CN105241497A - Transformer monitoring system and fault diagnosis method - Google Patents
Transformer monitoring system and fault diagnosis method Download PDFInfo
- Publication number
- CN105241497A CN105241497A CN201510607872.9A CN201510607872A CN105241497A CN 105241497 A CN105241497 A CN 105241497A CN 201510607872 A CN201510607872 A CN 201510607872A CN 105241497 A CN105241497 A CN 105241497A
- Authority
- CN
- China
- Prior art keywords
- transformer
- oil
- fault
- monitoring
- monitoring unit
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
Abstract
The invention proposes a transformer monitoring system and a fault diagnosis method. The system comprises an oil gas color pattern on-line monitoring unit, an oil micro-moisture monitoring unit, a temperature monitoring unit, a partial discharge monitoring unit, a grounding current on-line monitoring unit, a transformer, a load voltage adjustment switch vibration monitoring unit, and a main station computer. The system and method can employ a plurality of monitoring units and methods to monitor and diagnose a fault of the transformer, and improve the accuracy of monitoring.
Description
Technical field
The present invention relates to power system transformer field, be specifically related to a kind of transformer monitoring systems and method for diagnosing faults.
Background technology
Along with transferring electricity from the west to the east, north and south supplies mutually, on national network propelling, and the scale of China's electrical network and the scale of electric power transfer are all in continuous expansion, and power grid security problem highlights further.U.S.A adds " 8.14 " has a power failure on a large scale, West Europe " 11.4 " influential power outage in the world of having a power failure on a large scale etc. in recent years all causes tremendous economic and loses and create severe social influence.According to China the Eleventh Five-Year Plan period Power network fault statistics show, electrical equipment fault accounts for power grid accident cause first.Therefore, the safe operation of power equipment is the first system of defense avoiding electrical network major accident, and power transformer is the key equipment in this first system of defense.
Status monitoring is the technological means obtaining equipment running status, is the important information source of maintenance and O&M.It adopts informationization technology, the power transmission and transforming equipment of transformer station is installed the monitoring device of wireless sensor networks, many reference amounts monitoring is carried out to equipment, real-time data collection, multiple information synthesis analysis, realize the intellectuality of power transmission and transforming equipment, informationization, reach the object of equipment running status oneself perception, automatic fault diagnosis.Condition Monitoring Technology is for based on the maintenance of state or a kind of technology of foreseeability maintenance service, and its development comes from the technical demand of repair based on condition of component for electrical network equipment state acquisition of information, analysis, judge.Therefore, the Condition Monitoring Technology of at present development all for a certain class equipment, a certain embody rule and developing, mainly concentrates in the concrete equipment state overhauling application of the aspects such as generating set, converting equipment, transmission line of electricity.Owing to lacking the unified platform, all kinds of monitoring device is done things in his own way, isolated operation, and resource can not be shared.Online monitoring data cannot organically combine with other important state amounts simultaneously, fails effectively to play its auxiliary diagnosis effect.
Summary of the invention
Problem existing at least part of solution prior art, the present invention proposes a kind of transformer monitoring systems, comprise: oil dissolved gas chromatogram monitoring unit, micro-water monitoring unit in oil, temperature monitoring unit, partial discharge monitoring unit, ground current on-line monitoring unit, transformer and on-load voltage regulating switch vibration monitoring unit, master station computer, wherein
Oil dissolved gas chromatogram monitoring unit, utilize the content of gas chromatogram monitoring sensor to hydrogen contained in transformer oil, carbon monoxide, carbon dioxide, acetylene, ethene, methane, ethane to calculate, and the result of calculating is transferred to master station computer by RS232 communication protocol;
Micro-water monitoring unit in oil, utilizes moisture transducer to calculate contained humidity content in transformer oil, and the result of calculating is transferred to master station computer;
Temperature monitoring unit, adopt optic fiber thermometer to transformer top layer and bottom oil temperature, and case temperature carries out on-line checkingi, and testing result is transferred to master station computer;
Partial discharge monitoring unit, adopts external high frequency antenna as type UHF sensor, the ultrahigh frequency voltage signal that monitoring inside transformer shelf depreciation produces, and monitoring result is transferred to master station computer;
Ground current on-line monitoring unit, reflects iron core situation by electric current size variation on monitoring iron core grounding line, and monitoring result is transferred to master station computer;
Transformer and on-load voltage regulating switch vibration monitoring unit, utilize piezoelectric acceleration vibration transducer to monitor the Vibration Condition of transformer and on-load voltage regulating switch, and monitoring result be transferred to master station computer.
Wherein,
Superheated steam drier, transformer overload, short trouble, the discharge fault of oil dissolved gas chromatogram monitoring unit monitoring transformer;
The Superheated steam drier of temperature monitoring unit monitoring transformer.
The discharging fault of partial discharge monitoring unit monitoring transformer;
Ground current on-line monitoring unit monitoring multipoint earthing of iron core fault;
Transformer and on-load voltage regulating switch vibration monitoring unit monitoring load ratio bridging switch mechanical faults and iron core and winding mechanical fault.
Wherein, the monitoring result that master station computer reception oil dissolved gas chromatogram monitoring unit and shelf depreciation monitoring means transmit, the acetylene content monitored when oil dissolved gas chromatogram monitoring unit accounts for 20% to 70% of total hydrocarbon ratio, hydrogen content accounts for 30% to 90% of hydrocarbon total amount simultaneously, and, the content of carbon monoxide, carbon dioxide, ethene, methane, ethane is in predetermined normal range, and partial discharge monitoring unit display partial discharges fault, at this moment master station computer determination transformer generation shelf depreciation fault
Wherein, master station computer receives the monitoring result that in oil dissolved gas chromatogram monitoring unit and oil, micro-water monitoring unit transmits, the hydrogen content monitored when oil dissolved gas chromatogram monitoring unit exceeds preset range, the content of carbon monoxide, carbon dioxide, acetylene, ethene, methane, ethane is in predetermined normal range simultaneously, and, moisture in oil in micro-water monitoring unit display oil exceeds preset range, Moisture high UCL at this moment master station computer determination transformer oil.
Wherein, the monitoring result that master station computer reception oil dissolved gas chromatogram monitoring unit and temperature monitoring unit transmit, both the methane monitored when oil dissolved gas chromatogram monitoring unit and acetylene total amount exceeds 80% of total hydrocarbon, the temperature of each monitoring point of inside transformer of at this moment master station computer analysis temperature monitoring means monitoring, determines whether Superheated steam drier occurs.
Wherein, master station computer receives the monitoring result that temperature monitoring unit and transformer and on-load voltage regulating switch vibration monitoring unit transmit, if the monitoring result display local overheating fault of temperature monitoring unit and vibration monitoring unit display winding generation vibration deformation, then master station computer determination winding deforms fault.
Wherein, oil dissolved gas chromatogram monitoring unit is connected with master station computer by RS232 serial ports, the operating system of described oil dissolved gas chromatogram monitoring unit is built-in Linux, described oil dissolved gas chromatogram monitoring unit passes through serial interface management module management to the operation of RS232 serial ports, comprise serial ports is set attribute, serial ports open the reading with serial ports, the serial ports of described serial interface management module reads implementation procedure and comprises:
Add the header file used in the serial ports fetch program;
Define a new int type process ID number;
Write newly-built subprocess function, read the Monitoring Data in serial ports respective file and output in newly-built data.txt file;
Slept 1 second by programming parent process, and carry out terminator process by the kill function of system function;
The data.txt file storing Monitoring Data is opened with read-only mode;
Read the first row data from data.txt file, and stored in buffer memory buf;
Judge whether the data that a certain moment reads start with "--", if not then again read next line, until read complete data by the complete adaptation function of character string;
Concrete numerical value in the data line mated completely is extracted air-monitor storage of array;
Realize opening a new files jc.txt file with read-write mode;
The numerical value write jc..txt file stored in air-monitor array is supplied other equipment calls.
The present invention also proposes a kind of Diagnosis Method of Transformer Faults, uses described transformer monitoring systems to carry out fault diagnosis to transformer, comprising:
By described oil dissolved gas chromatogram monitoring unit, the gas in transformer oil is sampled, and by the absolute gas production rate of following formulae discovery:
γ
αrepresent the absolute gas production rate of certain gas, C
t, 2represent that second time sampling records certain gas concentration in oil, C
t, 1represent that first time sampling records certain gas concentration in oil, Δ t represents the actual run time in twice sample interval, the gross mass of m indication transformer oil, and ρ represents the density of oil;
When the content of total hydrocarbon is greater than first threshold and is less than 3 times of first threshold, and when the absolute gas production rate of total hydrocarbon is less than Second Threshold, determine that transformer exists fault but can continue to run; When 1 to 2 times that absolute gas production rate is Second Threshold, at this moment determine the round of visits that will shorten transformer;
When the content of total hydrocarbon is greater than 3 times of first threshold and the absolute gas production rate of total hydrocarbon is greater than 3 times of Second Threshold, at this moment illustrate that transformer has serious fault, notifies that staff takes corresponding measure immediately;
Wherein, total hydrocarbon refers to hydrocarbon gas all in transformer oil; The absolute gas production rate of total hydrocarbon refers to the absolute gas production rate sum of hydrocarbon gas all in transformer oil.
By the relative gas production rate of following formulae discovery:
γ
rrepresent the relative gas production rate of certain gas, C
t, 2represent that second time sampling records certain gas concentration in oil, C
t, 1represent that first time sampling records certain gas concentration in oil, Δ t represents the actual run time in twice sample interval;
When the relative gas production rate of total hydrocarbon is greater than 10%, determine to shorten the sense cycle to transformer.
Described Diagnosis Method of Transformer Faults, supporting vector machine model is adopted to diagnose transformer fault, choose hydrogen, acetylene, ethene, methane, ethane five kinds of failure gas as characteristic gas, the matrix formed is as the input vector of supporting vector machine model, the column vector of input matrix is above five kinds of characteristic gas, and the dimension of column vector is 5; The row vector of input matrix is the raw data collected, and the dimension of row vector is the number of raw data.
Wherein, the training process of described supporting vector machine model comprises:
First identical method is adopted to be normalized training set and test set, using training set as the training sample of support vector machine, Training Support Vector Machines is carried out by constantly optimizing kernel functional parameter, if the accuracy of fault diagnosis result does not reach requirement, then need to reselect the parameter area of kernel function, until the accuracy of diagnostic result reaches requirement, now be met the supporting vector machine model of requirement, finally verify that whether the support vector machine of training is correct to the diagnostic result of fault with test set.
Wherein, adopt described supporting vector machine model to carry out transformer fault diagnosis specifically to comprise:
(1) electric power transformer oil all 600 with clear failure conclusion is obtained, oil sample sample is divided into training set and test set, and oil sample sample is classified according to fault type, according to fault type, oil sample sample is divided into six classes, represent with " 1 ", " 2 ", " 3 ", " 4 ", " 5 ", " 0 " respectively, wherein, in " 1 " expression, cryogenic overheating, " 2 " expression hyperthermia and superheating, " 3 " represent that low-yield electric discharge, " 4 " expression high-energy discharge, " 5 " represent shelf depreciation, " 0 " expression normal condition; Wherein training set comprises 400 samples, and test set comprises 200 samples;
(2) oil sample sample is converted into the matrix of 600X5, and adopts identical method, respectively training set and test set are normalized;
(3) suitable kernel function is selected, first input larger data search model and adopt grid data service Selection parameter penalty factor c and kernel function δ roughly, then on the basis of rough search, reasonably reduce data search scope, utilize grid data service accurately to select optimal parameter c and δ;
(4) utilize training set sample training supporting vector machine model, and whether reach requirement with test set sample predictions diagnostic result, if not, then return the parameter area with reselecting kernel function to step (3);
(5) diagnostic result is obtained by needing the transformer oil sample data of diagnosis to substitute in described supporting vector machine model.
Wherein, the span utilizing grid data service to arrange penalty factor c is [2
-10, 2
10], stepping is 0.4; The span of kernel functional parameter δ is [2
-10, 2
10], stepping is 0.4, and by training support vector machine, the best value of penalty factor c is 0.83282, and the best value of kernel functional parameter δ is 0.39227, and the accuracy rate of support vector machine classifier Selection parameter is 77.5536%.
Wherein, the span utilizing grid data service to arrange penalty factor c is [2
-10, 2
0], stepping 0.2; The span of kernel functional parameter δ is [2
-10, 2
0], stepping 0.2, through training support vector machine, the best value of penalty factor c is 0.40421, and kernel functional parameter δ is best, and value is 1.00231, and the accuracy rate of support vector machine classifier Selection parameter is 93.1196%.
Wherein, utilize grid data service that the span [2 of penalty factor c is set
0, 2
10], stepping 0.2; The span of kernel functional parameter δ is [2
0, 2
10], stepping 0.2, through Training Support Vector Machines, the best value of penalty factor c is 1.2986, and kernel functional parameter δ is best, and value is 1.4093, and the accuracy rate of support vector machine classifier Selection parameter is 96.088%.
Wherein, the span utilizing grid data service to arrange penalty factor c is [2
0, 2
10], stepping 0.2, the span of kernel functional parameter δ is [2
-10, 2
0], stepping 0.2.By Training Support Vector Machines, the best value of penalty factor c is 23.2312, and kernel functional parameter δ is best, and value is 0.025102, and the rate of accuracy reached of support vector machine classifier Selection parameter is to 96.6598%.
Wherein, support vector machine adopts the support vector machine based on particle group optimizing, and the modeling process based on the support vector machine of particle group optimizing is:
(1) initialization population, is optimized the kernel function δ of particle swarm support vector machine and penalty factor c by the method for adjustment population inertia weight ω, makes parameter c and δ form particulate, i.e. (c, a δ), and set maximal rate as V
max, the initial position representing each particulate with pbest, represents fine-grained best initial position in population with gbest;
(2) evaluate the fitness of each particulate, calculate the optimal location of each particulate;
(3) adaptive value of each particulate after optimizing and its history optimal location pbest are compared, if current adaptive value is better than optimal location, then using adaptive value as the current desired positions pbest of particle;
(4) adaptive value of each particulate and the history optimal location gbest of colony's particulate after optimization are compared, if adaptive value is better than the history optimal location gbest of colony's particulate, then using the optimal location gbest of adaptive value as colony's particulate;
(5) speed and the position of current particulate is adjusted according to modified particle swarm optiziation;
(6) when adaptive value satisfies condition, iteration terminates, otherwise returns second step continuation Optimal Parameters, after the 6th step completes, will the parameter c of optimization the best and δ, so just can obtain optimal supporting vector machine model, carry out failure prediction with this model.
Wherein, if Population Size N=20, inertia weight ω=0.9, aceleration pulse C
1=1.4, C
2=1.6, Training Support Vector Machines, the best value obtaining penalty factor c be 3.8326, kernel function δ best value be 0.50433.Through particle cluster algorithm optimization, the classification accuracy of SVM classifier reaches 98.9234%.
Wherein, support vector machine adopts the support vector machine based on genetic algorithm, and the modeling process based on the support vector machine of genetic algorithm is:
(1) initialization population, generate the individuality of some as initial population, every bar chromosome is made up of (c, δ), and wherein c is penalty factor, and δ is kernel function;
(2) selected target function carries out support vector machine training to initial population, using the square error of support vector machine as objective function, calculates the fitness of each individuality;
(3) carry out Selecting operation, crossing operation, mutation operator obtain population of new generation, support vector machine training is carried out to the new population produced;
(4) if the new population produced meets termination rules, then output has the individuality of maximum adaptation degree as optimized parameter, predicts, otherwise increase evolutionary generation with optimized parameter, proceeds to step (3) and continues to perform.
Wherein, use the SVM prediction coiling hot point of transformer temperature based on particle cluster algorithm, gather the hot spot temperature of winding (θ of transformer
hst), top-oil temperature (θ
tot), active loss (P
l), reactive loss (R
l), load current (I) and environment temperature (T
amb), humidity (H), wind speed (v
w), as the inputoutput data set of the supporting vector machine model based on particle cluster algorithm, by the top-oil temperature (θ of transformer
tot), active loss (P
l), reactive loss (R
l), load current (I) and environment temperature (T
amb), humidity (H), wind speed (v
w) as the input of the supporting vector machine model based on particle cluster algorithm, the hot spot temperature of winding (θ of transformer
hst) as exporting.
, wherein, based on particle cluster algorithm in supporting vector machine model parameter optimisation procedure, choosing population quantity is 20, and the restriction factor of change in displacement is taken as c
1=c
2=2, it is 0.9 carry out stopping by for linear decrease to 0.4 that inertia weight factor ω gets initial value, stopping algebraically is 200, choose the fitness function of square error (RMSE) as particle cluster algorithm, optimum SVM parameter value after particle cluster algorithm optimization is, penalty factor c=17.0021, kernel function δ=0.5962, fitness function value is: RMSE=0.0039823.
A kind of transformer monitoring systems that the present invention proposes and method, can adopt multiple monitoring means and monitoring method to carry out monitoring, diagnosing to transformer fault, improve the accuracy of monitoring.
Accompanying drawing explanation
Fig. 1 is the schematic block diagram of the transformer monitoring systems that the present invention proposes;
Embodiment
Below in conjunction with accompanying drawing of the present invention, technical scheme of the present invention is clearly and completely described.Here will be described exemplary embodiment in detail, its sample table shows in the accompanying drawings.When description below relates to accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawing represents same or analogous key element.Embodiment described in following exemplary embodiment does not represent all embodiments consistent with the present invention.On the contrary, they only with as in appended claims describe in detail, the example of apparatus and method that aspects more of the present invention are consistent.
Power transformer is the important electric power primary equipment realizing delivery of electrical energy and electric pressure conversion according to energy of electromagnetic field transfer principle.According to the difference of kind and capacity, the structure of transformer also has certain difference, and the application carries out analysis discussion for oil-immersed and high-voltage transformer.Oil-filled transformer chief component has: other structures such as transformer body, regulator, fuel tank and cooling device, protective device and insulating sleeve.
(1) transformer body device
Iron core, winding, lead-in wire and certain insulating element constitute the device body part of transformer.Wherein iron core forms closed magnetic loop, and transformer produces alternating magnetic field by the winding be wrapped on primary side iron core, by the flux path closed, magnetic field energy is delivered to secondary side.At secondary side winding output terminal, induce corresponding electromotive force according to winding coil apart from number situation, complete electrical energy transfer and voltage transformation.
(2) regulator
Shunting switch, as the voltage adjusting device of transformer, by changing the method for the active coil resistance number of transformer, can realize the adjustment to transformer running parameter (capacity, voltage, electric current).According to during action whether with load, shunting switch can be divided into load ratio bridging switch and no-load tapping switch two kinds.
(3) fuel tank and cooling device
Fuel tank is the shell of transformer, and the iron core and winding that are soaked by transformer oil are equipped with in inside, and transformer oil mainly plays insulation and thermolysis.High-power transformer is equipped with two fuel tanks usually, and one is body fuel tank, inner installation iron core and winding; Another one is on-load voltage regulation fuel tank, and shunting switch is equipped with in inside.This is because may produce arc spark in on-load voltage regulation process, the quality of deteriorated transformer oil, so separate itself and body apparatus.According to the difference of capacity, fuel tank can be divided into box fuel tank, bell type tank and enclosure-type fuel tank three kinds of primary structures.According to whether installing cooling fan and other auxiliary heat dissipation mechanisms, Cooling Methods of Transformers can be divided into: oil mode (ONAN); Oil mode (ONAF); Forced oil-circulation air cooling way (OFAF); Forced oil circulation water-cooling mode (OFWF); Forced guidance oil circulating air cooling mode (ODAF); Forced guidance oil circulation water-cooling ODWF) mode.
(4) protective device
Protective device comprises oil conservater, pressure-relief vent, dehydrating breather, gas relay, oil purifier, and the sensor of other types and secondary device.
(5) insulating sleeve
Insulating sleeve is the primary insulation device of oil-filled transformer, and transformer extension line is drawn out to fuel tank outside through insulating sleeve.Insulating sleeve possesses enough electrical strengths and good thermal stability, makes extension line over the ground and casing insulation, plays the effect of anchor leg simultaneously.Insulating sleeve inside and transformer body fuel tank communicate, and inside is full of transformer oil.
According to the difference of criteria for classification, transformer fault can be divided into many types.Can be divided into according to abort situation: external fault and internal fault.Wherein internal fault is from being divided into electric fault and hot stall again in nature, and hot stall can show as the inner local overheating of oil tank of transformer usually, temperature obviously raises.Different according to hot source point, hot stall can be divided into four kinds of situations: when hot(test)-spot temperature is lower than 150 DEG C, for slightly overheated; When hot(test)-spot temperature is between 150-300 DEG C, it is cryogenic overheating; When hot(test)-spot temperature is between 300-700 DEG C, for middle temperature is overheated; Focus higher than the situation of 700 DEG C, then belongs to hyperthermia and superheating.Electric fault is often referred to inside transformer under the effect of high electric field intensity, and the fault of deterioration occurs the insulating property of insulating material.Different according to the energy density of electric discharge, electric fault is divided into again partial discharges fault, spark discharge fault and high energy arc discharge fault three types.Be divided into according to fault loop: fault, magnetic-circuit fault and oil path failure.Difference according to the agent structure of fault can be divided into: winding failure, iron core fault, annex fault and oily fault.Other faults also comprise cutting-out of voltage changer fault, Leakage in Transformer fault, oil flow electrification fault, false protection fault.And the most serious on transformer impact itself, at present occurrence probability the highest be cutting-out of voltage changer fault, also there is Leakage in Transformer fault, oil flow electrification fault, false protection fault etc. simultaneously.All these dissimilar faults, have what may reflect is hot stall, have what may reflect is electric fault, what have not only may reflect that overheating fault was simultaneously but also there is discharge fault, and Leakage in Transformer fault may not exist the feature of heat or electric fault in the ordinary course of things.
According to related statistics display, in transformer fault type, Superheated steam drier accounts for 62%; High-energy discharge fault accounts for 17%; Overheated double high-energy discharge fault accounts for 11%; Spark discharge fault accounts for 8%; To make moist or partial discharges fault accounts for 2%.In transformer fault position, on-load voltage regulation fault accounts for 40%, and winding coil fault accounts for 30%, and sleeve pipe fault accounts for 14%, and fuel tank and iron core fault account for 6% and 5% respectively, and other auxiliary faults account for 5%.According in above statistics, the situation of different faults type ratio and major failure type, herein intelligent on-line monitoring system is built while ensureing gas-monitoring function by institute, for reducing the scope of tracing trouble, temperature monitoring system is installed to monitor Superheated steam drier, vibration monitor system is installed to monitor on-load voltage regulating switch (OLTC) and iron core winding mechanical faults, ground current on-line monitoring system is installed to monitor multipoint earthing of iron core fault, partial discharge monitoring system is installed to determine partial discharges fault.
See Fig. 1, a kind of transformer monitoring systems that the present invention proposes, comprise oil dissolved gas chromatogram monitoring unit, micro-water monitoring unit in oil, temperature monitoring unit, partial discharge monitoring unit, ground current on-line monitoring unit, transformer and on-load voltage regulating switch vibration monitoring unit, master station computer, wherein
Oil dissolved gas chromatogram monitoring unit, utilize the content of gas chromatogram monitoring sensor to hydrogen contained in transformer oil, carbon monoxide, carbon dioxide, acetylene, ethene, methane, ethane to calculate, and the result of calculating is transferred to master station computer by RS232 communication protocol;
Micro-water monitoring unit in oil, utilizes moisture transducer to calculate contained humidity content in transformer oil, and the result of calculating is transferred to master station computer;
Temperature monitoring unit, adopt optic fiber thermometer to transformer top layer and bottom oil temperature, and case temperature carries out on-line checkingi, and testing result is transferred to master station computer;
Partial discharge monitoring unit, adopts external high frequency antenna as type UHF sensor, the ultrahigh frequency voltage signal that monitoring inside transformer shelf depreciation produces, and monitoring result is transferred to master station computer;
Ground current on-line monitoring unit, reflects iron core situation by electric current size variation on monitoring iron core grounding line, and monitoring result is transferred to master station computer;
Transformer and on-load voltage regulating switch vibration monitoring unit, utilize piezoelectric acceleration vibration transducer to monitor the Vibration Condition of transformer and on-load voltage regulating switch, and monitoring result be transferred to master station computer.
Superheated steam drier, transformer overload, short trouble, the discharge fault of oil dissolved gas chromatogram monitoring unit monitoring transformer.
Superheated steam drier: under inside transformer overheat condition, a large amount of heats can make transformer oil and the rotten oxidation of fraction solids insulating material, discharges gas.Main initiation reason comprises that transformer overload causes overheating fault, transformer sudden short circuit causes Superheated steam drier etc.Transformer overload: winding current overrate during transformer overload, when radiating efficiency deficiency, winding high temperature can cause insulating oil to decompose generation gas.Short trouble: during transformer sudden short circuit, its winding may bear the electric current of decades of times and ratings, produces very large heat, causes the impaired release gas of insulating material.Discharge fault: during inside transformer generation discharge fault, insulating oil or local solid insulation can be subject to rotten candle, and dielectric loss increases, and discharges corresponding failure gas.
The Superheated steam drier of temperature monitoring unit monitoring transformer.
The discharging fault of partial discharge monitoring unit monitoring transformer.
Discharging fault: comprise the various discharge faults such as shelf depreciation, spark discharge, arc discharge.
Ground current on-line monitoring unit monitoring multipoint earthing of iron core fault.
Transformer and on-load voltage regulating switch vibration monitoring unit monitoring load ratio bridging switch mechanical faults and iron core and winding mechanical fault.
Load ratio bridging switch mechanical faults: vibration monitoring unit according to the Vibration Condition detected by loaded switch course of action, can carry out spectrum analysis, judge according to analysis result to loaded switch duty, then monitoring equipment fault.
Iron core and winding mechanical fault: when iron core or winding generation mechanical faults, when such as compressing problem, displacement fault, winding deformation, the mechanical vibration of specific frequency spectrum can occur transformer, and vibration monitoring unit can be monitored the type fault.
The fault type that each monitoring means is diagnosed is all not identical, and level of abstraction is higher, and the monitoring result of an independent unit cannot be utilized to carry out accurate in detail judgement and location to fault.But simultaneously, often lap is had again to the diagnosis of fault between different monitoring means, the conformability of different units can be utilized to monitor the fault coverage reducing diagnosis, such as when oil dissolved gas chromatogram monitoring system discovery Superheated steam drier, can analyze in conjunction with the Monitoring Data of ground current monitoring system, get rid of or be confirmed whether that there occurs multipoint earthing of iron core fault causes insulation overheat, thus producing corresponding gas.In this way, intelligent monitor system can reduce the scope of fault further, directly can determine fault type in some cases.
The monitoring result that master station computer reception oil dissolved gas chromatogram monitoring unit and shelf depreciation monitoring means transmit, the acetylene content monitored when oil dissolved gas chromatogram monitoring unit accounts for 20% to 70% of total hydrocarbon ratio, hydrogen content accounts for 30% to 90% of hydrocarbon total amount simultaneously, and, the content of carbon monoxide, carbon dioxide, ethene, methane, ethane is in predetermined normal range, and partial discharge monitoring unit display partial discharges fault, at this moment master station computer determination transformer generation shelf depreciation fault.
Acetylene produces relevant with discharging fault, and when inside transformer generation arc discharge, acetylene content generally accounts for 20% to 70% of total hydrocarbon ratio, and hydrogen accounts for 30 to 90% of hydrocarbon total amount simultaneously.If other compositions do not exceed standard, and acetylene growth rate is very fast, and become oil dissolved gas principal ingredient and content overproof time, be then likely that equipment short circuit in winding or shunting switch switching etc. produce discharging fault.Now can the office of cooperation put monitoring means and carry out determinacy diagnosis to whether there occurs partial discharges fault.By monitoring partial discharges fault in conjunction with two monitoring means, monitoring result is more reliable.
Master station computer receives the monitoring result that in oil dissolved gas chromatogram monitoring unit and oil, micro-water monitoring unit transmits, the hydrogen content monitored when oil dissolved gas chromatogram monitoring unit exceeds preset range, the content of carbon monoxide, carbon dioxide, acetylene, ethene, methane, ethane is in predetermined normal range simultaneously, and, moisture in oil in micro-water monitoring unit display oil exceeds preset range, Moisture high UCL at this moment master station computer determination transformer oil.
Inside transformer watered and wetting is a kind of Hidden fault, the corresponding increase of its characteristic gas hydrogen content, if oil dissolved gas chromatogram monitoring unit display hydrogen content exceeds standard, and other gas contents are without when significantly increasing, and roughly can predict in transformer and may contain moisture.Whether now for determining fault further or getting rid of, micro-water monitoring system can be coordinated to export data analysis, confirming that hydrogen content exceeds standard is caused by oil, micro-water content exceeds standard.By monitoring Water in oil situations in conjunction with two monitoring means, monitoring result is more reliable.
The monitoring result that master station computer reception oil dissolved gas chromatogram monitoring unit and temperature monitoring unit transmit, both the methane monitored when oil dissolved gas chromatogram monitoring unit and acetylene total amount exceeds 80% of total hydrocarbon, the temperature of each monitoring point of inside transformer of at this moment master station computer analysis temperature monitoring means monitoring, determines whether Superheated steam drier occurs.
When inside transformer generation Superheated steam drier, the insulating oil decomposition of thermal source place can produce a large amount of methane and ethene, and the two total amount generally can account for more than 80% of total hydrocarbon, and along with the rising of trouble spot temperature, its proportion also increases thereupon.Now can coordinate temperature monitoring unit, selective analysis is carried out to the temperature that inside transformer different monitoring points gathers, be whether hotly carry out determinacy diagnosis to fault type, and determine thermal source trouble spot, and then made suitable solution.
Master station computer receives the monitoring result that temperature monitoring unit and transformer and on-load voltage regulating switch vibration monitoring unit transmit, if the monitoring result display local overheating fault of temperature monitoring unit and vibration monitoring unit display winding generation vibration deformation, then master station computer determination winding deforms fault.
Transformer Winding tap region and Jiu Jie district install imbalance and neighbouring winding wire cake can be caused to be out of shape, the horizontal stray field produced makes winding when cutting-out of voltage changer, the electric power that line cake is subject to will increase a lot, causes winding produce distortion and damage, particularly tap region line cake.When load ratio bridging switch generation tap section short trouble, winding can be deformed into wavy, affects the unobstructed of oil duct, reduces insulation system heat dispersion.Vibration monitor system now can be coordinated to diagnose, when there is local overheating fault in transformer, if vibration monitor system exports data display may there is vibration deformation in winding, then can estimate winding and to deform fault, and carry out further confirmatory work and maintenance.
Oil dissolved gas chromatogram monitoring unit of the present invention is connected with master station computer by RS232 serial ports,
Wherein, the operating system of described oil dissolved gas chromatogram monitoring unit is built-in Linux, described oil dissolved gas chromatogram monitoring unit passes through serial interface management module management to the operation of RS232 serial ports, comprise serial ports is set attribute, serial ports open the reading with serial ports, the serial ports of described serial interface management module reads implementation procedure and comprises:
Add the header file used in the serial ports fetch program;
Define a new int type process ID number;
Write newly-built subprocess function, read the Monitoring Data in serial ports respective file and output in newly-built data.txt file;
Slept 1 second by programming parent process, and carry out terminator process by the kill function of system function;
The data.txt file storing Monitoring Data is opened with read-only mode;
Read the first row data from data.txt file, and stored in buffer memory buf;
Judge whether the data that a certain moment reads start with "--", if not then again read next line, until read complete data by the complete adaptation function of character string;
Concrete numerical value in the data line mated completely is extracted air-monitor storage of array;
Realize opening a new files jc.txt file with read-write mode;
The numerical value write jc..txt file stored in air-monitor array is supplied other equipment calls.
The present invention will according to transformer fault qualitative classification situation (Superheated steam drier, discharging fault, mechanical faults and other faults), to how on the basis of wide in range property tracing trouble character, the method for refinement fault type and trouble location proposes solution.The fault type can monitor selected different monitoring subsystems or trouble location are determined; And different faults monitoring system is carried out many-sided system-level combination, set up malfunction monitoring overlay region, construct the fault monitoring system of refinement from system cooperation and connected mode.
The invention allows for a kind of Diagnosis Method of Transformer Faults, by oil dissolved gas chromatogram monitoring unit, the gas in transformer oil is sampled, and by the absolute gas production rate of following formulae discovery:
γ
αrepresent the absolute gas production rate of certain gas, C
t, 2represent that second time sampling records certain gas concentration in oil, C
t, 1represent that first time sampling records certain gas concentration in oil, Δ t represents the actual run time in twice sample interval, the gross mass of m indication transformer oil, and ρ represents the density of oil;
When the content of total hydrocarbon is greater than first threshold and is less than 3 times of first threshold, and when the absolute gas production rate of total hydrocarbon is less than Second Threshold, determine that transformer exists fault but can continue to run; When 1 to 2 times that absolute gas production rate is Second Threshold, at this moment determine the round of visits that will shorten transformer;
When the content of total hydrocarbon is greater than 3 times of first threshold and the absolute gas production rate of total hydrocarbon is greater than 3 times of Second Threshold, at this moment illustrate that transformer has serious fault, notifies that staff takes corresponding measure immediately;
Wherein, total hydrocarbon refers to hydrocarbon gas all in transformer oil; The absolute gas production rate of total hydrocarbon refers to the absolute gas production rate sum of hydrocarbon gas all in transformer oil.
By the relative gas production rate of following formulae discovery:
γ
rrepresent the relative gas production rate of certain gas, C
t, 2represent that second time sampling records certain gas concentration in oil, C
t, 1represent that first time sampling records certain gas concentration in oil, Δ t represents the actual run time in twice sample interval;
When the relative gas production rate of total hydrocarbon is greater than 10%, determine to shorten the sense cycle to transformer.
The invention allows for a kind of Diagnosis Method of Transformer Faults, adopt support vector machine to diagnose transformer fault.
The present invention chooses hydrogen, acetylene, ethene, methane, ethane five kinds of failure gas as characteristic gas, and the matrix of formation is as the input vector of support vector machine.The column vector of input matrix is above five kinds of characteristic gas, and the dimension of column vector is 5; The row vector of input matrix is the 600 groups of raw data collected, and the dimension of row vector is 600.Therefore, input vector is the matrix of 600X5.
The supporting vector machine model setting up transformer fault diagnosis is crucial and difficult point.Because sample data difference is very large, need sample data to be normalized.Before process sample, need play and sample data is divided into two parts, a part is as training set, remaining as test set.The process of its model training is: first adopt identical method to be normalized training set and test set, using training set as the training sample of support vector machine, Training Support Vector Machines is carried out by constantly optimizing kernel functional parameter, if the accuracy of fault diagnosis result does not reach requirement, then need to reselect the parameter area of kernel function, until the accuracy of diagnostic result reaches requirement, now be optimal support vector machine, finally verify that whether the vector machine of training is correct to the diagnostic result of fault with test set.
The specific implementation step of support vector machine in transformer fault diagnosis can be expressed as follows:
(1) all of the electric power transformer oil with clear failure conclusion is obtained, oil sample sample is divided into training set and test set, and oil sample sample is classified according to fault type, according to fault type, oil sample sample is divided into six classes, represent with " 1 ", " 2 ", " 3 ", " 4 ", " 5 ", " 0 " respectively, wherein, in " 1 " expression, cryogenic overheating, " 2 " expression hyperthermia and superheating, " 3 " represent that low-yield electric discharge, " 4 " expression high-energy discharge, " 5 " represent shelf depreciation, " 0 " expression normal condition;
(2) oil sample sample is converted into the matrix of 600X5, and adopts identical method, respectively training set and test set are normalized;
(3) suitable kernel function is selected, first input larger data search model and adopt grid data service Selection parameter penalty factor c and kernel function δ roughly, then on the basis of rough search, reasonably reduce data search scope, utilize grid data service accurately to select optimal parameter c and δ;
(4) utilize training set sample training based on the data model of support vector machine, and whether reach requirement with test set sample predictions diagnostic result, if not, then return the parameter area with reselecting kernel function to step (3);
(5) the oil sample data of diagnosis need be played to substitute in model and to obtain diagnostic result.
Analyze oil sample sample, the most difference between people's value with minimum value of characteristic gas is very large, and the degree of redying that the difference between oil sample sample value is large, variation range extensively will increase calculating, causes the imbalance of training, and the training time is long.In order to reduce by oil sample sample on training the impact that causes, just needs oil sample sample data carry out a change process, training set sample adopt identical method for normalizing with test set sample palpus.
In order to make support reach higher classification accuracy to the machine sorter of establishing, avoiding the situation occurring that in learning process " crossing study " one-tenth person " owes study ", selecting cross validation Support Vector Machines Optimized, adopt grid data service to select optimum kernel functional parameter.Its principle is that oil sample sample is divided into two parts, and get a part wherein as training set, a remaining part is as test set.First with training set sample, support vector machine is trained, grid data service is utilized to select to obtain optimized parameter, construct suitable decision function, verify with test set sample again and train the supporting vector machine model that obtains, using Zhi Chixiang Hair machine sorter to the accuracy rate of breakdown judge as evaluating the performance index supported to machine sorter.
Sample data pre-service
The present invention have collected 600 groups of oil dissolved gas data, and often organizing data has clear and definite fault conclusion.According to IEC60599 rule, Power Transformer Faults type is divided into following five types: hyperthermia and superheating, high-energy discharge, low-yield electric discharge, middle cryogenic overheating, shelf depreciation, represent with " 1 ", " 2 ", " 3 ", " 4 ", " 5 ", " 0 " respectively, wherein, in " 1 " expression, cryogenic overheating, " 2 " expression hyperthermia and superheating, " 3 " represent that low-yield electric discharge, " 4 " expression high-energy discharge, " 5 " represent shelf depreciation, " 0 " expression normal condition.According to above six kinds of labels, oil sample sample is divided into training set sample and test set sample, wherein training set has 400 samples, and remaining 200 samples are as test set.The 600 groups of oil sample sample datas arranging acquisition are converted into the matrix of 600X5, as the input data of support vector machine,
Grid data service chooses optimal parameter c, δ
Oil sample sample data after normalized is imported database, adopts grid data service to select kernel function optimal parameter δ and penalty factor c.Consider the factors such as oil sample sample data type, data volume, select 10 folding cross-validation methods.600 oil sample sample datas are divided into 10 groups, get 8 combinations wherein and as training set, remaining as test set, through training, by acquisition 10 subseries accuracy rate, finally get the performance index of arithmetic average as support vector machine classifier of 10 subseries accuracys rate.
The span that grid data service arranges penalty factor c is [2
-10, 2
10], stepping is 0.4; The span of kernel functional parameter δ is [2
-10, 2
10], stepping is 0.4.By training support vector machine, the best value of penalty factor c is 0.83282, and the best value of kernel functional parameter δ is 0.39227, and the accuracy rate of support vector machine classifier Selection parameter is 77.5536%.
Reduce the hunting zone of grid data service, training is proceeded to support vector machine, to find optimum parameter, improve the accuracy rate of support vector machine classifier Selection parameter.The span that grid data service arranges penalty factor c is [2
-10, 2
0], stepping 0.2; The span of kernel functional parameter δ is [2
-10, 2
0], stepping 0.2.Through training support vector machine, the best value of penalty factor c is 0.40421, and kernel functional parameter δ is best, and value is 1.00231, and the accuracy rate of support vector machine classifier Selection parameter is 93.1196%.
Grid data service arranges the span [2 of penalty factor c
0, 2
10], stepping 0.2; The span of kernel functional parameter δ is [2
0, 2
10], stepping 0.2.Through Training Support Vector Machines, the best value of penalty factor c is 1.2986, and kernel functional parameter δ is best, and value is 1.4093, and the accuracy rate of support vector machine classifier Selection parameter is 96.088%.
In order to analyze the impact that penalty factor c and kernel functional parameter δ trains support vector machine classifier, changing hunting zone, continuing Training Support Vector Machines.The span that grid data service arranges penalty factor c is [2
0, 2
10], stepping 0.2, the span of kernel functional parameter δ is [2
-10, 2
0], stepping 0.2.By Training Support Vector Machines, the best value of penalty factor c is 23.2312, and kernel functional parameter δ is best, and value is 0.025102, and the rate of accuracy reached of support vector machine classifier Selection parameter is to 96.6598%.
Be not difficult to draw by above analysis: kernel functional parameter δ value obtains excessive or too small all can causing oil sample sample " owing study " or " crossing study ".Penalty factor c plays a part to regulate maximum class interval and minimize training mistake, when support vector machine classifier is classified, if when penalty factor c value obtains larger, the generalization ability of support vector machine is poor; If when penalty factor c value is less, the generalization ability of support vector machine is better.If when the value of penalty factor c exceedes certain numerical value, the complexity of support vector machine will be strengthened, and make it reach maximal value needed for data space.Even if the expanded range of penalty factor c, the training accuracy rate of support vector machine will constantly change, but the test accuracy rate of support vector machine no longer changes.
The best value adopting grid data service to obtain penalty factor c is 23.2312, and kernel functional parameter δ is best, and value is 0.025102, and the rate of accuracy reached of support vector machine classifier Selection parameter is to 96.6598%.The satisfactory support vector machine classifier prediction test set utilizing training to obtain, test set 200 oil sample samples are input to support vector machine classifier, and the classification accuracy of support vector machine to test set sample reaches 93.36%.
A kind of Diagnosis Method of Transformer Faults that the present invention also proposes, wherein support vector machine can also be adopt the support vector machine based on particle group optimizing, and the modeling process based on the support vector machine of particle group optimizing is:
(1) initialization population, is optimized the kernel function δ of particle swarm support vector machine and penalty factor c by the method for adjustment population inertia weight ω, makes parameter c and δ form particulate, i.e. (c, a δ), and set maximal rate as V
max, the initial position representing each particulate with pbest, represents fine-grained best initial position in population with gbest;
(2) evaluate the fitness of each particulate, calculate the optimal location of each particulate;
(3) adaptive value of each particulate after optimizing and its history optimal location pbest are compared, if current adaptive value is better than optimal location, then using adaptive value as the current desired positions pbest of particle;
(4) adaptive value of each particulate and the history optimal location gbest of colony's particulate after optimization are compared, if adaptive value is better than the history optimal location gbest of colony's particulate, then using the optimal location gbest of adaptive value as colony's particulate;
(5) speed and the position of current particulate is adjusted according to modified particle swarm optiziation;
(6) when adaptive value satisfies condition, iteration terminates, otherwise returns second step continuation Optimal Parameters, after the 6th step completes, will the parameter c of optimization the best and δ, so just can obtain optimal supporting vector machine model, carry out failure prediction with this model.
If Population Size N=20, inertia weight ω=0.9, aceleration pulse C
1=1.4, C
2=1.6, Training Support Vector Machines, the best value obtaining penalty factor c be 3.8326, kernel function δ best value be 0.50433.Through particle cluster algorithm optimization, the classification accuracy of SVM classifier reaches 98.9234%.
A kind of Diagnosis Method of Transformer Faults that the present invention also proposes, wherein support vector machine can also be adopt the support vector machine based on genetic algorithm, and the modeling process based on the support vector machine of genetic algorithm is:
(1) initialization population, generate the individuality of some as initial population, every bar chromosome is made up of (c, δ), and wherein c is penalty factor, and δ is kernel function;
(2) selected target function carries out support vector machine training to initial population, using the square error of support vector machine as objective function, calculates the fitness of each individuality;
(3) carry out Selecting operation, crossing operation, mutation operator obtain population of new generation, support vector machine training is carried out to the new population produced;
(4) if the new population produced meets termination rules, then output has the individuality of maximum adaptation degree as optimized parameter, predicts, otherwise increase evolutionary generation with optimized parameter, proceeds to step (3) and continues to perform;
The c value that in the present invention, said method obtains is 50, δ value when being 0.52, and classification accuracy is 94.5%.
A kind of Diagnosis Method of Transformer Faults that the present invention also proposes, uses the SVM prediction coiling hot point of transformer temperature based on particle cluster algorithm.Gather the hot spot temperature of winding (θ of transformer
hst), top-oil temperature (θ
tot), active loss (P
l), reactive loss (R
l), load current (I) and environment temperature (T
amb), humidity (H), wind speed (v
w), as the inputoutput data set of the supporting vector machine model based on particle cluster algorithm, by the top-oil temperature (θ of transformer
tot), active loss (P
l), reactive loss (R
l), load current (I) and environment temperature (T
amb), humidity (H), wind speed (v
w) as the input of the supporting vector machine model based on particle cluster algorithm, the hot spot temperature of winding (θ of transformer
hst) as exporting.
For choosing of SVM parameter, generally by virtue of experience choose, or given range utilize grid search to find CV meaning under globally optimal solution, if travel through grid search in a wider context, will at substantial computational resource, be unfavorable for the real-time estimate of hot spot temperature of winding.Adopt didactic particle cluster algorithm, optimizing is carried out to SVM parameter, follow optimum example by particle in solution space to search for, choose most parameter corresponding to high-accuracy, if most high-accuracy correspondence organizes parameter more, then choose relatively little one group of penalty factor c as optimized parameter, if corresponding minimum penalty factor c organize parameter more, then the one group of parameter choosing first group of minimum penalty factor searching corresponding is optimum SVM parameter.
Based on particle cluster algorithm in supporting vector machine model parameter optimisation procedure, choosing population quantity is 20, and the restriction factor of change in displacement is taken as c
1=c
2=2, inertia weight factor ω get initial value be 0.9 carry out by for linear decrease to 0.4 stop, stop algebraically be 200, choose the fitness function of square error (RMSE) as particle cluster algorithm.Optimum SVM parameter value after particle cluster algorithm optimization is, penalty factor c=17.0021, kernel function δ=0.5962, fitness function value is: RMSE=0.0039823.The parameter value that particle cluster algorithm optimization obtains is sent back to SVM model just can rebuild regression model and predict hot spot temperature of winding.
Support vector regression method is applied to the hot spot temperature of winding prediction of oil-filled transformer by the present invention, first according to each factor affecting coiling hot point of transformer temperature, selects suitable characteristic parameter to set up support vector regression model.Based on the Historical Monitoring data of these characteristic parameters, wherein implicit and between hot(test)-spot temperature relation is carried out " intelligent learning ", sets up the distinctive temperature model of transformer.Model parameter adopts particle cluster algorithm (ParticleSwarmOptimization, PSO) be optimized, utilize the model after optimizing to carry out real-time estimate to the hot spot temperature of winding of transformer, and carry out contrasting with the accuracy of verification model and applicability with its measured result.Then the hot spot temperature of winding utilizing actual motion power transformer to monitor continuously, Real-time Load electric current and the information such as the on-site Weather information of transformer and site environment temperature, the parameter of application particle cluster algorithm to support vector regression model is optimized, and realizes the real-time estimate to Winding in Power Transformer hot(test)-spot temperature.
A kind of transformer monitoring systems that the present invention proposes and method for diagnosing faults, can the various status information of Real-Time Monitoring transformer, and diagnoses the fault of transformer accordingly, improves transformer fault diagnosis efficiency.
Those skilled in the art, at consideration instructions and after putting into practice invention disclosed herein, will easily expect other embodiment of the present invention.The application is intended to contain any modification of the present invention, purposes or adaptations, and these modification, purposes or adaptations are followed general principle of the present invention and comprised the undocumented common practise in the art of the present invention or conventional techniques means.
Should be understood that, the present invention is not limited to precision architecture described above and illustrated in the accompanying drawings, and can carry out various amendment and change not departing from its scope.Scope of the present invention is only limited by appended claim.
Claims (10)
1. a transformer monitoring systems, comprising: oil dissolved gas chromatogram monitoring unit, partial discharge monitoring unit, master station computer, wherein,
Oil dissolved gas chromatogram monitoring unit, utilize the content of gas chromatogram monitoring sensor to hydrogen contained in transformer oil, carbon monoxide, carbon dioxide, acetylene, ethene, methane, ethane to calculate, and the result of calculating is transferred to master station computer by RS232 communication protocol;
Partial discharge monitoring unit, adopts external high frequency antenna as type UHF sensor, the ultrahigh frequency voltage signal that monitoring inside transformer shelf depreciation produces, and monitoring result is transferred to master station computer.
2. transformer monitoring systems as claimed in claim 1, wherein,
Superheated steam drier, transformer overload, short trouble, the discharge fault of oil dissolved gas chromatogram monitoring unit monitoring transformer;
The discharging fault of partial discharge monitoring unit monitoring transformer.
3. transformer monitoring systems as claimed in claim 1, wherein, the monitoring result that master station computer reception oil dissolved gas chromatogram monitoring unit and shelf depreciation monitoring means transmit, the acetylene content monitored when oil dissolved gas chromatogram monitoring unit accounts for 20% to 70% of total hydrocarbon ratio, hydrogen content accounts for 30% to 90% of hydrocarbon total amount simultaneously, and, carbon monoxide, carbon dioxide, ethene, methane, the content of ethane is in predetermined normal range, and partial discharge monitoring unit display partial discharges fault, at this moment master station computer determination transformer generation shelf depreciation fault.
4. transformer monitoring systems as claimed in claim 1, wherein, oil dissolved gas chromatogram monitoring unit is connected with master station computer by RS232 serial ports, the operating system of described oil dissolved gas chromatogram monitoring unit is built-in Linux, described oil dissolved gas chromatogram monitoring unit passes through serial interface management module management to the operation of RS232 serial ports, comprise serial ports is set attribute, serial ports open the reading with serial ports, the serial ports of described serial interface management module reads implementation procedure and comprises:
Add the header file used in the serial ports fetch program;
Define a new int type process ID number;
Write newly-built subprocess function, read the Monitoring Data in serial ports respective file and output in newly-built data.txt file;
To be slept the people 1 second by programming parent process, and carry out terminator process by the kill function of system function;
The data.txt file storing Monitoring Data is opened with read-only mode;
Read the first row data from data.txt file, and stored in buffer memory buf;
Judge whether the data that a certain moment reads start with "--", if not then again read next line, until read complete data by the complete adaptation function of character string;
Concrete numerical value in the data line mated completely is extracted air-monitor storage of array;
Realize opening a new files jc.txt file with read-write mode;
The numerical value write jc..txt file stored in air-monitor array is supplied other equipment calls.
5. a Diagnosis Method of Transformer Faults, uses transformer monitoring systems as claimed in claim 1 to carry out fault diagnosis to transformer, comprising:
By described oil dissolved gas chromatogram monitoring unit, the gas in transformer oil is sampled, and by the absolute gas production rate of following formulae discovery:
wherein,
γ
αrepresent the absolute gas production rate of certain gas, C
t, 2represent that second time sampling records certain gas concentration in oil, C
t, 1represent that first time sampling records certain gas concentration in oil, Δ t represents the actual run time in twice sample interval, the gross mass of m indication transformer oil, and ρ represents the density of oil;
When the content of total hydrocarbon is greater than first threshold and is less than 3 times of first threshold, and when the absolute gas production rate of total hydrocarbon is less than Second Threshold, determine that transformer exists fault but can continue to run; When 1 to 2 times that absolute gas production rate is Second Threshold, at this moment determine the round of visits that will shorten transformer;
When the content of total hydrocarbon is greater than 3 times of first threshold and the absolute gas production rate of total hydrocarbon is greater than 3 times of Second Threshold, at this moment illustrate that transformer has serious fault, notifies that staff takes corresponding measure immediately;
Wherein, total hydrocarbon refers to hydrocarbon gas all in transformer oil; The absolute gas production rate of total hydrocarbon refers to the absolute gas production rate sum of hydrocarbon gas all in transformer oil.
By the relative gas production rate of following formulae discovery:
wherein,
γ
rrepresent the relative gas production rate of certain gas, C
t, 2represent that second time sampling records certain gas concentration in oil, C
t, 1represent that first time sampling records certain gas concentration in oil, Δ t represents the actual run time in twice sample interval;
When the relative gas production rate of total hydrocarbon is greater than 10%, determine to shorten the sense cycle to transformer.
6. Diagnosis Method of Transformer Faults as claimed in claim 5, supporting vector machine model is adopted to diagnose transformer fault, choose hydrogen, acetylene, ethene, methane, ethane five kinds of failure gas as characteristic gas, the matrix formed is as the input vector of supporting vector machine model, the column vector of input matrix is above five kinds of characteristic gas, and the dimension of column vector is 5; The row vector of input matrix is the raw data collected, and the dimension of row vector is the number of raw data.
7. Diagnosis Method of Transformer Faults as claimed in claim 6, wherein, the training process of described supporting vector machine model comprises:
First identical method is adopted to be normalized training set and test set, using training set as the training sample of support vector machine, Training Support Vector Machines is carried out by constantly optimizing kernel functional parameter, if the accuracy of fault diagnosis result does not reach requirement, then need to reselect the parameter area of kernel function, until the accuracy of diagnostic result reaches requirement, now be met the supporting vector machine model of requirement, finally verify that whether the support vector machine of training is correct to the diagnostic result of fault with test set.
8. Diagnosis Method of Transformer Faults as claimed in claim 7, wherein, adopts described supporting vector machine model to carry out transformer fault diagnosis and specifically comprises:
(1) electric power transformer oil all 600 with clear failure conclusion is obtained, oil sample sample is divided into training set and test set, and oil sample sample is classified according to fault type, according to fault type, oil sample sample is divided into six classes, represent with " 1 ", " 2 ", " 3 ", " 4 ", " 5 ", " 0 " respectively, wherein, in " 1 " expression, cryogenic overheating, " 2 " expression hyperthermia and superheating, " 3 " represent that low-yield electric discharge, " 4 " expression high-energy discharge, " 5 " represent shelf depreciation, " 0 " expression normal condition; Wherein training set comprises 400 samples, and test set comprises 200 samples;
(2) oil sample sample is converted into the matrix of 600X5, and adopts identical method, respectively training set and test set are normalized;
(3) suitable kernel function is selected, first input larger data search model and adopt grid data service Selection parameter penalty factor c and kernel function δ roughly, then on the basis of rough search, reasonably reduce data search scope, utilize grid data service accurately to select optimal parameter c and δ;
(4) utilize training set sample training supporting vector machine model, and whether reach requirement with test set sample predictions diagnostic result, if not, then turn back to the parameter area that step (3) reselects kernel function;
(5) diagnostic result is obtained by needing the transformer oil sample data of diagnosis to substitute in described supporting vector machine model.
9. Diagnosis Method of Transformer Faults as claimed in claim 8, wherein, the span utilizing grid data service to arrange penalty factor c is [2
-10, 2
10], stepping is 0.4; The span of kernel functional parameter δ is [2
-10, 2
10], stepping is 0.4, and by training support vector machine, the best value of penalty factor c is 0.83282, and the best value of kernel functional parameter δ is 0.39227, and the accuracy rate of support vector machine classifier Selection parameter is 77.5536%.
10. Diagnosis Method of Transformer Faults as claimed in claim 8, wherein, the span utilizing grid data service to arrange penalty factor c is [2
-10, 2
0], stepping 0.2; The span of kernel functional parameter δ is [2
-10, 2
0], stepping 0.2, through training support vector machine, the best value of penalty factor c is 0.40421, and kernel functional parameter δ is best, and value is 1.00231, and the accuracy rate of support vector machine classifier Selection parameter is 93.1196%.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510607872.9A CN105241497B (en) | 2015-09-23 | 2015-09-23 | A kind of transformer monitoring systems and method for diagnosing faults |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510607872.9A CN105241497B (en) | 2015-09-23 | 2015-09-23 | A kind of transformer monitoring systems and method for diagnosing faults |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105241497A true CN105241497A (en) | 2016-01-13 |
| CN105241497B CN105241497B (en) | 2018-03-23 |
Family
ID=55039224
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510607872.9A Expired - Fee Related CN105241497B (en) | 2015-09-23 | 2015-09-23 | A kind of transformer monitoring systems and method for diagnosing faults |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105241497B (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108414684A (en) * | 2018-01-31 | 2018-08-17 | 贵州乌江水电开发有限责任公司东风发电厂 | A kind of for transformer status assessment and method for diagnosing faults |
| CN109030998A (en) * | 2018-07-19 | 2018-12-18 | 浙江浙能常山天然气发电有限公司 | A kind of intelligent transformer monitoring system based on three shaft vibration technologies |
| CN109060892A (en) * | 2018-06-26 | 2018-12-21 | 西安交通大学 | SF based on graphene composite material sensor array6Decompose object detecting method |
| CN110197222A (en) * | 2019-05-29 | 2019-09-03 | 国网河北省电力有限公司石家庄供电分公司 | A method of based on multi-category support vector machines transformer fault diagnosis |
| CN110441500A (en) * | 2019-08-15 | 2019-11-12 | 昆明理工大学 | A kind of predicting model for dissolved gas in transformer oil method based on shot and long term memory network |
| CN110857869A (en) * | 2018-08-23 | 2020-03-03 | 国网天津市电力公司 | Power transformer on-line state monitoring and fault diagnosis system |
| CN112461942A (en) * | 2020-10-10 | 2021-03-09 | 广西电网有限责任公司电力科学研究院 | High-voltage bushing capacitor core fault diagnosis method and system |
| CN112926620A (en) * | 2021-01-21 | 2021-06-08 | 国网山东省电力公司检修公司 | Transformer bushing partial discharge defect type identification method and system |
| CN113063314A (en) * | 2021-03-23 | 2021-07-02 | 哈尔滨工程大学 | Fault diagnosis method for gun launching system based on SVM (support vector machine) and GA-SVM (genetic algorithm-support vector machine) |
| CN113933695A (en) * | 2021-10-12 | 2022-01-14 | 云南电网有限责任公司电力科学研究院 | Method for online judging vacuum on-load tap-changer fault based on gas component identification |
| CN115290798A (en) * | 2022-09-13 | 2022-11-04 | 国网河北省电力有限公司电力科学研究院 | Stability performance monitoring method and terminal of transformer oil chromatographic on-line monitoring device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6675635B2 (en) * | 2002-04-17 | 2004-01-13 | Caterpillar Inc | System and method for determining oil quality |
| CN2736782Y (en) * | 2004-08-09 | 2005-10-26 | 华北电力大学（北京） | Voltage transformer local discharge ultrahigh frequency (UHF) online monitoring device |
| US7219536B2 (en) * | 2005-04-29 | 2007-05-22 | Honeywell International Inc. | System and method to determine oil quality utilizing a single multi-function surface acoustic wave sensor |
| CN102539283A (en) * | 2011-01-06 | 2012-07-04 | 北京盈胜泰科技术有限公司 | System for monitoring ingredient change of insulating oil of transformer |
| CN202748346U (en) * | 2012-08-21 | 2013-02-20 | 广西电网公司电力科学研究院 | Intelligent on-line monitoring device for dissolved gas in power transformer insulating oil |
-
2015
- 2015-09-23 CN CN201510607872.9A patent/CN105241497B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6675635B2 (en) * | 2002-04-17 | 2004-01-13 | Caterpillar Inc | System and method for determining oil quality |
| CN2736782Y (en) * | 2004-08-09 | 2005-10-26 | 华北电力大学（北京） | Voltage transformer local discharge ultrahigh frequency (UHF) online monitoring device |
| US7219536B2 (en) * | 2005-04-29 | 2007-05-22 | Honeywell International Inc. | System and method to determine oil quality utilizing a single multi-function surface acoustic wave sensor |
| CN102539283A (en) * | 2011-01-06 | 2012-07-04 | 北京盈胜泰科技术有限公司 | System for monitoring ingredient change of insulating oil of transformer |
| CN202748346U (en) * | 2012-08-21 | 2013-02-20 | 广西电网公司电力科学研究院 | Intelligent on-line monitoring device for dissolved gas in power transformer insulating oil |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108414684A (en) * | 2018-01-31 | 2018-08-17 | 贵州乌江水电开发有限责任公司东风发电厂 | A kind of for transformer status assessment and method for diagnosing faults |
| CN109060892A (en) * | 2018-06-26 | 2018-12-21 | 西安交通大学 | SF based on graphene composite material sensor array6Decompose object detecting method |
| CN109060892B (en) * | 2018-06-26 | 2020-12-25 | 西安交通大学 | SF based on graphene composite material sensor array6Method for detecting decomposition product |
| CN109030998A (en) * | 2018-07-19 | 2018-12-18 | 浙江浙能常山天然气发电有限公司 | A kind of intelligent transformer monitoring system based on three shaft vibration technologies |
| CN110857869A (en) * | 2018-08-23 | 2020-03-03 | 国网天津市电力公司 | Power transformer on-line state monitoring and fault diagnosis system |
| CN110197222A (en) * | 2019-05-29 | 2019-09-03 | 国网河北省电力有限公司石家庄供电分公司 | A method of based on multi-category support vector machines transformer fault diagnosis |
| CN110441500A (en) * | 2019-08-15 | 2019-11-12 | 昆明理工大学 | A kind of predicting model for dissolved gas in transformer oil method based on shot and long term memory network |
| CN112461942A (en) * | 2020-10-10 | 2021-03-09 | 广西电网有限责任公司电力科学研究院 | High-voltage bushing capacitor core fault diagnosis method and system |
| CN112926620A (en) * | 2021-01-21 | 2021-06-08 | 国网山东省电力公司检修公司 | Transformer bushing partial discharge defect type identification method and system |
| CN113063314A (en) * | 2021-03-23 | 2021-07-02 | 哈尔滨工程大学 | Fault diagnosis method for gun launching system based on SVM (support vector machine) and GA-SVM (genetic algorithm-support vector machine) |
| CN113063314B (en) * | 2021-03-23 | 2022-03-22 | 哈尔滨工程大学 | Fault diagnosis method for gun launching system based on SVM (support vector machine) and GA-SVM (genetic algorithm-support vector machine) |
| CN113933695A (en) * | 2021-10-12 | 2022-01-14 | 云南电网有限责任公司电力科学研究院 | Method for online judging vacuum on-load tap-changer fault based on gas component identification |
| CN115290798A (en) * | 2022-09-13 | 2022-11-04 | 国网河北省电力有限公司电力科学研究院 | Stability performance monitoring method and terminal of transformer oil chromatographic on-line monitoring device |
| CN115290798B (en) * | 2022-09-13 | 2023-10-31 | 国网河北省电力有限公司电力科学研究院 | Stability performance monitoring method and terminal of transformer oil chromatographic online monitoring device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105241497B (en) | 2018-03-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105259435A (en) | Transformer monitoring device and fault diagnosis method | |
| CN105182161A (en) | Transformer monitoring system and method | |
| CN105241497A (en) | Transformer monitoring system and fault diagnosis method | |
| CN106908674B (en) | Transformer state evaluation method based on multi-state quantity prediction | |
| Azmi et al. | Evolution of transformer health index in the form of mathematical equation | |
| Taha et al. | Optimal ratio limits of rogers' four-ratios and IEC 60599 code methods using particle swarm optimization fuzzy-logic approach | |
| Dong et al. | A novel maintenance decision making model of power transformers based on reliability and economy assessment | |
| CN109490726A (en) | Electric power transformer insulated state evaluating method based on Clouds theory | |
| CN106841846A (en) | A kind of transformer state analysis and fault diagnosis method and system | |
| Lee et al. | New methods of DGA diagnosis using IEC TC 10 and related databases Part 2: application of relative content of fault gases | |
| Yang et al. | State evaluation of power transformer based on digital twin | |
| CN105425768B (en) | A kind of second power equipment monitoring device and method | |
| Bello et al. | Optimal settings for multiple groups of smart inverters on secondary systems using autonomous control | |
| CN104655948A (en) | Novel multistage fault diagnosis method for power transformer | |
| CN105467971A (en) | Electric power secondary equipment monitoring system and method | |
| CN108629491A (en) | A kind of change of current change repair quality comprehensive estimation method | |
| Gautam et al. | Identifying transformer oil criticality using fuzzy logic approach | |
| CN108845074A (en) | Oil-immersed transformer method for evaluating state | |
| Tippannavar et al. | Smart Transformer-An Analysis of Recent Technologies for Monitoring Transformer | |
| Scatiggio et al. | Health Index: the last frontier of TSO’s asset management | |
| Xiao et al. | Application of geometrical observation of security region in distribution network enhancement | |
| Cui et al. | Transformer hot spot temperature prediction using a hybrid algorithm of support vector regression and information granulation | |
| Wu et al. | CNN-MLP-based transformer digital twin model construction and fault diagnosis and condition evaluation analysis | |
| Zhang et al. | Deep Primary and Secondary Fusion Transformer Based on Internet of Things Technology | |
| CN113792636B (en) | Extra-high voltage direct current transmission converter transformer recessive fault prediction method based on fault wave recording |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180323 Termination date: 20180923 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |