[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

CN110308339B - Frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rule - Google Patents

Frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rule Download PDF

Info

Publication number
CN110308339B
CN110308339B CN201910291044.7A CN201910291044A CN110308339B CN 110308339 B CN110308339 B CN 110308339B CN 201910291044 A CN201910291044 A CN 201910291044A CN 110308339 B CN110308339 B CN 110308339B
Authority
CN
China
Prior art keywords
evidence
failure
capacitor
value
reference value
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
Application number
CN201910291044.7A
Other languages
Chinese (zh)
Other versions
CN110308339A (en
Inventor
高海波
廖林豪
徐晓滨
林治国
何业兰
盛晨兴
王�琦
武美君
杨再明
徐宏东
胡海斌
左文博
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University of Technology WUT
Original Assignee
Wuhan University of Technology WUT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan University of Technology WUT filed Critical Wuhan University of Technology WUT
Priority to CN201910291044.7A priority Critical patent/CN110308339B/en
Publication of CN110308339A publication Critical patent/CN110308339A/en
Application granted granted Critical
Publication of CN110308339B publication Critical patent/CN110308339B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • 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
    • G01R31/08—Locating faults in cables, transmission lines, or networks
    • G01R31/088—Aspects of digital computing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/15—Correlation function computation including computation of convolution operations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Computing Systems (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • Algebra (AREA)
  • Computer Hardware Design (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Testing Electric Properties And Detecting Electric Faults (AREA)
  • Fixed Capacitors And Capacitor Manufacturing Machines (AREA)

Abstract

The invention relates to a voltage source type frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rules. The input of the established evidence reasoning fault diagnosis model is two-dimensional voltage characteristics, and the output is reliability distribution of failure fault grades. And establishing a fault reliability distribution matrix according to a given voltage characteristic reference value, converting the input into a diagnosis evidence by using the matrix, calculating a reliability factor of the diagnosis evidence, fusing the two-dimensional input activated evidence, and estimating the level of the capacitance failure fault from a fusion result. And (3) carrying out parameter optimization on the established diagnosis model, and estimating the fault level of the capacitor of the test sample based on the optimized diagnosis model to realize fault diagnosis of the capacitor.

Description

Frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rule
Technical Field
The invention relates to the field of power system fault diagnosis, in particular to a voltage source type frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rules.
Background
With the continuous development of electric power systems, more and more application occasions are provided, and huge economic benefits are brought to the development of various industries. The capacitor is used as a basic element of the voltage source type frequency converter, plays roles of filtering, energy storage, tuning and the like, and particularly plays a role of stabilizing the voltage of the direct current bus on the direct current bus. As the operating time becomes longer, the capacitor ages and the capacitance decreases. At the moment, the voltage source type frequency converter still can work, but the performance can be gradually reduced, the tiny fault is difficult to identify and is difficult to activate the system protection, and the service life of the equipment is influenced after long-term operation. Therefore, the voltage source type frequency converter direct-current bus capacitor is effectively monitored, the abnormal working condition of the voltage source type frequency converter can be found in time, and the voltage source type frequency converter is maintained and overhauled in a targeted mode according to the abnormal occurrence degree.
Disclosure of Invention
The invention aims to provide a method for detecting failure fault levels of a direct-current bus capacitor of a voltage source type frequency converter based on evidence reasoning rules.
In order to achieve the purpose, the fault diagnosis method of the direct-current bus capacitor of the voltage source type frequency converter based on the evidence reasoning rule comprises the following steps:
step (1) collecting a voltage signal of a direct current bus of a voltage source type frequency converter;
extracting a characteristic value of a voltage signal, setting three capacitance failure fault levels to form a complete sample set, and randomly dividing the sample set into a training set and a testing set;
converting the data in the training set into a reliability distribution matrix corresponding to a given reference value;
converting the reliability distribution matrix to obtain evidence of the failure grade of the capacitor;
step (5) calculating a reliability factor of the evidence;
step (6) performing weighted fusion on the evidence of the capacitor failure fault grade according to the reliability factor, and initially estimating and training the capacitor failure fault grade;
step (7) an optimization objective function is set according to the estimated value of the capacitor failure fault level and the mean square error of the set capacitor failure fault level, and optimized parameters are determined to establish an evidence reasoning parameter optimization model;
and (8) estimating the failure level of the sample capacitor in the test set through an evidence reasoning parameter optimization model.
In connection with the above technical scheme, in the step (1), a voltage source type converter is taken as an example; establishing three capacitance failure levels, wherein the capacitance level 1 of the capacitor is C17500 e-6F; level 2 capacitance of C190% of (i), i.e. C26750 e-6F; level 3 capacitance of C180% of (i), i.e. C36000 e-6F; collecting 8 seconds of direct current bus voltage instantaneous value V at sampling frequency of 50kHz in the running process of equipmentDC。
3. The fault diagnosis method according to claim 1, wherein in the step (2), the voltage signal V of the step (1) is obtainedDCVoltage peak-to-peak value V ofPPPresetting a window length to obtain V of multiple data windowsPPAverage value to obtain VP(t); will VDCCalculating a root mean square value V by a preset number of electric periods and a certain number of sampling pointsRMSAnd obtaining V after normalization treatmentR(t)；
Figure GDA0002893668290000021
Figure GDA0002893668290000022
Wherein
Figure GDA0002893668290000023
Are respectively input characteristic signals VP(t) and VR(t) minimum and maximum values; the failure grade of the capacitor is recorded as C (t), and C (t) is [1,2,3](ii) a Will VP(t)、VR(T) and C (T) are expressed as a sample set TA＝{[VP(t),VR(t),C(t)]1,2, …, Ts }, where [ V ═ tP(t),VR(t),C(t)]Is a sample vector.
Following the above technical scheme, in step (3), the sample set T is sampledAData of [ V ]P(t),VR(t),C(t)]Converting a given reference value into a reliability distribution matrix; wherein the reference value set D ═ D of the capacitance failure fault level n1, …, N, wherein N is the reference value number of the capacitor failure fault grade; voltage signal ViInput set of reference values
Figure GDA0002893668290000024
i＝P,R，JiAs a voltage signal ViThe number of reference values of (a); the method comprises the following steps:
(4.1) calculation of Vi(t) to a reference value
Figure GDA0002893668290000025
Is formulated as
Figure GDA0002893668290000026
Figure GDA0002893668290000027
αi,j'＝0j'＝1,...,Ji,j'≠j,j+1 (4-1c)
αi,jRepresents Vi(t) to a reference value
Figure GDA0002893668290000028
The confidence of (2);
(4.2) similarly calculating C (t) vs. reference value DnIs formulated as
TO(C(t))＝{(Dn,γn)|n＝1,...,N} (4-2a)
Figure GDA0002893668290000029
γn'＝0n'＝1,...,N,n'≠n,n+1 (4-2c)
γnDenotes C (t) vs. reference value DnThe confidence of (2);
(4.3) sampling the sample set T according to the steps (4.1) and (4.2)ACalculating a confidence distribution (α) for all data sets in the data seti,jγn,αi,j+1γn,αi,jγn+1,αi,j+1γn+1) In which α isi,jγnRepresents a sample pair (V)i(t), C (t) wherein the input values correspond to reference values
Figure GDA0002893668290000031
While C (t) corresponds to the reference value DnThe comprehensive similarity of (2); and constructing Table 1, wherein an,jRepresents all Vi(t) corresponding to the reference value
Figure GDA0002893668290000032
And C (t) corresponds to the reference value DnSample pair (V)i(t), C (t) the sum of the integrated similarities,
Figure GDA0002893668290000033
denotes all C (t) corresponding reference values DnThe sum of the pair of integrated similarities of (a),
Figure GDA0002893668290000034
represents all Vi(t) corresponding to the reference value
Figure GDA0002893668290000035
Of the sample pair of comprehensive similarity, and
Figure GDA0002893668290000036
TABLE 1 (V)iReference value confidence coefficient distribution statistical table of (t), C (t)
Figure GDA0002893668290000037
In step (4), according to table 1 in step (3), the likelihood normalization is performed on the table by the following formula:
Figure GDA0002893668290000038
obtaining a reference value
Figure GDA0002893668290000039
Evidence of (A) is
Figure GDA00028936682900000310
Then, an evidence table 2 was constructed based on the formulae (5-1) and (5-2),
TABLE 2 inputs ViEvidence table of
Figure GDA00028936682900000311
Figure GDA0002893668290000041
Following the above protocol, in step (5), evidence
Figure GDA0002893668290000042
The reliability factor of (2) is calculated from the spearman rank correlation coefficient, and the formula is as follows
di＝Vi(t)-C(t) (6-1)
Figure GDA0002893668290000043
Wherein d isiIs the difference between the voltage characteristic value and the failure level of the capacitor, riIs a reliability factor of the evidence.
According to the technical scheme, in the step (6), the evidence table 2 and the reliability factor obtained in the steps (4) and (5) are utilized, and the failure level of the capacitor is estimated through an evidence reasoning rule
Figure GDA0002893668290000044
The method comprises the following steps:
(7.1) for input Vi(t), each set of evidence corresponding to the input
Figure GDA0002893668290000045
And
Figure GDA0002893668290000046
will be activated, then the value V is inputi(t) the final evidence can be obtained
Figure GDA0002893668290000047
And
Figure GDA0002893668290000048
weighting to obtain:
ei＝{(Dn,pn,i),n＝1,...,N} (7-1a)
Figure GDA0002893668290000049
(7.2) obtaining V from the formulae (7-1a) and (7-1b)p(t) and VREvidence of (t) e1And e2Setting an initial evidence weight wi＝riAnd fusing the images by using an evidence reasoning rule to obtain a fusion result, wherein the formula is as follows:
O(V(t))＝{(Dn,pn,e(2)),n＝1,...,N} (7-2a)
Figure GDA00028936682900000410
Figure GDA00028936682900000411
(7.3) estimating the failure level of the capacitor by the formula (7-3) according to the fusion result of the step (7.2)
Figure GDA0002893668290000051
Figure GDA0002893668290000052
In step (7), the mean square error between the estimated value of the capacitance failure level and the true value is set as an objective function:
Figure GDA0002893668290000053
s.t.0≤wi≤1,i＝1,2 (8-2a)
Figure GDA0002893668290000054
simultaneous determination of parameter sets
Figure GDA0002893668290000055
wiWeight indicating evidence, and other parameters are set to
Figure GDA0002893668290000056
And establishing an optimization model to optimize the objective function, and obtaining an optimal parameter set P after training.
And (5) acquiring the direct current bus voltage value of the existing voltage source type frequency converter in the step (8), and repeating the steps (3) to (6) again to obtain a more accurate estimated value of the capacitor failure fault level
Figure GDA0002893668290000057
According to the method for identifying the failure fault grades of the direct-current bus capacitor of the voltage source type frequency converter based on evidence reasoning, the voltage characteristic value is extracted and three failure fault grades of the capacitor are set according to the collected direct-current voltage data; converting the sample data into a confidence coefficient distribution form of the reference value by using a given input and output reference value, and constructing a confidence coefficient distribution table according to the confidence coefficient distribution form; then, acquiring evidences of failure grades of each capacitor according to the table, and establishing an evidence matrix table; calculating a reliability factor of the evidence according to the input and output Spireman grade correlation coefficient; fusing evidences according to an evidence reasoning rule and obtaining an estimation result of the failure level of the capacitor; establishing a target function training optimization parameter set, bringing in a training sample, and repeating the steps to obtain an estimation result of the capacitance failure fault level; therefore, the diagnosis of the failure fault level of the direct current bus capacitor of the voltage source type frequency converter is realized.
Drawings
FIG. 1 is a flow chart of a fault diagnosis method for a DC bus capacitor of a voltage source type frequency converter based on evidence reasoning rules according to the present invention;
FIG. 2 is a voltage signal collected during the practice of the method of the present invention;
FIG. 3a is a schematic of the peak-to-peak voltage in the practice of the method of the present invention;
FIG. 3b is a schematic diagram of the RMS value of the voltage in the practice of the method of the invention;
FIG. 3c is a schematic diagram of the corresponding voltage failure fault levels in the practice of the method of the present invention;
FIG. 4 is a graph of the capacitance failure fault level and the true capacitance failure fault level of the test set estimated by the optimized model in the method embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Taking an NXI _07305 type converter as an example, the fault diagnosis method of the direct-current bus capacitor of the voltage source type frequency converter based on the evidence reasoning rule comprises the following steps:
(1) establishing three capacitance failure levels, wherein the capacitance level 1 of the capacitor is C17500 e-6F; level 2 capacitance of C190% of (i), i.e. C26750 e-6F; level 3 capacitance of C180% of (i), i.e. C36000 e-6F. Collecting 8 seconds of direct current bus voltage instantaneous value V at sampling frequency of 50kHz in the running process of equipmentDC。
(2) Obtaining step (1) VDCVoltage peak-to-peak value V ofPP(Peak-to-Peak value, as shown in FIG. 3 a) and take V for a window of 400 data points with a window length of 1000 data points at 0.02 secondsPPAverage value to obtain VP(t); will VDCCalculating a root mean square value V by using 100 electric periods and 1000 sampling pointsRMS(RMS value, as shown in FIG. 3 b), and then normalized to obtain VR(t)；
Figure GDA0002893668290000061
Figure GDA0002893668290000062
Wherein
Figure GDA0002893668290000063
Are respectively input characteristic signals VP(t) and VR(t) minimum and maximum values; the failure grade of the capacitor is recorded as C (t), and C (t) is [1,2,3]As shown in fig. 3 c; will VP(t)、VR(T) and C (T) are expressed as a sample set TA＝{[VP(t),VR(t),C(t)]1,2, …, Ts 960; wherein [ V ]P(t),VR(t),C(t)]Is a sample vector;
(3) set VP(t) set of reference values A1-7.68, 8.06,8.47, 8.70,9.13,9.55,10.03,10.48,10.95 }; total J19 reference values; vR(t) set of reference values A20.07,0.14,0.18,0.25,0.27,0.35,0.46,0.40,0.64, and J in total29 reference values; the reference value set D of c (t) {1,2,3 }; n is 3 reference values in total; will TAData in sample set [ V ]P(t),VR (t),C(t)]Converting the given reference value into a tabular form of reference value confidence, comprising the following steps:
(3.1) calculation of Vi(t) to a reference value
Figure GDA0002893668290000064
Is formulated as
Figure 191769DEST_PATH_IMAGE002
Figure GDA0002893668290000066
αi,j'＝0j'＝1,...,Ji,j'≠j,j+1 (3-1c)
αi,jRepresents Vi(t) to a reference value
Figure GDA0002893668290000067
The confidence of (2);
(3.2) calculating C (t) for reference value DnIs formulated as
TO(C(t))＝{(Dn,γn)|n＝1,...,N} (3-2a)
Figure GDA0002893668290000071
γn'＝0n'＝1,...,N,n'≠n,n+1 (3-2c)
γnDenotes C (t) vs. reference value DnThe confidence of (2);
sample data [ V ] is exemplified hereP(t),VR(t),C(t)]＝[7.9726,0.2233,1]To illustrate, V is obtained by using the reference values given in (3) and the formulae (3-1) and (3-2)P(t) confidence of corresponding reference value is alphaP,10.2387 and αP,2＝0.7613；VR(t) confidence of corresponding reference value is alphaR,30.4218 and αR,40.5782; c (t) confidence of corresponding reference value is gamma 11 and γ 20. Thereby obtaining VP(t) Integrated confidence distribution (α)P,3γ1,αP,4γ1,αP,3γ2,αP,4γ2) Is (0.2387,0.7613,0,0) and VR(t) Integrated confidence distribution (α)R,3γ1,αR,4γ1,αR,3γ2,αR,4γ2) Is (0.4218,0.5782,0, 0).
(4) According to the steps (3.1) and (3.2), the sample set T is obtainedAThe confidence distributions were calculated for all data samples in (1) and tables 5 and 6 were constructed:
TABLE 5 (V)PReference value confidence statistical table of (t), C (t)
Figure GDA0002893668290000072
TABLE 6 sample pairs (V)RReference value confidence statistical table of (t), C (t)
Figure GDA0002893668290000073
(5) From the reference value confidence statistics calculated in tables 5 and 6, the tables are likelihood normalized by the following formula:
Figure GDA0002893668290000074
obtaining a reference value
Figure GDA0002893668290000081
Evidence of (A) is
Figure GDA0002893668290000082
Then, according to the formulae (5-1) and (5-2), the evidences Table 7 and Table 8 can be constructed.
TABLE 7 inputs VPEvidence matrix table of
Figure GDA0002893668290000083
TABLE 8 inputs VREvidence matrix table of
Figure GDA0002893668290000084
(6) Evidence (evidence)
Figure GDA0002893668290000085
The reliability factor of (2) is determined by the spearman rank correlation coefficient, and the formula is as follows
di＝Vi(t)-C(t) (6-1)
Figure GDA0002893668290000086
Wherein d isiIs the difference between the voltage characteristic value and the failure level of the capacitor, riIs a reliability factor of the evidence. Calculated to obtain r1＝0.8897，r2＝0.4117。
(7) Estimating initial capacitance failure fault level by using the evidence matrix table obtained in the step (5) and the step (6) and the reliability factor of the evidence through an evidence reasoning rule
Figure GDA0002893668290000087
The method comprises the following steps:
(7.1) for input Vi(t), each set of evidence corresponding to the input
Figure GDA0002893668290000088
And
Figure GDA0002893668290000089
will be activated, then the value V is inputi(t) the final evidence can be obtained
Figure GDA00028936682900000810
And
Figure GDA00028936682900000811
weighting to obtain:
ei＝{(Dn,pn,i),n＝1,...,N} (7-1a)
Figure GDA00028936682900000812
(7.2) obtaining V Using formulae (7-1a) and (7-1b)P(t) and VREvidence of (t) e1And e2Setting an initial evidence weight wi＝riFusing the two by using an evidence reasoning rule to obtain a fused result
O(V(t))＝{(Dn,pn,e(2)),n＝1,...,N} (7-2a)
Figure GDA0002893668290000091
Figure GDA0002893668290000092
(7.3) estimating the failure level of the capacitor according to the fusion obtained in the step (7.2) and the expression (7-3)
Figure GDA0002893668290000093
Figure GDA0002893668290000094
Sample [ V ] is exemplified in (3)P(t),VR(t),C(t)]＝[8.4764,0.6073,1]Explanation, from the obtained VP(t) confidence of corresponding reference value is alphaP,10.2387 and αP,20.7613 proof of activation
Figure GDA0002893668290000095
Figure GDA0002893668290000096
And
Figure GDA0002893668290000097
obtaining evidence e from formula (7-1)P＝(0.9363,0.0637,0)；VR(t) confidence of corresponding reference value is alphaR,30.4218 and αR,40.5782 proof of activation
Figure GDA0002893668290000098
And
Figure GDA0002893668290000099
Figure GDA00028936682900000910
obtaining evidence e according to formula (7-1)R＝(0.3251,0.3252, 0.34970); the fusion result obtained by using the evidence of the fusion of formula (7-2) is O (V (t) { (D)1,0.8657),(D2,0.1085),(D30.0258) and an estimate of the capacitance failure level by equation (7-3) is。
Figure GDA00028936682900000911
Therein, the
Figure GDA00028936682900000912
May be classified as a capacitive failure fault class 1.
Obtaining a sample set TAThe estimated value of the capacitor failure level can be obtained, and T can be obtainedAMean square error of estimated value and true value of capacitor failure fault level is msetraining＝0.1491。
(8) Setting the mean square error of the estimated value and the true value of the capacitance failure fault level as an objective function:
Figure GDA0002893668290000101
s.t.0≤wi≤1,i＝1,2 (8-2a)
Figure GDA0002893668290000102
simultaneous determination of parameter sets
Figure GDA0002893668290000103
wiWeight indicating evidence, and other parameters are set to
Figure GDA0002893668290000104
And establishing an optimization model to optimize the objective function. Obtaining an optimal parameter set P after training to obtain a sample set T of an optimization modelAHas a mean square error of
Figure GDA0002893668290000105
Collection of direct-current bus voltage value set T of existing voltage source type frequency converterERepeating step (3) to EStep (6) can obtain a set TECapacitance failure fault level estimation value
Figure GDA0002893668290000106
The final capacitive failure level estimation results are shown in fig. 4. Computing a set TEMean square error mse of initial capacitance failure fault level estimated value and true valuetesting1.3324. Then calculating the mean square error between the estimated value and the true value of the capacitance failure fault level of the optimized model
Figure GDA0002893668290000107
It can be seen that the accuracy of the capacitance failure fault level obtained through the optimized model estimation is high.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.

Claims (9)

1. A fault diagnosis method of a direct-current bus capacitor of a voltage source type frequency converter based on evidence reasoning rules comprises the following steps:
step (1) collecting a voltage signal of a direct current bus of a voltage source type frequency converter;
extracting a characteristic value of a voltage signal, setting three capacitance failure fault levels to form a complete sample set, and randomly dividing the sample set into a training set and a testing set;
converting the data in the training set into a reliability distribution matrix corresponding to a given reference value;
converting the reliability distribution matrix to obtain evidence of the failure grade of the capacitor;
step (5) calculating a reliability factor of the evidence;
step (6) performing weighted fusion on the evidence of the capacitor failure fault grade according to the reliability factor, and initially estimating and training the capacitor failure fault grade;
step (7) an optimization objective function is set according to the estimated value of the capacitor failure fault level and the mean square error of the set capacitor failure fault level, optimization parameters are determined, and an evidence reasoning parameter optimization model is established;
and (8) estimating the failure level of the sample capacitor in the test set through an evidence reasoning parameter optimization model.
2. The fault diagnosis method according to claim 1, wherein in the step (1), a voltage source converter is taken as an example; establishing three capacitance failure levels, wherein the capacitance level 1 of the capacitor is C17500 e-6F; level 2 capacitance of C190% of (i), i.e. C26750 e-6F; level 3 capacitance of C180% of (i), i.e. C36000 e-6F; collecting 8 seconds of direct current bus voltage instantaneous value V at sampling frequency of 50kHz in the running process of equipmentDC。
3. The fault diagnosis method according to claim 1, wherein in the step (2), the voltage signal V of the step (1) is obtainedDCVoltage peak-to-peak value V ofPPPresetting a window length to obtain V of multiple data windowsPPAverage value to obtain VP(t); will VDCCalculating a root mean square value V by a preset number of electric periods and a certain number of sampling pointsRMSAnd obtaining V after normalization treatmentR(t)；
Figure FDA0002893668280000011
Figure FDA0002893668280000012
Figure FDA0002893668280000013
Wherein
Figure FDA0002893668280000014
Are respectively input characteristic signals VP(t) and VR(t) minimum and maximum values; the failure level of the capacitor is marked as C (t), C (t)＝[1,2,3](ii) a Will VP(t)、VR(T) and C (T) are expressed as a sample set TA＝{[VP(t),VR(t),C(t)]1,2, …, Ts }, where [ V ═ tP(t),VR(t),C(t)]Is a sample vector.
4. The failure diagnosing method according to claim 3, wherein, in the step (3), the sample set T is setAData of [ V ]P(t),VR(t),C(t)]Converting a given reference value into a reliability distribution matrix; wherein the reference value set D ═ D of the capacitance failure fault leveln1, …, N, wherein N is the reference value number of the capacitor failure fault grade; voltage signal ViInput set of reference values
Figure FDA0002893668280000021
JiAs a voltage signal ViThe number of reference values of (a); the method comprises the following steps:
(4.1) calculation of Vi(t) to a reference value
Figure FDA0002893668280000022
Is formulated as
Figure FDA0002893668280000023
Figure FDA0002893668280000024
αi,j'＝0 j'＝1,...,Ji,j'≠j,j+1 (4-1c)
αi,jRepresents Vi(t) to a reference value
Figure FDA0002893668280000025
The confidence of (2);
(4.2) similarly calculating C (t) vs. reference value DnIs formulated as
TO(C(t))＝{(Dn,γn)|n＝1,...,N} (4-2a)
Figure FDA0002893668280000026
γn'＝0 n'＝1,...,N,n'≠n,n+1 (4-2c)
γnDenotes C (t) vs. reference value DnThe confidence of (2);
(4.3) sampling the sample set T according to the steps (4.1) and (4.2)ACalculating a confidence distribution (α) for all data sets in the data seti,jγn,αi,j+1γn,αi,jγn+1,αi,j+1γn+1) In which α isi,jγnRepresents a sample pair (V)i(t), C (t) wherein the input values correspond to reference values
Figure FDA0002893668280000027
While C (t) corresponds to the reference value DnThe comprehensive similarity of (2); and constructing Table 1, wherein an,jRepresents all Vi(t) corresponding to the reference value
Figure FDA0002893668280000028
And C (t) corresponds to the reference value DnSample pair (V)i(t), C (t) the sum of the integrated similarities,
Figure FDA0002893668280000029
denotes all C (t) corresponding reference values DnThe sum of the pair of integrated similarities of (a),
Figure FDA00028936682800000210
represents all Vi(t) corresponding to the reference value
Figure FDA00028936682800000211
Of the sample pair of comprehensive similarity, and
Figure FDA00028936682800000212
TABLE 1 (V)iReference value confidence coefficient distribution statistical table of (t), C (t)
Figure FDA00028936682800000213
Figure FDA0002893668280000031
5. The failure diagnosing method according to claim 4, wherein in the step (4), according to the table 1 in the step (3), the table is likelihood-normalized by the following formula:
Figure FDA0002893668280000032
obtaining a reference value
Figure FDA0002893668280000033
Evidence of (A) is
Figure FDA0002893668280000034
Then, an evidence table 2 was constructed based on the formulae (5-1) and (5-2),
TABLE 2 inputs ViEvidence table of
Figure FDA0002893668280000035
6. The failure diagnosis method according to claim 5, wherein in step (5), the evidence is
Figure FDA0002893668280000036
The reliability factor of (2) is calculated from the spearman rank correlation coefficient, and the formula is as follows
di＝Vi(t)-C(t) (6-1)
Figure FDA0002893668280000037
Wherein d isiIs the difference between the voltage characteristic value and the failure level of the capacitor, riIs a reliability factor of the evidence.
7. The fault diagnosis method according to claim 6, wherein in the step (6), the evidence table 2 and the reliability factor obtained in the steps (4) and (5) are used for estimating the fault level of the capacitance failure through an evidence reasoning rule
Figure FDA0002893668280000041
The method comprises the following steps:
(7.1) for input Vi(t), each set of evidence corresponding to the input
Figure FDA0002893668280000042
And
Figure FDA0002893668280000043
will be activated, then the value V is inputi(t) the final evidence can be obtained
Figure FDA0002893668280000044
And
Figure FDA0002893668280000045
weighting to obtain:
ei＝{(Dn,pn,i),n＝1,...,N} (7-1a)
Figure FDA0002893668280000046
(7.2) obtaining V from the formulae (7-1a) and (7-1b)p(t) and VREvidence of (t) e1And e2Setting an initial evidence weight wi＝riAnd fusing the images by using an evidence reasoning rule to obtain a fusion result, wherein the formula is as follows:
O(V(t))＝{(Dn,pn,e(2)),n＝1,...,N} (7-2a)
Figure FDA0002893668280000047
(7.3) estimating the failure level of the capacitor by the formula (7-3) according to the fusion result of the step (7.2)
Figure FDA0002893668280000048
Figure FDA0002893668280000049
8. The fault diagnosis method according to claim 7, wherein in step (7), the mean square error between the estimated value of the capacitance failure fault level and the true value is set as an objective function:
Figure FDA00028936682800000410
s.t.0≤wi≤1,i＝1,2 (8-2a)
Figure FDA00028936682800000411
simultaneous determination of parametersCollection
Figure FDA00028936682800000412
wiWeight indicating evidence, and other parameters are set to
Figure FDA00028936682800000413
And establishing an optimization model to optimize the objective function, and obtaining an optimal parameter set P after training.
9. The fault diagnosis method according to claim 1, wherein, in step (8), the voltage value of the direct current bus of the existing voltage source type frequency converter is collected, and the steps (3) to (6) are repeated again to obtain a more accurate estimated value of the failure level of the capacitor
Figure FDA0002893668280000051
CN201910291044.7A 2019-04-11 2019-04-11 Frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rule Active CN110308339B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910291044.7A CN110308339B (en) 2019-04-11 2019-04-11 Frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rule

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910291044.7A CN110308339B (en) 2019-04-11 2019-04-11 Frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rule

Publications (2)

Publication Number Publication Date
CN110308339A CN110308339A (en) 2019-10-08
CN110308339B true CN110308339B (en) 2021-05-25

Family

ID=68074555

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910291044.7A Active CN110308339B (en) 2019-04-11 2019-04-11 Frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rule

Country Status (1)

Country Link
CN (1) CN110308339B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110850206A (en) * 2019-11-13 2020-02-28 武汉理工大学 Inverter capacitor aging fault diagnosis method based on confidence rule reasoning

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112668164A (en) * 2020-12-18 2021-04-16 武汉大学 Transformer fault diagnosis method and system for inducing ordered weighted evidence reasoning
CN113033078B (en) * 2021-03-05 2022-06-03 国网安徽省电力有限公司 Construction method, system and early warning method of fault early warning model of relay protection equipment

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103592554A (en) * 2013-12-03 2014-02-19 武汉大学 On-line monitoring system and method of 35kV high voltage shunt capacitor
CN105488270A (en) * 2015-11-27 2016-04-13 国家电网公司 Multiattribute comprehensive method for structural fault diagnosis of transformer
CN105825271A (en) * 2016-03-21 2016-08-03 南京邮电大学 Satellite fault diagnosis and prediction method based on evidential reasoning (ER)
CN105923014A (en) * 2016-04-27 2016-09-07 杭州电子科技大学 Track longitudinal irregularity amplitude value estimation method based on evidential reasoning rule
CN109086483A (en) * 2018-06-29 2018-12-25 广东工业大学 A kind of evidence fusion and Method of Set Pair Analysis of the assessment of transformer ageing state

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA2242069A1 (en) * 1998-06-25 1999-12-25 Postlinear Management Inc. Possibilistic expert systems and process control utilizing fuzzy logic

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103592554A (en) * 2013-12-03 2014-02-19 武汉大学 On-line monitoring system and method of 35kV high voltage shunt capacitor
CN105488270A (en) * 2015-11-27 2016-04-13 国家电网公司 Multiattribute comprehensive method for structural fault diagnosis of transformer
CN105825271A (en) * 2016-03-21 2016-08-03 南京邮电大学 Satellite fault diagnosis and prediction method based on evidential reasoning (ER)
CN105923014A (en) * 2016-04-27 2016-09-07 杭州电子科技大学 Track longitudinal irregularity amplitude value estimation method based on evidential reasoning rule
CN109086483A (en) * 2018-06-29 2018-12-25 广东工业大学 A kind of evidence fusion and Method of Set Pair Analysis of the assessment of transformer ageing state

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于证据推理规则的信息融合故障诊断方法;徐晓滨 等;《控制理论与应用》;20150930;第32卷(第9期);第1170-1182页 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110850206A (en) * 2019-11-13 2020-02-28 武汉理工大学 Inverter capacitor aging fault diagnosis method based on confidence rule reasoning

Also Published As

Publication number Publication date
CN110308339A (en) 2019-10-08

Similar Documents

Publication Publication Date Title
CN110308339B (en) Frequency converter direct-current bus capacitor fault diagnosis method based on evidence reasoning rule
JP6141235B2 (en) How to detect anomalies in time series data
CN103310051B (en) Board information terminal Failure Rate Forecasting Method in a kind of life cycle management
WO2016107246A1 (en) Wavelet noise reduction and relevance vector machine-based method for predicting remaining life of lithium battery
CN110852476A (en) Passenger flow prediction method and device, computer equipment and storage medium
CN108304685A (en) A kind of non-linear degradation equipment method for predicting residual useful life and system
KR102067344B1 (en) Apparatus and Method for Detecting Abnormal Vibration Data
CA2931624A1 (en) Systems and methods for event detection and diagnosis
CN112149860A (en) Automatic anomaly detection method and system
CN102521534B (en) Intrusion detection method based on crude entropy property reduction
CN110942137A (en) Power grid information operation and maintenance monitoring method based on deep learning
CN117034043A (en) Intelligent building comprehensive energy consumption monitoring method and system based on multi-energy Internet of things
KR20170085369A (en) Method and apparatus for managing of battery capacity
CN107220500B (en) Bayesian reliability evaluation method for performance degradation test based on inverse Gaussian process
CN114290960A (en) Method and device for acquiring battery health degree of power battery and vehicle
CN103646013B (en) Multiple fault reconstruction method based on covariance matrix norm approximation
CN112380992B (en) Method and device for evaluating and optimizing accuracy of monitoring data in machining process
CN115081331A (en) Wind turbine generator running state abnormity detection method based on state parameter reconstruction error
CN111046790A (en) Pump bearing fault diagnosis method
CN110532629A (en) A kind of method for predicting residual useful life of multi-modal degenerative process
CN114266457A (en) Method for detecting different loss inducement of distribution line
CN109165396A (en) A kind of equipment remaining life prediction technique of failure evolution trend
CN109946542B (en) Inverter direct-current loop capacitance fault detection method based on long-time and short-time information fusion
CN117675230A (en) Knowledge-graph-based oil well data integrity identification method
CN110986267A (en) Detection method and device for abnormal sound of air conditioner and air conditioner

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant