CN108399579A - The intelligently parsing system of substation equipment monitoring data signal - Google Patents

The intelligently parsing system of substation equipment monitoring data signal Download PDF

Info

Publication number
CN108399579A
CN108399579A CN201810034859.2A CN201810034859A CN108399579A CN 108399579 A CN108399579 A CN 108399579A CN 201810034859 A CN201810034859 A CN 201810034859A CN 108399579 A CN108399579 A CN 108399579A
Authority
CN
China
Prior art keywords
event
signal
substation
monitoring
equipment
Prior art date
Application number
CN201810034859.2A
Other languages
Chinese (zh)
Inventor
王伟臣
赵向阳
郝达智
杨晓静
黄志刚
冯长有
王健
王伟力
马超
王晶
赵士朗
马钢
王鑫
路树森
张�杰
王飞跃
Original Assignee
国网天津市电力公司
中国科学院自动化研究所
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 国网天津市电力公司, 中国科学院自动化研究所 filed Critical 国网天津市电力公司
Priority to CN201810034859.2A priority Critical patent/CN108399579A/en
Publication of CN108399579A publication Critical patent/CN108399579A/en

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING; COUNTING
    • G06Q—DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTING PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • Y04S10/60—

Abstract

The present invention relates to the intelligently parsing fields of monitoring data signal, and in particular to a kind of intelligently parsing system of substation equipment monitoring data signal.It is intended to provide a kind of intelligently parsing system, be changed into event all announcement mode to realize data information by traditional all announcement mode one by one and establishes correspondence and inner link between notification and monitored device.Intelligently parsing system proposed by the present invention, pass through the study and analysis of the monitoring data signal that the intelligence learning algorithm of Bayesian neural network acquires monitoring system, realize analysis and management of the auxiliary monitoring management personnel to monitoring signal, simultaneously according to the intelligent monitoring information Disposal Strategies library of foundation and network topology structure library, in conjunction with specific monitoring signal, realize automatic sensing, the analysis and decision when warning information is sent out, mitigate the pressure of staff, it reduces it and largely repeats cumbersome workload, the way to manage of high efficient and reliable is provided for monitoring management.

Description

The intelligently parsing system of substation equipment monitoring data signal

Technical field

The present invention relates to the intelligently parsing fields of monitoring data signal, and in particular to a kind of substation equipment monitoring data letter Number intelligently parsing system.

Background technology

With science and technology and economical continuous development and progress, power equipment scale increasingly increases, and to the steady of power equipment The requirement for determining reliability service is also higher and higher.With the development of microcomputer data processing, on-line monitoring technique is got over To be more widely applied, therefore it is also more and more comprehensively to the operation monitoring data of power equipment.The numerous and disorderly Various types of data to differ Oligomerisation information platform of supervising, is directly presented to monitoring personnel, also increasingly increases the work disposal Capability Requirement of monitoring personnel.Cause How this mitigates the pressure of monitoring personnel, realize data present it is kinder and more approachable, more clear and easy to understand, more it is intuitive understand be It is highly important.

Current substation equipment monitoring information management system mainly by various converging informations together, then be pocessed And it is presented to staff.The exhibition method of information is that monitoring information directly by magnanimity is sent into the eye of monitoring personnel, including The warning information etc. of a large amount of each switch tool and protection equipment, need timely to analyze after these a large amount of signal acquisitions and Processing, this is very arduous for monitoring personnel.Simultaneously when a certain failure or event occur for monitoring device, also can Simultaneous others signal, these other signals are since caused by some associate devices, monitoring personnel is also It needs to analyze these factors, undoubtedly increases difficulty.

Invention content

In order to solve the above problem in the prior art, the present invention proposes a kind of substation equipment monitoring data signal Intelligently parsing system alleviates the operating pressure of monitoring management personnel, reduces workload, improves the efficiency of management.

The present invention is intended to provide a kind of intelligently parsing method, to realize that data information is changed by traditional all announcement mode one by one For event all announcement mode, and establish the correspondence and inner link between notification and monitored device.Pass through pattra leaves Auxiliary monitoring is realized in the study and analysis for the monitoring data signal that the intelligence learning algorithm of this neural network acquires monitoring system Analysis and management of the administrative staff to monitoring signal, while according to the monitoring information Disposal Strategies library of foundation and network topology structure Library, in conjunction with specific monitoring signal, the automatic sensing, analysis and decision when realization warning information is sent out are for substation face from now on The problem for a large amount of monitoring signal processing analysis faced, the help monitoring management personnel which is undoubtedly capable of high efficient and reliable realize Management to power system monitor information provides the way to manage of high efficient and reliable for monitoring management.

The present invention proposes a kind of intelligently parsing system of substation equipment monitoring data signal, including：Standardization mould Block, data-signal parsing module and monitoring information Disposal Strategies library；

The standardization module, the conversion for carrying out format and content to monitoring data signal to be resolved are defeated Go out standardization treated monitoring data signal to be resolved, including：The affiliated substation of signal time, signal, signal generating apparatus Title and signal type；

The data-signal parsing module is Bayesian neural network, for to be resolved after the standardization Monitoring data signal is parsed, and one or more monitor events corresponding to the monitoring data signal are obtained；

Monitoring information Disposal Strategies library, for the analysis result according to the Bayesian neural network, query monitor Data-signal corresponding signal paraphrase, causes consequence and Disposal Strategies at the analysis of causes.

Preferably, the Bayesian neural network, the history monitoring data after training is standardization with data set are believed Number.

Preferably, the Bayesian neural network, including input layer, mode layer, summation layer and output layer；

The input layer, is configured to：The monitoring data signal after the standardization is received, and passes to node net Network, the neuron number of input layer network are equal with the dimension of monitoring data signal after the standardization；

The mode layer, is configured to：From input layer obtain input data feature vector, calculate input data feature vector with The matching relationship of each monitor event in training set；The neuron number of mode layer is the group number of training sample；

The summation layer, is configured to：Each monitor event only corresponds to a summation layer unit；Each summation layer unit only with The pattern layer units of its corresponding one or more monitor event are connected, and by the output result of the pattern layer units connected It is added, to obtain the estimated probability density function of the monitor event；

The output layer, is configured to：The probability density function for receiving each monitor event exported from summation layer, passes through normalizing Change is handled, and obtains the probability Estimation of each monitor event；Select a neuron with maximum a posteriori probability density as entire The output of system.

Preferably, monitoring information Disposal Strategies library, establishing criteriaization treated history monitoring data signal, and its Corresponding signal paraphrase, the analysis of causes cause consequence and Disposal Strategies to build.

Preferably, the intelligently parsing system further includes：Network topology structure library；

The network topology structure library, the relevance for inquiring relevance and signal between each equipment；It is described Network topology structure library is the double-layer network topological structure built according to the electric connecting relation of each equipment in substation, including： Topological structure is inscribed in the external topological structure of substation and substation.

Preferably, intelligently parsing system further includes：It is grouped display module；

The grouping display module, for being packed into prison to the query result grouping in monitoring information Disposal Strategies library Event information list is controlled, and is sent in display equipment and shows；And/or according to the monitoring number to be resolved after the standardization It is believed that number, the relevant information in the network topology structure library is inquired, and be sent in display equipment and show.

Preferably, the network topology structure library using layering grouping form by list by the way of carry out storage and It calls.

Preferably, external topological structure of the substation, including power transformation station name, substation's attribute, transformer substation voltage etc. Interval type between grade, connection type, external power transformation station name and external substation；

Wherein：

Substation's attribute, including intelligence and conventional substation；

The transformer substation voltage grade, including 500KV, 330KV, 220KV, 110KV and 66KV；

The connection type, including double female double separated times, single mother partition, 3/2 wiring and single mother；

Interval type between the external substation, including main transformer body, main transformer high-pressure side, main transformer medium voltage side, main transformer are low Press side, line segregation, capacitor standoff, reactor interval and busbar interval.

Preferably, topological structure is inscribed in the substation, is the connection topology shape between each equipment inside substation Formula.

Preferably, topological structure is inscribed in the substation, and construction method is：

Endpoint to equipment all in substation and tie point Unified number；

The equipment list of description device attribute is set, the port table of port attribute is described and describes the node of tie point attribute Table obtains the substation and topological structure is inscribed；

Wherein：

The equipment, including：Power cable, transformer, switch, busbar, capacitor and reactor；

The equipment list, including：Number, device name and the equipment endpoint number of equipment；

The port table, including：Port name, port numbering and the tie point number being connected with the port；

The node table, including：Nodename, node serial number and the port numbering being connected with the node.

Preferably, the monitoring data signal, including：The status information of multiple equipment in substation and equipment alarm letter Breath.

Preferably, the multiple equipment in the substation, including：Main transformer include first set protection, second set protect and Measure and control device；Main transformer high side switch includes breaker, disconnecting link position, GIS device, spring mechanism of circuit breaker, mechanism letter extremely Number, control box, control loop state and measure and control device；Main transformer low-pressure side switch includes breaker, disconnecting link position, GIS device, breaks Road device spring mechanism, mechanism abnormal signal, control box, control loop state and measure and control device；Busbar includes disconnecting link position, mother If；Circuit includes disconnecting link position, SF6 breakers, GIS device, SF6 mutual inductors, Hydraulic Mechanism of Circuit-breaker, breaker pneumatic motor Structure, mechanism abnormal signal, control loop state, control box, protective device, measure and control device, merges list at spring mechanism of circuit breaker Member, intelligent terminal and intelligent cell.

Preferably, the monitor event is line events, main transformer event, main transformer high side switch event, main transformer low-pressure side Switch events or busbar event；

Wherein,

The line events, be disconnecting link position event, SF6 breakers event, GIS device event, SF6 mutual inductors event, Hydraulic Mechanism of Circuit-breaker event, pneumatic mechanism for circuit breaker event, spring mechanism of circuit breaker event, mechanism abnormal signal event, control Loop state event processed, control box event, protective device event, measure and control device event, combining unit event, intelligent terminal event Or intelligent cell event；

The main transformer event protects event, second set of protection event or measure and control device event for first set；

Become on high-tension side breaker event, disconnecting link position event, GIS device thing based on the main transformer high side switch event Part, spring mechanism of circuit breaker event, mechanism abnormal signal event, control box event, control loop state event or measure and control device Event；

The main transformer low-pressure side switch events are breaker event, disconnecting link position event, the GIS device of main transformer low-pressure side Event, spring mechanism of circuit breaker event, mechanism abnormal signal event, control box event, control loop state event or observing and controlling dress Set event；

The busbar event sets event for disconnecting link position event or mother.

Beneficial effects of the present invention：

The present invention passes through the rule that are standardized monitoring data signal, establish between monitoring signal and monitor event Then model, i.e. Bayesian neural network model, establish monitoring information Disposal Strategies library, establish substation network topological structure library with And intelligently parsing result is packaged grouping display and is convenient for monitoring personnel analysis management.The signal of system combination Bayesian neural network Analysis result and monitoring information Disposal Strategies library generate the associated equipment situation of monitoring signal, caused by impact analysis and disposition Strategy protocol realizes automatic sensing, analysis and decision when monitoring information is sent out, processing of the auxiliary monitoring personnel to monitoring information Analysis, is greatly reduced operating pressure, reduces workload, improve the efficiency of management of monitoring management personnel.

Description of the drawings

Fig. 1 is the composition schematic diagram of the intelligently parsing system of substation equipment monitoring data signal in the embodiment of the present invention；

Fig. 2 is the structure of Bayesian neural network and its connection relation with standardization module in the embodiment of the present invention Schematic diagram；

Fig. 3 is substation network topological structure schematic diagram in the embodiment of the present invention；

Fig. 4 is Bayesian neural network resolving schematic diagram in the embodiment of the present invention.

Specific implementation mode

The preferred embodiment of the present invention described with reference to the accompanying drawings.It will be apparent to a skilled person that this A little embodiments are only used for explaining the technical principle of the present invention, it is not intended that limit the scope of the invention.

Intelligent algorithm and artificial study provide an opportunity to the intelligently parsing of substation equipment monitoring information.By counting greatly According to excavation calculate and study can learn and find the relevance in data between, therefore electric power is handled using intelligent algorithm Equipment mass data has huge advantage.

The present invention the intelligently parsing system for substation equipment monitoring data signal, by monitoring data signal into Row standardization establishes rule model between monitoring signal and monitor event, establishes monitoring information Disposal Strategies library, establishes Substation network topological structure library and intelligently parsing result are packaged grouping display and are convenient for monitoring personnel analysis management.

Monitoring system carries out the power grid in administrative area the acquisition of real time monitoring and information, and information collection is with per minute Section is time point, acquires information and calculates and show, and is stored to every 5~10 minutes acquisition information.

The embodiment of the intelligently parsing system of a kind of substation equipment monitoring data signal of the present invention, as shown in Figure 1, packet It includes：Standardization module 10, Bayesian neural network (i.e. data-signal parsing module) 20, monitoring information Disposal Strategies library 30, network topology structure library 40 and grouping display module 50.

The standardization module 10, the conversion for carrying out format and content to monitoring data signal to be resolved, Outputting standardization treated monitoring data signal to be resolved, including：The affiliated substation of signal time, signal, signal generation are set Standby title and signal type.

The Bayesian neural network 20, for being solved to the monitoring data signal to be resolved after the standardization Analysis, obtains one or more monitor events corresponding to the monitoring data signal.

Monitoring information Disposal Strategies library 30, for the analysis result according to the Bayesian neural network 20, inquiry The corresponding signal paraphrase of monitoring data signal, causes consequence and Disposal Strategies at the analysis of causes.Monitoring information Disposal Strategies library 30, be according to after standardization history monitoring data signal and its corresponding signal paraphrase, the analysis of causes, cause consequence With Disposal Strategies structure.

The network topology structure library 40, the relevance for inquiring relevance and signal between each equipment；The knot Structure library is the double-layer network topological structure built according to the electric connecting relation of each equipment in substation, including：Substation is external Topological structure is inscribed in topological structure and substation.The network topology structure library 40 passes through list in the form of layering is grouped Mode stored and called.

The grouping display module 50, for being packaged shape to the query result grouping in monitoring information Disposal Strategies library 30 At monitor event list, and it is sent in display equipment and is shown；And/or according to the prison to be resolved after the standardization Data-signal is controlled, the relevant information inquired in the network topology structure library 40 flushes in display equipment.Monitoring management personnel It can be executed corresponding in conjunction with the connection relation shown by network topology structure library according to the query result in Disposal Strategies library Processing strategy.

In the present embodiment, the Bayesian neural network 20, including input layer, mode layer, summation layer and output layer.

The input layer, is configured to：The monitoring data signal after the standardization is received, and passes to node net Network, the neuron number of input layer network are equal with the dimension of monitoring data signal after the standardization.

The mode layer, is configured to：From input layer obtain input data feature vector, calculate input data feature vector with The matching relationship of each monitor event in training set；The neuron number of mode layer is the group number of training sample.

The summation layer, is configured to：Each monitor event only corresponds to a summation layer unit；Each summation layer unit only with The pattern layer units of its corresponding one or more monitor event are connected, and are not connect with other units in mode layer. Therefore, summation layer simply will to one's name correspond to event pattern layer units output addition, and with belong to other events The unit output of mode layer is unrelated.Exporting for layer of summing is proportional to the estimation of probability density function of each event based on kernel.

The output layer, is configured to：The probability density function for receiving each monitor event exported from summation layer, passes through normalizing Change is handled, and obtains the probability Estimation of each monitor event；The output decision-making level of network is made of simple threshold value discriminator, by Select a neuron with maximum a posteriori probability density as the defeated of whole system in the estimated probability density of each event Go out.Output layer is a kind of competition neurons, and each neuron corresponds respectively to a data type i.e. monitor event, it receive from The probability density function of each monitor event of summation layer output, maximum that neuron output of probability density function is 1, i.e. institute That corresponding monitor event is the monitor event classification of sample to be identified, and the output of other neurons is all 0.

In the present embodiment, the external topological structure of substation, including power transformation station name, substation's attribute, substation's electricity Interval type etc. between pressure grade, connection type, external power transformation station name and external substation.

Wherein：

Substation's attribute, including intelligence and conventional substation；

The transformer substation voltage grade, including 500KV, 330KV, 220KV, 110KV and 66KV etc.；

Described connection type, including double female double separated times, single mother partition, 3/2 wiring and single mother etc.；

Interval type between the external substation, including main transformer body, main transformer high-pressure side, main transformer medium voltage side, main transformer are low Press side, line segregation, capacitor standoff, reactor interval and busbar interval etc..

In the present embodiment, topological structure is inscribed in the substation, and the connection between each equipment inside substation is opened up Flutter form.Its construction method is：

By the endpoint of equipment all in substation and tie point Unified number；

The equipment list of description device attribute is set, the port table of port attribute is described and describes the node of tie point attribute Table obtains the substation and topological structure is inscribed；

Wherein：

The equipment, including：Power cable, transformer, switch, busbar, capacitor and reactor etc.；The equipment list, Including：Number, device name and equipment endpoint number of equipment etc.；The port table, including：Port name, port numbering and Tie point number being connected with the port etc.；The node table, including：Nodename, node serial number and it is connected with the node Port numbering etc..

In the present embodiment, the monitoring data signal, including：The status information of each equipment and equipment alarm in substation Information.

Multiple equipment in the substation, including：Main transformer includes first set protection, second set of protection and observing and controlling dress It sets；Main transformer high side switch includes breaker, disconnecting link position, GIS device, spring mechanism of circuit breaker, mechanism abnormal signal, operation Case, control loop state and measure and control device；Main transformer low-pressure side switch includes breaker, disconnecting link position, GIS device, breaker bullet Spring mechanism, mechanism abnormal signal, control box, control loop state and measure and control device；Busbar includes that disconnecting link position, mother set；Circuit Including disconnecting link position, SF6 breakers, GIS device, SF6 mutual inductors, Hydraulic Mechanism of Circuit-breaker, pneumatic mechanism for circuit breaker, breaker Spring mechanism, mechanism abnormal signal, control loop state, control box, protective device, measure and control device, combining unit, intelligent terminal With intelligent cell etc..

In the present embodiment, the monitor event is low for line events, main transformer event, main transformer high side switch event, main transformer Press side switch events or busbar event etc.；

Wherein,

The line events, be disconnecting link position event, SF6 breakers event, GIS device event, SF6 mutual inductors event, Hydraulic Mechanism of Circuit-breaker event, pneumatic mechanism for circuit breaker event, spring mechanism of circuit breaker event, mechanism abnormal signal event, control Loop state event processed, control box event, protective device event, measure and control device event, combining unit event, intelligent terminal event Or intelligent cell event etc.；

The main transformer event protects event, second set of protection event or measure and control device event etc. for first set；

Become on high-tension side breaker event, disconnecting link position event, GIS device thing based on the main transformer high side switch event Part, spring mechanism of circuit breaker event, mechanism abnormal signal event, control box event, control loop state event or measure and control device Event etc.；

The main transformer low-pressure side switch events are breaker event, disconnecting link position event, the GIS device of main transformer low-pressure side Event, spring mechanism of circuit breaker event, mechanism abnormal signal event, control box event, control loop state event or observing and controlling dress Set event etc.；

The busbar event sets event etc. for disconnecting link position event or mother.

In the present embodiment, Bayesian neural network 20 is trained to obtain by using a large amount of history monitoring datas, Its input be history monitoring signal data include signal time x1, power station x2, signal name x3 and signal type x4 described in signal, It is exported as the monitor event y1 to ym belonging to monitoring signal.It is obtained by the training of a large amount of monitoring signal data to history Bayesian network just can be used for analysis and processing to acquiring monitoring signal in real time.

Bayesian neural network 20 realizes the reasoning and judging to monitoring signal according to Bayes decision rule, realizes to reality When signal dissection process.As shown in Fig. 2, Bayesian neural network is mainly by input layer, mode layer, summation layer and output layer four Layer composition.First by the time x1 of the monitoring signal obtained after monitor signals in real time standardization, affiliated substation x2, signal The input feature value of type x4 is sent into input layer after generating device name x3 and believing.The time of monitoring signal is the moon in year-- -when-cellular, it is 53 minutes at 3 points in afternoons of on May 4th, 2017 for the time that signal occurs, such as 2017-05-04-15-53； Affiliated substation is the power transformation station name generated where the equipment of the signal, and title is named according to grid equipment generic data model Specification is named input；The entitled device name that the signal occurs of signal generating apparatus, the title is according to electric system portion The device name number input that subset Unified number criterion carries out, if breaker is numbered with 4-digit number, front two represents electricity Grade is pressed, latter two represent junction style；Signal type is signal status information or warning information, according to alarm grade scale Carry out specification, such as switch changed position, exception.

Then feature vector carries out dissection process by input layer Dietary behavior layer, and mode layer passes through connection weight and input Layer connection, calculates the matching degree of input feature value and each monitor event in training set, i.e. similarity, its distance is sent into Gaussian function obtains the output of mode layer.In mode layer, the neuron number of mode layer is input training sample group number.Then Layer of summing is cumulative to get to the posterior probability values of monitor event by the mode layer output of the event of to one's name class, then exports Layer using with maximum a posteriori probability density monitor event as the output of signal resolution result, i.e., most by probability density function Big that neuron output is 1, i.e., that corresponding monitor event is the monitor event classification of sample to be identified, other The output of neuron is all 0.The monitor event that generation monitoring signal is belonged to is obtained by the parsing, and is sorted out to corresponding Monitor event in.In conjunction with the analysis of knowledge and topological structure of electric library in monitoring information Disposal Strategies library, can be had Impact analysis and Disposal Strategies caused by associated equipment situation occurs for the monitoring signal in body monitor event, signal occurs.

It is parsed by topological structure of electric library, the electric connecting relation and topology that can be obtained between each equipment of power grid are tied Structure relationship, so as to obtain the relevance of relevance and signal between each equipment.According to monitoring information Disposal Strategies library In concrete signal included signal paraphrase, signal reason, cause consequence and disposition principle, you can in conjunction with specific monitor event Monitoring signal and monitoring information Disposal Strategies library and topological structure of electric library, realize automatic sensing when warning information is sent out, Analysis and decision.

For status signal and the action of a large amount of switchgears and breaker apparatus of substation's power system monitor information monitoring Signal realizes the intelligently parsing and aid decision of signal using the network method.

The signal of Bayesian neural network input includes main transformer for a large amount of monitoring signals of breaker and isolated switchgear Breaker signal, main transformer disconnecting link signal, busbar disconnecting link signal or line-breaker signal etc..These switchgears of all monitorings Actuating signal be form input after taking standardization, specifically set for signal time, the affiliated substation of signal, signal Standby title and signalizing activity type.Since the range of signal of these above-mentioned equipment is wide, quantity is more, relationship is complicated, using artificial point Analysis processing is very time-consuming and laborious.Simultaneously because the effect of the electrical connection and logical connection between each equipment of power grid so that this There are certain relevances for actuating signal between a little switchgears, therefore when a certain switchgear is acted because of failure, Other switchgear linkage actions can also be led to because of chain reaction simultaneously, so as to cause occurring a large amount of prison on monitoring personnel screen Control signal, therefore sex chromosome mosaicism be associated in order to solve above-mentioned signal, by study to above-mentioned a large amount of historical signal and Training easily quickly can help monitoring personnel to solve the problems, such as this.

Below by carrying out illustrating the nerve net for study and signal resolution to topological structure of electric as shown in Figure 3 Network assists monitoring personnel into the course of work of row information parsing and decision：

Corresponding Bayesian neural network structure as shown in figure 4, the network by the training to training sample data then Processing parsing is carried out to the signal of actual needs parsing.Therefore its training process is with the increase of training samples, the god of mode layer It gradually increases, is also gradually increased with the neuron of the increase of analysis result classification its output layer, and training net through first number Network speed is fast, and without contents such as complex parameters adjustment, output accuracy reliability is high.Specially network input layer is that the power grid is opened up Flutter the switchgear actuating signal occurred in structure, it is contemplated that associated time of origin will not differ greatly between each signal, Training data is not very more simultaneously, therefore the input of its signal is the four-dimension, the first dimension data be signal time for according to when- Cellular inputs；Second dimension data of input is the affiliated substation of equipment；The third dimension data of input is signal generating apparatus name Claim, i.e. the number of equipment；Key equipment in Fig. 3 networks, i.e., three breaker numbers are 5021,5022 and 5023；Input Fourth dimension data are type of action when signal occurs for the equipment, including the opening and closing state signal of breaker and disconnecting switch Position signal.

Then mode layer calculates input layer input data and exports the matching degree of opening and closing state signal, that is, similar Degree.Its calculating process is similar between input signal and matched signal to describe using dist functions calculating Euclidean distance first Degree；Then nonlinear transformation is made to the Euclidean distance of calculating according to radbas Gaussian functions, followed by netprod quadrature networks Input function obtains the output result of mode layer to all each calculating nonlinear transformation result quadratures.

Then summation layer each unit is only connected with the mode unit of respective classes after training, layer each unit meter of summing Shown in the conditional probability such as formula (1) for calculating such analysis result：

C in formulaiFor analysis result classification, X is switchgear sample of signal to be resolved, XiFor the pattern sample of classification i, m is Input vector dimension, σ are smoothing parameter, and n is the pattern sample size of classification i.The wherein direct shadow of value size of smoothing parameter It rings and arrives final classifying quality, smoothing parameter choosing method chooses a suitable fixed value according to hands-on experience.

Output layer number of nodes is equal to the classification number (for 3) of analysis result, according to the estimation to all kinds of probability in above formula, uses Bayes classification rule, selection wherein " risk " is minimum, i.e., posterior probability is maximum parses as final as a result, its rule and method For shown in formula (2)：

P(X|Ci)P(Ci)>P(X|Cj)P(Cj) (2)

Wherein, i ≠ j.

Prior probability is denoted as P (C), meets formula (3)：

Wherein, n is the total sample number of training set, ncThe total sample number for being C for analysis result.It is such as public then to export analysis result Shown in formula (4)：

Y (X)=Ci (4)

Output layer realizes the screening of above-mentioned maximum a posteriori probability according to competitive function compet, to obtain final solution Analyse result.Wherein the node weights of each network layer of network and threshold value are completed during network training, wherein pattern The weights of layer are corresponding trained input signal values, and the selection of threshold value then depends on the selection of smoothing parameter, therefore threshold value can It is incorporated experience into when with according to hands-on and test to adjust.It can be seen that carrying out signal using the Bayesian neural network Parsing and identification avoid weight and threshold value and constantly update the trouble of adjustment, while can ensure that its precision and accuracy have again Very high confidence level makes other intelligence learning algorithms have prodigious advantage.

Assuming that certain substation has configuration of power network shown in Fig. 3, the monitor supervision platform of the substation is in 2017-05-04- 15-53 (year-month-day-when-point) get many monitoring alarm signals include North China it is stable/500KV.5062 switches, China Northern is stable/500KV.5061 switches, Tianjin, North China north suburb/500KV.5023 switches, Tianjin, North China north suburb/500KV.5022 switches, The separating brake of the switchgears such as North China Wu Zhuan/500KV.5013 switches acts alarm signal, normalized treated data such as table Shown in 1：

Table 1

Parse to obtain monitor event by Bayesian neural network be：The north suburbs Tianjin/500KV.5022/ separating brakes are always latched, That is, above-mentioned a series of alarm signal is always to be latched this event by the north suburbs Tianjin/500KV.5022/ separating brakes to cause 's.Then, we can be found signal paraphrase by Disposal Strategies library according to this event, the analysis of causes, cause consequence and disposition Strategy, as shown in table 2：

Table 2

By above-mentioned real case illustrate that this patent provides well to substation equipment monitoring data signal Intelligently parsing method can be good at realizing the intelligently parsing to supervisory control of substation information data, by this method to substation Mass data processing and analysis, realize aid decision and the analysis of warning information, greatly improve current monitor personnel to prison The present situation of information processing and management is controlled, monitoring management staff is reduced and largely repeats cumbersome work, provided for monitoring management The way to manage of high efficient and reliable.

Those skilled in the art should be able to recognize that, side described in conjunction with the examples disclosed in the embodiments of the present disclosure Method step and module, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate electronics The interchangeability of hardware and software generally describes each exemplary composition and step according to function in the above description Suddenly.These functions are executed with electronic hardware or software mode actually, depend on technical solution specific application and design about Beam condition.Those skilled in the art can use different methods to achieve the described function each specific application, but It is that such implementation should not be considered as beyond the scope of the present invention.

So far, it has been combined preferred embodiment shown in the drawings and describes technical scheme of the present invention, still, this field Technical staff is it is easily understood that protection scope of the present invention is expressly not limited to these specific implementation modes.Without departing from this Under the premise of the principle of invention, those skilled in the art can make the relevant technologies feature equivalent change or replacement, these Technical solution after change or replacement is fallen within protection scope of the present invention.

Claims (13)

1. a kind of intelligently parsing system of substation equipment monitoring data signal, which is characterized in that including：Standardization mould Block, data-signal parsing module and monitoring information Disposal Strategies library；
The standardization module, the conversion for carrying out format and content to monitoring data signal to be resolved, output mark Standardization treated monitoring data signal to be resolved, including：The affiliated substation of signal time, signal, signal generating apparatus title And signal type；
The data-signal parsing module is Bayesian neural network, for the monitoring to be resolved after the standardization Data-signal is parsed, and one or more monitor events corresponding to the monitoring data signal are obtained；
Monitoring information Disposal Strategies library, for the analysis result according to the Bayesian neural network, query monitor data Signal corresponding signal paraphrase, causes consequence and Disposal Strategies at the analysis of causes.
2. intelligently parsing system according to claim 1, which is characterized in that the Bayesian neural network training data History monitoring data signal after integrating as standardization.
3. intelligently parsing system according to claim 2, which is characterized in that the Bayesian neural network, including input Layer, mode layer, summation layer and output layer；
The input layer, is configured to：The monitoring data signal after the standardization is received, and passes to meshed network, it is defeated The neuron number for entering node layer network is equal with the dimension of monitoring data signal after the standardization；
The mode layer, is configured to：Input data feature vector is obtained from input layer, calculates input data feature vector and training The matching relationship for each monitor event concentrated；The neuron number of mode layer is the group number of training sample；
The summation layer, is configured to：Each monitor event only corresponds to a summation layer unit；Each summation layer unit is only right with it The pattern layer units for the one or more monitor events answered are connected, and by the output result phase of the pattern layer units connected Add, to obtain the estimated probability density function of the monitor event；
The output layer, is configured to：The probability density function for receiving each monitor event exported from summation layer, at normalization Reason, obtains the probability Estimation of each monitor event；Select a neuron with maximum a posteriori probability density as whole system Output.
4. intelligently parsing system according to claim 1, which is characterized in that monitoring information Disposal Strategies library, foundation History monitoring data signal and its corresponding signal paraphrase after standardization, cause consequence and Disposal Strategies at the analysis of causes Structure.
5. intelligently parsing system according to claim 1, which is characterized in that further include：Network topology structure library；
The network topology structure library, the relevance for inquiring relevance and signal between each equipment；The network Topological structure library is the double-layer network topological structure built according to the electric connecting relation of each equipment in substation, including：Power transformation Topological structure is inscribed in stand external topological structure and substation.
6. intelligently parsing system according to claim 1, which is characterized in that further include：It is grouped display module；
The grouping display module, for being packed into monitoring thing to the query result grouping in monitoring information Disposal Strategies library Part information list, and be sent in display equipment and show；And/or believed according to the monitoring data to be resolved after the standardization Number, the relevant information in the network topology structure library is inquired, and be sent in display equipment and show.
7. intelligently parsing system according to claim 5, which is characterized in that the network topology structure library is using layering The form of grouping is stored and is called by way of list.
8. intelligently parsing system according to claim 7, which is characterized in that the external topological structure of substation, including Between power transformation station name, substation's attribute, transformer substation voltage grade, connection type, external power transformation station name and external substation Interval type；
Wherein,
Substation's attribute, including intelligence and conventional substation；
The transformer substation voltage grade, including 500KV, 330KV, 220KV, 110KV and 66KV；
The connection type, including double female double separated times, single mother partition, 3/2 wiring and single mother；
Interval type between the external substation, including main transformer body, main transformer high-pressure side, main transformer medium voltage side, main transformer low-pressure side, Line segregation, capacitor standoff, reactor interval and busbar interval.
9. intelligently parsing system according to claim 7, which is characterized in that topological structure is inscribed in the substation, to become The connection topological form between each equipment inside power station.
10. intelligently parsing system according to claim 9, which is characterized in that topological structure, structure is inscribed in the substation Construction method is：
Endpoint to equipment all in substation and tie point Unified number；
The equipment list of description device attribute is set, describes the port table of port attribute and describes the node table of tie point attribute, is obtained Topological structure is inscribed to the substation；
Wherein,
The equipment, including：Power cable, transformer, switch, busbar, capacitor and reactor；
The equipment list, including：Number, device name and the equipment endpoint number of equipment；
The port table, including：Port name, port numbering and the tie point number being connected with the port；
The node table, including：Nodename, node serial number and the port numbering being connected with the node.
11. the intelligently parsing system according to any one of claim 1-10, which is characterized in that the monitoring data letter Number, including：The status information and equipment alarm information of multiple equipment in substation.
12. intelligently parsing system according to claim 11, which is characterized in that the multiple equipment in the substation, packet It includes：Main transformer includes first set protection, second set of protection and measure and control device；Main transformer high side switch includes breaker, disconnecting link position It sets, GIS device, spring mechanism of circuit breaker, mechanism abnormal signal, control box, control loop state and measure and control device；Main transformer low pressure Side switch includes breaker, disconnecting link position, GIS device, spring mechanism of circuit breaker, mechanism abnormal signal, control box, control loop State and measure and control device；Busbar includes that disconnecting link position, mother set；Circuit includes disconnecting link position, SF6 breakers, GIS device, SF6 mutual Sensor, Hydraulic Mechanism of Circuit-breaker, pneumatic mechanism for circuit breaker, spring mechanism of circuit breaker, mechanism abnormal signal, control loop state, Control box, protective device, measure and control device, combining unit, intelligent terminal and intelligent cell.
13. intelligently parsing system according to claim 12, which is characterized in that the monitor event is line events, master Change event, main transformer high side switch event, main transformer low-pressure side switch events or busbar event；
Wherein,
The line events are disconnecting link position event, SF6 breakers event, GIS device event, SF6 mutual inductors event, open circuit Device hydraulic mechanism event, spring mechanism of circuit breaker event, mechanism abnormal signal event, controls back pneumatic mechanism for circuit breaker event Line state event, control box event, protective device event, measure and control device event, combining unit event, intelligent terminal event or intelligence It can unit event；
The main transformer event protects event, second set of protection event or measure and control device event for first set；
Become based on the main transformer high side switch event on high-tension side breaker event, disconnecting link position event, GIS device event, Spring mechanism of circuit breaker event, mechanism abnormal signal event, control box event, control loop state event or measure and control device thing Part；
The main transformer low-pressure side switch events, be the breaker event of main transformer low-pressure side, disconnecting link position event, GIS device event, Spring mechanism of circuit breaker event, mechanism abnormal signal event, control box event, control loop state event or measure and control device thing Part；
The busbar event sets event for disconnecting link position event or mother.
CN201810034859.2A 2018-01-15 2018-01-15 The intelligently parsing system of substation equipment monitoring data signal CN108399579A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810034859.2A CN108399579A (en) 2018-01-15 2018-01-15 The intelligently parsing system of substation equipment monitoring data signal

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810034859.2A CN108399579A (en) 2018-01-15 2018-01-15 The intelligently parsing system of substation equipment monitoring data signal

Publications (1)

Publication Number Publication Date
CN108399579A true CN108399579A (en) 2018-08-14

Family

ID=63094853

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810034859.2A CN108399579A (en) 2018-01-15 2018-01-15 The intelligently parsing system of substation equipment monitoring data signal

Country Status (1)

Country Link
CN (1) CN108399579A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109687592A (en) * 2019-02-14 2019-04-26 国网山东省电力公司泰安供电公司 Substation information display control method, device and electronic equipment

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN102638100A (en) * 2012-04-05 2012-08-15 华北电力大学 District power network equipment abnormal alarm signal association analysis and diagnosis method
CN103425743A (en) * 2013-07-17 2013-12-04 上海金自天正信息技术有限公司 Steam pipe network prediction system based on Bayesian neural network algorithm
CN104657913A (en) * 2015-02-13 2015-05-27 广东电网有限责任公司电力科学研究院 Intelligent warning system based on all-data model

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
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
CN102638100A (en) * 2012-04-05 2012-08-15 华北电力大学 District power network equipment abnormal alarm signal association analysis and diagnosis method
CN103425743A (en) * 2013-07-17 2013-12-04 上海金自天正信息技术有限公司 Steam pipe network prediction system based on Bayesian neural network algorithm
CN104657913A (en) * 2015-02-13 2015-05-27 广东电网有限责任公司电力科学研究院 Intelligent warning system based on all-data model

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109687592A (en) * 2019-02-14 2019-04-26 国网山东省电力公司泰安供电公司 Substation information display control method, device and electronic equipment

Similar Documents

Publication Publication Date Title
Xie et al. Dimensionality reduction of synchrophasor data for early event detection: Linearized analysis
Ardakanian et al. On identification of distribution grids
CN103487749B (en) Mechanical state of high-voltage circuit breaker on-line monitoring and diagnosis system and method
Reddy et al. Smart fault location for smart grid operation using RTUs and computational intelligence techniques
Wen et al. Probabilistic approach for fault-section estimation in power systems based on a refined genetic algorithm
Chen et al. On-line fault diagnosis of distribution substations using hybrid cause-effect network and fuzzy rule-based method
Ebron et al. A neural network approach to the detection of incipient faults on power distribution feeders
CN106443310B (en) A kind of transformer fault detection method based on SOM neural network
Dalstein et al. Neural network approach to fault classification for high speed protective relaying
CN102934312B (en) Energy production system and control thereof
CN100470417C (en) Fault diagnostic system and method for under industrial producing process small sample condition
Thang et al. Analysis of power transformer dissolved gas data using the self-organizing map
Guo et al. Probabilistic framework for assessing the accuracy of data mining tool for online prediction of transient stability
US8588984B2 (en) Rotor angle stability prediction using post disturbance voltage trajectories
Tang et al. Condition monitoring and assessment of power transformers using computational intelligence
Rajapakse et al. Rotor angle instability prediction using post-disturbance voltage trajectories
CN101464964B (en) Pattern recognition method capable of holding vectorial machine for equipment fault diagnosis
Zhang et al. Fault localization in electrical power systems: A pattern recognition approach
CN103245881B (en) Power distribution network fault analyzing method and device based on tidal current distribution characteristics
Amraee et al. Transient instability prediction using decision tree technique
CN102361014B (en) State monitoring and fault diagnosis method for large-scale semiconductor manufacture process
Kezunovic A survey of neural net applications to protective relaying and fault analysis
CN105425118B (en) A kind of Multi-information acquisition fault arc detection method and device
CN105512474B (en) A kind of method for detecting abnormality of Transformer's Condition Monitoring data
CN103914064A (en) Industrial process fault diagnosis method based on multiple classifiers and D-S evidence fusion

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination