CN109683591B - Underwater propeller fault degree identification method based on fusion signal time domain energy and time-frequency entropy - Google Patents
Underwater propeller fault degree identification method based on fusion signal time domain energy and time-frequency entropy Download PDFInfo
- Publication number
- CN109683591B CN109683591B CN201811609960.2A CN201811609960A CN109683591B CN 109683591 B CN109683591 B CN 109683591B CN 201811609960 A CN201811609960 A CN 201811609960A CN 109683591 B CN109683591 B CN 109683591B
- Authority
- CN
- China
- Prior art keywords
- fault
- time
- signal
- propeller
- fusion
- 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
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0243—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model
- G05B23/0254—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults model based detection method, e.g. first-principles knowledge model based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Complex Calculations (AREA)
Abstract
The invention discloses a method for identifying the fault degree of an underwater propeller based on fusion signal time domain energy and time-frequency entropy, which organically fuses two single-aspect fault information, namely speed signal fault information of an underwater robot, propeller control signal fault information and the like, so as to obtain more comprehensive fusion fault information, extracts multi-domain fault characteristics, such as time domain energy, time-frequency entropy and the like, from the fusion fault information, is used for constructing a fault sample, and finally classifies the fault sample based on a support vector field description algorithm to obtain the fault degree of the underwater propeller. The mapping relation between the extracted fault characteristics and the fault degree is unique, the classification of the fault degree of the propeller can be realized, and the classification precision reaches over 95 percent.
Description
Technical Field
The invention belongs to the technology of underwater robot thrusters, and particularly relates to an underwater thruster fault degree identification method based on fusion of signal time domain energy and time-frequency entropy.
Background
The propeller is an important component of the underwater robot power system. The propeller faults will cause the change of dynamic signals such as speed signals of the underwater robot, propeller control voltage signals and the like, and then singular behaviors are generated in the dynamic signals. According to the phenomenon, fault characteristics are extracted from singular behaviors of dynamic signals, fault samples are constructed, a fault classification model is built based on the fault samples, and the fault degree of the underwater robot is diagnosed. The known wavelet energy identification method, a chinese patent with application number 201410705681.1, extracts wavelet energy features from two single aspects of an underwater robot speed signal and a propeller control voltage signal, respectively, for propeller fault degree identification, and does not organically fuse the two types of fault information. In addition, the method extracts the fault characteristics from the signal time domain and the signal frequency domain, and the mapping relation between the obtained fault characteristics and the fault degree is not unique.
Disclosure of Invention
The purpose of the invention is as follows: the invention aims to solve the defects in the prior art and provides a method for identifying the fault degree of an underwater propeller based on fusion signal time domain energy and time-frequency entropy, which fuses two single-aspect fault information, namely speed signal fault information of an underwater robot, propeller control signal fault information and the like, extracts time domain energy characteristics and time-frequency entropy characteristics from the fusion characteristics for constructing fault samples, and finally classifies the fault samples based on a support vector field description algorithm to obtain the fault degree of the underwater propeller.
The technical scheme is as follows: the invention discloses an underwater propeller fault degree identification method based on fusion signal time domain energy and time-frequency entropy, which comprises the following steps of:
first, the length of L is adopted1The time window (for example, the value can be 300) is used for respectively intercepting the speed signal of the underwater robot and the control voltage change rate signal of the propeller;
the second step, the data obtained in the first step is processed with the conventional wavelet decomposition, the wavelet approximate component is extracted, the wavelet approximate component is processed with the conventional correction Bayes operation, and the operation result d is obtainedSA(n), n is a signal data number, n is 1,2, …, L1；
Thirdly, taking the failure of the propeller at the nth time as a focal element BnEstablishing a fault evidence identification framework theta ═ B1,B2,…,BnCalculating a reliability distribution function m (B) of the fault evidence through formula (1)n) Then m (B) is addedn) Confidence distribution function m instantiated as evidence of speed signal faultU(Bn) Control signal fault evidence credibility distribution function mC(Bn) M isU(Bn)、mC(Bn) Substituting the formulas (2) to (3) for fusion to obtain a fusion signal fault evidence credibility distribution function mF(Bn)；
In the formula (d)SA(n) is the result of Bayesian calculation for wavelet correction of dynamic signal of underwater robot, i1As temporary variable, i1＝1，2，…，N5，N5For the length of the time window, i.e. N5＝L1＝300，i2And j2Is the number of the coke element, namely the value is 1 to L1Positive integer of (1), K5Is an intermediate process variable;
fourthly, extracting fusion signal time domain energy fault characteristics FTP：
Function m is distributed to fusion signal fault evidence credibilityF(Bn) Performing a convolution calculation, i.e. mconv(n)＝mF(Bn)*mF(Bn) Determining the convolution calculation result mconv(n) summing the data between two adjacent minimum value points, taking the result as the time domain energy of the region between the two minimum value points, and taking m as the time domain energy of the region between the two minimum value pointsconvThe maximum value of the time domain energy in (n) is used as the time domain energy of the fusion signalBarrier characteristic value FTP；
Fifthly, extracting the time-frequency entropy fault characteristics of the smooth pseudo-Vigrener-Weili distribution of the fusion signal:
calculating a fusion signal credibility distribution function m by adopting a smooth pseudo-Wigner-Willi distribution algorithm as shown in a formula (4)F(Bn) The smoothed pseudo-wigner-willi spectrum SPWVD (n, m), and the shannon entropy F of the smoothed pseudo-wigner-willi spectrum SPWVD (n, m) is calculatedTFHAs shown in the formulas (5) to (6), the obtained result is used as the fusion signal time-frequency entropy fault characteristic value FTFH；
p(n,m)＝|SPWVD(n,m)|/∑∑|SPWVD(n,m)| (5)
FTFH＝-∑∑p(n,m)log2 p(n,m) (6)
Wherein SPWVD (n, m) is a smooth pseudo-Vigrener-Weiley spectrum, n is the time beat, and n is 1-L1M is a frequency band number, m is 1 to N4An integer of between, N4Usually 256 and 512, e.g. N4＝512；
h(k1)、g(l2) Are all gaussian functions;
wherein k is1∈[-(K1-1),(K1-1)]，l2∈[-(L2-1),(L2-1)]，K1、L2Are respectively not more than (N)4)/4、(N4) The maximum integer of/5, z (n) is the analytic value of the velocity signal, z x (n) is conjugate to z (n), i is an imaginary number, i.e., i2＝-1，FTFHThe fusion signal time-frequency entropy fault characteristic value is obtained;
sixthly, fusing signal time domain energy fault characteristics FTPFusion signal time-frequency entropy fault feature FTFHThe component fault samples x ═ FTP FTFH]TThe length in the step one is L1The time window function slides to the right, and each time the time window function slides by one time beat, a group of underwater machines is obtainedRepeating the first step to the sixth step to obtain a fault sample, and moving N6A time beat, obtain N6A fault sample, N6Is any positive integer, which is the number of failed samples, in principle, the larger the better, e.g. taking N6＝100；
Seventhly, establishing a hypersphere model S by adopting a conventional support vector field description method to obtain a hypersphere center C and a hypersphere model radius R of the hypersphere model S;
eighthly, classifying the failure degree of the propeller: obtaining the fault degree of the propeller as lambdaqThe underwater robot dynamic signal data of the time, Q is the propeller fault degree grade, Q is 1,2,3, …, Q, Q is the propeller fault degree grade quantity;
establishing a plurality of hypersphere models according to the contents of the first step to the seventh step q1,2,3, …, Q, and each hypersphere modelqS, using a super ball centerqC and radius of hypersphereqR is described;
calculating test sample to hypersphere modelqS sphere centerqGeneralized distance of CqD, by the formulaqε＝qD/qR-computing test sample-to-hypersphere modelqRelative distance of SqEpsilon; relative distanceqHypersphere model corresponding to epsilon minimum valueqDegree of failure λ represented by SqI.e. degree of propeller failure λq。
Has the advantages that: the method organically fuses the fault information of different aspects to obtain more comprehensive fused fault information, extracts multi-domain fault characteristics such as time domain energy, time-frequency entropy and the like from the fused fault information, has a unique mapping relation between the fault characteristics and the fault degree, can realize classification of the fault degree of the propeller, and has the classification precision of more than 95%.
Drawings
FIG. 1 is an overall flow chart of the present invention;
FIG. 2 is a time domain waveform diagram of a dynamic signal of an underwater robot in an embodiment;
FIG. 3 is a graph of a fused signal fault signature distribution in an embodiment;
FIG. 4 is a sample distribution diagram of propeller failure in an embodiment;
FIG. 5 is a graph showing the relative distance from a failure sample with a failure degree of 0% to each single type hyper-sphere model in the example;
FIG. 6 is a graph showing the relative distance from a failure sample of 10% failure degree to each single type hyper-sphere model in the example;
FIG. 7 is a graph showing the relative distance from a failure sample of 20% failure degree to each single type hyper-sphere model in the example;
FIG. 8 is a graph showing the relative distance from a failure sample of 30% failure degree to each single type hyper-sphere model in the example;
fig. 9 is a graph showing the relative distance from a failure sample of 40% failure degree to each single type hypersphere model in the example.
Detailed Description
The technical solution of the present invention is described in detail below, but the scope of the present invention is not limited to the embodiments.
As shown in figure 1 of the drawings, in which,
the invention discloses an underwater propeller fault degree identification method based on fusion signal time domain energy and time-frequency entropy, which comprises the following steps of:
first, the length of L is adopted1Intercepting a speed signal of the underwater robot and a control voltage change rate signal of a propeller respectively by a time window of 300;
the second step, the data obtained in the first step is processed with the conventional wavelet decomposition, the wavelet approximate component is extracted, the wavelet approximate component is processed with the conventional correction Bayes operation, and the operation result d is obtainedsA(n); n is a signal data serial number, n is 1,2, …, L1；
Thirdly, taking the failure of the propeller at the nth time as a focal element BnEstablishing a fault evidence identification framework theta ═ B1,B2,…,BnCalculating a reliability distribution function m (B) of the fault evidence through formula (1)n) Then m (B) is addedn) Confidence distribution function m instantiated as evidence of speed signal faultU(Bn) And control signal failure evidence canConfidence score function mC(Bn) M isU(Bn)、mC(Bn) Substituting the formulas (1) to (3) for fusion to obtain a fusion signal fault evidence credibility distribution function mF(Bn)；
In the formula (d)SA(N) is the result of Bayesian calculation for wavelet modification of dynamic signals of the underwater robot, N5For the length of the time window, i.e. N5＝300；
Fourthly, extracting fusion signal time domain energy fault characteristics FTP：
Function m is distributed to fusion signal fault evidence credibilityF(Bn) Performing a convolution calculation, i.e. mconv(n)＝mF(Bn)*mF(Bn) Determining the convolution calculation result mconv(n) summing the data between two adjacent minimum value points, taking the result as the time domain energy of the region between the two minimum value points, and taking m as the time domain energy of the region between the two minimum value pointsconvThe maximum value of the time domain energy in (n) is used as the fusion signal time domain energy fault characteristic value FTE；
Fifthly, extracting the time-frequency entropy fault characteristics of the smooth pseudo-Vigrener-Weili distribution of the fusion signal:
calculating a fusion signal credibility distribution function m by adopting a smooth pseudo-Wigner-Willi distribution algorithm as shown in a formula (4)F(Bn) The smoothed pseudo-wigner-willi spectrum SPWVD (n, m), and the shannon entropy F of the smoothed pseudo-wigner-willi spectrum SPWVD (n, m) is calculatedTFHAs shown in equations (5) to (6)Showing that the obtained result is taken as a fusion signal time-frequency entropy fault characteristic value FTFH；
p(n,m)＝|SPWVD(n,m)|/∑∑|SPWVD(n,m)| (5)
FTFH＝-∑∑p(n,m)log2 p(n,m) (6)
Wherein SPWVD (n, m) is a smooth pseudo-Vigrener-Weiley spectrum, n is the time beat, and n is 1-L1M is a frequency band number, m is 1 to N4An integer of between, N4Usually 256 and 512, in this embodiment, N4＝512，h(k1)、g(l2) Are all gaussian functions;
wherein k is1∈[-(K1-1),(K1-1)]，l2∈[-(L2-1),(L2-1)]，K1、L2Are respectively not more than (N)4)/4、(N4) Maximum integer of/5, z (n) is the analytic value of the velocity signal, z x (n) is conjugate to z (n), FTFHThe fusion signal time-frequency entropy fault characteristic value is obtained;
sixthly, fusing signal time domain energy fault characteristics FTEFusion signal time-frequency entropy fault feature FTFHThe component fault samples x ═ FTE FTFH]TSliding the time window function to the right, obtaining a group of underwater robot dynamic signal data every time the time window function slides by a time beat, repeating the first step to the sixth step to obtain a fault sample, and moving N6A time beat until N is obtained6A fault sample;
seventhly, establishing a hypersphere model S by adopting a conventional support vector field description method to obtain a hypersphere center C and a hypersphere model radius R of the hypersphere model S;
eighthly, classifying the failure degree of the propeller: obtaining the fault degree of the propeller as lambdaqThe underwater robot dynamic signal data of the time, Q is the propeller fault degree grade, Q is 1,2,3, …, Q, Q is the propeller fault rangeNumber of degree grades;
establishing a plurality of hypersphere models according to the contents of the first step to the seventh step q1,2,3, …, Q, and each hypersphere modelqS, using a super ball centerqC and radius of hypersphereqR is described;
calculating test sample to hypersphere modelqS sphere centerqGeneralized distance of CqD, by the formulaqε＝qD/qR-computing test sample-to-hypersphere modelqRelative distance of SqEpsilon; relative distanceqHypersphere model corresponding to epsilon minimum valueqDegree of failure λ represented by SqI.e. degree of propeller failure λq。
Example 1:
as shown in fig. 2(a), the underwater robot starts from a standstill, the speed of the underwater robot and the control voltage of the propeller gradually increase, the underwater robot starts to operate at a steady speed of 0.3m/s from the 101 th time beat, and the propeller has a power loss fault at the 250 th time beat, wherein the fault degrees are respectively 0%, 10%, 20%, 30% and 40% until the experiment is finished. As shown by oval boxes in fig. 2(b) and 2(c), the speed signal of the underwater robot and the propeller control voltage change rate signal form a singular signal in 250 th to 350 th time beats.
With a length L1Capturing the speed signal and the control voltage change rate signal data from 101 th to 400 th in fig. 2(b) and 2(c) by a time window of 300, extracting fusion signal time domain energy fault features and time-frequency entropy fault features from the captured data, constructing fault samples, moving the time window to the right by 100 time beats, obtaining a group of fault samples every time beat, obtaining 500 fault samples in total, wherein the distribution of the fault features is shown in fig. 3, and the distribution of the fault samples is shown in fig. 4.
In fig. 3(a), the time-frequency entropy features corresponding to a larger fault degree are always smaller than the time-frequency entropy features corresponding to a smaller fault degree, in fig. 3(b), the time-domain energy features corresponding to a larger fault degree are always larger than the time-domain energy features corresponding to a fault degree, and the data in fig. 3 indicate that the fault degrees and the fault features are in a monotonous corresponding relationship, one fault feature only corresponds to one fault degree, and the mapping relationship between the fault features and the fault degrees is unique.
From the fault samples shown in fig. 4, 50% of the fault samples are randomly selected as training samples, the remaining 50% of the fault samples are used as test samples, and the fault sample division results are shown in table 1.
Table 1 fault sample partitioning results
The training samples in table 1 are used to establish a single-class hypersphere model for each fault degree, the test sample corresponding to a certain fault degree is used as the target sample of the fault degree, the test samples corresponding to other fault degrees are used as the non-target samples of the fault degree, the classification performance of the single-class hypersphere model corresponding to the fault degree is calculated, the index AUC is used to describe, and the result is shown in table 2. In table 2, AUC is an area enclosed by the coordinate axis under the ROC curve, ROC is a receiver operation characteristic curve, the larger AUC is, the better the classifier effect is, and the extremum of AUC is 1.
TABLE 2 AUC of fusion signal single-class hypersphere model
As can be seen from table 2, the AUC of the single-class hypersphere model corresponding to the failure degree of 10% is 0.98, and the AUC of the remaining single-class hypersphere models are 1, which indicates that the AUC of all the single-class hypersphere models is higher than 0.95, and the classification effect of the single-class hypersphere model established by the invention is better.
The relative distance from the test sample corresponding to the failure degree of 0% to the single-type hypersphere model corresponding to each failure degree is calculated, and the result is shown in fig. 5. In the same manner, the relative distances from the test samples corresponding to the failure degrees of 10%, 20%, 30% and 40% to the single-type hypersphere model corresponding to each failure degree are calculated, and the results are shown in fig. 6 to 9. In fig. 5 to 9, the single type hypersphere model with the smallest relative distance is used as the type of the test sample. And dividing the number of the correct classification samples by the total number of the test samples to obtain the classification precision of the classification model of the propeller fault degree. According to the results shown in fig. 5 to 9, the classification accuracy of the fault degree of the propeller is calculated to be 95.2% and more than 95%, and the classification effect is good.
In addition, according to the fault feature extraction and fault degree classification process, time domain energy and time-frequency entropy features are extracted from the speed signals, fault samples are constructed, a fault classification model is built, fault degree classification is carried out, a single-class hypersphere model AUC is obtained, and the fault degree classification precision is 90.0% as shown in Table 3. By adopting the same process, time domain energy and time-frequency entropy characteristics are extracted from the control voltage change rate signal, a fault sample is constructed, a fault classification model is established, fault degree classification is carried out, a single-class hypersphere model AUC is obtained, and the fault degree classification precision is 72.4% as shown in Table 3.
TABLE 3 AUC of speed signal and control signal single-class hypersphere model
Comparing data in table 2 and table 3, it can be seen that, when the AUC of the fusion signal single-class hypersphere model is compared with the AUC of the speed signal single-class hypersphere model, the AUC of the fusion signal single-class hypersphere model is greater than those of the speed signal single-class hypersphere model, and the rest is equal, when the AUC of the fusion signal single-class hypersphere model is compared with that of the control signal single-class hypersphere model, the AUC of the fault degree 40% is equal to that of the control signal single-class hypersphere model, and the rest is greater, which indicates that the classification performance of the fusion signal single-class hypersphere model is better than that of.
In addition, in the aspect of classification precision, the classification precision of the fusion signal fault degree classification model is 95.2 percent, is greater than the classification precision of the speed signal fault degree classification model by 90.0 percent, and is greater than the classification precision of the control signal fault degree classification model by 72.4 percent.
Claims (1)
1. A method for identifying the fault degree of an underwater propeller based on fusion of signal time domain energy and time-frequency entropy is characterized by comprising the following steps: the method comprises the following steps:
first, the length of L is adopted1The time windows respectively intercept the speed signal of the underwater robot and the control voltage change rate signal of the propeller;
second, the data obtained in the first step is processed with conventional wavelet decomposition to extract approximate component s of waveletA(n) for wavelet approximation component sA(n) carrying out conventional correction Bayes operation to obtain an operation result dsA(n)；
Thirdly, taking the failure of the propeller at the nth time as a focal element BnEstablishing a fault evidence identification framework theta ═ B1,B2,L,BnCalculating a reliability distribution function m (B) of the fault evidence through formula (1)n) Then m (B) is addedn) Confidence distribution function m instantiated as evidence of speed signal faultU(Bn) Control signal fault evidence credibility distribution function mC(Bn) M isU(Bn)、mC(Bn) Substituting the formulas (2) to (3) for fusion to obtain a fusion signal fault evidence credibility distribution function mF(Bn)；
In the formula (d)SA(n) is the result of Bayesian calculation for wavelet correction of dynamic signal of underwater robot, i1As temporary variable, i1＝1，2，…，N5，N5Is the length of the time window, i.e. N5＝L1，i2And j2Is the number of the coke element, namely the value is 1 to L1Positive integer of (1), K5Is an intermediate process variable;
fourthly, extracting fusion signal time domain energy fault characteristics FTP：
Function m is distributed to fusion signal fault evidence credibilityF(Bn) Performing a convolution calculation, i.e. mconv(n)＝mF(Bn)*mF(Bn) Determining the convolution calculation result mconv(n) summing the data between two adjacent minimum value points, taking the result as the time domain energy of the region between the two minimum value points, and taking m as the time domain energy of the region between the two minimum value pointsconvThe maximum value of the time domain energy in (n) is used as the fusion signal time domain energy fault characteristic value FTP；
Fifthly, extracting the time-frequency entropy fault characteristics of the smooth pseudo-Vigrener-Weili distribution of the fusion signal:
calculating a fusion signal credibility distribution function m by adopting a smooth pseudo-Wigner-Willi distribution algorithm as shown in a formula (4)F(Bn) The smoothed pseudo-wigner-willi spectrum SPWVD (n, m), and the shannon entropy F of the smoothed pseudo-wigner-willi spectrum SPWVD (n, m) is calculatedTFHAs shown in the formulas (5) to (6), the obtained result is used as the fusion signal time-frequency entropy fault characteristic value FTFH；
p(n,m)＝|SPWVD(n,m)|/∑∑|SPWVD(n,m)| (5)
FTFH＝-∑∑p(n,m)log2p(n,m) (6)
Wherein SPWVD (n, m) is a smooth pseudo-Vigrener-Weiley spectrum, n is the time beat, and n is 1-L1M is an integer ofM is a frequency band number of 1 to N4An integer of h (k) between1)、g(l2) Are all gaussian window functions;
wherein k is1∈[-(K1-1),(K1-1)]，l2∈[-(L2-1),(L2-1)]，K1、L2Are respectively not more than (N)4)/4、(N4) The maximum integer of/5, z (n) is the analytic value of the velocity signal, z x (n) is conjugate to z (n), i is an imaginary number, i.e., i2＝-1，FTFHThe fusion signal time-frequency entropy fault characteristic value is obtained;
sixthly, fusing signal time domain energy fault characteristics FTPFusion signal time-frequency entropy fault feature FTFHThe component fault samples x ═ FTP FTFH]TThe length in the step one is L1Sliding the time window function to the right, obtaining a group of underwater robot dynamic signal data every time the time window function slides by a time beat, repeating the first step to the sixth step to obtain a fault sample, and moving N6A time beat, obtain N6A fault sample, N6Is any positive integer;
seventhly, establishing a hypersphere model S by adopting a conventional support vector field description method to obtain a hypersphere center C and a hypersphere model radius R of the hypersphere model S;
eighthly, classifying the failure degree of the propeller: obtaining the fault degree of the propeller as lambdaqThe underwater robot dynamic signal data of the time, Q is the propeller fault degree grade, Q is 1,2,3, …, Q, Q is the propeller fault degree grade quantity;
establishing a plurality of hypersphere models according to the contents of the first step to the seventh stepq1,2,3, …, Q, and each hypersphere modelqS, using a super ball centerqC and radius of hypersphereqR is described;
calculating test sample to hypersphere modelqS sphere centerqGeneralized distance of CqD, by the formulaqε＝qD/qR-computing test sample-to-hypersphere modelqRelative distance of SqEpsilon; relative distanceqEpsilon minimum value correspondenceSuper ball modelqDegree of failure λ represented by SqI.e. degree of propeller failure λq。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811609960.2A CN109683591B (en) | 2018-12-27 | 2018-12-27 | Underwater propeller fault degree identification method based on fusion signal time domain energy and time-frequency entropy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811609960.2A CN109683591B (en) | 2018-12-27 | 2018-12-27 | Underwater propeller fault degree identification method based on fusion signal time domain energy and time-frequency entropy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109683591A CN109683591A (en) | 2019-04-26 |
| CN109683591B true CN109683591B (en) | 2021-03-19 |
Family
ID=66190017
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811609960.2A Active CN109683591B (en) | 2018-12-27 | 2018-12-27 | Underwater propeller fault degree identification method based on fusion signal time domain energy and time-frequency entropy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109683591B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112015160B (en) * | 2019-05-31 | 2021-10-22 | 北京新能源汽车股份有限公司 | Fault temperature determination method and device |
| CN110716532A (en) * | 2019-10-17 | 2020-01-21 | 哈尔滨工程大学 | Underwater robot propeller weak fault identification method based on wavelet packet energy and FFT |
| CN112034826B (en) * | 2020-09-11 | 2021-09-24 | 江苏科技大学 | Method for identifying fault degree of underwater propeller based on least square method |
| CN112051836B (en) * | 2020-09-11 | 2021-09-24 | 江苏科技大学 | Underwater robot propeller state monitoring method based on multi-core model |
| CN113655778B (en) * | 2021-08-16 | 2024-03-26 | 江苏科技大学 | Underwater propeller fault diagnosis system and method based on time-frequency energy |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104503432A (en) * | 2014-11-27 | 2015-04-08 | 哈尔滨工程大学 | Autonomous underwater robot fault identification method based on wavelet energy |
| CN107132760A (en) * | 2017-03-27 | 2017-09-05 | 哈尔滨工程大学 | The underwater robot state monitoring method described based on fuzzy support vector domain |
| CN107133376A (en) * | 2017-03-27 | 2017-09-05 | 哈尔滨工程大学 | A kind of weak fault degree Forecasting Methodology of autonomous type underwater robot propeller based on grey forecasting model |
| CN109033965A (en) * | 2018-06-22 | 2018-12-18 | 江苏科技大学 | A kind of underwater robot propeller failure time-frequency characteristics Enhancement Method |
-
2018
- 2018-12-27 CN CN201811609960.2A patent/CN109683591B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104503432A (en) * | 2014-11-27 | 2015-04-08 | 哈尔滨工程大学 | Autonomous underwater robot fault identification method based on wavelet energy |
| CN107132760A (en) * | 2017-03-27 | 2017-09-05 | 哈尔滨工程大学 | The underwater robot state monitoring method described based on fuzzy support vector domain |
| CN107133376A (en) * | 2017-03-27 | 2017-09-05 | 哈尔滨工程大学 | A kind of weak fault degree Forecasting Methodology of autonomous type underwater robot propeller based on grey forecasting model |
| CN109033965A (en) * | 2018-06-22 | 2018-12-18 | 江苏科技大学 | A kind of underwater robot propeller failure time-frequency characteristics Enhancement Method |
Non-Patent Citations (4)
| Title |
|---|
| 旋转机械设备关键部件故障诊断与预测方法研究;沈长青;《中国博士学位论文全文数据库工程科技Ⅱ辑》;20141030(第10期);全文 * |
| 水下机器人推进器故障特征分离与故障程度辨识方法研究;殷宝吉;《中国博士学位论文全文数据库信息科技辑》;20180630(第6期);全文 * |
| 自主式水下机器人故障特征增强方法;张铭钧，刘维新，殷宝吉，王玉甲;《哈尔滨工程大学学报》;20140930;第35卷(第9期);全文 * |
| 随机干扰下ＡＵＶ推进器故障特征提取与融合;张铭钧,殷宝吉,刘维新,王玉甲;《华中科技大学学报》;20150630;第43卷(第6期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109683591A (en) | 2019-04-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109683591B (en) | Underwater propeller fault degree identification method based on fusion signal time domain energy and time-frequency entropy | |
| CN109410242B (en) | Target tracking method, system, equipment and medium based on double-current convolutional neural network | |
| CN109727246B (en) | Comparative learning image quality evaluation method based on twin network | |
| CN110826630A (en) | Radar interference signal feature level fusion identification method based on deep convolutional neural network | |
| CN110728360A (en) | Micro-energy device energy identification method based on BP neural network | |
| CN109696906B (en) | Underwater robot propeller fault diagnosis method based on wavelet correction Bayes convolution energy | |
| CN109740057B (en) | Knowledge extraction-based enhanced neural network and information recommendation method | |
| CN105160303A (en) | Fingerprint identification method based on mixed matching | |
| Theodoridis et al. | Cross-modal variational alignment of latent spaces | |
| CN112014108A (en) | Bearing fault diagnosis method based on LMD and improved PSO (particle swarm optimization) optimized BP (Back propagation) neural network | |
| Salekshahrezaee et al. | Feature extraction for class imbalance using a convolutional autoencoder and data sampling | |
| Pérez-Ortiz et al. | Machine learning techniques to select Be star candidates-An application in the OGLE-IV Gaia south ecliptic pole field | |
| CN115687970A (en) | Method for improving electromyographic signal identification accuracy | |
| CN105184654A (en) | Public opinion hotspot real-time acquisition method and acquisition device based on community division | |
| Mamun et al. | Bangla handwritten digit recognition approach with an ensemble of deep residual networks | |
| Zha et al. | LT-SEI: Long-tailed specific emitter identification based on decoupled representation learning in low-resource scenarios | |
| CN109685823B (en) | Target tracking method based on deep forest | |
| Yin et al. | A multi-scale graph convolutional neural network framework for fault diagnosis of rolling bearing | |
| CN109190471B (en) | Attention model method for video monitoring pedestrian search based on natural language description | |
| Lin et al. | Domestic activities clustering from audio recordings using convolutional capsule autoencoder network | |
| CN117606801A (en) | Cross-domain bearing fault diagnosis method based on multi-characterization self-adaptive network | |
| CN109919057B (en) | Multi-mode fusion gesture recognition method based on efficient convolutional neural network | |
| CN104504181B (en) | A kind of signal envelope extracting method based on sparse recovery | |
| CN112488238B (en) | Hybrid anomaly detection method based on countermeasure self-encoder | |
| US20140343944A1 (en) | Method of visual voice recognition with selection of groups of most relevant points of interest |
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 |