CN103678941B - The Forecasting Methodology of electrode air gap breakdown voltage - Google Patents
The Forecasting Methodology of electrode air gap breakdown voltage Download PDFInfo
- Publication number
- CN103678941B CN103678941B CN201310752299.1A CN201310752299A CN103678941B CN 103678941 B CN103678941 B CN 103678941B CN 201310752299 A CN201310752299 A CN 201310752299A CN 103678941 B CN103678941 B CN 103678941B
- Authority
- CN
- China
- Prior art keywords
- breakdown voltage
- electrode
- air gap
- voltage
- field signature
- 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.)
- Active
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses the Forecasting Methodology of a kind of electrode air gap breakdown voltage, first, measure the breakdown voltage value of different structure typical electrode the air gap, and it is interval interval with breakdown voltage to define withstanding voltage;Then, different structure typical electrode gap carried out electric Field Calculation and extracts Field signature amount, building training sample set;Then, build breakdown voltage forecast model based on training sample set, this breakdown voltage forecast model with Field signature amount be input, interval with withstanding voltage and breakdown voltage is interval for output;Finally, the breakdown voltage of breakdown voltage forecast model prediction electrode the air gap is used.The present invention is simple to operate, it was predicted that accuracy is high, and the cycle is short, low cost, it is adaptable to engineer applied, it was predicted that result can be used to instruct transmission of electricity power transformation engineering design.
Description
Technical field
The present invention relates to project of transmitting and converting electricity external insulation field, especially relate to a kind of electrode air gap breakdown voltage
Forecasting Methodology.
Background technology
The flash-over characteristic of the air gap is the important evidence of high-voltage testing room external insulation design.For between air
The flash-over characteristic of gap, research worker is mainly to typical electrode the air gap and actual project of transmitting and converting electricity the air gap two
Substantial amounts of experimental study has been carried out in class gap.For electrode the air gap, typically by typical electrode is put
Electric test, grasps the fundamental characteristics that the air gap punctures, and utilizes hitting of gap factor predictive engine the air gap
Wear characteristic, thus instruct project of transmitting and converting electricity to design.On this basis, and by engineering air gap discharge test
Check the reasonability that Insulation Coordination designs.But, rely on hitting of gap factor prediction Practical Project the air gap
Wear voltage accuracy inadequate, even if using real model experiment to be also difficult to exhaustive actual project of transmitting and converting electricity the air gap, and
The deficiency that real model experiment there is also cycle length, cost is high.
For solving the problems referred to above, the air gap discharge mechanism and scale-model investigation need to be carried out in a deep going way, and set up test and
The research means that emulation combines, the breakdown voltage of application simulation model prediction Practical Project the air gap, it is achieved
The minute design of project of transmitting and converting electricity external insulation, reduces experiment work amount required in external insulation design simultaneously.
The air gap discharge mechanism research mainly by setting up the air gap electric discharge observation method, is carried out the air gap and is put
Electric test is observed, and measures the critical physical parameter in discharge process and analyzes observed result, and disclosing each stage puts
Motor is managed, and explains the macroscopic properties that the air gap electric discharge presents.On this basis, the number of flash-over characteristic is set up
Learn phantom, the breakdown voltage of the air gap is predicted.But, the most still there are many key physical to join
Number cannot accurate description and measurement, such as, lack heightened awareness to multiple abscess regional space CHARGE DISTRIBUTION rule,
By supposing that multiple abscess geometry and region electric field constant calculate distribution of space charge and there is bigger error.Lack
Understanding to leader channel characteristic parameter, assumes that leader channel has the characteristic similar to electric arc, and unit at present
Length leader charge density is constant, it is impossible to accurately reflect the pass between guide's characteristic parameter and thermal ionization degree
System.Additionally, partial discharge mechanism is not perfect, such as, multiple abscess is thermal ionization to the main mechanism that guide converts,
Assume the critical temperature that the critical temperature that guide initiates is desorbed equal to anion at present, not yet obtain what test was measured
Confirm.The problems referred to above result in current mathematic simulated mode and are difficult to the most pre-of Air Gap Breakdown Voltage
Survey.
Summary of the invention
The deficiency existed for prior art, the invention provides that a kind of process is simple, precision is high, the cycle is short,
The Forecasting Methodology of the electrode air gap breakdown voltage of low cost.
For solving above-mentioned technical problem, the present invention adopts the following technical scheme that:
The Forecasting Methodology of electrode air gap breakdown voltage, including step:
Step 1, measures the breakdown voltage value of different structure typical electrode the air gap, uses meteorologic parameter value to repair
Destructive positive voltage value, and define, according to revised breakdown voltage value, the withstanding voltage that each typical electrode is corresponding respectively
Interval interval with breakdown voltage;
Step 2, loads withstanding voltage interval and the breakdown potential nip of its correspondence respectively to different structure typical electrode
Between, and electrode the air gap is carried out the Field signature quantity set that the electric Field Calculation each on-load voltage of acquisition is corresponding, and structure
Build training sample set;
Step 3, based on training sample set, uses artificial intelligence's mathematical method to build breakdown voltage forecast model,
This breakdown voltage forecast model is to input, to represent tolerance with the Field signature quantity set of electrode the air gap to be predicted
A and B in voltage range and breakdown voltage interval is output;
Step 4, uses the prediction of breakdown voltage forecast model to treat the breakdown voltage of electrode the air gap, and this step is entered
One step includes sub-step:
4.1 pairs of electrode on-load voltages to be predicted, carry out electric Field Calculation and obtain Field signature amount electrode the air gap
Collect and input breakdown voltage forecast model, on-load voltage initial value sets itself;
If 4.2 breakdown voltage forecast model output A, rising on-load voltage, repeated execution of steps 4.1, until
Breakdown voltage forecast model output B, on-load voltage now is the breakdown potential of electrode the air gap to be predicted
Pressure;
If 4.3 breakdown voltage forecast model output B, reduction on-load voltage, repeated execution of steps 4.1, until
Breakdown voltage forecast model output A, is now output as the on-load voltage that the breakdown voltage interval limit of B is corresponding
It is the breakdown voltage of electrode the air gap to be predicted.
Above-mentioned withstanding voltage is interval and breakdown voltage interval be respectively [(100%-a) V, 100%V), [100%V,
(100%+a) V], a is rule of thumb manually set;V is revised breakdown voltage value.
Step 2 use FEM calculation tool ANSYS electrode the air gap is carried out electric Field Calculation.
As preferably, the Field signature quantity set obtained is carried out dimension-reduction treatment, and be normalized in step 2.
Described dimension-reduction treatment method is:
Field signature amount in Field signature quantity set is divided into M class, the most only selects wherein N class
Field signature amount, N < M;Or, use correlation analysis method that the Field signature amount in Field signature quantity set is entered
Row dimension-reduction treatment;Or, use PCA that the Field signature amount in Field signature quantity set is carried out at dimensionality reduction
Reason.
Breakdown voltage forecast model constructed in step 3 is supporting vector machine model or neural network model.
Described supporting vector machine model is SVC, LIBSVM or LSSVM workbox.
The penalty factor c and kernel functional parameter g of supporting vector machine model is optimized, particularly as follows:
Thick-refined net search method or genetic algorithm or particle cluster algorithm is used to take different penalty factor c and core letter
Number parameters g, use k-folding cross-validation method obtain different predicting the outcome, take the best punishment of prediction effect because of
Sub-c and kernel functional parameter g is as optimized parameter, thus obtains the supporting vector machine model after optimization.
Compared with prior art, the present invention has the following advantages and beneficial effect:
1, predicting electrode air gap breakdown voltage based on describable physical quantity, robustness is more preferable.
2, the electrode the air gap discharge physics process of complexity is avoided, it was predicted that process is simple, accuracy is high,
It is applicable to engineer applied.
3, the measured data only needing a small amount of Air Gap Breakdown Voltage and measurement ambient parameter is pre-to breakdown voltage
Survey model is trained, low cost, and the cycle is short;And the mathematical physics of traditional prediction Air Gap Breakdown Voltage
Model needs finely to measure critical physical parameter, not only required measurement apparatus expensive, and predetermined period is long.
4, predicting the outcome of present invention acquisition can be used to instruct transmission of electricity power transformation engineering design.
Accompanying drawing explanation
Fig. 1 is a kind of particular flow sheet of the inventive method;
Fig. 2 is the prediction principle figure of the inventive method;
Fig. 3 is the searching process of different parameters optimization method in the embodiment of the present invention;
Fig. 4 is one group of ball gap test sample breakdown voltage predictive value and the contrast of test value in the embodiment of the present invention.
Detailed description of the invention
Below by embodiment, and combine accompanying drawing, technical solution of the present invention is further described in detail.
Seeing Fig. 1, the present invention includes step:
Step 1, carries out pressure test respectively and obtains breakdown voltage value different structure typical electrode the air gap,
Use meteorologic parameter correction breakdown voltage value according to standard GB/T/T 16927.1-2011, and according to revising after
Breakdown voltage value define that withstanding voltage corresponding to each electrode is interval and breakdown voltage is interval respectively.
The withstanding voltage of definition is interval and breakdown voltage interval is designated as-1 and 1 respectively, predicts mould as breakdown voltage
The output of type.In the present embodiment, and definition [(100%-a) V, 100%V) it is that withstanding voltage is interval, definition
[100%V, (100%+a) V] is that breakdown voltage is interval.A rule of thumb value, in being originally embodied as, a
Take 10%;V is revised breakdown voltage value.
Step 2, loads withstanding voltage interval and the breakdown potential nip of its correspondence respectively to different structure typical electrode
Between, electrode the air gap is carried out electric Field Calculation and obtains the Field signature quantity set that each on-load voltage is corresponding, as instruction
Practice sample set.
Withstanding voltage interval and breakdown voltage that electrode loads definition successively are interval, use FEM calculation instrument
ANSYS carries out electric Field Calculation to electrode the air gap, and extracts Field signature amount acquisition Field signature quantity set,
It is shown in Table 1, using Field signature quantity set as training sample set.
In table 1, listed Field signature amount sum is 50, and these Field signature amounts are from being spatially divided into whole district
Territory, discharge channel, electrode surface and discharge path 4 category feature amount;These Field signature amounts again may be used from dimension
(dimension is: V/m), electric field energy (dimension is J), (dimension is energy density: J/m to be divided into electric field3)、
(dimension is surface area: m2), (dimension is length: m), (dimension is electric-force gradient: V/m2) and ratio
Parameter (dimensionless) 7 category feature amount.
Table 1 Field signature quantity set
For reducing computational complexity, improving computational efficiency, Field signature quantity set is dropped by this detailed description of the invention
Dimension processes, and concrete dimension-reduction treatment method can take following one according to actual needs: only consider a certain class or
A few class Field signature amounts;Or use correlation analysis method or PCA that Field signature quantity set is carried out dimensionality reduction
Process.After the Field signature quantity set normalization after dimension-reduction treatment, as the input of training sample.
Correlation analysis method and PCA are dimension-reduction treatment methods ripe in the art.Dependency
Analytic process is a kind of structural feature subset method, uses statistical correlation method, selects the feature strong with exporting dependency
Amount, rejects the characteristic quantity weak with exporting dependency, rejects the characteristic quantity that between characteristic quantity, dependency is strong simultaneously.Main one-tenth
Dividing analytic process is a kind of transform characteristics space law, big more than 85-95% or eigenvalue according to accumulative variance contribution ratio
Main constituent is chosen in 1.
Step 3, based on training sample set, uses artificial intelligence's mathematical method to build breakdown voltage forecast model,
This breakdown voltage forecast model is to input, to represent tolerance with the Field signature quantity set of electrode the air gap to be predicted
Voltage range and the numerical value-1 in breakdown voltage interval and 1 are output.
Breakdown voltage forecast model can use supporting vector machine model or neural network model.In this specific embodiment
Supporting vector machine model is used to build breakdown voltage forecast model.
Supporting vector machine model type, such as, optional SVC, LIBSVM is selected according to Field signature quantity set
Or LSSVM workbox.The present embodiment selects LIBSVM workbox, because LIBSVM workbox can have
Effect solves classification problem and cross validation parameter selects.The breakdown voltage forecast model built, by electricity characteristic quantity
Collection is constituted, including input and output, with the electrode to be predicted Field signature quantity set after dimension-reduction treatment for input,
Withstanding voltage with step 1 definition is interval and breakdown voltage is interval for output.
Use cross validation thought and thick-refined net search method Support Vector Machines Optimized model parameter, refer specifically to excellent
Change the penalty factor c and kernel functional parameter g of supporting vector machine model.The present embodiment uses thick-refined net search
Method takes different penalty factor c and kernel functional parameter g, uses k-folding cross-validation method to obtain different prediction knots
Really, take prediction effect best, the parameter that i.e. error is minimum as optimal value, thus obtain the support after optimization to
Amount machine model.Originally, in being embodied as, k takes 3.Except thick-refined net search method, it would however also be possible to employ genetic algorithm
Or particle cluster algorithm chess game optimization supporting vector machine model parameter.
Step 4, uses the breakdown voltage of the breakdown voltage forecast model prediction electrode the air gap after optimizing.
The air gap unknown to breakdown voltage loads initial voltage U0, carries out electric Field Calculation and extracts electric field spy
Levy quantity set, the Field signature quantity set extracted is carried out dimension-reduction treatment and after normalization, inputs breakdown voltage prediction mould
Type.If breakdown voltage forecast model output-1, then after raising on-load voltage, i.e. on-load voltage U=U0+dU,
The air gap is carried out electric Field Calculation again, and extracts Field signature quantity set input breakdown voltage forecast model, until
Breakdown voltage forecast model output 1, on-load voltage U now is breakdown voltage Uc of the air gap.If
Breakdown voltage forecast model output 1, then reduce on-load voltage, i.e. on-load voltage U=U0-dU, then to air between
Gap carries out electric Field Calculation, and extracts Field signature quantity set input breakdown voltage forecast model, until breakdown voltage is pre-
Surveying model output-1, on-load voltage corresponding to breakdown voltage interval limit being now output as 1 is breakdown voltage,
I.e. breakdown voltage Uc of the air gap is current on-load voltage U and dU sum.This Principle of Process can be found in figure
2, dU can sets itself and adjustment.
Electrode air gap breakdown voltage prediction is transformed into classification problem by regression problem by the present invention, i.e. passes through core
Regression problem is transformed to higher-dimension Hilbert space by the non-linear map in Function feature space, then at height
Dimension carries out optimal hyperlane classification to the sample after mapping in Hilbert space, the principle of algorithm of support vector machine
Can be found in " Algorithmic Design & Analysis of support vector machine " that Yang Xiaowei and Hao Zhifeng writes with internal arithmetic process.
The inventive method is applicable to the electrode of arbitrary structures.
Technical solution of the present invention and beneficial effect thereof is further illustrated below as a example by ball gap.
(1) training sample set and test sample collection are set up.
By ball gap is carried out pressure test, it is thus achieved that the breakdown potential that different-diameter D is corresponding with the ball gap of spacing d
Pressure value, and breakdown voltage value is modified, it is thus achieved that sample set, it is shown in Table 2, wherein, font-weight data are
Training sample set, totally 15 training samples, remaining is test sample collection, 5 groups of totally 25 test samples, its
In, test sample integrates 1 as the ball gap breakdown voltage Value Data of D=6.25cm, d=1.5~2.2cm;Test sample
Integrate 2 as the ball gap breakdown voltage Value Data of d=2.2cm, D=6.25~50cm;Test sample integrate 3 as D=10cm,
The ball gap breakdown voltage Value Data of d=2.6~4.0cm;Test sample integrates 4 as d=4.0cm, D=12.5~75cm
Ball gap breakdown voltage Value Data;Test sample integrates 5 and punctures as the ball gap of d=4.5~8.0cm, D=10~75cm
Voltage value data.In table 2, D is electrode diameter, and d is electrode gap distance.
(2) Field signature quantity set extracts.
Ball gap carried out electric Field Calculation and extracts Field signature quantity set, being shown in Table 1.It is respectively adopted different dimensionality reduction sides
Formula carries out dimension-reduction treatment to Field signature quantity set.Use correlational analysis method that Field signature quantity set is down to 26 dimensions.
Use PCA to take the main constituent more than 95% of the accumulative variance contribution ratio, Field signature quantity set is down to
15 dimensions.Reject Wr25d、Wr25d、Vr25d、Vr7d、Wr25d、Wr7d、L25、L7、Lr25、Lr7With electric field ladder
Field signature quantity set is down to 33 dimensions by degree class class Field signature amount.Only reject Wr25d、Wr25d、Vr25d、Vr7d、
Wr25d、Wr7d、L25、L7、Lr25And Lr7Field signature quantity set is down to 40 dimensions by Field signature amount.
3) using the Field signature quantity set after dimension-reduction treatment as the input of breakdown voltage forecast model, and to puncturing
Voltage-prediction model parameter is optimized.
Based on 3 folding cross validation thoughts, it is respectively adopted thick-refined net search method, genetic algorithm, population calculation
The penalty factor c and kernel functional parameter g of method chess game optimization supporting vector machine model, the process of 3 kinds of optimization methods
Seeing the searching process that Fig. 3, Fig. 3 (a) are thick-refined net search method, cross validation accuracy rate is 97.7778%;
Table 2 training sample and test sample
Fig. 3 (b) is the fitness curve of genetic algorithm, and population scale is 20, and terminating algebraically is 100, and intersection is tested
Card accuracy rate is 97.7778%;Fig. 3 (c) is the fitness curve of particle cluster algorithm, and population scale is 30,
Terminating algebraically is 200, and Studying factors c1 takes 1.5, and Studying factors c2 takes 1.7, and cross validation accuracy rate is
97.7778%.Genetic algorithm and particle cluster algorithm are with classification accuracy as fitness function.
The optimizing result of 3 kinds of optimization methods, optimal time and error criterion are shown in Table 3, and wherein, SSE is error
Quadratic sum, MSE is mean square error, and MAPE is mean absolute percentage error, and MSPE is mean square percentage
Ratio error.From table 3 it can be seen that the result of 3 kinds of optimization methods is more or less the same, and genetic algorithm and population
Algorithm is heuritic approach, and each optimizing result is the most different, therefore slightly-refined net search method is the preferred of the present invention
Optimization method.
The different optimizing result of optimization method of table 3, optimal time and error criterion
Slightly-refined net search method | Genetic algorithm | Particle cluster algorithm | |
Optimum penalty factor c | 415.8732 | 382.173 | 386.166 |
Optimum kernel functional parameter g | 0.2973 | 0.319357 | 0.324232 |
Optimal time (s) | 16.8 | 18.0 | 176.74 |
SSE (test sample collection 1-4) | 108.5158 | 105.1686 | 108.9318 |
MSE (test sample collection 1-4) | 0.5209 | 0.5128 | 0.5219 |
MAPE (test sample collection 1-4) | 0.018 | 0.018 | 0.0185 |
MSPE (test sample collection 1-4) | 0.0055 | 0.0055 | 0.0055 |
MAPE (test sample collection 5) | 0.072 | 0.074 | 0.074 |
Optimized parameter under table 4 different characteristic dimension and error criterion
Under different characteristic dimension, use breakdown voltage forecast model that thick-refined net search method optimization obtains
Excellent parameter and error criterion are shown in Table 4.From table 4, it can be seen that under 26 dimension Field signature quantity sets, it was predicted that precision is
Height, Fig. 4 is shown in the now breakdown voltage predictive value of test sample collection 5 and the contrast of test value.
Specific embodiment described above is only to present invention spirit explanation for example.Technology belonging to the present invention
Described specific embodiment can be made various amendment or supplements or use class by the technical staff in field
As mode substitute, but without departing from the spirit of the present invention or surmount model defined in appended claims
Enclose.
Claims (7)
1. the Forecasting Methodology of electrode air gap breakdown voltage, it is characterised in that include step:
Step 1, measures the breakdown voltage value of different structure typical electrode the air gap, uses meteorologic parameter value to repair
Destructive positive voltage value, and define, according to revised breakdown voltage value, the withstanding voltage that each typical electrode is corresponding respectively
Interval interval with breakdown voltage;
Step 2, loads withstanding voltage interval and the breakdown potential nip of its correspondence respectively to different structure typical electrode
Between, and electrode the air gap is carried out the Field signature quantity set that the electric Field Calculation each on-load voltage of acquisition is corresponding, and structure
Build training sample set;
Step 3, based on training sample set, uses artificial intelligence's mathematical method to build breakdown voltage forecast model,
This breakdown voltage forecast model is to input, to represent tolerance with the Field signature quantity set of electrode the air gap to be predicted
A and B in voltage range and breakdown voltage interval is output;
Constructed breakdown voltage forecast model is supporting vector machine model or neural network model;
Step 4, uses the prediction of breakdown voltage forecast model to treat the breakdown voltage of electrode the air gap, and this step is entered
One step includes sub-step:
4.1 pairs of electrode on-load voltages to be predicted, carry out electric Field Calculation and obtain Field signature amount electrode the air gap
Collect and input breakdown voltage forecast model, on-load voltage initial value sets itself;
If 4.2 breakdown voltage forecast model output A, rising on-load voltage, repeated execution of steps 4.1, until
Breakdown voltage forecast model output B, on-load voltage now is the breakdown potential of electrode the air gap to be predicted
Pressure;
If 4.3 breakdown voltage forecast model output B, reduction on-load voltage, repeated execution of steps 4.1, until
Breakdown voltage forecast model output A, is now output as the on-load voltage that the breakdown voltage interval limit of B is corresponding
It is the breakdown voltage of electrode the air gap to be predicted.
2. the Forecasting Methodology of electrode air gap breakdown voltage as claimed in claim 1, it is characterised in that:
Described withstanding voltage is interval and breakdown voltage interval be respectively [(100%-a) V, 100%V),
[100%V, (100%+a) V], a is rule of thumb manually set;V is revised breakdown voltage value.
3. the Forecasting Methodology of electrode air gap breakdown voltage as claimed in claim 1, it is characterised in that:
Step 2 use FEM calculation tool ANSYS electrode the air gap is carried out electric Field Calculation.
4. the Forecasting Methodology of electrode air gap breakdown voltage as claimed in claim 1, it is characterised in that:
The Field signature quantity set obtained in step 2 is carried out dimension-reduction treatment, and is normalized.
5. the Forecasting Methodology of electrode air gap breakdown voltage as claimed in claim 4, it is characterised in that:
Described dimension-reduction treatment method is:
Field signature amount in Field signature quantity set is divided into M class, the most only selects wherein N class
Field signature amount, N < M;Or, use correlation analysis method that the Field signature amount in Field signature quantity set is entered
Row dimension-reduction treatment;Or, use PCA that the Field signature amount in Field signature quantity set is carried out at dimensionality reduction
Reason.
6. the Forecasting Methodology of electrode air gap breakdown voltage as claimed in claim 1, it is characterised in that:
Described supporting vector machine model is SVC, LIBSVM or LSSVM workbox.
7. the Forecasting Methodology of electrode air gap breakdown voltage as claimed in claim 1, it is characterised in that:
The penalty factor c and kernel functional parameter g of supporting vector machine model is optimized, particularly as follows:
Thick-refined net search method or genetic algorithm or particle cluster algorithm is used to take different penalty factor c and core letter
Number parameters g, use k-folding cross-validation method obtain different predicting the outcome, take the best punishment of prediction effect because of
Sub-c and kernel functional parameter g is as optimized parameter, thus obtains the supporting vector machine model after optimization.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310752299.1A CN103678941B (en) | 2013-12-31 | 2013-12-31 | The Forecasting Methodology of electrode air gap breakdown voltage |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310752299.1A CN103678941B (en) | 2013-12-31 | 2013-12-31 | The Forecasting Methodology of electrode air gap breakdown voltage |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103678941A CN103678941A (en) | 2014-03-26 |
CN103678941B true CN103678941B (en) | 2016-08-17 |
Family
ID=50316471
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310752299.1A Active CN103678941B (en) | 2013-12-31 | 2013-12-31 | The Forecasting Methodology of electrode air gap breakdown voltage |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103678941B (en) |
Families Citing this family (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104391233B (en) * | 2014-11-21 | 2018-03-20 | 华北电力大学(保定) | A kind of the air gap breakdown test platform and the air gap breakdown test method |
CN104880650A (en) * | 2015-05-26 | 2015-09-02 | 武汉大学 | Air gap breakdown voltage prediction method |
CN105678014B (en) * | 2016-01-30 | 2018-11-23 | 武汉大学 | A kind of energy storage characteristic of gas medium and discharge voltage prediction technique |
CN106021848B (en) * | 2016-05-04 | 2018-12-21 | 国网重庆市电力公司电力科学研究院 | A kind of extra high voltage direct current transmission line ground total electric field shielding analysis method |
CN105977948B (en) * | 2016-05-16 | 2018-07-10 | 山东大学 | Multivariable effect dive towards single-pole reclosing supplies arc characteristic Forecasting Methodology |
CN107085172A (en) * | 2017-04-11 | 2017-08-22 | 华南理工大学 | A kind of Oil-Paper Insulation breakdown voltage Forecasting Methodology based on SVMs |
CN107315903B (en) * | 2017-05-23 | 2020-05-08 | 浙江大学 | Intelligent electric field analysis system |
CN108090311B (en) * | 2018-01-24 | 2021-11-02 | 武汉大学 | Air breakdown voltage prediction method based on path feature set and support vector regression |
CN107992713B (en) * | 2018-01-24 | 2021-02-19 | 武汉大学 | Combined air gap breakdown voltage prediction method |
CN108414897B (en) * | 2018-01-24 | 2019-08-02 | 武汉大学 | A kind of helicopter livewire work combination the air gap discharge voltage prediction technique |
CN110929434B (en) * | 2018-09-19 | 2022-04-15 | 武汉大学 | Simulation calculation method for development process of combined gap impulse discharge |
CN109408937B (en) * | 2018-10-17 | 2023-04-28 | 南方电网科学研究院有限责任公司 | Method and device for calculating gap coefficient of air gap impulse discharge voltage |
CN109799454A (en) * | 2018-12-29 | 2019-05-24 | 西安交通大学 | Generator stator insulation residual breakdown strength prediction technique based on particle group optimizing |
CN110348104A (en) * | 2019-07-04 | 2019-10-18 | 西安交通大学 | The flashover voltage of the insulator prediction technique based on Support vector regression |
CN111077418B (en) * | 2019-12-25 | 2021-08-10 | 华南理工大学 | Virtual experiment system and method for simulating breakdown voltage of non-uniform electric field |
CN111896851B (en) * | 2020-08-04 | 2023-04-14 | 云南电力技术有限责任公司 | Method and system for predicting breakdown voltage of short air gap and storage medium |
CN114325278B (en) * | 2022-03-16 | 2022-07-08 | 广东电网有限责任公司东莞供电局 | Air gap tolerance voltage measuring and calculating method and device, computer equipment and medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7379924B1 (en) * | 2004-12-01 | 2008-05-27 | Advanced Micro Devices, Inc. | Quantifying and predicting the impact of line edge roughness on device reliability and performance |
CN101187686A (en) * | 2007-12-19 | 2008-05-28 | 中国电力科学研究院 | Tube bus corona inception voltage and extinction voltage test judging method |
-
2013
- 2013-12-31 CN CN201310752299.1A patent/CN103678941B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7379924B1 (en) * | 2004-12-01 | 2008-05-27 | Advanced Micro Devices, Inc. | Quantifying and predicting the impact of line edge roughness on device reliability and performance |
CN101187686A (en) * | 2007-12-19 | 2008-05-28 | 中国电力科学研究院 | Tube bus corona inception voltage and extinction voltage test judging method |
Also Published As
Publication number | Publication date |
---|---|
CN103678941A (en) | 2014-03-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103678941B (en) | The Forecasting Methodology of electrode air gap breakdown voltage | |
CN103675629B (en) | Based on the electrode discharge inception voltage Forecasting Methodology of Field signature amount | |
Eisenhower et al. | A methodology for meta-model based optimization in building energy models | |
Qiu et al. | A prediction method for breakdown voltage of typical air gaps based on electric field features and support vector machine | |
CN110598854A (en) | GRU model-based transformer area line loss rate prediction method | |
CN106443381A (en) | Dynamic electric arc model building method and system | |
CN103246762A (en) | Method of comprehensive evaluation for simulation credibility of electric propulsion system | |
CN110442911B (en) | High-dimensional complex system uncertainty analysis method based on statistical machine learning | |
Qiu et al. | Energy storage features and a predictive model for switching impulse flashover voltages of long air gaps | |
CN113743651A (en) | Electronic voltage transformer error state prediction method based on machine learning and SVM algorithm | |
CN110147835A (en) | Resisting shear strength of reinforced concrete beam-column joints prediction technique based on grad enhancement regression algorithm | |
Qiu et al. | A method for breakdown voltage prediction of short air gaps with atypical electrodes | |
CN114583767B (en) | Data-driven wind power plant frequency modulation response characteristic modeling method and system | |
Ge et al. | Computation of breakdown voltage of long rod-plane air gaps in large temperature and humidity range under positive standard switching impulse voltage | |
CN105808962A (en) | Assessment method considering voltage probabilities of multiple electric power systems with wind power output randomness | |
Shen et al. | Nonlinear optimization strategy for state of power estimation of lithium-ion batteries: A systematical uncertainty analysis of key impact parameters | |
CN108197381B (en) | Parameter identification method based on optimizing spatial shape analysis | |
CN108233356B (en) | Consistency evaluation method and evaluation platform for photovoltaic inverter controller | |
CN106055761B (en) | A kind of load model parameters method of calibration based on decision tree classification | |
CN105868544A (en) | A method of analyzing insulating material performance by using three-parameter Weibull distribution to process flashover voltages | |
CN116090390A (en) | FINFET device direct current characteristic prediction method based on deep learning | |
Sha et al. | Prediction of commercial building lighting energy consumption based on EPSO-BP | |
Liu et al. | Machine learning approaches for estimating building energy consumption | |
Neuner | Model Predictive control for electric Drives | |
CN105842535A (en) | Harmonic wave main characteristic group screening method based on similar characteristic fusion |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
CB03 | Change of inventor or designer information |
Inventor after: Huang Daochun Inventor after: Ruan Jiangjun Inventor after: Shu Shengwen Inventor after: Pu Ziheng Inventor before: Ruan Jiangjun Inventor before: Shu Shengwen Inventor before: Huang Daochun Inventor before: Pu Ziheng |
|
COR | Change of bibliographic data | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |