hidden danger is formed for the life safety of personnel on a ship and the navigation safety of the ship, so that the development of the fault reconstruction technology research of the ship MVDC system has urgent practical requirements. The core of the fault reconstruction of the ship MVDC system is that when faults such as short circuit and circuit break occur in the system, the power supply of important loads is recovered to the maximum extent on the premise of meeting the operation constraint of the MVDC system by adjusting the power distribution topology, and meanwhile, the operation safety requirement of the MVDC system is ensured. At present, the prior art scheme still has a plurality of bottlenecks to be broken through, namely firstly, the traditional topology identification method cannot feed back the switching state of the breaker in real time, the fault point positioning delay or misjudgment is easy to cause, the reconstruction efficiency is influenced, secondly, most of the existing reconstruction methods are single in target, only aim at maximizing the load recovery degree, and do not take account of minimizing the switching operation times and generator load balancing, the switching loss increase or generator overload is easy to cause, thirdly, the performance of a core algorithm is provided with a short board, a discrete Binary Particle Swarm Optimization (BPSO) is easy to fall into a local optimal solution, the convergence of a Simulated Annealing (SA) algorithm is slow, the overall searching and local optimizing balance capacity of the existing hybrid algorithm (such as SA-BPSO) is insufficient, the complex fault scene is difficult to adapt, and fourthly, the scene suitability is insufficient, aiming at the specific 'generator-branch-load' combined fault scene of a ship MVDC system, and the existing method lacks a targeted reconstruction strategy, and the completeness and suitability of the fault recovery scheme are difficult to meet the actual running requirements. Disclosure of Invention In order to solve the problems of slow fault information extraction, single reconstruction target, easy local optimization of algorithm, poor scene suitability and the like in the prior art, the invention provides a fault reconstruction method of a ship medium-voltage direct current power system, which can rapidly and accurately extract fault information and position a fault area, a multi-target optimization reconstruction scheme is carried out on the basis so as to maximize important load recovery, minimize switching operation times and ensure load balance of a generator, and a core algorithm effectively avoids trapping in local optimal solution by improving global searching capability and convergence speed, so that a plurality of complex fault scenes can be flexibly adapted and a safe and reliable complete power supply recovery scheme can be finally output. The invention also relates to a fault reconstruction system of the ship medium-voltage direct-current power system. The technical scheme of the invention is as follows: A fault reconstruction method of a ship medium voltage direct current power system comprises the following steps: The method comprises the steps of establishing an expansion incidence matrix, namely, based on the topological structure of a ship medium voltage direct current power system, taking a generator, a main switchboard and a port and starboard contact bus as main nodes, taking a regional switchboard as sub-nodes, taking a load as a load node, taking a circuit breaker as a branch, constructing a node-branch diagram together, constructing an expansion incidence matrix based on the node-branch diagram by utilizing graph theory, wherein the node-branch diagram reflects the running state