CN105067956A - Anti-colony-algorithm-based distribution network fault positioning method - Google Patents

Abstract

The invention discloses an anti-colony-algorithm-based distribution network fault positioning method. The method comprises: an anti colony algorithm evaluation function is constructed; and fault positioning is carried out by using an anti colony algorithm. When the anti colony algorithm is used for carrying out fault positioning, the evaluation function serves as a basis of an evaluation solution performance; and the good performance expresses the selected path is short. When equipment with the fault in the distribution network is diagnosed, a hypothesis is found and all uploaded RTU or FTU information can be explained well; and thus deviation of the corresponding FTU or RTU information and actual uploaded information is minimized by finding the hypothesis, thereby realizing fault positioning of the distribution network.

Description

A kind of method based on ant group algorithm Distribution Network Failure location

Technical field

The present invention relates to distribution network failure positioning field, particularly a kind of method based on ant group algorithm Distribution Network Failure location.

Background technology

Distribution Fault Location System is the disposal system of a real-time online, all failure messages come from outdoor FTU communicator mostly, its work under bad environment, difference variation scope are large, and be mostly contained on electric power terminal or in power distribution cabinet, the disturbing factors such as high voltage, electric current, thunder and lightning be born; The communication point of power distribution network many and dispersion, be difficult to adopt same communication mode to deal with problems, in actual applications, the general communication modes all adopting mixing, in addition switching node loosen, the existence of the factor such as the erroneous judgement of FTU itself, distribution network failure information is disturbed or the possibility of losing certainly exists.For distribution network failure positional matrix algorithm, often occur erroneous judgement, cause accident expanded range when the information upload of each FTU is wrong, power off time extends, and directly reduces power supply reliability.Therefore study that a kind of to have compared with the algorithm of strong fault tolerance be problem in the urgent need to address in distribution network failure Position Research.

Because the optimizing algorithm being applicable to distribution network failure location should possess higher fault-tolerance, cause calculated amount can be larger than direct algorithm, speed will be slower than matrix algorithms, develop speed and the strong optimizing algorithm of fault-tolerant ability has important value.In existing mathematical algorithm, ant group algorithm has stronger robustness, and the number of the pheromones that optimizing Route Dependence discharges in ant, has stronger fault-tolerant ability, and be imbued with the features such as greedy heuristic search, be applicable to solve distribution network failure and locate such combinatorial optimization problem.Effectively can solve the best of breed of each equipment state in power distribution network like this, the failure message making it to upload with FTU is the most identical.

Summary of the invention

Given this, the object of this invention is to provide a kind of assist trouble positioning system.

The object of the invention is to be achieved through the following technical solutions.

Based on a method for ant group algorithm Distribution Network Failure location, feature of the present invention comprises the following steps:

S1. evaluation function is constructed;

Distribution Network Failure locating structure evaluation function is the foundation evaluating the performance of separating, the expression that performance is good is selected " short " this principle in path, find out a hypothesis, all RTU or FTU information uploaded can be explained, namely find out a hypothesis make the information of FTU or RTU corresponding to it and actual information upload deviation minimum, be namely the equipment that breaks down in diagnosis power distribution network, be constructed as follows evaluation function:

$F_{it}\left(S_B\right)=\underset{j=1}{\overset{N}{\Σ}}|I_j-I_j^{*}\left(S_B\right)|+w*\underset{j=1}{\overset{N}{\Σ}}\left|S_B\left(i\right)\right|$

Wherein, F _it(S _b) fitness corresponding to each solution; S _bfor the state value (1 is the malfunction of equipment, and 0 is the normal condition of equipment) of equipment each in power distribution network; N is the quantity of feeder line in power distribution network; I _jit is the out-of-limit signal of electric current (having during fault current is 1, otherwise is 0) at each block switch place in power distribution network; for the expectation state of measuring control point each in distribution, it is the function of each feeder line state of section, is determined by real-time Network topology result, and having during fault overcurrent is 1, is 0 during non-fault excess current; W is weight coefficient (value 0 ~ 1); Ask for the fitness value of each switch in fault loop according to the evaluation function of structure, minimum is trouble spot;

The state evaluation function focusing on each equipment in the principles of construction diagnosis power distribution network of Utilization assessment function of the present invention, can be used for Single Point of Faliure and directly locates;

S2. ant group algorithm of sampling carries out localization of fault;

Ant group algorithm is adopted to solve distribution network failure orientation problem herein.In the fault of actual power distribution network, according to distribution network failure number, Single Point of Faliure type and multipoint fault type can be divided into.The probability that Single Point of Faliure occurs is far longer than the probability of multipoint fault, and the probability that in multipoint fault, the element (equipment) of more than 3 breaks down simultaneously is very low;

1) when diagnosing Single Point of Faliure type, in order to improve computing velocity, do not adopt ant group algorithm herein, and adopt the method for traversal, the evaluation function value in distribution network failure situation is solved successively;

2) when diagnosing multipoint fault type, consider that the probability of 2 simultaneous faultss is greater than the probability of 3 simultaneous faultss generations, first carry out optimizing to two points fault, next is only three point failures and carries out optimizing; When supposing two points fault occurs in power distribution network, the trouble spot under adopting ant group algorithm to obtain the minimum value of evaluation function and assumed conditions; Assuming that during generation three point failure, in like manner;

3) finally compare the evaluation function value of above-mentioned three kinds of situations, supposition situation that wherein evaluation function value is minimum is set up, and equipment state combination corresponding to this evaluation function value is globally optimal solution;

Therefore, the ant group algorithm calculation procedure for two points fault is as follows:

1. roulette

When iterating to t time, according to the pheromones amount in each equipment state " 1 ", distribute each self-corresponding numerical intervals, then according to the pheromones amount in all devices state " 1 " and, produce one and be greater than 0 and the random number being less than this and value, the equipment state " 1 " that the numerical intervals that random number drops on is corresponding is selected; When the fault supposed is two points fault, need select the equipment that two states are " 1 ", the state of all the other equipment is all set to " 0 ", and this creates the terminal one may separate; When the fault supposed is three point failure, in like manner;

2. Calculation Estimation functional value

According to the equipment 0-1 state SB that wheel disc method produces, adopt the expectation state computing method of distribution network failure position monitoring and control point to obtain I* (SB), calculate all ant i (i=1,2,, evaluation function value F (i) of power distribution network m);

3. the release of pheromones

When iterating to t time, according to the size of the evaluation function value F of all ants, release pheromone in corresponding each equipment state may be separated each.The amount of release pheromone is:

$\mathrm{\Δ}\mathrm{\τ}\left(t\right)=\frac{Q}{F}$

4. the volatilization of pheromones

When iterating to t time, after all ants complete circulation, the pheromones amount after t circulation adjusts according to following formula

τ(t+1)＝ρ·τ(t)+Δτ(t)

5. convergence criterion

After iterating to t time, the pheromones intensity in each equipment state is compared; If assuming that fault is two points fault, meet when there being the pheromones intensity on two equipment states " l "

$\begin{array}{cc}{\mathrm{\τ}}_i>10{\mathrm{\τ}}_j& \∀j\end{array}$

In formula, i represents that two equipment states are the numbering of " l ", and j represents that all the other equipment states are the numbering of " 0 ";

When meeting formula (12), two equipment states are the pheromones of " l " when being the pheromones of " 0 " considerably beyond state of other, are judged as that these two equipment are faulty equipments, otherwise continue iteration.Assuming that when fault is three point failure, in like manner.

The invention has the beneficial effects as follows, ant group algorithm is a kind of general bionic Algorithm, and this algorithm comes from the research to ant group group behavior.It is a kind of many agent algorithms based on colony in essence, is also a kind of new heuritic approach of imitating ant working method simultaneously.When utilizing ant group algorithm to carry out localization of fault, evaluation function is the foundation evaluating the performance of separating, the expression selected " path is short " that performance is good; And diagnose the equipment broken down in power distribution network namely to find out a hypothesis, all RTU or FTU information uploaded can be explained, namely find out a hypothesis make the information of FTU or RTU corresponding to it and actual information upload deviation minimum, realize the localization of fault of power distribution network.This method obtains FTU information, and the analytical approach setting up a set of scientific system positions trouble spot, and to obtain distribution network failure be accurately position, changes the coarse determination methods of artificial line walking in the past.

Accompanying drawing explanation

In order to make the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, the present invention is described in further detail, wherein:

Fig. 1 is the basic calculating flow process of distribution network failure location;

Fig. 2 is the calculation process of the ant group algorithm that point failure adopts.

Embodiment

Below with reference to accompanying drawing, the preferred embodiments of the present invention are described in detail; Should be appreciated that preferred embodiment only in order to the present invention is described, instead of in order to limit the scope of the invention.

See Fig. 1, Fig. 2, based on the method for ant group algorithm Distribution Network Failure location, comprise the following steps:

S1. evaluation function is constructed;

Distribution Network Failure locating structure evaluation function is the foundation evaluating the performance of separating, the expression that performance is good is selected " short " this principle in path, find out a hypothesis, all RTU or FTU information uploaded can be explained, namely find out a hypothesis make the information of FTU or RTU corresponding to it and actual information upload deviation minimum, be namely the equipment that breaks down in diagnosis power distribution network, be constructed as follows evaluation function:

$F_{it}\left(S_B\right)=\underset{j=1}{\overset{N}{\Σ}}|I_j-I_j^{*}\left(S_B\right)|+w*\underset{j=1}{\overset{N}{\Σ}}\left|S_B\left(i\right)\right|$

Wherein, F _it(S _b) fitness corresponding to each solution; S _bfor the state value (1 is the malfunction of equipment, and 0 is the normal condition of equipment) of equipment each in power distribution network; N is the quantity of feeder line in power distribution network; I _jit is the out-of-limit signal of electric current (having during fault current is 1, otherwise is 0) at each block switch place in power distribution network; for the expectation state of measuring control point each in distribution, it is the function of each feeder line state of section, is determined by real-time Network topology result, and having during fault overcurrent is 1, is 0 during non-fault excess current; W is weight coefficient (value 0 ~ 1); Ask for the fitness value of each switch in fault loop according to the evaluation function of structure, minimum is trouble spot;

The state evaluation function focusing on each equipment in the principles of construction diagnosis power distribution network of Utilization assessment function of the present invention, can be used for Single Point of Faliure and directly locates;

S2. ant group algorithm of sampling carries out localization of fault;

Ant group algorithm is adopted to solve distribution network failure orientation problem herein.In the fault of actual power distribution network, according to distribution network failure number, Single Point of Faliure type and multipoint fault type can be divided into; The probability that Single Point of Faliure occurs is far longer than the probability of multipoint fault, and the probability that in multipoint fault, the element (equipment) of more than 3 breaks down simultaneously is very low;

1) when diagnosing Single Point of Faliure type, in order to improve computing velocity, do not adopt ant group algorithm herein, and adopt the method for traversal, the evaluation function value in distribution network failure situation is solved successively;

2) when diagnosing multipoint fault type, consider that the probability of 2 simultaneous faultss is greater than the probability of 3 simultaneous faultss generations, first carry out optimizing to two points fault, next is only three point failures and carries out optimizing.When supposing two points fault occurs in power distribution network, the trouble spot under adopting ant group algorithm to obtain the minimum value of evaluation function and assumed conditions; Assuming that during generation three point failure, in like manner;

3) finally compare the evaluation function value of above-mentioned three kinds of situations, supposition situation that wherein evaluation function value is minimum is set up, and equipment state combination corresponding to this evaluation function value is globally optimal solution;

Therefore, as follows for the ant group algorithm calculation procedure of two points fault herein:

1) roulette

When iterating to t time, according to the pheromones amount in each equipment state " 1 ", distribute each self-corresponding numerical intervals, then according to the pheromones amount in all devices state " 1 " and, produce one and be greater than 0 and the random number being less than this and value, the equipment state " 1 " that the numerical intervals that random number drops on is corresponding is selected.When the fault supposed is two points fault, need select the equipment that two states are " 1 ", the state of all the other equipment is all set to " 0 ", and this creates the terminal one may separate; When the fault supposed is three point failure, in like manner;

2) Calculation Estimation functional value

According to the equipment 0-1 state SB that wheel disc method produces, adopt the expectation state computing method of distribution network failure position monitoring and control point to obtain I* (SB), calculate all ant i (i=1,2,, evaluation function value F (i) of power distribution network m);

3) release of pheromones

When iterating to t time, according to the size of the evaluation function value F of all ants, release pheromone in corresponding each equipment state may be separated each.The amount of release pheromone is:

$\mathrm{\Δ}\mathrm{\τ}\left(t\right)=\frac{Q}{F}$

4) volatilization of pheromones

When iterating to t time, after all ants complete circulation, the pheromones amount after t circulation adjusts according to following formula

τ(t+1)＝ρ·τ(t)+Δτ(t)

5) convergence criterion

After iterating to t time, the pheromones intensity in each equipment state is compared; If assuming that fault is two points fault, meet when there being the pheromones intensity on two equipment states " l "

$\begin{array}{cc}{\mathrm{\τ}}_i>10{\mathrm{\τ}}_j& \∀j\end{array}$

In formula, i represents that two equipment states are the numbering of " l ", and j represents that all the other equipment states are the numbering of " 0 ";

When meeting formula (12), two equipment states are the pheromones of " l " when being the pheromones of " 0 " considerably beyond state of other, are judged as that these two equipment are faulty equipments, otherwise continue iteration; Assuming that when fault is three point failure, in like manner.

The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, obviously, those skilled in the art can carry out various change and modification to the present invention and not depart from the spirit and scope of the present invention.Like this, if these amendments of the present invention and modification belong within the scope of the claims in the present invention and equivalent technologies thereof, then the present invention is also intended to comprise these change and modification.

Claims (1)

1., based on a method for ant group algorithm Distribution Network Failure location, it is characterized in that, comprise the following steps:

S1. evaluation function is constructed;

The equipment broken down in diagnosis power distribution network, is constructed as follows evaluation function:

$F_{it}\left(S_B\right)=\underset{j=1}{\overset{N}{\Σ}}|I_j-I_j^{*}\left(S_B\right)|+w*\underset{j=1}{\overset{N}{\Σ}}\left|S_B\left(i\right)\right|$

Wherein, F _it(S _b) fitness corresponding to each solution; S _bfor the state value of equipment each in power distribution network, 1 is the malfunction of equipment, and 0 is the normal condition of equipment; N is the quantity of feeder line in power distribution network; I _jbe the out-of-limit signal of electric current at each block switch place in power distribution network, having during fault current is 1, otherwise is 0; for the expectation state of measuring control point each in distribution, it is the function of each feeder line state of section, is determined by real-time Network topology result, and having during fault overcurrent is 1, is 0 during non-fault excess current; W is weight coefficient, value 0 ~ 1; Ask for the fitness value of each switch in fault loop according to the evaluation function of structure, minimum is trouble spot;

S2. ant group algorithm of sampling carries out localization of fault;

1) when diagnosing Single Point of Faliure type, adopting the method for traversal, the evaluation function value in distribution network failure situation is solved successively;

2) when diagnosing multipoint fault type, first optimizing being carried out to two points fault, secondly optimizing being carried out to three point failures; When supposing two points fault occurs in power distribution network, the trouble spot under adopting ant group algorithm to obtain the minimum value of evaluation function and assumed conditions; Assuming that during generation three point failure, in like manner;

3) finally compare the evaluation function value of above-mentioned three kinds of situations, supposition situation that wherein evaluation function value is minimum is set up, and equipment state combination corresponding to this evaluation function value is globally optimal solution;

Therefore, the ant group algorithm calculation procedure for two points fault is as follows:

1. roulette

When iterating to t time, according to the pheromones amount in each equipment state " 1 ", distribute each self-corresponding numerical intervals, then according to the pheromones amount in all devices state " 1 " and, produce one and be greater than 0 and the random number being less than this and value, the equipment state " 1 " that the numerical intervals that random number drops on is corresponding is selected; When the fault supposed is two points fault, need select the equipment that two states are " 1 ", the state of all the other equipment is all set to " 0 ", and this creates the terminal one may separate; When the fault supposed is three point failure, in like manner;

2. Calculation Estimation functional value

According to the equipment 0-1 state SB that wheel disc method produces, adopt the expectation state computing method of distribution network failure position monitoring and control point to obtain I* (SB), calculate all ant i (i=1,2,, evaluation function value F (i) of power distribution network m);

3. the release of pheromones

When iterating to t time, according to the size of the evaluation function value F of all ants, release pheromone in corresponding each equipment state may be separated each; The amount of release pheromone is:

4. the volatilization of pheromones

When iterating to t time, after all ants complete circulation, the pheromones amount after t circulation adjusts according to following formula

τ(t+1)＝ρ·τ(t)+Δτ(t)

5. convergence criterion

After iterating to t time, the pheromones intensity in each equipment state is compared.If assuming that fault is two points fault, meet when there being the pheromones intensity on two equipment states " l "

$\begin{array}{cc}{\mathrm{\τ}}_i>10{\mathrm{\τ}}_j& \∀j\end{array}$

In formula, i represents that two equipment states are the numbering of " l ", and j represents that all the other equipment states are the numbering of " 0 ";

When meeting formula (12), two equipment states are the pheromones of " l " when being the pheromones of " 0 " considerably beyond state of other, are judged as that these two equipment are faulty equipments, otherwise continue iteration; Assuming that when fault is three point failure, in like manner.