CN114236374A - Real-time diagnosis method for open circuit fault of rectifier - Google Patents

Real-time diagnosis method for open circuit fault of rectifier Download PDF

Info

Publication number
CN114236374A
CN114236374A CN202111522592.XA CN202111522592A CN114236374A CN 114236374 A CN114236374 A CN 114236374A CN 202111522592 A CN202111522592 A CN 202111522592A CN 114236374 A CN114236374 A CN 114236374A
Authority
CN
China
Prior art keywords
fault
rectifier
phase
sampling
follows
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202111522592.XA
Other languages
Chinese (zh)
Other versions
CN114236374B (en
Inventor
田里思
胡彬
杨涛
唐超权
邓先明
余强
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.)
China University of Mining and Technology CUMT
Original Assignee
China University of Mining and Technology CUMT
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 China University of Mining and Technology CUMT filed Critical China University of Mining and Technology CUMT
Priority to CN202111522592.XA priority Critical patent/CN114236374B/en
Publication of CN114236374A publication Critical patent/CN114236374A/en
Application granted granted Critical
Publication of CN114236374B publication Critical patent/CN114236374B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01RMEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/327Testing of circuit interrupters, switches or circuit-breakers

Abstract

The invention discloses a real-time diagnosis method for rectifier open circuit faults, which comprises the following steps: s1, constructing a three-phase rectifier system model, and adding a fault diagnosis unit into the system; s2, respectively collecting three-phase currents i under the health condition and various faults by simulating fault generationa、ib、ic(ii) a S3, processing the three-phase current data by adopting a current data reconstruction algorithm, and eliminating the phase difference between the three-phase currents; s4, introducing a multi-scale entropy algorithm and further processing the data; s5, using the processed current data, detects the occurrence of open circuit fault and completes the classification and identification of open circuit fault. The method can diagnose 21 types of single-tube open-circuit faults and composite-tube open-circuit faults in real time; the invention does not need additional sensors and has low cost.

Description

Real-time diagnosis method for open circuit fault of rectifier
Technical Field
The invention relates to a power electronic converter diagnosis method, in particular to a real-time diagnosis method for rectifier open circuit faults.
Background
The development of power electronic technology has been widely used worldwide for more than sixty years, and is an important technical support for the development of power electronic converters. Meanwhile, the application scale of the power electronic converter is continuously enlarged, the structure is more complex, and higher requirements on safety and reliability are provided. Therefore, the health of the power electronic converter is critical to the stable operation of the entire system. The rectifier is generally provided with a fast fuse, and when a short-circuit fault of a device occurs, the short-circuit fault can be converted into an open-circuit fault through the fast fuse. The open-circuit fault of the power tube can not only damage the converter, but also cause system breakdown, and cause serious accidents. Therefore, the method has important significance for real-time fault detection of the power electronic converter while ensuring the system performance.
The fault diagnosis technology is applied to monitoring the state in the industrial control process, detecting the running state of a system, gradually measuring the health state of the system through fault diagnosis, and analyzing the fault of the system after detecting an abnormal condition.
Disclosure of Invention
The purpose of the invention is as follows: the invention aims to provide a real-time diagnosis method for open-circuit faults of a rectifier, which can judge the types and positions of the faults in real time.
The technical scheme is as follows: the real-time diagnosis method comprises the following steps:
s1, constructing a simulation model of the three-phase rectifier system comprising the fault diagnosis unit;
s2, simulating the generation of various single-tube open-circuit faults and double-tube composite open-circuit faults, and respectively collecting original three-phase current signals i under the healthy condition and various faultsa、ib、ic
S3, processing the original three-phase current data by adopting a data reconstruction algorithm;
s4, after the phase difference of the three-phase current data is eliminated, processing the three-phase current data by adopting a multi-scale entropy algorithm;
and S5, detecting the generation of open circuit faults according to the symmetry of the rectifier topological structure, and completing the classification and identification of single-tube open circuit faults and double-tube composite open circuit faults by using the processed three-phase current data.
Further, the implementation process of step S2 is as follows:
s2a, comparing the three-phase current signal i at the time k in one perioda、ib、icThe stored sampling linked list adopts the expression of time sequence as follows:
[in(k-L+1),in(k-L+2),...,in(k)]
wherein, L is the number of sampling points of the current signal in one period, and n is a, b, c;
s2b, acquiring current sampling data by adopting a sliding window method, wherein the number L of sampling points of a current signal in one period is as follows:
L=fs/f
wherein f issIs the sampling frequency, f is the frequency of the three-phase voltage source;
s2c, saving data in the current fundamental wave period, as follows:
at the moment k, the sampling chain table stores data in a current fundamental wave period, and the expression is as follows:
[in(k-L+1),in(k-L+2),...,in(k)]
at the moment of k +1, inserting a new sampling point into the current sampling chain table, deleting an old sampling point, and updating in real time, wherein the updated expression is as follows:
[in(k-L+2),in(k-L+3),...,in(k+1)]。
further, the specific implementation process of step S3 is as follows:
s3a, recording the position and phase angle of a bottom-to-top zero crossing point in one period of each phase current;
a certain phase current I in the phase current InPosition of bottom-to-top zero crossing within a cycleIs marked as FnThe phase angle is
Figure BDA0003408302560000021
S3b, when F is detectednThen, F is mixednThe previous sampling data is directly spliced to the end of the sliding window;
s3c, at time k, i ═ k-L,
Figure BDA0003408302560000022
wherein, L is the number of sampling points in one period;
if it satisfies
Figure BDA0003408302560000023
Then Fn=k-L;
If not satisfied with
Figure BDA0003408302560000024
Then FnUntil satisfied, at which time F is shifted to the rightnThe former sampling points are moved to the sampling chain table in sequence and then are filled up, thereby forming a new current chain table
Figure BDA0003408302560000025
Further, the implementation process of step S4 is as follows:
s4a, setting the data of the current signal collected from the shared current sensor as In(k) Length is N, mode dimension is m, and similarity tolerance is r;
from the collected data, a time series is constructed as follows:
In(k)=[in(k),in(k+1),...,in(k+m-1)],k=1,2,...,N-m+1;
for each value of k, calculate In(k) And all of In(j) The distance between:
Figure BDA0003408302560000031
for each k value, count d [ I ]n(k),In(j)]The ratio of the number < r to the total number of vectors N-m is noted
Figure BDA0003408302560000032
Figure BDA0003408302560000033
To find
Figure BDA0003408302560000034
The average value of all k values is recorded as Bm(r):
Figure BDA0003408302560000035
Adding 1 to the model dimension m to obtain m +1, repeating the operation process, and calculating Bm+1(r) to obtain the sequence In(k) The corresponding sample entropy values are as follows:
SampEn(m,r)=-ln(Bm+1(r)/Bm(r));
s4b, original time sequence I is processed according to the scale factor taun(k) Coarse grained variation was performed, the expression is as follows:
Figure BDA0003408302560000036
the length of the obtained new time series is N/tau, and for each coarse grained time series, the corresponding sample entropy SampEn (m, r, N) is calculated.
Further, the implementation process of step S5 is as follows:
s5a, in the working process of the rectifier, if a power tube on a bridge arm of the rectifier breaks down, a part of sampling points which are constantly zero appear in the corresponding phase current, and the expression is as follows:
Figure BDA0003408302560000037
wherein, Kth1Is a selected threshold;
the percentage P of the interval with the sampling value of zero in one current periodnComprises the following steps:
Figure BDA0003408302560000041
s5b, using Ra、Rb、RcThe difference relation between the two represents the symmetry of the topological structure, and in a healthy state, PnThe difference between two is approximately 0, if any one exceeds the set threshold value KdetAnd judging that the rectifier has an open-circuit fault:
Figure BDA0003408302560000042
the FD-1 represents that the rectifier has open-circuit fault, and the FD-0 represents that the state of the rectifier is normal;
Figure BDA0003408302560000043
Pa、Pb、Pcthe intervals with sampling values of 0 in a sampling period of the three-phase currents a, b and c respectively account for the percentage;
s5c, according to PnJudging a fault bridge arm:
Pn>Kth1,Sn=0
wherein S isn0 represents that the n-phase bridge arm has open-circuit fault;
under the healthy condition, if an open-circuit fault occurs, half-wave partial waveform loss occurs to the corresponding bridge arm phase current, which is expressed as:
Figure BDA0003408302560000044
KL2、KL1respectively representing the positive half-wave length and the negative half-wave length in one current fundamental wave period; LP represents that an upper pipe of one bridge arm fails, LN represents that a lower pipe of one bridge arm fails;
if the upper pipe and the lower pipe of the same bridge arm simultaneously break down, the positive half-wave length and the negative half-wave length are equal, the ratio of the zero part of the waveform is increased, and the complexity MSE of the three-phase current is respectively calculated by adopting a multi-scale entropy algorithma、MSEb、MSEcThen, the threshold K is setth2And comparing, positioning the fault, and judging as follows:
when MSEa、MSEb、MSEcOne value of which is constantly greater than the threshold value Kth2And then the double-tube open circuit fault of the same bridge arm occurs.
Compared with the prior art, the invention has the following remarkable effects: 1. the method reconstructs the acquired three-phase current data, detects the occurrence of the fault and positions the fault bridge arm through a symmetric function, and finally further determines the position of the fault tube by combining the calculation of multi-scale entropy and the comparison of the lengths of positive and negative half-waves of the fault phase current, can detect the occurrence of the fault in a quarter period, can complete fault positioning in one period, realizes the rapid detection and positioning of the fault, is suitable for online monitoring, and has high diagnosis precision; 2. the invention does not need additional sensors, has low cost and good practical significance.
Drawings
FIG. 1 is a topology diagram of a rectifier system with a fault diagnosis unit;
FIG. 2 is a schematic flow chart of the present invention;
FIG. 3(a) is a diagram of an analog power transistor T according to the present invention1A result graph of the rectified output voltage, the three-phase current and the fault detection characteristic quantity of the open-circuit fault,
(b) for simulating the power tube T of the invention1Fault location characteristic quantity of open circuit fault, multi-scale entropy value and result diagram of fault detection and fault location;
FIG. 4(a) is a diagram of an analog power transistor T according to the present invention1T2Rectified output voltage, three-phase with open-circuit faultA resulting map of current and fault detection characteristics,
(b) for simulating the power tube T of the invention1T2Fault location characteristic quantity of open circuit fault, multi-scale entropy value and result diagram of fault detection and fault location;
FIG. 5(a) is a diagram of an analog power transistor T according to the present invention1T3A result graph of the rectified output voltage, the three-phase current and the fault detection characteristic quantity of the open-circuit fault,
(b) for simulating the power tube T of the invention1T3Fault location characteristic quantity of open circuit fault, multi-scale entropy value and result diagram of fault detection and fault location;
FIG. 6(a) is a diagram of an analog power transistor T according to the present invention1T4A result graph of the rectified output voltage, the three-phase current and the fault detection characteristic quantity of the open-circuit fault,
(b) for simulating the power tube T of the invention1T4Fault location characteristic quantity of open circuit fault, multi-scale entropy value and result graph of fault detection and fault location.
Detailed Description
The invention is described in further detail below with reference to the drawings and the detailed description.
Fig. 1 shows a topology of a rectifier system with a fault diagnosis unit.
Fig. 2 is a flow chart of the real-time diagnosis method of the present invention, which includes the following steps:
step one, constructing a model of the three-phase rectifier system shown in fig. 1, wherein an adopted control scheme is vector control (VFOC) based on virtual flux linkage orientation, and the model mainly comprises the following parts: the three-phase power supply voltage at the power grid side is respectively ua、ub、ucThe frequency is f, and the filter inductance of the three-phase current is L respectivelya、Lb、LcThe resistances are respectively Ra、Rb、RcOutput voltage measurement value is Ud6 power tubes T1-T6, each power tube corresponding to a freewheeling diode respectively D1-D6, a DC bus filter capacitor C and a resistive load RLAnd a fault diagnosis module.
Simulating the generation of various single-tube open-circuit faults and double-tube composite open-circuit faults, and respectively acquiring original three-phase current signals i under the healthy condition and various faultsa、ib、ic. The method specifically comprises the following steps:
(2a) time-series processing of phase currents
The three-phase current has periodicity, the phase current in one period can be regarded as a time sequence, and then the three-phase current signal i is applied to the k time in one perioda、ib、icThe representation of the saved sample linked list is as follows:
[in(k-L+1),in(k-L+2),...,in(k)],(n=a,b,c) (1)
in the formula (1), L is the number of sampling points in one period.
(2b) Obtaining current sample data
Acquiring current sampling data by adopting a sliding window method, and recording the number of sampling points of a current signal in one period as L:
L=fs/f (2)
in the formula (2), fsTo sample frequency, f is the frequency of the three-phase voltage source.
(2c) Preserving data over a period of a current fundamental
At time k, the linked list holds data for one current fundamental cycle:
[in(k-L+1),in(k-L+2),...,in(k)],(n=a,b,c) (3)
at the moment of k +1, inserting a new sampling point into the current chain table, deleting an old sampling point, and updating in real time:
[in(k-L+2),in(k-L+3),...,in(k+1)],(n=a,b,c) (4)。
and step three, in order to eliminate the phase difference between the three-phase currents, respectively processing the original three-phase current data by adopting a current data reconstruction algorithm. The method specifically comprises the following steps:
(3a) in a healthy stateAnd the phase difference between the three-phase currents in one period is 2 pi/3. With a certain phase current InFor example, the bottom-up zero crossing position within a cycle is denoted as FnThe phase angle is
Figure BDA0003408302560000071
(3b) When F is detectednThen, F is mixednThe previous sampled data is spliced directly to the end of the sliding window.
(3c) At time k, i-k-L,
Figure BDA0003408302560000072
if formula (5) is satisfied:
Figure BDA0003408302560000073
in the formula (5), L is the number of sampling points in one period;
then Fn=k-L;
If formula (5) is not satisfied, FnUntil equation (5) is satisfied, at which point F is shifted to the rightnThe former sampling points are moved to the linked list in sequence and then are filled up, so as to form a new current linked list
Figure BDA0003408302560000074
For original three-phase current data ia、ib、icThe phase difference of the three-phase current is eliminated by processing the three-phase current respectively through the steps (3a) to (3 c).
And step four, after the phase difference of the three-phase current data is eliminated, subsequently detecting the fault and finishing the classification identification. The invention adopts a multi-scale entropy algorithm to further process the three-phase current data. The method specifically comprises the following steps:
(4a) setting data of current signal collected from shared current sensor as In(k) The length is N, m is the mode dimension, and r is the similarity tolerance. Firstly, construct according to the collected dataThe time sequence was as follows:
In(k)=[in(k),in(k+1),...,in(k+m-1)],k=1,2,...,N-m+1 (6)
calculating I for each value of kn(k) And all of In(j) The distance between:
Figure BDA0003408302560000075
for each k value, count d [ I ]n(k),In(j)]The ratio of the number < r to the total number of vectors N-m is noted
Figure BDA0003408302560000076
Figure BDA0003408302560000077
To find
Figure BDA0003408302560000078
The average value of all k values is recorded as Bm(r):
Figure BDA0003408302560000081
Repeating the above operation processes (6) - (9) after adding 1 to the mode dimension m to become m +1, and calculating Bm+1(r) to obtain the sequence In(k) The corresponding sample entropy values are as follows:
SampEn(m,r)=-ln(Bm+1(r)/Bm(r)) (10)
(4b) the multi-scale entropy is based on the sample entropy. First, the original time sequence I is subjected to scale factor taun(k) Coarse graining change is carried out, a non-overlapping formula is selected, tau data are jumped every time, the tau data are taken for averaging to generate new data, and the formula is as follows:
Figure BDA0003408302560000082
the resulting new time series has a length of N/τ and for each coarse grained time series the corresponding sample entropy SampEn (m, r, N) is calculated.
And step five, detecting the generation of open-circuit faults according to the symmetry of the topological structure of the rectifier, and completing the classification and identification of 21 single-tube open-circuit faults and double-tube composite open-circuit faults of the rectifier in total by utilizing the processed three-phase current data. The method specifically comprises the following steps:
(5a) in the working process of the rectifier, if a power tube on a bridge arm of the rectifier breaks down, a part of constant zero sampling points exist in corresponding phase current, and a fault diagnosis algorithm is designed on the basis.
Figure BDA0003408302560000083
In the formula, Kth1Is a selected threshold; characteristic quantity Mn(k) 1 denotes this sampling point in(k) The value of (b) can be regarded as 0, so that the number of sampling points of which the current is 0 in one period can be counted conveniently.
Calculating the percentage P of the interval with the sampling value of zero in one current periodnUsed in the subsequent step (5b) of calculating R as followsn
Figure BDA0003408302560000084
Wherein n is a, b, c.
(5b) By means of Ra,Rb,RcThe difference relationship between the two adjacent topological structures represents the symmetry of the topological structure, the difference between every two adjacent topological structures in a healthy state is approximately 0 as long as any one of the two adjacent topological structures exceeds a set threshold value KdetThen, it can be determined that the open-circuit fault occurs in the rectifier:
Figure BDA0003408302560000091
FD-1 represents that the rectifier has an open circuit fault, and FD-0 represents that the rectifier is in a normal state.
Wherein the content of the first and second substances,
Figure BDA0003408302560000092
Pa、Pb、Pcthe interval with the sampling value of 0 in one sampling period of the three-phase current of a, b and c respectively accounts for percentage.
(5c) Only when the open-circuit fault occurs on the bridge arm, the interval with zero half-wave appears, and the interval can be determined according to PnJudging a fault bridge arm:
Pn>Kth1,Sn=0 (15)
wherein S isnAnd 0 represents that the n-phase bridge arm has open circuit faults.
Under the healthy condition, the lengths of the positive half-wave and the negative half-wave are almost equal, and the half-wave partial waveform loss of the corresponding bridge arm phase current is caused by the open-circuit fault:
Figure BDA0003408302560000093
KL2、KL1respectively representing the positive half-wave length and the negative half-wave length in one current fundamental wave period; LP represents that the upper pipe of one bridge arm fails, and LN represents that the lower pipe of one bridge arm fails. The upper and lower positions of the fault can be judged from the upper position.
If the positive half-wave length and the negative half-wave length are equal after the upper pipe and the lower pipe of the same bridge arm simultaneously break down, the proportion of the zero part of the waveform is increased, and the upper pipe and the lower pipe of the same bridge arm cannot be judged to simultaneously break down according to the formula (16), so that the multi-scale entropy algorithm mentioned above is added: calculate MSE separatelya、MSEb、MSEcAnd is in contact with the set threshold kth2In comparison, special positioning is performed according to the complexity of the current signal. Through a multi-scale entropy algorithm, the complexity of three-phase current signals can be quantitatively measured, and MSE is used as a calculation resulta、MSEb、MSEcAnd (4) showing. Obtaining a threshold value K in the process of simulation and experimentth2The effective range of (1). By observing experimental phenomena, MSE can occur only when double-pipe open circuit fault of the same bridge arm occursa、MSEb、MSEcOne value of which is constantly greater than the threshold value Kth2(ii) a Under other conditions, the three values are not constantly larger than Kth2. Judging a fault bridge arm according to a formula (15), and comparing MSEa、MSEb、MSEcAnd a threshold value Kth2And the discrimination of double-tube open-circuit faults of the same bridge arm can be realized. So far, the classification and identification of 21 open-circuit faults can be realized.
Fig. 1 shows a topology of a rectifier system with a fault diagnosis unit, which is composed of the following parts: the three-phase power supply voltage at the power grid side is respectively ua、ub、ucThree-phase current filter inductor La、Lb、LcResistance Ra、Rb、RcOutputting the measured voltage value Ud6 power tubes T1-T6A freewheel diode D corresponding to each power transistor1-D6A DC bus filter capacitor C and a resistive load RLAnd a fault diagnosis module.
Different fault types are simulated by shielding the driving signals of the power tube, and the open-circuit fault diagnosis table of the rectifier power tube is obtained by processing the driving signals through the first step to the fifth step of the method and is shown in table 1.
TABLE 1 open-circuit fault diagnosis table for rectifier power tube
Figure BDA0003408302560000101
Figure BDA0003408302560000111
The rectifier single tube open circuit fault and double tube composite open circuit fault are 21 types in total, and all are listed in table 1. Wherein S isn0 represents that the n-phase bridge arm has open circuit fault, and is represented by the formula (15)The zero point interval ratio PnIs determined.
After the bridge arm with the fault is judged, the upper position and the lower position of the fault can be judged by utilizing the relation between the positive half wave and the negative half wave described in the formula (16), wherein LP represents an upper tube fault, LN represents a lower tube fault, and the single-tube open-circuit fault and the double-tube composite open-circuit fault of different bridge arms can be accurately judged.
If the upper and lower tubes of the same bridge arm have composite open-circuit faults, the positive and negative half-wave lengths are equal, the proportion of the zero part of the waveform is increased, and the simultaneous open-circuit faults of the upper and lower tubes of the same bridge arm cannot be judged according to a formula (16), so that a multi-scale entropy algorithm is added to calculate MSE (mean Square error)a、MSEb、MSEcAnd a set threshold value Kth2And comparing, namely performing special positioning according to the complexity of the current signal. Therefore, 21 types of real-time diagnosis of the single-tube open-circuit fault and the double-tube composite open-circuit fault of the rectifier are realized.
The effectiveness and reliability of the invention are verified through simulation: and establishing a vector control rectifier system, and then realizing fault diagnosis by adding the fault diagnosis algorithm flow. Because the power tube can be equivalent to an open-circuit fault (keeping the freewheeling diode normally conducting) when being disconnected, the conduction and the disconnection of the power tube can be realized by controlling the trigger pulse of the power tube when in simulation design, so that the normal operation and the open-circuit fault of the power tube can be simulated.
The value ranges of the parameters in the examples are shown in table 2.
TABLE 2 value ranges of the parameters
Figure BDA0003408302560000112
Since all 21 fault types are classified into 4 categories in table 1, a representative fault is selected from each of the 4 categories for simulation verification. Respectively with a power tube T1、T1T2、T1T3And T1T4Verification was performed with open circuit faults as representative, and the results are shown in fig. 3 to 6.
The rectifier gives an output voltage of 300V and a load of 80 omega. The rectified output voltage U is given in sequence in FIG. 3(a)dWaveform of (1), three-phase current ia、ib、icAnd a fault detection characteristic quantity Ra、Rb、RcA graph of results of (1); the fault location characteristic quantity P is given in sequence in FIG. 3(b)a、Pb、PcMulti-scale entropy MSEa、MSEb、MSEcAnd a result graph of fault detection FD and fault localization FI.
As shown in FIG. 3(a), is T1Open circuit failure, 0.51s shut off power tube T1After 0.51s the phase current is distorted, Rb、RcGradually increases over a threshold value KdetThe topological structure of the three-phase bridge arm is no longer symmetrical, and FD is 1, that is, the fault is detected. After the occurrence of the detected fault, Pa>Kth1→SaWhen the waveform of the phase current on the a-phase arm is zero, 0 indicates that the a-phase arm has a fault. MSE, as shown in FIG. 3(b)a、MSEb、MSEcNot constantly above threshold Kth2The situation that the single bridge arm double-tube fault does not occur is explained. Comparing the positive and negative half-wave lengths, partial absence of phase current half-wave, AL1<AL2→ LP, thus recognizing T1Open circuit failure. FI was located to 1, consistent with the corresponding results in Table 1.
Shown in FIG. 4(a) is T1T2Open circuit failure, 0.50s shut off power tube T1After 0.50s the phase current is distorted, Rb、RcGradually increases over a threshold value KdetThe topological structure of the three-phase bridge arm is no longer symmetrical, and FD is 1, that is, the fault is detected. After the occurrence of the detected fault, Pa>Kth1→SaWhen the waveform of the phase current on the a-phase arm is zero, 0 indicates that the a-phase arm has a fault. MSE as shown in FIG. 4(b)bIs constantly greater than the threshold value Kth2In the case where there is a double-pipe fault in the same bridge arm, T is recognized1T2Open circuit failure. FI was located to 7, consistent with the corresponding results in Table 1.
FIGS. 5(a) and 5(b) are T1T3The verification results of open circuit failure, T in FIGS. 6(a) and 6(b)1T4And (5) verifying the open-circuit fault. Analysis method and analysis T1Open circuit fault, T1T2The open circuit fault process is similar and will not be described in detail.
According to the verification process, the method can realize the rapid detection and diagnosis of the composite open-circuit fault of the single tube and the double tube of the rectifier, and has higher accuracy.

Claims (5)

1. A real-time diagnosis method for rectifier open circuit fault is characterized by comprising the following steps:
s1, constructing a simulation model of the three-phase rectifier system comprising the fault diagnosis unit;
s2, simulating the generation of various single-tube open-circuit faults and double-tube composite open-circuit faults, and respectively collecting original three-phase current signals i under the healthy condition and various faultsa、ib、ic
S3, processing the original three-phase current data by adopting a data reconstruction algorithm;
s4, after the phase difference of the three-phase current data is eliminated, processing the three-phase current data by adopting a multi-scale entropy algorithm;
and S5, detecting the generation of open circuit faults according to the symmetry of the rectifier topological structure, and completing the classification and identification of single-tube open circuit faults and double-tube composite open circuit faults by using the processed three-phase current data.
2. The method for real-time diagnosis of rectifier open circuit fault according to claim 1, wherein the step S2 is implemented as follows:
s2a, comparing the three-phase current signal i at the time k in one perioda、ib、icThe stored sampling linked list adopts the expression of time sequence as follows:
[in(k-L+1),in(k-L+2),...,in(k)]
wherein, L is the number of sampling points of the current signal in one period, and n is a, b, c;
s2b, acquiring current sampling data by adopting a sliding window method, wherein the number L of sampling points of a current signal in one period is as follows:
L=fs/f
wherein f issIs the sampling frequency, f is the frequency of the three-phase voltage source;
s2c, saving data in the current fundamental wave period, as follows:
at the moment k, the sampling chain table stores data in a current fundamental wave period, and the expression is as follows:
[in(k-L+1),in(k-L+2),...,in(k)]
at the moment of k +1, inserting a new sampling point into the current sampling chain table, deleting an old sampling point, and updating in real time, wherein the updated expression is as follows:
[in(k-L+2),in(k-L+3),...,in(k+1)]。
3. the method for real-time diagnosis of the open circuit fault of the rectifier according to claim 1, wherein the step S3 is implemented as follows:
s3a, recording the position and phase angle of a bottom-to-top zero crossing point in one period of each phase current;
a certain phase current I in the phase current InThe position of the bottom-to-top zero crossing point in one period is recorded as FnThe phase angle is
Figure FDA0003408302550000029
S3b, when F is detectednThen, F is mixednThe previous sampling data is directly spliced to the end of the sliding window;
s3c, at time k, i ═ k-L,
Figure FDA0003408302550000021
wherein, L is the number of sampling points in one period;
if it satisfies
Figure FDA0003408302550000022
Then Fn=k-L;
If not satisfied with
Figure FDA0003408302550000023
Then FnUntil satisfied, at which time F is shifted to the rightnThe former sampling points are moved to the sampling chain table in sequence and then are filled up, thereby forming a new current chain table
Figure FDA0003408302550000024
4. The method for real-time diagnosis of rectifier open circuit fault according to claim 1, wherein the step S4 is implemented as follows:
s4a, setting the data of the current signal collected from the shared current sensor as In(k) Length is N, mode dimension is m, and similarity tolerance is r;
from the collected data, a time series is constructed as follows:
In(k)=[in(k),in(k+1),...,in(k+m-1)],k=1,2,...,N-m+1;
for each value of k, calculate In(k) And all of In(j) The distance between:
Figure FDA0003408302550000025
for each k value, count d [ I ]n(k),In(j)]The ratio of the number < r to the total number of vectors N-m is noted
Figure FDA0003408302550000026
Figure FDA0003408302550000027
To find
Figure FDA0003408302550000028
The average value of all k values is recorded as Bm(r):
Figure FDA0003408302550000031
Adding 1 to the model dimension m to obtain m +1, repeating the operation process, and calculating Bm+1(r) to obtain the sequence In(k) The corresponding sample entropy values are as follows:
SampEn(m,r)=-ln(Bm+1(r)/Bm(r));
s4b, original time sequence I is processed according to the scale factor taun(k) Coarse grained variation was performed, the expression is as follows:
Figure FDA0003408302550000032
the length of the obtained new time series is N/tau, and for each coarse grained time series, the corresponding sample entropy SampEn (m, r, N) is calculated.
5. The method for real-time diagnosis of rectifier open circuit fault according to claim 4, wherein said step S5 is implemented as follows:
s5a, in the working process of the rectifier, if a power tube on a bridge arm of the rectifier breaks down, a part of sampling points which are constantly zero appear in the corresponding phase current, and the expression is as follows:
Figure FDA0003408302550000033
wherein, Kth1Is a selected threshold;
the percentage P of the interval with the sampling value of zero in one current periodnComprises the following steps:
Figure FDA0003408302550000034
s5b, in healthy state, PnThe difference between two is approximately 0, if any one exceeds the set threshold value KdetAnd judging that the rectifier has an open-circuit fault:
Figure FDA0003408302550000035
the FD-1 represents that the rectifier has open-circuit fault, and the FD-0 represents that the state of the rectifier is normal;
Figure FDA0003408302550000041
Pa、Pb、Pcthe intervals with sampling values of 0 in a sampling period of the three-phase currents a, b and c respectively account for the percentage;
s5c, according to PnJudging a fault bridge arm:
Pn>Kth1,Sn=0
wherein S isn0 represents that the n-phase bridge arm has open-circuit fault;
under the healthy condition, if an open-circuit fault occurs, half-wave partial waveform loss occurs to the corresponding bridge arm phase current, which is expressed as:
Figure FDA0003408302550000042
KL2、KL1respectively representing the positive half-wave length and the negative half-wave length in one current fundamental wave period; LP represents that a pipe on one bridge arm is in fault, LN represents one bridgeThe underarm tube fails;
if the upper pipe and the lower pipe of the same bridge arm simultaneously break down, the positive half-wave length and the negative half-wave length are equal, the ratio of the zero part of the waveform is increased, and the complexity MSE of the three-phase current is respectively calculated by adopting a multi-scale entropy algorithma、MSEb、MSEcThen, the threshold K is setth2And comparing, positioning the fault, and judging as follows:
when MSEa、MSEb、MSEcOne value of which is constantly greater than the threshold value Kth2And then the double-tube open circuit fault of the same bridge arm occurs.
CN202111522592.XA 2021-12-13 2021-12-13 Real-time diagnosis method for open-circuit fault of rectifier Active CN114236374B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202111522592.XA CN114236374B (en) 2021-12-13 2021-12-13 Real-time diagnosis method for open-circuit fault of rectifier

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111522592.XA CN114236374B (en) 2021-12-13 2021-12-13 Real-time diagnosis method for open-circuit fault of rectifier

Publications (2)

Publication Number Publication Date
CN114236374A true CN114236374A (en) 2022-03-25
CN114236374B CN114236374B (en) 2023-11-14

Family

ID=80755471

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111522592.XA Active CN114236374B (en) 2021-12-13 2021-12-13 Real-time diagnosis method for open-circuit fault of rectifier

Country Status (1)

Country Link
CN (1) CN114236374B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115267606A (en) * 2022-07-26 2022-11-01 合肥工业大学 Open-circuit fault diagnosis method for IGBT (insulated Gate Bipolar translator) in ANPC (adaptive neural network) type three-level grid-connected inverter

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101413989A (en) * 2008-12-01 2009-04-22 南京航空航天大学 Rectification circuit failure diagnosis method of electric excitation biconvex electrode electric machine based on exciting current
TW201301150A (en) * 2011-06-24 2013-01-01 Univ Nat Taiwan Normal Training method for fault detection system
CN104390781A (en) * 2014-11-26 2015-03-04 中国矿业大学 Gear fault diagnosis method based on LMD and BP neural network
CN106124913A (en) * 2016-06-23 2016-11-16 华中科技大学 A kind of open-circuit fault of power tubes detection, localization method
WO2017116582A1 (en) * 2015-12-31 2017-07-06 Brainscope Company, Inc. System and methods for neurological monitoring and assisted diagnosis
CN107301243A (en) * 2017-07-07 2017-10-27 西安电子科技大学 Switchgear fault signature extracting method based on big data platform
CN110441695A (en) * 2019-08-07 2019-11-12 南京佑创汽车研究院有限公司 A kind of battery pack multiple faults error comprehensive diagnosis method combined based on model and signal processing
CN112034394A (en) * 2020-11-03 2020-12-04 中南大学 Rectifier open-circuit fault diagnosis method based on current half-wave difference and electronic equipment
CN112183400A (en) * 2020-09-30 2021-01-05 福州大学 Novel latent fault feature extraction method and system for distribution transformer
CN113567851A (en) * 2021-07-28 2021-10-29 南京航空航天大学 Motor driving system and fault diagnosis method for switching tube thereof

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101413989A (en) * 2008-12-01 2009-04-22 南京航空航天大学 Rectification circuit failure diagnosis method of electric excitation biconvex electrode electric machine based on exciting current
TW201301150A (en) * 2011-06-24 2013-01-01 Univ Nat Taiwan Normal Training method for fault detection system
CN104390781A (en) * 2014-11-26 2015-03-04 中国矿业大学 Gear fault diagnosis method based on LMD and BP neural network
WO2017116582A1 (en) * 2015-12-31 2017-07-06 Brainscope Company, Inc. System and methods for neurological monitoring and assisted diagnosis
CN106124913A (en) * 2016-06-23 2016-11-16 华中科技大学 A kind of open-circuit fault of power tubes detection, localization method
CN107301243A (en) * 2017-07-07 2017-10-27 西安电子科技大学 Switchgear fault signature extracting method based on big data platform
CN110441695A (en) * 2019-08-07 2019-11-12 南京佑创汽车研究院有限公司 A kind of battery pack multiple faults error comprehensive diagnosis method combined based on model and signal processing
CN112183400A (en) * 2020-09-30 2021-01-05 福州大学 Novel latent fault feature extraction method and system for distribution transformer
CN112034394A (en) * 2020-11-03 2020-12-04 中南大学 Rectifier open-circuit fault diagnosis method based on current half-wave difference and electronic equipment
CN113567851A (en) * 2021-07-28 2021-10-29 南京航空航天大学 Motor driving system and fault diagnosis method for switching tube thereof

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
DUAN HAO 等: "Fault Diagnosis of PMSG Wind Power Generation System Based on LMD and MSE", 《COMPLEXITY》, pages 1 - 11 *
张波德 等: "基于无监督学习的MMC子模块开路故障诊断方法", 《电力系统保护与控制》, vol. 49, no. 12, pages 98 - 105 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115267606A (en) * 2022-07-26 2022-11-01 合肥工业大学 Open-circuit fault diagnosis method for IGBT (insulated Gate Bipolar translator) in ANPC (adaptive neural network) type three-level grid-connected inverter
CN115267606B (en) * 2022-07-26 2024-04-05 合肥工业大学 Open-circuit fault diagnosis method for IGBT in ANPC type three-level grid-connected inverter

Also Published As

Publication number Publication date
CN114236374B (en) 2023-11-14

Similar Documents

Publication Publication Date Title
CN105891658B (en) A kind of diagnostic method of open-circuit fault of power tubes of inverter
CN107942219B (en) A kind of open-circuit fault diagnostic method of three-phase four-wire system inverter transistor
CN111077471B (en) NPC three-level inverter open-circuit fault diagnosis method based on instantaneous frequency
CN110907758B (en) Small current ground fault line selection method covering CT polarity self-correction
CN109375029B (en) Open-circuit fault diagnosis method and system for switching device of two-level converter system
CN110426664B (en) Comprehensive diagnosis method for open-circuit fault and current sensor fault of three-phase three-wire system inverter power tube with two current sensors
CN105891659B (en) A kind of wind electric converter open-circuit fault diagnostic method
WO2021184820A1 (en) Two-level three-phase voltage source inverter bridge arm open-circuit fault diagnosis method
WO2015169265A1 (en) Phase current integration method for diagnosing fault in switched reluctance motor power converter
CN116879662B (en) Transformer fault detection method based on data analysis
CN110470465B (en) Circuit breaker testing method and system based on vibration signal analysis
CN103983889B (en) Diagnosis method for active power filter switching device open-circuit fault based on model reference analysis
CN111157827B (en) Method for detecting state of direct current converter valve based on port current time sequence characteristics
CN107102247B (en) A kind of CRH3 type EMU traction inverter IGBT open-circuit fault diagnostic method
CN106841949B (en) Method and device for monitoring stator insulation of three-phase asynchronous alternating current motor on line
CN113358993B (en) Online fault diagnosis method and system for multi-level converter IGBT
CN113534035B (en) Tiny fault diagnosis method for current sensor of new energy electric vehicle driving system
CN114236374B (en) Real-time diagnosis method for open-circuit fault of rectifier
CN110275082B (en) Grounding diagnosis method, system and device for main circuit of converter
CN109725213A (en) Buck converter fault detection method based on reverse Kalman filter
CN105093054A (en) Method for fast diagnosing direction connection of big power rectifier switch tube online
CN112444759A (en) Open-circuit fault diagnosis method and system for power switch device of three-phase three-level rectifier
CN112731193B (en) NPC inverter multi-class fault diagnosis method based on sliding-mode observer
CN112485595B (en) Power distribution network ground fault line selection protection method and device
CN113156336A (en) Method and device for identifying single-tube open-circuit fault of Vienna rectifier in two stages and storage medium

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