CN102497024B - Intelligent warning system based on integer programming - Google Patents

Intelligent warning system based on integer programming Download PDF

Info

Publication number
CN102497024B
CN102497024B CN201110425428.7A CN201110425428A CN102497024B CN 102497024 B CN102497024 B CN 102497024B CN 201110425428 A CN201110425428 A CN 201110425428A CN 102497024 B CN102497024 B CN 102497024B
Authority
CN
China
Prior art keywords
protection
switch
fault
action
state
Prior art date
Application number
CN201110425428.7A
Other languages
Chinese (zh)
Other versions
CN102497024A (en
Inventor
田华
葛立青
笃峻
朱广名
陈宏辉
张学强
叶睆
陈海波
都海坤
Original Assignee
广东电网公司茂名供电局
南京南瑞继保电气有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广东电网公司茂名供电局, 南京南瑞继保电气有限公司 filed Critical 广东电网公司茂名供电局
Priority to CN201110425428.7A priority Critical patent/CN102497024B/en
Publication of CN102497024A publication Critical patent/CN102497024A/en
Application granted granted Critical
Publication of CN102497024B publication Critical patent/CN102497024B/en

Links

Abstract

The invention relates to an intelligent warning system based on integer programming. The intelligent warning system adopts a signal acquisition module, a diagnostic startup judgment module, a fault analysis module, a failure report generating module and an analytical result comprehensive display module to warn; the intelligent warning system can be implemented through the following steps of: 1) network modelling, filling primary equipment of a transformer substation and a connected relation of the transformer substation into a database; and generating a structural model meeting IEC (international electrotechnical commission) 61970 standard of an electric system factory station and a system, and then using the model to generate a network device library through a filling library operation; 2) protected information modeling: carrying out protected information modeling according to the hierarchical relationship of the primary equipment, a protective device, protected information and a trip breaker, building an incidence relation between a protection main device, the protected information and a breaker motion, and protecting the main device namely the primary equipment; and 3) carrying out fault diagnosis. The intelligent warning system in the invention can be capable of rapidly and effectively carrying out detailed analysis on a single or complicated failure and protection motion analysis in a failure process.

Description

Intelligent warning system based on integer programming

Technical field

The present invention relates to Intelligent transformer station intelligent warning system.

Background technology

The development that science and technology is with rapid changepl. never-ending changes and improvements, automation of transformation substations degree is more and more higher, and amount of information is also increasing; The abundant of amount of information certainly will bring very large help to the analysis of data, but this also becomes a double-edged sword, the back side of this double-edged sword is " recklessly spending charming eye ",, in short time, magnanimity information is to the vision of operator and psychological impact, the information often previously having poured in also has enough time to finish watching, and is buried in oblivion again between soon by a collection of new signal, and operations staff on duty is difficult to concentrate on crucial points fast, be unfavorable for rapid failure judgement, provide correct processing judgement.

Along with the intelligentized propelling of electrical network, the intellectualized technology of transformer station becomes an important ring of intelligent grid technology gradually.The function on the monitoring backstage of transformer station, can be divided into " prison " and " control " two large classes.And " prison " has no longer been confined to traditional function of directly adopting Various types of data, should further expand the intelligent data analysis function strengthening in transformer station, make full use of the Various types of data that transformer station's end gathers, carry out intellectual analysis, strengthen the deep processing ability of warning information, and the information interaction between realization and scheduling station.

Accident analysis based on the original alarm event of automated system in station is crucial; and the intelligent warning system of application is just done simple screening to alarm event at present; filter; after hierarchical classification, carry out compared with shallow hierarchy accident analysis; as the analysis based on protection action chain algorithm, be not well positioned to meet the comparatively senior demands such as complex fault differentiation and protection motion analysis.Therefore, an accident analysis algorithm that can carry out some preliminary protection action behaviors and the evaluation of switch motion behavior correctness is significant, can additional related personnel carry out further malfunction elimination.Parser based on integer programming, may meet this requirement.This algorithm is by relevant supposition variable (the master-failure state of fault; protection operate condition; switch trip state) as input; according to certain protection professional knowledge; calculate the expectation state of protection action and the expectation state of switch trip; and compare with actual state, the hypothesis tallying with the actual situation is most just as the result of failure diagnosis.

Parser based on integer programming can embody the deviation between hypothesis state and virtual condition with a Zero-one integer programming model description, clear concept, but when solving this model iterations along with the scale of problem once index rise.Therefore, use traversal search method to solve this model and can only be confined to some small-scale problems, after the scale of problem surpasses to a certain degree, can only rely on some has suggestive searching algorithm (tabu search algorithm) and obtains rapidly a convergence solution, but this convergence solution might not be the best, it can only be the optimal solution meeting under quick accident analysis condition.

Parser based on protection action chain is a kind of quick power system fault analysis method, its core logic is for collecting the action event of the same fault zone of section reflection sometime together, according to time sequencing, form a protection action sequence, be referred to as to protect action chain, then utilize local topology and trip protection logic, the tripping operation coverage rate of each protection action in analytical sequence, each action sequence represents a fault, so finally obtaining this method of accident analysis result, to have logic simple, realize the feature of being easy to, but also there is the strict time accuracy that relies on action event, fault-tolerance is short of to some extent, in addition, the method is to Analysis of Complex fault, as overstep tripping etc. also lacks rigorous theoretical foundation, accuracy is also short of to some extent.

Summary of the invention

The present invention seeks to, the intelligent alarm method and system of proposition based on integer programming, overcome the problem of prior art, proposing a kind of accident analysis to system has had the intelligent alarm method and system of rigorous theoretical foundation, system can adopt the adaptive fault solution searching algorithm of intelligence, can adapt to different problem scales, maximum raising the speed of accident analysis.

Technical scheme of the present invention is: the intelligent warning system based on integer programming, and adopt and comprise signal acquisition module, diagnosis starts judge module, failure analysis module, the overview display module of Trouble Report generation module and analysis result; Wherein failure analysis module comprises action chain Algorithm Analysis module and the analysis module based on integer programming; The step of carrying out alarm is as follows

1), network modelling, transforming plant primary equipment and annexation thereof are inserted to database, user power utilization power element builds plant stand or system line chart, by the structural model that meets IEC61970 standard of this mapping mode generating power system plant stand and system, then with this model, by filling out storehouse, operate generating network equipment library, this network equipment storehouse is not only used as network modelling, and in failure diagnostic process, or the data medium of network topology work;

2), protection information modeling: according to primary equipment, protective device, protection information, the hierarchical relationship of trip breaker carries out protection information modeling, set up the incidence relation between the action of protection main equipment---protection information---circuit breaker, protection main equipment refers to primary equipment;

3) step of failure diagnosis:

3-1) by signal acquisition module, the warning information of substation level collection is carried out to the demonstration of the classification based on priority; Gather Real-time Alarm message and the real-time section information of electrical network;

3-2) scanning external alarm information, finds and meets the signal that starts failure diagnosis criterion; Diagnosis starts judge module work, and the criterion that starts failure diagnosis is defined as the combination of following two conditions:

A: the appearance of breaker open operation or emergency stop valve trip signal

B: the protection alarm action message that triggers tripping operation

As long as there is a condition to meet in these two conditions, meet criterion requirement, i.e. criterion E=A ∪ B;

3-3) find after the warning information that meets criterion, this warning information is labeled as to article one information of this fault period alert analysis;

3-4) according to fault warning ending-criterion, judge that whether this warning information is this fault period the last item warning information, is defined as fault warning ending-criterion the combination of following condition:

F: do not have new alarm to occur after this alarm event occurs in a period of time Tdelta;

G: although there is new alarm to occur after this alarm event occurs in a period of time Tdelta, new alarm event time of origin surpasses certain time difference range Tdura apart from this fault period article one alarm event time of origin;

Wherein time difference Tdelta and time period threshold values Tdura all can be defined by the user, and general, the value of these two threshold values can be carried out in following interval: Tdelta ∈ [10,60], and unit: second, Tdura ∈ [2,10], unit: minute.

As long as there is a condition to meet in these two conditions, judgement meets criterion requirement, i.e. criterion H=F ∪ G;

3-5) failure analysis module work, comprise that action chain Algorithm Analysis module and the analysis module based on integer programming start, first carry out the preliminary treatment of failure diagnosis, warning information to failure diagnosis information Biao Biaonei filters, only retain the protection warning information that can trigger tripping operation, automatics action message and circuit breaker displacement information;

3-6) for the warning information event after filtering, carry out the failure diagnosis based on integer programming, analysis result is sent to alarm window,

3-7) the overview display module of Trouble Report generation module and analysis result work, the demonstration of fault diagnosis result and the preservation of fault diagnosis report.

Wherein, the workflow of the analysis module based on integer programming is:

1) determine suspected fault, the malfunction of protection and according to the mistake action and rejecting action state of moving state and switch, and determine thus fault hypothesis phasor H;

Described hypothesis vector H=[D, R, B], D={d wherein 0, d 1, d 2... d m-1represent the malfunction of each main equipment, and 0 represents normally, 1 represents fault, R={rf 0, rm 0, rf 1, rm 1, rf 2, rm 2... rf n-1, rm n-1represent the mistake action and rejecting action state of each protection, rf i=1 represents i protection action malfunction, rm i=1 represents i protection tripping.B={bf 0, bm 0, bf 1, bm 1, bf 2, bm 2... bf q-1, bm q-1show the mistake action and rejecting action state of each switch, bf i=1 represents i switch malfunction, bm i=1 represents i switch failure.

H=[D, R, B] be some variablees that suspected fault region is relevant, specifically determine that method is as follows:

1-1) determining of suspected fault, D's determines

S d = S d b ∪ S d r

D like this kbe S din element E kmalfunction

1-2) the malfunction of protection and according to the determining of moving state, R's determines

for suspected fault element E kassociated all protections (comprising reclosing)

The protection here can be defined as: if element E kassociated main protection has many covers, can by they with or relation be merged into E kmain protection.Element E kbackup protection, breaker fail protection, reclosing is also processed according to the method.Rf like this i, rm ibe S rin protection r imalfunction state and tripping state.

1-3) determining of the mistake action and rejecting action state of switch, B's determines

for suspected fault element E kassociated all switches

Bf like this i, bm ibe S bin switch b imalfunction state and tripping state.

2) dimension of judgement H surpass that convenient search can bear maximum dimension, if surpassed, be to adopt TABU search optimal solution; Otherwise adopt traversal search optimal solution.

Beneficial effect of the present invention is:

(1) Mathematical Modeling that the present invention can be based on strict and protection professional knowledge principle are carried out the structure of intelligent warning system; the accident analysis of system has had rigorous theoretical foundation; this external system adopts intelligence; adaptive fault solution searching algorithm; can adapt to different problem scales, maximum raising the speed of accident analysis.

(2) intelligent warning system, in substation operation, adopts intelligent local topology searching algorithm, and the mode of connection that can be different according to transformer station produces corresponding tripping operation strategy and carries out the differentiation of fault.Like this, system just can adapt to the various operational modes of transformer station, has suitable versatility.

(3) system adopts more intelligent and comprehensive resultant fault result exhibition scheme, mainly comprises and Real-time Alarm two kind exhibition methods online based on picture.Like this, related personnel can check accident analysis result simpler and easy and intuitively, is convenient to historical statistics and the analysis of fault.

In a word, the present invention can adapt to different problem scales, maximum raising the speed of accident analysis.Be better displaying accident analysis result, system adopts more intelligent and comprehensive resultant fault result exhibition scheme, mainly comprises and Real-time Alarm two kind exhibition methods online based on picture.The present invention can fast, effectively carry out the protection motion analysis in labor and failure process to single or complex fault.

Accompanying drawing explanation

Fig. 1 is entire system structure chart;

Fig. 2 is the whole information flow chart of signal acquisition module;

Fig. 3 is the accident analysis algorithm basic flow sheet based on integer programming;

Fig. 4 is the runnable interface figure of the intelligent warning system based on integer programming;

Fig. 5 is the fault bulletin figure after intelligent warning system accident analysis.

Embodiment

Entire system structure chart as shown in Figure 1, is mainly divided into signals collecting, and diagnosis starts judgement, accident analysis, and Trouble Report generates, the modules such as the overview display of analysis result; Fast trouble analysis system nucleus module is based on the accident analysis logic of algorithm of the same race not.Fast trouble analysis system can be in conjunction with the feature of problem, and scale comes switching manual or that automatically carry out various analysis modules to provide optimum accident analysis result.Here failure analysis module is: the analysis module based on integer programming.

1. signal acquisition module

The data of organizing have been only data.” transformer station end, by man-machine interface, carries out the warning information of substation level collection the demonstration of the classification based on priority.Without artificial treatment, in real time the warning information of current period is shown disaggregatedly and is pointed out user to occur at present which other warning information of level, need to first process those alarms at present, very clear.

The collection of real time data is divided into two classes: Real-time Alarm event message, the real-time section information of electrical network; Consider the difference (the interface difference that native system data source and third party scada/EMS data source cause) of data source simultaneously, meet the real-time data acquisition demand from the different types of data in different pieces of information source.In warning message information, enter after the filterEvent table of database; also need to these data screen (screen out non-protection switch action event, screen out excessive signal of the time difference, as the SOE alarm occurring before a day), sequence chronologically, etc. preliminary treatment work.

The Real-time Alarm information table of scada ALM information bank, as eventLog, SOELog, the warning information event table of third party's system (protection waits secondary device action message, switch motion information); Meet and start/finish the criterion that the relevant warning information of fault gathers; The setting in message image data source, the setting in section image data source in real time;

(1) the mutual need that carry out Real-time Alarm event acquisition from native system and third party's system meet real-time needs;

(2) acquisition phase of alarm event should can, by filtration and the shielding of warning information in the test/maintenance/situations such as preposition simulation, be avoided the erroneous judgement of subsequent analysis in source side (SCADA side and third party's system side) as far as possible.

(3) to the criterion setting of startup and end data collection, should there is the restriction of zone of reasonableness.

As shown in Figure 2, the signal acquisition module of the intelligent warning system based on integer programming mainly comprises two functions to the whole information flow of signal acquisition module: and intelligent alarm analysis module alternative events information and and the mutual section information in real time of intelligent alarm analysis module.

The signal acquisition module of the intelligent warning system based on integer programming, its event information source can be the event of PCS9700 scada system acquisition, also can be the event of PCS9000 scada system acquisition, and can be third party's DSN, as D5000 system etc., intelligent warning system based on integer just can move in kinds of platform like this, the data source of three types is via different transmission paths, finally by unified data-interface, delivered to the Real-time Alarm information source table of intelligent warning system, after alarm event preliminary treatment, put in the rapid failure diagnosis information table of intelligent warning system, used for the follow-up accident analysis based on integer programming.

The signal acquisition module of the intelligent warning system based on integer programming, its real-time section source can be the real-time section of PCS9700s system acquisition, also can be the real-time section of PCS9000 system acquisition, and can be the real-time profile data source of third party's system, as D5000 system etc., intelligent warning system based on integer just can move in kinds of platform like this, the data source of three types is delivered to via different data access interfaces in the profile data storehouse of intelligent warning system, used for the follow-up accident analysis based on integer programming.

2. diagnosis starts module

The startup of failure diagnosis is mainly triggered by following two kinds of signals:

1 switch motion,

2 protection actions;

3. the analysis module based on integer programming

Accident analysis algorithm basic procedure based on integer programming is as Fig. 3:

Hypothesis vector H=[D, R, B],, D={d wherein 0, d 1, d 2... d m-1represent the malfunction of each main equipment, and 0 represents normally, 1 represents fault, R={rf 0, rm 0, rf 1, rm 1, rf 2, rm 2... rf n-1, rm n-1represent the mistake action and rejecting action state of each protection, rf i=1 represents i protection action malfunction, rm i=1 represents i protection tripping.B={bf 0, bm 0, bf 1, bm 1, bf 2, bm 2... bf q-1, bm q-1show the mistake action and rejecting action state of each switch, bf i=1 represents i switch malfunction, bm i=1 represents i switch failure.

H=[D, R, B] be some variablees that suspected fault region is relevant, specifically determine that method is as follows:

3.1D determines

S d = S d b ∪ S d r

D like this kbe S din element E kmalfunction

3.2R determines

for suspected fault element E kassociated all protections (comprising reclosing)

The protection here can be defined as: if element E kassociated main protection has many covers, can by they with or relation be merged into E kmain protection.Element E kbackup protection, breaker fail protection, reclosing is also processed according to the method.Rf like this i, rm ibe S rin protection r imalfunction state and tripping state.

3.3B determines

for suspected fault element E kassociated all switches

Bf like this i, bm ifor S bin switch b imalfunction state and tripping state.

3.4 target function

What target function was described is by hypothesis vector H=[D, R, E] expectation state that draws with virtual condition [R r, B r] between difference, specific as follows:

W wherein 1can be set to 100, w 2can be set to 1.The essence of malfunction analysis procedure is exactly to search for Δ (H) to be worth optimum hypothesis vector H like this. represent to move under its protection actual conditions, represent to move under its switch actual conditions. represent to move in its protection expectation situation, represent to move in its switch expectation situation.

3.5 are not counted in the expectation state of protection and switch failure and malfunction

A(r i) be protection r iexpectation state in the situation that being not counted in tripping and malfunction, A (b i) be switch b iexpectation state in the situation that being not counted in tripping and malfunction.

3.5.1 main protection

A(r i)=d k

D kfor r ithe master status of protecting

3.5.2 backup protection

D kfor r ithe state of all main equipments that the main equipment topology of protecting is associated, i 0, i 1... i p-1, p altogether.What the symbol here " V " represented is " or " computing, be also like this in below.

3.5.3 breaker fail protection

be all can tripping switch b jthe expectation state of protection, i 0, i 1... i x-1, x altogether, the symbol here " " represent be " with " computing, be also like this in below.

3.5.4 reclosing

be all can tripping switch b jthe expectation state of route protection (have start reclosing function), i 0, i 1... i y-1, y altogether.

3.5.5 switch

be all can tripping switch b jthe expectation state of protection (not being reclosing), i 0, i 1... i z-1, z altogether.

3.6 count the expectation state of protection and switch failure and malfunction

for protection r iexpectation state in the situation that counting tripping and malfunction, for switch b iexpectation state in the situation that counting tripping and malfunction.

3.6.1 protection and reclosing

3.6.2 switch

3.6.3 the hypothesis of contradiction

Rm i=rf i=1 or bm i=bf i=1

A(r i)=0 and rm i=1, or A (b i)=0 and bm i=1

A(r i)=1 and rf i=1, or A (b i)=1 and bf i=1

Meeting the hypothesis of one of above condition, is the hypothesis of contradiction, does not carry out the calculating of target function value, directly gets rid of.In the time of program actual motion, by getting rid of in the hypothesis that counts protection and switch failure and malfunction expectation state, do not carry out the calculating of target function, to improve process analysis efficiency.

The search of 3.7 optimum hypothesis

3.7.1 the coding of hypothesis vector

Hypothesis vector H=[D, R, B]=[d 0, d 1, d 2... d m-1.

rf 0,rm 0,rf 1,rm 1,rf 2,rm 2...rf n-1,rm n-1

bf 0,bm 0,bf 1,bm 1,bf 2,bm 2...bf q-1,bm q-1]

Can be with the binary number representation of a m+2n+2q position, hypothesis vector H has 2 like this m+2n+2qthe situation of kind, 2 m+2n+2qthe dimension that just represents hypothesis phasor.

3.7.2 the traversal search of optimum hypothesis

If find optimum H with traversal search, searching times is 2 like this m+2n+2q, along with the scale of problem is index rising.Therefore, traversal search can only be confined to a small-scale problem.

3.7.3 the TABU search of optimum hypothesis

When m+2n+2q is larger, use traversal search, computing time will be very long, lost the meaning of rapid analysis.Therefore, need a quicker and intelligent searching algorithm to search for optimal solution.TABU search is exactly an algorithm that meets above condition.

TABU search (Tabu Search or Taboo Search are called for short TS), it is a kind of expansion to the search of local field, is a kind of overall Stepwise optimization algorithm, is a kind of simulation to human mind's process.TS algorithm by introduce one flexibly storage organization and corresponding taboo criterion avoid roundabout search, and absolve by despising criterion the excellent condition that some are avoided, and then guarantee that diversified effective exploration is finally to realize global optimization.

For target function herein, an approximately continuous function, does not have too high-frequency shake in fact, therefore utilizes tabu search algorithm to carry out the search of optimum hypothesis vector H, has suitable rapidity and accuracy

3.7.4 some concepts of TABU search

Mobile: movement is to refer to hypothesis vector to carry out the once operation of position, specifically comprises two kinds: unit moves and exchange movement.Unit moves a certain position to hypothesis vector referring to and carries out negate, mobile certain two exchange that are worth to hypothesis vector that refer to of exchange.These two kinds of movements be all meet symmetric.

Neighborhood: the neighborhood is here the concept of a narrow sense, concrete is exactly hypothesis vector to be carried out to all units move with pin and change mobile the vectorial set drawing.Be used for as candidate solution.

Taboo list: taboo list has comprised some and moved, these move can be so that current solution worsens.

Emission levels: certain moves and meets such condition in taboo list: after acting on current solution, can be better than optimal solution in the past, just think that this moves and meet emission levels.

The size of taboo list: the size of taboo list can be defined as

The taboo frequency: the maximum taboo frequency can be defined as

Worsen number of times: worsening number of times can be defined as

Maximum dimension: the figure place of the hypothesis vector maximum herein can be 22, like this N clmmAX:

Maximum iteration time: n is the problem of traversal search energy Real-time solution

The maximum of dimension

3.7.5 the flow process of TABU search

Its basic procedure is:

1. produce initial solution, iterations is set to k, worsens number of times k warse, current optimal solution is set to H best:

2. if k>=KMAX or k warse>=KMAX, exporting l is optimal solution, EP (end of program).Otherwise proceed to the 3rd step.

3. iterations is set to k=1.To in taboo list, over the entry of maximum taboo frequency KMAX, delete, the taboo frequency of remaining entry increases by 1 time, if the top unnecessary entry deletion in showing over S ' of taboo list length.Each element of right neighborhood carries out the calculating of target function and sorts and obtain candidate solution sequence according to the size of value

4., if be better than 1, will k '=0, proceeds to the 2nd step, otherwise proceeds to the 5th step.

5.k warse=k wars。If the movement producing not in taboo list or in taboo list but reached emission levels, is set to current solution and the movement of generation joined in taboo list and proceed to the 2nd step, otherwise investigating successively

4. Trouble Report generation module

System should, after accident roughly in 2 minutes, provide far, and report content should at least comprise:

● time of failure;

● faulty equipment;

● protection action message;

● automatic safety device action message;

● switch motion information;

● tripping switch;

● Judgement of failure: transient fault or permanent fault.

● fault is separate;

● protection exception behavior evaluation

5. the overview display module of analysis result

In the online and Real-time Alarm of the picture of man-machine interface, tackle fault message and carry out overview display, displaying content comprises:

● the labor content of the last fault of overview display.Comprise:

● the particular content of position of failure point, fault point analysis result: time of failure, faulty equipment, nature of trouble,, concrete rapid analysis process, protection action message, automatic safety device action message, switch motion information, malfunction switch, tripping switch, etc.;

● protection action behavior check results: anti-action protector for current, malfunction protective device actuating signal;

● protection action message and switch motion information:

● can pass through the associative operation menu item of analysis result, obtain the data source information of correlation analysis;

● alarm window should provide simple accident analysis result message information; And by this message information, call the function of the impression window of concrete fault diagnosis report.

Set up model:

1, network modelling, inserts database by primary equipment and annexation thereof.This work can utilize domestic very ripe artwork storehouse integrated technique, and in conjunction with pel, graphics edition instrument, by drawing, filling out, storehouse two step operations can complete.Force device first generates by Image editors, then, in graphics edition instrument, user power utilization power element builds plant stand or system line chart, by this mapping mode generating power system plant stand and system structure model, meets IEC61970 standard, then with this model, by filling out storehouse, operate generating network equipment library, this network equipment storehouse is not only as network modelling, in failure diagnostic process below, also by the data medium that is network topology work.So far, meeting the primary equipment network modelling work that this method requires completes.

2, protection information modeling: according to primary equipment; protective device, protection information, the hierarchical relationship of trip breaker carries out protection information modeling; set up the incidence relation between the action of protection main equipment---protection information---circuit breaker, protection main equipment refers to primary equipment.

The step of failure diagnosis

The 1st step: scanning external signal information, find and meet the signal that starts failure diagnosis criterion.We are defined as the criterion that starts failure diagnosis the combination of following two conditions:

A: the appearance of breaker open operation or emergency stop valve trip signal

B: the protection alarm action message that triggers tripping operation

As long as there is a condition to meet in these two conditions, meet criterion requirement, i.e. criterion E=A ∪ B;

The 2nd step: find after the warning information that meets criterion, this warning information is labeled as to article one information of this fault period alert analysis;

The 3rd step: according to fault warning ending-criterion, judge whether this warning information is this fault period the last item warning information, and we are defined as fault warning ending-criterion the combination of following condition:

F: do not have new alarm to occur after this alarm event occurs in a period of time Tdelta;

G: although there is new alarm to occur after this alarm event occurs in a period of time Tdelta, new alarm event time of origin surpasses certain time difference range Tdura apart from this fault period article one alarm event time of origin.

Wherein time difference Tdelta and time period threshold values Tdura all can be defined by the user, and general, the value of these two threshold values can be carried out in following interval: Tdelta ∈ [10,60], and unit: second, Tdura ∈ [2,10], unit: minute.

As long as there is a condition to meet in these two conditions, judgement meets criterion requirement, i.e. criterion H=F ∪ G.

The 4th step, enters the pretreatment stage of failure diagnosis, and the warning information of failure diagnosis information Biao Biaonei is filtered, and only retains the protection warning information that can trigger tripping operation, automatics action message and circuit breaker displacement information;

The 5th step: carry out the failure diagnosis based on integer programming for the warning information event after filtering, analysis result is sent to alarm window, as shown in Figure 4;

The 6th step: the demonstration of fault diagnosis result and the preservation of fault diagnosis report, as shown in Figure 5.

Claims (3)

1. the intelligent warning system based on integer programming, is characterized in that: adopt and comprise signal acquisition module, diagnosis starts judge module, failure analysis module, and the overview display module of Trouble Report generation module and analysis result is carried out alarm; Wherein failure analysis module comprises action chain Algorithm Analysis module and the analysis module based on integer programming; The step of carrying out alarm is as follows,
1) network modelling, inserts database by transforming plant primary equipment and annexation thereof; The structural model that meets IEC61970 standard of generating power system plant stand and system, then with this model, by filling out storehouse, operate generating network equipment library, this network equipment storehouse is not only used as network modelling, and in failure diagnostic process, or the data medium of network topology work;
2) protection information modeling: according to primary equipment, protective device, protection information, the hierarchical relationship of trip breaker carries out protection information modeling, set up the incidence relation between the action of protection main equipment---protection information---circuit breaker, protection main equipment is primary equipment;
3) step of failure diagnosis:
3-1) by signal acquisition module, the warning information of substation level collection is carried out to the demonstration of the classification based on priority; Gather Real-time Alarm message and the real-time section information of electrical network;
3-2) scanning external alarm information, finds and meets the signal that starts failure diagnosis criterion; Diagnosis starts judge module work, and the criterion that starts failure diagnosis is defined as the combination of following two conditions:
A: the appearance of breaker open operation or emergency stop valve trip signal
B: the protection alarm action message that triggers tripping operation
As long as there is a condition to meet in these two conditions, meet criterion requirement, i.e. criterion E=A ∪ B;
3-3) find after the warning information that meets criterion, this warning information is labeled as to article one information of this fault period alert analysis;
3-4) according to fault warning ending-criterion, judge that whether this warning information is this fault period the last item warning information, is defined as fault warning ending-criterion the combination of following condition:
F: do not have new alarm to occur after this alarm event occurs in a period of time Tdelta;
G: although there is new alarm to occur after this alarm event occurs in a period of time Tdelta, new alarm event time of origin surpasses certain time difference range Tdura apart from this fault period article one alarm event time of origin;
Wherein time difference Tdelta and time period threshold values Tdura are all defined by the user, and the value of these two threshold values is carried out in following interval: Tdelta ∈ [10,60], and unit: second, Tdura ∈ [2,10], unit: minute;
As long as there is a condition to meet in these two conditions, judgement meets criterion requirement, i.e. criterion H=F ∪ G;
3-5) failure analysis module work, comprise that action chain Algorithm Analysis module and the analysis module based on integer programming start, first carry out the preliminary treatment of failure diagnosis, warning information to failure diagnosis information Biao Biaonei filters, only retain the protection warning information that can trigger tripping operation, automatics action message and circuit breaker displacement information;
3-6) for the warning information event after filtering, carry out the failure diagnosis based on integer programming, analysis result is sent to alarm window,
3-7) the overview display module of Trouble Report generation module and analysis result work, the demonstration of fault diagnosis result and the preservation of fault diagnosis report.
2. the intelligent warning system based on integer programming according to claim 1, is characterized in that the workflow of the analysis module based on integer programming is:
1-1) determine suspected fault, the mistake action and rejecting action state of protection and the mistake action and rejecting action state of switch, and determine thus fault hypothesis phasor H;
Described hypothesis vector H=" D, R, B], D={d wherein 0, d 1, d 2.., d m-1represent the malfunction of each main equipment, and 0 represents normally, 1 represents fault, R={rf 0, rm 0, rf 1, rm 1, rf 2, rm 2..rf n-1, rm n-1represent the mistake action and rejecting action state of each protection, rf i=1 represents i protection action malfunction, rm i=1 represents i protection tripping; B={bf 0, bm 0, bf 1, bm 1, bf 2, bm 2..bf q-1bm q-1represent the mistake action and rejecting action state of each switch, bf i=1 represents i switch malfunction, bm i=1 represents i switch failure;
H=[D, R, B] be some variablees that suspected fault region is relevant, specifically determine that method is as follows:
1-1-1) determining of suspected fault, D's is definite,
for the relevant not live components of all trip switch two etale topologies },
element for the protection association of everything },
d like this kbe S din element E kmalfunction;
1-2) determining of the mistake action and rejecting action state of protection, R's is definite,
it is suspected fault element associated all protections, comprise reclosing;
The protection is here defined as: if element E kassociated main protection has many covers, by their with or relation be merged into E kmain protection; Element E kbackup protection, breaker fail protection, reclosing is also processed according to the method; Rf like this i, rm ibe S rin protection r imalfunction state and tripping state;
1-3) determining of the mistake action and rejecting action state of switch, B's determines;
for suspected fault element E kassociated all switches
Bf like this i, bm ibe s bin switch b imalfunction state and tripping state.
3. the intelligent warning system based on integer programming according to claim 1, is characterized in that
1-1) setting up target function calculation expectation value target function and describe, is by hypothesis vector H=[D, R, B] expectation state [R that draws *, B *] and virtual condition [R r, B r] between difference, specific as follows:
W wherein 1be set to 100, w 2be set to 1, the essence of malfunction analysis procedure is exactly to search for Δ (H) to be worth optimum hypothesis vector H like this, represent to move under its protection actual conditions, represent to move under its switch actual conditions; represent to move in its protection expectation situation, represent to move in its switch expectation situation;
1-1-1) be not counted in the expectation state of protection and switch failure and malfunction
A(r i) be protection r iat the expectation state being not counted under tripping and maloperation situation, A (b i) be switch b iexpectation state in the situation that being not counted in tripping and malfunction;
1-1-1-1) main protection
A(r i)=d k
D kfor r ithe master status of protecting
1-1-1-2) backup protection
D kfor r ithe state of all main equipments that the main equipment topology of protecting is associated, i 0, i 1..i p-1, p altogether; What the symbol here " V " represented is inclusive-OR operation, is also like this in below;
1-1-1-3) breaker fail protection
all tripping switch b jthe expectation state of protection, i 0, i 1..i x-1, x altogether, what the symbol " " here represented is AND operation, is also like this in below;
1-1-1-4) reclosing
be all can tripping switch b jroute protection there is the expectation state that starts reclosing function, i 0, i 1..i y-1, y altogether;
1-1-1-5) switch
be all can tripping switch b iprotection be not the expectation state of reclosing, i 0, i 1..i z-1, z altogether;
1-1-2) count the expectation state of protection and switch failure and malfunction
for protection r iexpectation state in the situation that counting tripping and malfunction, for switch b iexpectation state in the situation that counting tripping and malfunction;
1-1-2-1) protection and reclosing
1-1-2-2) switch
1-1-2-3) the hypothesis of contradiction
Rm i=rf i=1 or bm i=bf i=1
A(r i)=0 and rm i=1, or A (b i)=O and bm i=1
A(r i)=1 and rf i=1, or A (b i)=1 and bf i=1
Meeting the hypothesis of one of above condition, is the hypothesis of contradiction, does not carry out the calculating of target function value, directly gets rid of;
1-2) dimension of judgement H surpass that convenient search can bear maximum dimension, if surpassed, be to adopt TABU search optimal solution; Otherwise adopt traversal search optimal solution.
CN201110425428.7A 2011-12-16 2011-12-16 Intelligent warning system based on integer programming CN102497024B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201110425428.7A CN102497024B (en) 2011-12-16 2011-12-16 Intelligent warning system based on integer programming

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201110425428.7A CN102497024B (en) 2011-12-16 2011-12-16 Intelligent warning system based on integer programming

Publications (2)

Publication Number Publication Date
CN102497024A CN102497024A (en) 2012-06-13
CN102497024B true CN102497024B (en) 2014-01-29

Family

ID=46188822

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201110425428.7A CN102497024B (en) 2011-12-16 2011-12-16 Intelligent warning system based on integer programming

Country Status (1)

Country Link
CN (1) CN102497024B (en)

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103020713A (en) * 2012-11-19 2013-04-03 山东大学 Intelligent substation fault diagnosis method combining topology and relay protection logic
CN102967780B (en) * 2012-11-21 2015-04-08 国电南瑞科技股份有限公司 Modeling method for substation intelligent alarm
CN103197168B (en) * 2013-02-21 2015-08-12 华东电网有限公司 Realize the method for fault diagnosis control based on the event set chain of causation in electric system
CN103208049B (en) * 2013-04-25 2016-10-05 国家电网公司 Abnormality alarming quick accident analysis method and system
CN103324128B (en) * 2013-05-30 2015-07-01 国家电网公司 Primary equipment fault warning comprehensive compression method in power dispatching automation system
CN103281410B (en) * 2013-06-14 2016-08-31 江苏省广电有线信息网络股份有限公司南京分公司 Broadcast television network intelligent obstacle pretreatment method and system
CN103545922B (en) * 2013-09-24 2015-10-28 许继集团有限公司 Based on the intelligent alarm inference method of many scene analysis
CN104316801B (en) * 2014-10-31 2015-09-02 国家电网公司 Based on the power system failure diagnostic method of Time Series Similarity coupling
CN106911492A (en) * 2015-12-23 2017-06-30 北京谊安医疗系统股份有限公司 A kind of Anesthesia machine warning message detection and display methods
CN105678337B (en) * 2016-01-12 2020-02-04 国网技术学院 Information fusion method in intelligent substation fault diagnosis
CN105700506B (en) * 2016-04-29 2018-02-06 南京信息职业技术学院 A kind of long-range transformer station temperature and humidity control system
CN107658980B (en) * 2017-09-29 2020-12-18 国网浙江省电力公司 Analysis method and system for rechecking power grid monitoring alarm information

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101316051A (en) * 2008-07-03 2008-12-03 绍兴电力局 Internetwork communication log analysis system and method based on IEC61850 transforming plant automatization system
CN102035260A (en) * 2010-12-17 2011-04-27 国电南瑞科技股份有限公司 Intelligent warning and fault analysis realizing method for intelligent transformer substation monitoring system

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101316051A (en) * 2008-07-03 2008-12-03 绍兴电力局 Internetwork communication log analysis system and method based on IEC61850 transforming plant automatization system
CN102035260A (en) * 2010-12-17 2011-04-27 国电南瑞科技股份有限公司 Intelligent warning and fault analysis realizing method for intelligent transformer substation monitoring system

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
刘伟等.智能变电站智能告警与辅助决策的实现.《电力系统保护与控制》.2011,第39卷(第15期),第146-150页.
变电站智能告警专家系统研究;姬书军等;《华东电力》;20110531;第39卷(第5期);第0773-0775页 *
姬书军等.变电站智能告警专家系统研究.《华东电力》.2011,第39卷(第5期),第0773-0775页.
智能变电站智能告警与辅助决策的实现;刘伟等;《电力系统保护与控制》;20110801;第39卷(第15期);第146-150页 *

Also Published As

Publication number Publication date
CN102497024A (en) 2012-06-13

Similar Documents

Publication Publication Date Title
CN103400302B (en) A kind of wind power base cascading failure risk perceptions method for early warning
Giri et al. The situation room: Control center analytics for enhanced situational awareness
CN102170124B (en) Early warning method of stable-state index of power quality
US6868311B2 (en) Method and system for on-line dynamical screening of electric power system
CN101577426B (en) Power system state estimator applicable to wide area measurement system
CN106124935B (en) Middle and low voltage network Fault Locating Method
CN103296677B (en) A kind of online bulk power grid recovers aid decision-making system
CN201532416U (en) Intelligent cable monitor
CN104578426B (en) Detection method for information comprehensive analysis and intelligent alarming application
CN101692585B (en) Anti-operation locking system of transformation substation automatic system
CN102638100B (en) District power network equipment abnormal alarm signal association analysis and diagnosis method
CN102496072B (en) System for estimating distributive state of intelligent transformer station
CN104297637B (en) The power system failure diagnostic method of comprehensive utilization electric parameters and time sequence information
CN104578414B (en) One kind realizes intelligent station protection act visualization method based on SVG
CN104950187B (en) A kind of thunder and lightning analysis and early warning method and system based on power grid GIS
CN203324773U (en) Wind farm centralized control system
CN1120556C (en) In-line monitoring and prevention control system with stable transient for power system
CN103278746B (en) Consider the implementation method of the electric network failure diagnosis system of uncertain information
CN103414251A (en) Substation network communication simulation system based on IEC 61850
CN103607042B (en) The distribution network failure processing method of long fault indicator for overhead lines towards outskirts of a town
CN101419267B (en) Electric network fault on-line rapid diagnosis and analysis method
CN103500249A (en) Visual relay protection setting calculation system and method
CN102193504B (en) Safety and stability control system modeling method in power system dynamic simulation
Guo et al. An analytic model-based approach for power system alarm processing employing temporal constraint network
CN104753178A (en) Power grid fault handling system

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant