WO2013042928A1 - 부분방전 결함유형 판정 방법 및 그 장치 - Google Patents
부분방전 결함유형 판정 방법 및 그 장치 Download PDFInfo
- Publication number
- WO2013042928A1 WO2013042928A1 PCT/KR2012/007482 KR2012007482W WO2013042928A1 WO 2013042928 A1 WO2013042928 A1 WO 2013042928A1 KR 2012007482 W KR2012007482 W KR 2012007482W WO 2013042928 A1 WO2013042928 A1 WO 2013042928A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- partial discharge
- defect type
- defect
- types
- type determination
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
- G01R31/1263—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation
- G01R31/1272—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation of cable, line or wire insulation, e.g. using partial discharge measurements
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02B—BOARDS, SUBSTATIONS OR SWITCHING ARRANGEMENTS FOR THE SUPPLY OR DISTRIBUTION OF ELECTRIC POWER
- H02B13/00—Arrangement of switchgear in which switches are enclosed in, or structurally associated with, a casing, e.g. cubicle
- H02B13/02—Arrangement of switchgear in which switches are enclosed in, or structurally associated with, a casing, e.g. cubicle with metal casing
- H02B13/035—Gas-insulated switchgear
- H02B13/065—Means for detecting or reacting to mechanical or electrical defects
Definitions
- the present invention relates to a method for determining a partial discharge defect type and a device thereof (Method for deciding defect type of partial discharge and apparatus), and more particularly to a new defect type determination method for reinforcing the weakness of the neural network defect type determination method.
- the present invention relates to a partial discharge defect type determination method and apparatus therefor.
- Ultra high frequency (UHF) partial discharge diagnosis online system and portable equipment which are currently used to prevent gas insulated switchgear (GIS) failure, are equipped with the determination method for recognizing the type of defect as the core of the software.
- GIS gas insulated switchgear
- This defect recognition judgment method uses a neural network to display the results in the form of probabilities that match with a predefined defect type, and the neural network calculates hundreds of calculations based on phase, discharge amount, and discharge frequency. More than one learning data are used to determine the defect type of the detected partial discharge signal.
- the Neural Network judgment method currently used has a fundamental problem that the accuracy of the defect type determination is significantly lowered and the other defects are judged in the parameter by the unlearned pattern. do.
- the present invention provides an applied partial discharge defect type determination method and apparatus.
- another object of the present invention is to provide a partial discharge defect type determination method and apparatus for improving the accuracy of the defect type determination by giving a weight to use in parallel with the neural network determination method in use.
- Partial discharge defect type determination method for achieving the above object, to form a virtual space corresponding to the n parameters generated based on the defect types that generate partial discharge, each of the defect types Arranging defect type models in the virtual space by combining parameter values of the n parameters with respect to the n-parameters; Extracting and arranging the partial discharge model in the virtual space by combining the parameter values, comparing the location information of the defect type models with the location information of the partial discharge model, and comparing the defect type models with the partial discharge model.
- the virtual space is characterized in that the n-dimensional space consisting of n axes corresponding to each of the n parameters.
- the disposing of the defect types may include arranging the parameter values for each defect type in the virtual space, and if the parameter values form a cluster having a constant distribution, the coordinate values of the intermediate positions of the clusters are determined for each defect type. And calculating the coordinate values calculated for each defect type as position information of each defect type model.
- the calculating of the probability values of the defect types may include calculating the probability value by calculating a ratio of relative distances for each of the defect types based on the total distance values with the defect types.
- the probability value is characterized in that the closer the distance value between the partial discharge model and the defect type model, the higher the probability value.
- the determining of the defect type may include determining a defect type for the partial discharge signal event by combining the probability values of the defect types and the probability values from the neural network based defect type determination result.
- the determining of the defect type may be performed by assigning different weights to probability values of the defect types and probability values from the neural network based defect type determination result.
- the determining of the defect type may include determining a defect type for the partial discharge signal event by adding a probability value of the defect types to which the weight is different and a probability value from the result of the neural network based defect type determination. It is characterized by.
- the parameters are generated based on the phase, the amount of discharge, and the number of discharges of the partial discharge signal measured corresponding to the defect types.
- the parameters are classified into three types of two-dimensional distributions of phase-discharge amount, phase-discharge frequency, and discharge-discharge frequency, with the phase, the discharge amount, and the discharge frequency as variables. do.
- phase-discharge amount and the phase-discharge frequency type parameters are classified as positive and negative since phases are dependent variables.
- the discharge amount-discharge frequency type parameters may be normalized by setting the maximum value to 100.
- the partial discharge defect type may include at least one of particles, floating, corona, cavities, and noise.
- the partial discharge defect type determination apparatus for achieving the above object, the virtual space implementation unit for forming a virtual space corresponding to the n parameters generated based on the defect types for generating a partial discharge, A partial discharge signal analysis unit extracting parameter values corresponding to the n parameters based on the partial discharge signal generated when the partial discharge signal event occurs; combining parameter values of the n parameters for each of the defect types By comparing the position information of the defect type models disposed in the virtual space and the parameter values of the n parameters for the partial discharge signal to compare the position information of the partial discharge model disposed in the virtual space to determine the defect type model Calculates a distance value between the field and the partial discharge model, and calculates the partial discharge from the distance value. And a calculation unit for calculating a probability value of the defect types for a signal event, and a defect type determination unit for determining a defect type for the partial discharge signal event based on the probability values of the defect types for the partial discharge signal event. It is done.
- the defect type determining unit receives a neural network based defect type determination result for the defect types and the partial discharge signal event, and determines a probability value of the defect types and probability values from the neural network based defect type determination result. Combine to determine a defect type for the partial discharge signal event.
- the defect type determination unit may assign different weights to probability values of the defect types and probability values from the neural network based defect type determination result.
- the defect type determining unit may determine a defect type for the partial discharge signal event by summing a probability value of the defect types to which the different weights are assigned and probability values from the neural network based defect type determination result. do.
- FIG. 1 is a diagram showing the basic concept of the partial discharge defect type determination apparatus according to the present invention.
- Fig. 2 is a block diagram showing the configuration of the partial discharge defect type determining apparatus according to the present invention.
- FIG 3 is an exemplary view showing an embodiment of parameters according to the present invention.
- FIG. 4 is an exemplary diagram referred to for describing an operation of calculating a parameter value for each defect type according to the present invention.
- FIG. 5 is an exemplary view showing a parameter value for each defect type according to the present invention.
- FIG. 6 is an exemplary view showing a defect type model disposed in a virtual space according to the present invention.
- FIG. 7 is an exemplary diagram referred to for explaining an operation of calculating a parameter value of a partial discharge signal according to the present invention.
- FIG 8 is an exemplary view showing a partial discharge model disposed in the virtual space according to the present invention.
- FIG 9 is an exemplary view referred to for explaining the operation of determining the partial discharge defect type according to the present invention.
- 10 and 11 are flowcharts showing an operation flow for determining a partial discharge defect type according to the present invention.
- the apparatus for determining a partial discharge defect type according to the present invention includes a sensor connected to a transformer externally, and determines a partial discharge defect type by analyzing signals detected by the sensors.
- Fig. 2 is a block diagram showing the configuration of the partial discharge defect type determining apparatus according to the present invention.
- the partial discharge defect type determining apparatus includes a sensor unit 110, an interface unit 120, a controller 130, a storage unit 140, and a defect type analysis unit 160. , A virtual space implementation unit 160, a partial discharge signal analysis unit 170, a calculation unit 180, and a defect type determination unit 180.
- the sensor unit 110 is connected to a transformer as shown in FIG. 1, and includes a plurality of sensors that detect a partial discharge signal from the transformer.
- the sensor detecting the partial discharge signal transmits the partial discharge signal event to the controller 130 together with the corresponding partial discharge signal.
- the defect type analyzing unit 160 generates n parameters for at least one or more defect types as a means for setting a reference value for determining a defect type of the partial discharge.
- the defect type analysis unit 160 generates n parameters based on the phase Phi, the discharge amount q, and the number of discharges n of the partial discharge signal measured corresponding to the defect types. The related detailed description will be described with reference to FIG. 3.
- the defect type analyzer 160 calculates n parameter values for each defect type by using the information of the defect types and the n parameters stored in the storage 140.
- a configuration for calculating n parameter values for each defect type will be described in more detail with reference to FIG. 4. Refer to FIG. 5 for parameter values for each defect type calculated through the process as shown in FIG. 4. Similarly, the parameter values for each defect type shown in FIG. 5 are stored in the storage 140.
- the virtual space implementation unit 150 forms a virtual space corresponding to the n parameters generated based on the defect types causing the partial discharge.
- the virtual space refers to an n-dimensional space composed of n axes corresponding to each of the n parameters.
- the virtual space implementation unit 150 forms a 25-dimensional virtual space consisting of 25 axes corresponding to each parameter. At this time, different parameters correspond to each of the 25 axes.
- the virtual space implementation unit 150 arranges the defect type models in the virtual space by combining parameter values corresponding to the n parameters for each of the defect types. At this time, the virtual space implementation unit 150 arranges each parameter value for each defect type in the virtual space, and calculates a coordinate value of an intermediate position for the cluster for each defect type when a cluster having a constant distribution of each parameter value is formed. . At this time, the defect type model is disposed at the coordinate value of the intermediate position.
- the virtual space implementation unit 150 stores the coordinate values calculated for each defect type in the storage unit 140 as position information of each defect type model.
- the defect type model is arranged for each defect type. Defect type model arrangement operation for each defect type will be described in more detail with reference to FIG. 6.
- the partial discharge signal analysis unit 170 analyzes the generated partial discharge signal when the partial discharge signal event occurs, and extracts parameter values corresponding to n parameters based on the analysis result. A detailed operation description thereof will be described with reference to FIG. 7.
- the virtual space implementation unit 150 combines the n parameter values extracted by the partial discharge signal analysis unit 170 to place the partial discharge model on the nth virtual space.
- a partial discharge model arrangement operation according to the partial discharge signal event will be described in more detail with reference to FIG. 8.
- the calculation unit 180 calculates a distance value between the defect type models and the partial discharge model by comparing the position information of the defect type models and the position information of the partial discharge model.
- the equation for calculating the distance value between the defect type models and the partial discharge model is shown in Equation 1 below.
- Equation 1 calculates the distance between the Corona defect type model and the partial discharge model, and c0p1, c0p2, c0p3, ..., c0p25 are intermediate in each parameter dimension of the Corona defect type. It shows the position coordinate value.
- e1p1, e1p2, e1p3, ..., e1p25 represent intermediate position coordinate values in each parameter dimension of the first partial discharge signal.
- the calculator 180 calculates a probability value of defect types for the partial discharge signal event from the calculated distance value.
- the equation for calculating the probability of defect types for the partial discharge signal event from the distance value is shown in Equation 2 below.
- Dp, Df, Dc, Dv, and Dn are distance values between the partial discharge model and each defect type model
- D major is a distance value between the partial discharge model and the major defect type.
- Equation 2] is not an absolute equation, and the probability value according to the present invention may be calculated based on the ratio of the distance values.
- the calculation unit 180 calculates a probability value by calculating a ratio of relative distances for each defect type based on the total distance values with the defect types.
- the defect type determination unit 180 determines a defect type for the partial discharge signal event based on the probability values of the defect types for the partial discharge signal event.
- the defect type determination unit 180 may determine a defect type for partial discharge by combining with a neural network based defect type determination technique.
- the defect type determination unit 180 receives a neural network-based defect type determination result for the partial discharge signal event from the outside, and calculates a probability value and a calculation unit 180 from the received neural network-based defect type determination result.
- the defect type for the partial discharge signal event can be determined by combining the probability values calculated by.
- the defect type determination unit 180 assigns different weights to the probability values of the defect types and the probability values from the neural network based defect type determination result, and assigns the weights to the probability values from the neural network based defect type determination result.
- the defect type for the partial discharge signal event is determined by adding the applied probability value and the probability value weighted to the probability values of the defect types calculated by the operation unit 180. In this regard, a more detailed configuration description will be referred to the embodiment of FIG. 9.
- FIG 3 is an exemplary view showing an embodiment of parameters according to the present invention.
- the defect type analysis unit 160 sets parameters based on the phase Phi, the discharge amount q, and the number of discharges n of the partial discharge signal measured corresponding to the defect types. Create
- the defect type analysis unit 160 uses the phase, the discharge amount, and the discharge frequency as variables, and includes the phase-discharge amount (Phi-q), the phase-discharge frequency (Phi-n), and the discharge amount-discharge It is classified into three types of two-dimensional distribution of the number qn.
- the parameters of the phase-discharge amount (Phi-q) and the phase-discharge frequency (Phi-n) type are classified into positive and negative since the phase is a dependent variable.
- the defect type analysis unit 160 sets the phase-discharge amount Phi-q and the phase-discharge frequency Phi-n to be positive and negative, respectively. Classify the maximum, minimum, and std. For each of the positive and negative characteristics. Create parameters for Dev., Skewness, and Kurtosis.
- the parameters of the discharge amount-discharge frequency (q-n) type do not consider the polarity since the dependent variable has a linear characteristic.
- the maximum value is set to 100 to normalize.
- the defect type analysis unit 160 does not classify the discharge amount-discharge frequency (q-n) type without any polarity classification. Create parameters for Dev., Skewness, and Kurtosis.
- the positive and negative polarities of the phase-discharge amount Phi-q, the positive and negative polarities of the phase-discharge frequency Phi-n, and the discharge amount are as follows. -Maximum, Minimum, Std. For each discharge frequency (qn). Generates parameters of Dev., Skewness, Kurtosis, so a total of 25 parameters are created.
- parameter generation criteria or the number of generated parameters can be changed according to the setting, and is not limited to any one criterion.
- FIG. 4 is an exemplary diagram referred to for describing an operation of calculating a parameter value for each defect type according to the present invention.
- the defect type analyzer 160 generates parameter values by applying a partial discharge signal based on a defect type stored in the storage 140 to each parameter.
- the signal is first related to the particle defect type. Apply to each parameter, and as a result yield n parameter values such as p1, p2, p3, ..., pn.
- This process performs the same process for each of Floating, Corona, Viod, and Noise, thereby calculating n parameter values for each defect type.
- FIG. 5 is an exemplary view showing a parameter value for each defect type according to the present invention.
- the defect type analysis unit 160 has parameter values corresponding to n parameters from the partial discharge signal based on particles, that is, p1, p2, p3, ..., pn. To calculate. Similarly, the defect type analyzing unit 160 calculates parameter values corresponding to n parameters, that is, f1, f2, f3,..., Fn, from the partial discharge signal based on floating. In addition, the defect type analysis unit 160 calculates n parameter values, that is, c1, c2, c3, ..., cn from the partial discharge signal based on corona, and the partial discharge based on the cavity.
- N parameter values i.e., v1, v2, v3, ..., vn
- n parameter values i.e., n1, n2, n3, ..., from the partial discharge signal based on noise.
- FIG. 6 is an exemplary view showing a defect type model disposed in a virtual space according to the present invention. As shown in FIG. 6, assuming that the number of parameters generated in FIG. 3 is 25, 25 axes in the virtual space do.
- the virtual space implementing unit 150 arranges 25 parameter values calculated for the particle defect type on each of 25 axes, and sets the parameter values on each axis. Place the particle defect type model at the combined intermediate position coordinates.
- Particle defect type model is denoted by p0.
- the virtual space implementation unit 150 arranges 25 parameter values calculated for the floating defect type in each of 25 axes, and sets parameter values in each axis. Place the floating defect type model at the combined intermediate position coordinates. Floating defect type models are denoted by f0.
- the virtual space implementation unit 150 performs the same process for each of Corona, Viod, and Noise, such as a Corona defect type model, a Viod defect type model, And a noise defect type model are placed in the virtual space, respectively.
- a Corona defect type model is represented by c0
- the void defect type model is represented by v0
- the noise defect type model is represented by n0.
- FIG. 7 is an exemplary diagram referred to for explaining an operation of calculating a parameter value of a partial discharge signal according to the present invention.
- the partial discharge signal analyzer 170 generates the parameter values by applying the partial discharge signal detected by the sensors to the respective parameters.
- the partial discharge signal analyzer 170 applies the corresponding partial discharge signals to n parameters, respectively, and calculates n parameter values such as e1, e2, e3, ..., en as a result.
- the virtual space implementation unit 150 combines n parameters for the partial discharge signal calculated in FIG. 7 and arranges a partial discharge model corresponding to the corresponding partial discharge signal event in the n-dimensional virtual space. An embodiment related to this will refer to FIG. 8.
- FIG 8 is an exemplary view showing a partial discharge model disposed in the virtual space according to the present invention.
- the virtual space implementation unit 150 includes a particle defect type model, a floating defect type model, a corona defect type model, a void defect type model, Then, the partial discharge model is disposed in the virtual space where the noise defect type models are arranged.
- the virtual space implementation unit 150 arranges the 25 parameter values calculated for the partial discharge signal on each of the 25 axes, and places the partial discharge model on the intermediate position coordinate values combining the parameter values on each axis.
- the partial discharge model is denoted by e0.
- the calculation unit 180 may determine a particle defect type model, a floating defect type model, a corona defect type model, a cavity defect type model based on a partial discharge model disposed in a virtual space. And distance values between noise defect type models.
- the distance value between the partial discharge model and each defect type model is calculated using Equation 1 described above.
- the distance value between the partial discharge model and the particle defect type model is Dp
- the distance value between the partial discharge model and the floating defect type model is Df
- the distance value between the partial discharge model and the Corona defect type model is defined as Dc
- the distance value between the partial discharge model and the void defect type model is Dv
- the distance value between the partial discharge model and the noise defect type model is defined as Dn.
- the distance values Dp, Df, Dc, Dv, and Dn between the partial discharge model and each defect type model are applied to calculate the probability value of each defect type for the partial discharge signal.
- FIG 9 is an exemplary view referred to for explaining the operation of determining the partial discharge defect type according to the present invention.
- the partial discharge defect type determining apparatus may determine a defect type for partial discharge in combination with a neural network based defect type determination technique. .
- the apparatus for determining the partial discharge defect type receives a neural network based defect type determination result for the corresponding partial discharge signal event from the outside, and calculates the received neural network based defect type determination result and the calculation unit 180.
- the probability type can be combined to determine the defect type for the partial discharge signal event.
- the neural network-based defect type determination result is described as being externally received, but the neural network-based defect type determination result is not received externally, and in the partial discharge defect type determination apparatus according to the present invention. It may be performed separately.
- the defect type determining unit 180 combines the probability values of the defect types calculated by the calculating unit 180 and the probability values from the neural network based defect type determination result to determine a defect type for the partial discharge signal event. Determine.
- the defect type determination unit 180 may assign different weights to the probability values of the defect types and the probability values from the neural network based defect type determination result.
- the defect type determination unit 180 may assign a weight of 0.6 to a probability value from a neural network-based defect type determination result and 0.4 to a probability value of defect types calculated by the calculator 180.
- the probability value from the neural network-based defect type determination result is 55% noise and the floating 45%
- the probability values of the defect types calculated by the calculation unit 180 are 70% floating
- the noise is 30%
- the probability value from the neural network based defect type determination result is 33% noise and 27% floating.
- the probability values of the defect types calculated by the calculator 180 may be 28% floating and 12% noise.
- the defect type determination unit 180 sums 27% of a floating value, which is a probability value from a neural network-based defect type determination result, and 28% of a floating value, which is a probability value of defect types calculated by the calculation unit 180. Finally, it determines 55% of floating.
- the defect type determination unit 180 sums a noise of 33%, which is a probability value from a neural network-based defect type determination result, and 12% of a noise, which is a probability value of defect types calculated by the calculation unit 180. Finally, the noise is determined to be 45%.
- 10 and 11 are flowcharts showing an operation flow for determining a partial discharge defect type according to the present invention.
- the partial discharge defect type determining apparatus generates a criterion for determining a defect type for partial discharge (S100).
- a defect type for partial discharge S100
- at least one of a particle, a floating, a corona, a cavity, and a noise may be included as a defect type standard.
- the apparatus for determining a partial discharge defect type generates a parameter for determining a defect type determined in step S100 (S110), and implements a virtual space having an axis corresponding to the number of parameters generated in step S110 (S120). .
- the apparatus for determining the partial discharge defect type is based on the partial discharge signal for the defect types such as Particle, Floating, Corona, Viod, and Noise.
- Parameter values are calculated for the parameters generated in the process (S130), and the corresponding defect type models are arranged in the virtual space of the process 'S120' by combining the respective parameter values (S140).
- the partial discharge defect type determination device calculates parameter values of parameters of the 'S110' process from the partial discharge signal (S210).
- the apparatus for determining the partial discharge defect type arranges the partial discharge model in the virtual space of the 'S120' process by combining each parameter value calculated in the 'S210' process (S220).
- the apparatus for determining the partial discharge defect type calculates a distance value between each of the defect type models disposed in the virtual space and the partial discharge model in step S140 (S230), and calculates a probability value for each defect type from the calculated distance value. To calculate (S240).
- the partial discharge defect type determination apparatus receives a neural network based defect type determination result (S250), and calculates a probability value for each defect type from the neural network based defect type determination result (S260).
- the apparatus for determining the partial discharge defect type assigns different weights to the probability values based on the neural network based determination result and the probability values calculated in the 'S240' process (S270), and thus the probability values based on the neural network based determination result and 'S240'.
- S280 the probability values calculated in the process
- S290 a defect type determination result for the partial discharge is generated (S290).
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Testing Relating To Insulation (AREA)
Abstract
Description
Claims (18)
- 부분방전을 발생시키는 결함유형들을 기준으로 생성된 n개의 파라미터들에 대응하는 가상 공간을 형성하고, 상기 결함유형들 각각에 대한 상기 n개의 파라미터들의 파라미터 값들을 조합하여 상기 가상 공간상에 결함유형모델들을 배치하는 단계;부분방전신호 이벤트 발생 시, 발생된 부분방전신호에 근거하여 상기 n개의 파라미터들에 해당하는 파라미터 값들을 추출하고, 상기 파라미터 값들을 조합하여 상기 가상 공간상에 부분방전모델을 배치하는 단계;상기 결함유형모델들의 위치 정보와 상기 부분방전모델의 위치 정보를 비교하여 상기 결함유형모델들과 상기 부분방전모델 사이의 거리값을 산출하는 단계;상기 산출하는 단계에서 산출된 상기 거리값으로부터 상기 부분방전신호 이벤트에 대한 상기 결함유형들의 확률값을 산출하는 단계; 및상기 부분방전신호 이벤트에 대한 상기 결함유형들의 확률값에 근거하여 상기 부분방전신호 이벤트에 대한 결함유형을 판정하는 단계;를 포함하는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 1에 있어서,상기 가상 공간은,상기 n개의 파라미터들 각각에 대응하는 n개의 축으로 이루어진 n차원 공간인 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 1에 있어서,상기 결함유형들을 배치하는 단계는,상기 가상 공간상에서 상기 결함유형별로 상기 파라미터 값들을 배치하고, 상기 파라미터 값들이 일정한 분포를 갖는 군집을 형성하면 상기 결함유형별로 해당 군집에 대한 중간 위치의 좌표값을 산출하는 단계; 및상기 결함유형별로 산출된 좌표값을 상기 각 결함유형모델의 위치 정보로 저장하는 단계;를 포함하는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 1에 있어서,상기 결함유형들의 확률값을 산출하는 단계는,상기 결함유형들과의 전체 거리값을 기준으로 상기 각 결함유형별로 상대적인 거리의 비를 계산하여 상기 확률값을 산출하는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 1에 있어서,상기 확률값은,상기 부분방전모델과 상기 결함유형모델 사이의 거리값이 가까울수록 높은 확률값을 갖는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 1에 있어서,상기 결함유형들과 상기 부분방전신호 이벤트에 대한 뉴럴 네트워크 기반의 결함유형 판정 결과를 수신하는 단계;를 더 포함하는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 6에 있어서,상기 결함유형을 판정하는 단계는,상기 결함유형들의 확률값과, 상기 뉴럴 네트워크 기반의 결함유형 판정 결과로부터의 확률값들을 조합하여 상기 부분방전신호 이벤트에 대한 결함유형을 판정하는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 6에 있어서,상기 결함유형을 판정하는 단계는,상기 결함유형들의 확률값과, 상기 뉴럴 네트워크 기반의 결함유형 판정 결과로부터의 확률값들에 서로 다른 가중치를 부여하는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 8에 있어서,상기 결함유형을 판정하는 단계는,상기 서로 다른 가중치를 부여한 상기 결함유형들의 확률값과, 상기 뉴럴 네트워크 기반의 결함유형 판정 결과로부터의 확률값들을 각각 합산하여 상기 부분방전신호 이벤트에 대한 결함유형을 판정하는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 1에 있어서,상기 파라미터들은,상기 결함유형들에 대응하여 측정된 부분방전신호의 위상, 방전량, 및 방전횟수를 기초로 하여 생성된 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 1에 있어서,상기 파라미터들은,상기 위상, 상기 방전량, 및 상기 방전횟수를 변수로 하여, 위상-방전량, 위상-방전횟수, 및 방전량-방전횟수의 세가지 타입의 2차원 분포로 분류되는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 11에 있어서,상기 위상-방전량 및 상기 위상-방전횟수 타입의 파라미터들은 위상이 종속변수이므로 정극성과 부극성으로 분류되는 것을 부분방전 결함유형 판정 방법.
- 청구항 11에 있어서,상기 방전량-방전횟수 타입의 파라미터들은 종속변수가 선형 특성을 가지므로 최대값을 100으로 하여 노멀라이징(Normalizing)하는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 청구항 1에 있어서,상기 부분방전 결함유형은,입자(Particle), 플로팅(Floating), 코로나(Corona), 공동(Viod), 및 노이즈(Noise) 중 적어도 어느 하나를 포함하는 것을 특징으로 하는 부분방전 결함유형 판정 방법.
- 부분방전을 발생시키는 결함유형들을 기준으로 생성된 n개의 파라미터들에 대응하는 가상 공간을 형성하는 가상공간 구현부;부분방전신호 이벤트 발생 시, 발생된 부분방전신호에 근거하여 상기 n개의 파라미터들에 해당하는 파라미터 값들을 추출하는 부분방전신호 분석부;상기 결함유형들 각각에 대한 상기 n개의 파라미터들의 파라미터 값들을 조합하여 상기 가상 공간상에 배치된 결함유형모델들의 위치 정보와 부분방전신호에 대한 상기 n개의 파라미터들의 파라미터 값들을 조합하여 상기 가상 공간상에 배치된 부분방전모델의 위치 정보를 비교하여 상기 결함유형모델들과 상기 부분방전모델 사이의 거리값을 산출하고, 상기 거리값으로부터 상기 부분방전신호 이벤트에 대한 상기 결함유형들의 확률값을 산출하는 연산부; 및상기 부분방전신호 이벤트에 대한 상기 결함유형들의 확률값에 근거하여 상기 부분방전신호 이벤트에 대한 결함유형을 판정하는 결함유형 판정부;를 포함하는 것을 특징으로 하는 부분방전 결함유형 판정 장치.
- 청구항 15에 있어서,상기 결함유형 판정부는,상기 결함유형들과 상기 부분방전신호 이벤트에 대한 뉴럴 네트워크 기반의 결함유형 판정 결과를 수신하고, 상기 결함유형들의 확률값과, 상기 뉴럴 네트워크 기반의 결함유형 판정 결과로부터의 확률값들을 조합하여 상기 부분방전신호 이벤트에 대한 결함유형을 판정하는 것을 특징으로 하는 부분방전 결함유형 판정 장치.
- 청구항 16에 있어서,상기 결함유형 판정부는,상기 결함유형들의 확률값과, 상기 뉴럴 네트워크 기반의 결함유형 판정 결과로부터의 확률값들에 서로 다른 가중치를 부여하는 것을 특징으로 하는 부분방전 결함유형 판정 장치.
- 청구항 15에 있어서,상기 결함유형 판정부는,상기 서로 다른 가중치를 부여한 상기 결함유형들의 확률값과, 상기 뉴럴 네트워크 기반의 결함유형 판정 결과로부터의 확률값들을 각각 합산하여 상기 부분방전신호 이벤트에 대한 결함유형을 판정하는 것을 특징으로 하는 부분방전 결함유형 판정 장치.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
GB1404833.4A GB2508560B (en) | 2011-09-19 | 2012-09-19 | Method and device for determining the defect type of a partial discharge |
US14/345,163 US9658272B2 (en) | 2011-09-19 | 2012-09-19 | Method and device for determining the defect type of a partial discharge |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
KR10-2011-0094159 | 2011-09-19 | ||
KR1020110094159A KR101268761B1 (ko) | 2011-09-19 | 2011-09-19 | 부분방전 결함유형 판정 방법 및 그 장치 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2013042928A1 true WO2013042928A1 (ko) | 2013-03-28 |
Family
ID=47914616
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/KR2012/007482 WO2013042928A1 (ko) | 2011-09-19 | 2012-09-19 | 부분방전 결함유형 판정 방법 및 그 장치 |
Country Status (4)
Country | Link |
---|---|
US (1) | US9658272B2 (ko) |
KR (1) | KR101268761B1 (ko) |
GB (1) | GB2508560B (ko) |
WO (1) | WO2013042928A1 (ko) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101515231B1 (ko) * | 2013-11-22 | 2015-04-24 | 주식회사 효성 | 가스 절연 기기의 부분방전 진단 방법 |
CN103777123A (zh) * | 2014-01-27 | 2014-05-07 | 国家电网公司 | 用于gis设备的局部放电故障综合诊断方法 |
JP6548454B2 (ja) * | 2015-05-28 | 2019-07-24 | 株式会社日立パワーソリューションズ | 電気機器の診断装置、電気機器の診断システム、電気機器の診断方法およびプログラム |
JP6952575B2 (ja) * | 2017-10-31 | 2021-10-20 | 株式会社東芝 | 部分放電診断装置 |
CN109991519B (zh) * | 2019-03-08 | 2021-11-16 | 上海交通大学 | 基于神经网络和无线传感阵列的局部放电测向方法及系统 |
CN111157850B (zh) * | 2020-01-15 | 2022-06-21 | 上海电力大学 | 一种基于均值聚类的电网线路故障识别方法 |
CN111932493B (zh) * | 2020-06-28 | 2024-06-07 | 北京国网富达科技发展有限责任公司 | 一种配电网局部放电超声波检测方法及系统 |
CN112684311B (zh) * | 2021-01-30 | 2023-04-07 | 国网上海市电力公司 | 用于变压器油纸绝缘局部放电类型识别的特征量提取方法 |
KR102621378B1 (ko) * | 2021-07-06 | 2024-01-08 | 한국전력공사 | Hvdc 케이블 부분방전 진단 시스템 및 방법 |
CN113533921B (zh) * | 2021-09-15 | 2021-12-17 | 北京普源瑞新仿真科技有限公司 | 局部放电检测方法及局部放电检测装置 |
CN114091340B (zh) * | 2021-11-26 | 2024-08-02 | 华中科技大学 | 一种基于多物理场直流局部放电模型的构建与判别方法 |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20120000228A (ko) * | 2010-06-25 | 2012-01-02 | 주식회사 케이디파워 | 부분방전 검출장치 및 검출방법 |
-
2011
- 2011-09-19 KR KR1020110094159A patent/KR101268761B1/ko active IP Right Grant
-
2012
- 2012-09-19 WO PCT/KR2012/007482 patent/WO2013042928A1/ko active Application Filing
- 2012-09-19 US US14/345,163 patent/US9658272B2/en active Active
- 2012-09-19 GB GB1404833.4A patent/GB2508560B/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20120000228A (ko) * | 2010-06-25 | 2012-01-02 | 주식회사 케이디파워 | 부분방전 검출장치 및 검출방법 |
Non-Patent Citations (2)
Title |
---|
SEONG-IL, KIM ET AL.: "A Study on Optimization of Partial Discharge Pattern Recognition using Genetic Algorithm", THE KOREAN INSTITUTE OF ELECTRICAL ENGINEERS, ELECTRICAL PROPERTY, APPLICATION PART SOCIETY, JOURNAL OF FALL CONFERENCE, 2006, pages 145 - 146 * |
SUN-GEUN, GOO ET AL.: "Auto-classification of UHF partial discharge signal without phase signal", THE KOREAN INSTITUTE OF ELECTRICAL ENGINEERS, JOURNAL OF SUMMER CONFERENCE, 2005, pages 2208 - 2210 * |
Also Published As
Publication number | Publication date |
---|---|
GB201404833D0 (en) | 2014-04-30 |
KR101268761B1 (ko) | 2013-05-29 |
KR20130030585A (ko) | 2013-03-27 |
GB2508560A (en) | 2014-06-04 |
GB2508560B (en) | 2017-05-03 |
US9658272B2 (en) | 2017-05-23 |
US20140372052A1 (en) | 2014-12-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2013042928A1 (ko) | 부분방전 결함유형 판정 방법 및 그 장치 | |
WO2021002549A1 (ko) | 딥러닝 기반의 자동차 부위별 파손정도 자동 판정 시스템 및 방법 | |
WO2022114653A1 (ko) | 데이터 경계 도출 시스템 및 방법 | |
WO2014193040A1 (ko) | 센싱 데이터 분석 시스템 및 방법 | |
WO2014204179A1 (en) | Method for verifying bad pattern in time series sensing data and apparatus thereof | |
WO2010041836A2 (en) | Method of detecting skin-colored area using variable skin color model | |
WO2022240163A1 (ko) | Mems 기반 보조 지진 관측망에서 지진을 감지하는 방법, 이를 수행하기 위한 기록 매체 및 장치 | |
WO2023043215A1 (ko) | 표준 운용 수준 평가 기반 산업 설비 운용 제어 장치 및 그 동작 방법 | |
CN111443091B (zh) | 电缆线路隧道工程缺陷判断方法 | |
WO2019045147A1 (ko) | 딥러닝을 pc에 적용하기 위한 메모리 최적화 방법 | |
WO2020032506A1 (ko) | 시각 감지 시스템 및 이를 이용한 시각 감지 방법 | |
WO2023128669A1 (ko) | 부분방전 모니터링 시스템 및 부분방전 모니터링 방법 | |
WO2021033969A1 (en) | Apparatus and method for detecting defective component using infrared camera | |
WO2021040396A1 (ko) | 온도 추정 모델 결정 방법 및 장치, 온도 추정 모델이 적용된 배터리 관리 시스템 | |
WO2020050456A1 (ko) | 설비 데이터의 이상 정도 평가 방법 | |
WO2015126058A1 (ko) | 암 예후 예측 방법 | |
WO2023163305A1 (ko) | 딥러닝 기반의 보행 패턴 검출 방법 및 이를 수행하는 컴퓨터 프로그램 | |
WO2023282500A1 (ko) | 슬라이드 스캔 데이터의 자동 레이블링 방법, 장치 및 프로그램 | |
WO2022181919A1 (ko) | 가상 현실 기반의 수술 환경을 제공하는 장치 및 방법 | |
WO2020209513A1 (ko) | 부분방전 진단 및 위치 검출 방법 | |
WO2020222373A1 (ko) | 실내 측위 장치 및 방법 | |
WO2022075516A1 (ko) | 절리면 안전성 평가 장치 | |
WO2020045903A1 (ko) | Cnn을 이용하여 크기 독립적으로 물체를 검출하는 방법 및 장치 | |
WO2023282620A1 (ko) | Hvdc 케이블 부분방전 진단 시스템 및 방법 | |
WO2023128668A1 (ko) | 부분방전 모니터링 시스템 및 부분방전 모니터링 방법 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 12834066 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14345163 Country of ref document: US |
|
ENP | Entry into the national phase |
Ref document number: 1404833 Country of ref document: GB Kind code of ref document: A Free format text: PCT FILING DATE = 20120919 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1404833.4 Country of ref document: GB |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 12834066 Country of ref document: EP Kind code of ref document: A1 |