CN116243683A - Method for diagnosing faults of propulsion system based on torque and multi-head self-encoder - Google Patents
Method for diagnosing faults of propulsion system based on torque and multi-head self-encoder Download PDFInfo
- Publication number
- CN116243683A CN116243683A CN202310244944.2A CN202310244944A CN116243683A CN 116243683 A CN116243683 A CN 116243683A CN 202310244944 A CN202310244944 A CN 202310244944A CN 116243683 A CN116243683 A CN 116243683A
- Authority
- CN
- China
- Prior art keywords
- torque
- encoder
- propeller
- estimated
- fault
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000003745 diagnosis Methods 0.000 claims abstract description 18
- 238000001514 detection method Methods 0.000 claims abstract description 9
- 238000013528 artificial neural network Methods 0.000 claims description 22
- 230000035772 mutation Effects 0.000 claims description 5
- 230000033228 biological regulation Effects 0.000 claims description 3
- 230000007423 decrease Effects 0.000 claims description 3
- 238000000354 decomposition reaction Methods 0.000 abstract description 4
- 230000008859 change Effects 0.000 description 4
- 239000000284 extract Substances 0.000 description 4
- 230000009467 reduction Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 239000003550 marker Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012847 principal component analysis method Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000013535 sea water Substances 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- 238000004804 winding Methods 0.000 description 1
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
- G05B23/0221—Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Control Of Electric Motors In General (AREA)
Abstract
The invention discloses a propulsion system fault diagnosis method based on torque and a multi-head self-encoder, which comprises the following steps: firstly, estimating the load of a motor and the torque of a propeller; then detecting faults based on a multi-head self-encoder, wherein the multi-head self-encoder comprises an independent encoder 1 and an independent encoder 2, inputting estimated motor load torque and control signals into the encoder 1 to extract a feature 1, inputting estimated propeller torque and control signals into the encoder 2 to extract a feature 2, connecting the feature 1 and the feature 2 and feeding the feature 2 into a decoder, and realizing fault detection through signal reconstruction; finally, based on a fault detection result, the method reduces the dimension by utilizing singular value decomposition, and inputs the dimension-reduced data into a clustering algorithm to realize the classification of faults.
Description
Technical Field
The invention belongs to the field of autonomous underwater vehicle fault diagnosis, and particularly relates to a propulsion system fault diagnosis method based on torque and a multi-head self-encoder.
Background
Autonomous Underwater Vehicles (AUVs) are becoming increasingly popular for wide range of applications and acceptance in defense, marine and industrial applications. AUVs must be secured for safety and reliability as an important tool for marine exploration. Therefore, the fault diagnosis technique has become one of the most important research subjects in this field. Currently, most commercial AUVs mainly use underwater propellers as actuators when navigating underwater, and therefore propeller failure is one of the most common failure sources for AUVs. In fact, in case of failure of the propulsion system of the AUV, not only is the task not completed, but the AUV itself is also at risk of being lost or destroyed. In order to avoid the situation, a timely and effective fault diagnosis strategy is researched, which is beneficial to reducing the damage risk of the AUV, avoiding the deep propagation of faults and has important significance for ensuring the safety of the AUV and improving the maneuverability in a complex marine environment.
Since the 90 s of the 20 th century, a great deal of research has been conducted on propeller fault diagnosis. Currently, methods for fault diagnosis of a propeller are classified into three main categories, including: based on analytical models, based on data-driven and based on hybrid approaches. However, the existing fault diagnosis method for the propulsion system mainly has the following problems: 1. the method for directly diagnosing the faults of the propeller mostly depends on feedback information of the propeller, and once the feedback information is wrong, the faults are wrongly diagnosed; 2. manual feature extraction may be required in fault location, and diagnosis efficiency is low.
Disclosure of Invention
Aiming at the defects of poor fault diagnosis precision, low efficiency and the like in the prior art, the invention provides a propulsion system fault diagnosis method based on a torque and multi-head convolution self-encoder so as to realize accurate diagnosis of faults.
The invention is realized by adopting the following technical scheme: a propulsion system fault diagnosis method based on torque and multi-head self-encoders comprises the following steps:
step A, load torque Q of motor m And propeller torque Q p Estimating;
step B, detecting faults based on a multi-head self-encoder, wherein the multi-head self-encoder comprises an independent encoder 1 and an independent encoder 2;
the motor load torque Q estimated in the step A is calculated m And a control signal s is input to the encoder 1 to extract its characteristic 1, and an estimated propeller torque Q is input to the encoder p And the control signal s is input into the encoder 2 to extract the characteristic 2, then the characteristic 1 and the characteristic 2 are connected and fed into the decoder, and fault detection is realized through signal reconstruction;
and C, classifying faults based on the fault detection result of the step B.
Further, in the step a, the propeller torque is estimated based on the improved neural network, specifically:
the loss function of the neural network is improved, the advancing speed of the equipment is deduced according to the torque equation of the propeller, and the output of the neural network is the propeller torque:
wherein 
wherein ,
representing an estimated torque coefficient calculated from the estimated torque of the neural network,/->
Representing an estimated propeller inflow speed, n, estimated from an estimated torque coefficient p Indicating the rotational speed of the propeller>
Representing the 5 th power of the propeller diameter, ρ represents the sea water density, ucos (θ) cos (ψ) represents the actual inflow velocity of the propeller, δ, calculated approximately from the measured values u Threshold representing AUV speed deviation, super parameter gamma PHY Control standard Loss NN Loss of PHY Balance between.
Further, in the step C:
if Q m Mutation of Q p And s remains the same trend, then diagnosing as a current fault;
if Q p Mutation of Q m If the trend of the motor speed is consistent with that of s, diagnosing the motor speed fault;
if Q m Near zero, the device speed decreases, the motor speed increases with the control signal s to maximum, then Q p If the propeller is at the maximum, diagnosing that the propeller is lost;
if the control signal s increases gradually with the deceleration of the propulsion system, the motor current increases to a maximum, Q p Zero, Q m If the maximum torque is increased, diagnosing that the propeller is wound;
if Q m and Qp At zero, the control signal s reaches a maximum value under the regulation of the control system of the device, and an open circuit fault is diagnosed.
Compared with the prior art, the invention has the advantages and positive effects that:
in the scheme, an extended state observer and a neural network based on physical guidance are utilized to estimate motor load and propeller torque respectively, a multi-head convolution self-encoder is utilized to extract characteristics of motor load and propeller torque respectively, and density-based spatial clustering with noise is utilized to realize the distinction of propeller current feedback faults, motor speed feedback faults and propeller loss, propeller breakage and open-circuit faults;
by introducing the multi-head convolution self-encoder, the two independent encoders respectively extract the characteristics of the motor load and the control signal and the propeller torque and the control signal, so that the characteristic aliasing caused when the single convolution encoder extracts the characteristics can be avoided; performing dimension reduction processing on the features extracted by the two encoders by utilizing singular value decomposition, inputting the two dimension reduced features into a clustering algorithm, and clustering the feature points to realize classification of different faults; the fault detection method does not depend on a large amount of fault data, and can realize the identification and the determination of various faults such as loss, falling-off, open circuit and the like of the propeller by only depending on the measurement data of the sensor carried by the AUV.
Drawings
FIG. 1 is a schematic flow chart of a fault diagnosis method according to an embodiment of the invention;
fig. 2 is a schematic diagram of a multi-headed convolutional self-encoder according to an embodiment of the present invention.
Detailed Description
In order that the above objects, features and advantages of the invention will be more readily understood, a further description of the invention will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced otherwise than as described herein, and therefore the present invention is not limited to the specific embodiments disclosed below.
In general, the torque of the propeller may be approximately represented by the motor load, or may be calculated using a propeller torque equation. The current and motor speed are mainly utilized in estimating the motor load, while the motion state of the AUV and motor speed are used in estimating the propeller torque. The motor load and the propeller torque may represent the state of the propeller from different angles. In order to improve the accuracy of propeller torque estimation and enhance generalization thereof, the invention adds a physical constraint item based on AUV speed on the basis of a conventional neural network so as to ensure that the estimated torque meets the speed constraint. The fault is then identified using a multi-headed self-encoder whose input and output residuals would exceed a set threshold when the fault occurs. After the faults are identified, the characteristics of motor load and propeller torque are automatically extracted by two independent encoders in the multi-head self-encoding self-encoder respectively, the singular value decomposition is utilized to reduce the dimension of the motor load and the propeller torque, and the dimension-reduced data are input into a clustering algorithm to realize the classification of the faults.
The embodiment provides a propulsion system fault diagnosis method based on torque and multi-head self-encoders, as shown in fig. 1, comprising the following steps:
step A, estimating the load of a motor and the torque of a propeller;
step B, fault detection is carried out based on a multi-head self-encoder, wherein the multi-head self-encoder comprises an independent encoder 1 and an independent encoder 2,
inputting the motor load torque and the control signal estimated in the step A into an encoder 1 to extract the characteristic 1 of the motor load torque and the control signal, inputting the estimated propeller torque and the control signal into an encoder 2 to extract the characteristic 2 of the propeller torque and the control signal, connecting the characteristic 1 and the characteristic 2 and feeding the characteristic into a decoder, and realizing fault detection through signal reconstruction;
and C, performing dimension reduction on the fault based on the fault identification result in the step B by utilizing singular value decomposition, and inputting the dimension reduced data into a clustering algorithm to realize fault classification.
Specifically, in step a, when estimating the load torque and the propeller torque, the following method is specifically adopted:
(1) Motor load torque estimation
In the embodiment, the load of the motor is estimated by using the extended state observer, and the motion equation of the motor is shown in formula (1):
wherein ,Te Is the electromagnetic torque of the motor, Q m Is the load torque of the motor, J M Is the rotational inertia of the motor, B v Is the friction coefficient of the motor and is characterized by comprising a friction coefficient of the motor,n is the motor speed. When the working state of the motor is stable, the electromagnetic torque of the motor can be represented by current, the load torque can be equivalent to a propeller, the acceleration of the motor is 0, and the rotating speed of the motor is constant.
Definition λ= (B) v n+Q m )/J M As a new variable, formula (1) is rewritten as formula (2):
where Cm is the motor torque coefficient and y is the system output.
The extended state observer designed according to the present embodiment of the formula (2) is as shown in the formula (3):
wherein ,
is an estimate of n>
Is an estimated value of lambda, k 1 and k2 Is the observation gain designed for the extended state observer. When the observer converges according to +.>
An observed motor load can be obtained.
(2) Estimating propeller torque based on improved neural network
Consider that an AUV has not only vertical but also horizontal motion. When the propeller is far from the water surface, its vertical displacement has little effect on the performance of the propeller, and therefore, the torque equation of the propeller mounted on the AUV is written as formula (4):
where g () is a torque polynomial, θ and T θ Is the change in pitch angle and period caused by pitch motion, ψ and T ψ Is the change in pitch angle and period due to yaw motion, h is the change in depth of the device (AUV), u represents the device speed, ω represents the wake coefficient of the propeller, n p The rotating speed of the propeller is represented,
represents the 5 th power of the propeller diameter and ρ represents the sea water density.
Approximating the propeller moment equation with a neural network is an effective method in complex sea conditions due to the non-linearity and uncertainty term in g (), however, one major limitation of using a neural network is that it depends on the available marker data and the trained neural network cannot be interpreted with physical principles. Especially when the data set does not cover all cases, false relations are easily found, which look good on the data set but cannot be summarized well beyond the available marker data. A more serious problem with neural networks is that their predictions lack consistency with known laws of physics. Therefore, the method improves the loss function of the neural network, so that the prediction result of the neural network is more in line with the physical rule. As can be seen from equation (4), when Q is obtained p At this time, the advance speed of the AUV can be derived from g (), and therefore (4) can be rewritten as (5).
wherein 
wherein ,
representing an estimated torque coefficient calculated from the estimated torque of the neural network,/->
Represents the estimated propeller inflow velocity estimated from the estimated torque coefficient, ucos (θ) cos (ψ) represents the actual propeller inflow velocity, δ, calculated approximately from the measured values u Threshold representing AUV speed deviation, super parameter gamma PHY Control standard Loss NN Loss of PHY Balance between.
According to the analysis, the embodiment defines a physical loss function term based on the AUV speed and combines the physical loss function term with a training target of a standard neural network, and in an actual system, the estimated inflow speed of the propeller should be consistent with the actual inflow speed, but an unknown disturbance, a factor estimated value such as parameter uncertainty and the like is considered to have a certain deviation from the actual value, so that an error term is introduced in the embodiment. The ReLU () function is used so that only deviations above the threshold are penalized, and the physical loss function is disabled when the estimated value deviates from the true value by less than the error term, and vice versa penalizing the neural network. This threshold is introduced because the physical model is affected by unknown factors.
The neural network contains 5 input variables, 2 hidden layers containing 10 nodes and 1 output variable. The 5 input variables include: u, θ, ψ, h, n p These can be obtained by sensors or simple calculations. It should be noted that T θ and Tψ It can be calculated from the sum so that they do not act as inputs to the neural network, the output of which is the propeller torque.
In the step B, when the fault is identified, the method for detecting the fault in this embodiment is a multi-head convolutional self-encoder, which reconstructs normal data with lower error and abnormal data with higher error, and then uses the magnitude of error to determine whether the fault exists. Encoder 1 extracts Q m And characteristic 1 of the control signal s, encoder 2 extracts Q p And feature 2 of the control signal s; these two features thenIs connected and fed into the decoder. Considering the fault-free condition, Q m and Qp Respectively, maintains the same trend as the control signal s. When a fault occurs, a different fault may result in Q m and Qp Different variations between them.
The multi-head automatic encoder structure according to the present embodiment is shown in fig. 2. Each encoder consists of two convolutional layers and two fully-concatenated layers. The decoder consists of two transposed convolutional layers and two fully-concatenated layers, outputting a sequence of the same size as the input. The model is trained to minimize the average absolute error between the input and output. Sequences with a shape of 100×1×3 are split into two sequences of 100×1×2. The first convolution layer moves in units of 4 points and filters an input sequence of 100×1×2 size by a kernel of 6×1×32 size. The second layer is a kernel of size 7×1×64, which moves in units of 4 points, and filters the output value of the first layer convolutional layer. The bottleneck layer is then flattened and output as a 25 node fully connected layer by inputting 128 nodes. Features of the two encoder outputs are stitched into a row of 50-dimensional features. The decoder receives the 50-dimensional features and outputs through a full connection layer having 128 and 384 nodes. The output layer is then despread, shifted by 4 points, and filtered through two transposed convolutional layers with kernel sizes of 7 x 1 x 32 and 6 x 1 x 3, respectively. After all operations are performed, a sequence of 100×1×3 in size is output.
In step C, when a fault is detected, a key issue is how to determine the cause of the fault, i.e. fault isolation:
when a current feedback fault occurs, Q m Abrupt change, Q, taking into account that the AUV speed is not affected by current feedback failure at this time p And s still maintain the same trend.
Contrary to the current feedback fault, when the motor rotation speed feedback fault occurs, Q p Mutations occur. At the same time Q m There will be slight fluctuations but the effect is limited to system losses. So Q is m And the trend of s remains the same.
For loss of propeller, slave et alAs can be seen from formula (1), Q m Should be close to zero. At this time, the speed of the AUV decreases, the control signal gradually increases, and the motor speed increases with the control signal until it reaches a maximum. Then Q p Increasing to a maximum torque.
When the propeller winding occurs, the motor speed is zero. As the AUV decelerates, the control signal gradually increases and the motor current increases to a maximum. At this time, as can be seen from equations (1) and (4), Q p Zero, Q m Will increase to maximum torque.
When an open circuit fault occurs, the propeller is disconnected from the battery. Current and motor speed are zero, thus Q m and Qp And also zero. The propeller control signal then reaches a maximum value under the regulation of the AUV control system.
The invention uses the characteristics extracted by two independent self-encoders as fault characteristics and realizes fault classification by a dimension reduction and clustering algorithm. The bottleneck feature in the multi-headed convolutional self-encoder is used to perform fault classification, the output features of the encoders 1 and 2 are reduced in dimension by a principal component analysis method in order to analyze patterns in the bottleneck feature, and each output feature is reduced from 25 dimensions to 1 dimension. The two one-dimensional features are combined into a two-dimensional feature and clustered using a noisy density-based spatial clustering algorithm. The noisy density-based spatial clustering algorithm uses a minimum number of adjacent m input points within the neighborhood to estimate the minimum density level. The detailed steps of the principal component analysis method and the density-based spatial clustering algorithm with noise are general methods and thus will not be described in detail. Thus, features extracted under unknown conditions (no or faulty) are classified as clusters or noise, each cluster being assumed to represent no or different faults, and other clustering methods may also be used to implement fault classification, including k-means clustering, fuzzy clustering, etc.
The present invention is not limited to the above-mentioned embodiments, and any equivalent embodiments which can be changed or modified by the technical content disclosed above can be applied to other fields, but any simple modification, equivalent changes and modification made to the above-mentioned embodiments according to the technical substance of the present invention without departing from the technical content of the present invention still belong to the protection scope of the technical solution of the present invention.
Claims (3)
1. The propulsion system fault diagnosis method based on the torque and the multi-head self-encoder is characterized by comprising the following steps of:
step A, load torque Q of motor m And propeller torque Q p Estimating;
step B, detecting faults based on a multi-head self-encoder, wherein the multi-head self-encoder comprises an independent encoder 1 and an independent encoder 2;
the motor load torque Q estimated in the step A is calculated m And a control signal s is input to the encoder 1 to extract its characteristic 1, and an estimated propeller torque Q is input to the encoder p And the control signal s is input into the encoder 2 to extract the characteristic 2, then the characteristic 1 and the characteristic 2 are connected and fed into the decoder, and fault detection is realized through signal reconstruction;
and C, classifying faults based on the fault detection result of the step B.
2. The torque and multi-headed self-encoder based propulsion system fault diagnosis method of claim 1, wherein: in the step A, the propeller torque is estimated based on an improved neural network, specifically:
the loss function of the neural network is improved, the advancing speed of the equipment is deduced according to the torque equation of the propeller, and the output of the neural network is the propeller torque:
wherein 
wherein ,
representing an estimated torque coefficient calculated from the estimated torque of the neural network,/->
Represents the estimated propeller inflow velocity estimated from the estimated torque coefficient, ucos (θ) cos (ψ) represents the actual propeller inflow velocity, δ, calculated approximately from the measured values u Threshold representing AUV speed deviation, super parameter gamma PHY Control standard Loss NN Loss of PHY Balance between.
3. The torque and multi-headed self-encoder based propulsion system fault diagnosis method of claim 1, wherein: in the step C:
if Q m Mutation of Q p And s remains the same trend, then diagnosing as a current fault;
if Q p Mutation of Q m If the trend of the motor speed is consistent with that of s, diagnosing the motor speed fault;
if Q m Near zero, the device speed decreases, the motor speed increases with the control signal s to maximum, then Q p If the propeller is at the maximum, diagnosing that the propeller is lost;
if the control signal s increases gradually with the deceleration of the propulsion system, the motor current increases to a maximum, Q p Zero, Q m If the maximum torque is increased, diagnosing that the propeller is wound;
if Q m and Qp At zero, the control signal s reaches a maximum value under the regulation of the control system of the device, and an open circuit fault is diagnosed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310244944.2A CN116243683B (en) | 2023-03-15 | 2023-03-15 | Method for diagnosing faults of propulsion system based on torque and multi-head self-encoder |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310244944.2A CN116243683B (en) | 2023-03-15 | 2023-03-15 | Method for diagnosing faults of propulsion system based on torque and multi-head self-encoder |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116243683A true CN116243683A (en) | 2023-06-09 |
| CN116243683B CN116243683B (en) | 2024-02-13 |
Family
ID=86629424
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310244944.2A Active CN116243683B (en) | 2023-03-15 | 2023-03-15 | Method for diagnosing faults of propulsion system based on torque and multi-head self-encoder |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116243683B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117233520A (en) * | 2023-11-16 | 2023-12-15 | 青岛澎湃海洋探索技术有限公司 | AUV propulsion system fault detection and evaluation method based on improved Sim-GAN |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7002318B1 (en) * | 2004-09-23 | 2006-02-21 | General Motors Corporation | Position sensor fault tolerant control for automotive propulsion system |
| CN111721535A (en) * | 2020-06-23 | 2020-09-29 | 中国人民解放军战略支援部队航天工程大学 | Bearing fault detection method based on convolution multi-head self-attention mechanism |
| US20200386811A1 (en) * | 2019-06-05 | 2020-12-10 | Wuhan University | Power electronic circuit fault diagnosis method based on extremely randomized trees and stacked sparse auto-encoder algorithm |
| CN113807444A (en) * | 2021-09-22 | 2021-12-17 | 湖南工业大学 | Fault diagnosis method based on constraint-confrontation convolutional self-coding memory fusion network |
| CN114035550A (en) * | 2021-11-15 | 2022-02-11 | 江苏科技大学 | Autonomous underwater robot executing mechanism fault diagnosis method based on ESO |
| CN114372494A (en) * | 2022-01-13 | 2022-04-19 | 湖南大学 | Air path fault diagnosis method for aviation turboprop engine based on attention mechanism |
| CN114528869A (en) * | 2022-01-05 | 2022-05-24 | 东南大学 | Fan blade damage identification method |
| CN114993677A (en) * | 2022-05-11 | 2022-09-02 | 山东大学 | Rolling bearing fault diagnosis method and system based on unbalanced small sample data |
| CN115118197A (en) * | 2022-06-22 | 2022-09-27 | 青岛澎湃海洋探索技术有限公司 | AUV propeller system fault diagnosis method |
| US20220341375A1 (en) * | 2020-02-04 | 2022-10-27 | Japan Aerospace Exploration Agency | Failure diagnostic system for spacecraft liquid propulsion system and failure diagnostic method for spacecraft liquid propulsion system |
| CN115407197A (en) * | 2022-08-25 | 2022-11-29 | 淮阴工学院 | Motor fault diagnosis method based on multi-head sparse self-encoder and Goertzel analysis |
-
2023
- 2023-03-15 CN CN202310244944.2A patent/CN116243683B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7002318B1 (en) * | 2004-09-23 | 2006-02-21 | General Motors Corporation | Position sensor fault tolerant control for automotive propulsion system |
| US20200386811A1 (en) * | 2019-06-05 | 2020-12-10 | Wuhan University | Power electronic circuit fault diagnosis method based on extremely randomized trees and stacked sparse auto-encoder algorithm |
| US20220341375A1 (en) * | 2020-02-04 | 2022-10-27 | Japan Aerospace Exploration Agency | Failure diagnostic system for spacecraft liquid propulsion system and failure diagnostic method for spacecraft liquid propulsion system |
| CN111721535A (en) * | 2020-06-23 | 2020-09-29 | 中国人民解放军战略支援部队航天工程大学 | Bearing fault detection method based on convolution multi-head self-attention mechanism |
| CN113807444A (en) * | 2021-09-22 | 2021-12-17 | 湖南工业大学 | Fault diagnosis method based on constraint-confrontation convolutional self-coding memory fusion network |
| CN114035550A (en) * | 2021-11-15 | 2022-02-11 | 江苏科技大学 | Autonomous underwater robot executing mechanism fault diagnosis method based on ESO |
| CN114528869A (en) * | 2022-01-05 | 2022-05-24 | 东南大学 | Fan blade damage identification method |
| CN114372494A (en) * | 2022-01-13 | 2022-04-19 | 湖南大学 | Air path fault diagnosis method for aviation turboprop engine based on attention mechanism |
| CN114993677A (en) * | 2022-05-11 | 2022-09-02 | 山东大学 | Rolling bearing fault diagnosis method and system based on unbalanced small sample data |
| CN115118197A (en) * | 2022-06-22 | 2022-09-27 | 青岛澎湃海洋探索技术有限公司 | AUV propeller system fault diagnosis method |
| CN115407197A (en) * | 2022-08-25 | 2022-11-29 | 淮阴工学院 | Motor fault diagnosis method based on multi-head sparse self-encoder and Goertzel analysis |
Non-Patent Citations (2)
| Title |
|---|
| 李炜;李青朋;毛海杰;龚建兴;: "基于T-S模糊神经网络模型的编码器故障软闭环容错控制方法", 计算机应用, no. 12 * |
| 蒋爱国;符培伦;谷明;王金江;: "基于多模态堆叠自动编码器的感应电机故障诊断", 电子测量与仪器学报, no. 08 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117233520A (en) * | 2023-11-16 | 2023-12-15 | 青岛澎湃海洋探索技术有限公司 | AUV propulsion system fault detection and evaluation method based on improved Sim-GAN |
| CN117233520B (en) * | 2023-11-16 | 2024-01-26 | 青岛澎湃海洋探索技术有限公司 | AUV propulsion system fault detection and evaluation method based on improved Sim-GAN |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116243683B (en) | 2024-02-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Alessandri et al. | Fault detection of actuator faults in unmanned underwater vehicles | |
| Liu et al. | Review on fault diagnosis of unmanned underwater vehicles | |
| CN116243683B (en) | Method for diagnosing faults of propulsion system based on torque and multi-head self-encoder | |
| CN111898285B (en) | Underwater unmanned autonomous vehicle cluster operation efficiency evaluation method | |
| CN114021441A (en) | Ship motion attitude prediction method based on CNN-BilSTM | |
| CN116243682B (en) | Underwater propeller fault detection method based on improved generation of countermeasure network | |
| CN108764122A (en) | A kind of optimization of AUV control systems and fault monitoring method based on deep learning | |
| Zhu et al. | Information fusion fault diagnosis method for unmanned underwater vehicle thrusters | |
| CN115118197A (en) | AUV propeller system fault diagnosis method | |
| Perera | Handling big data in ship performance and navigation monitoring | |
| Gao et al. | Unsupervised fault diagnosis framework for underwater thruster system using estimated torques and multi-head convolutional autoencoder | |
| Hassani et al. | Adaptive wave filtering for dynamic positioning of marine vessels using maximum likelihood identification: Theory and experiments | |
| CN117932279A (en) | Ship working condition online identification and analysis method and system based on deep learning | |
| Wu et al. | An overview of developments and challenges for unmanned surface vehicle autonomous berthing | |
| CN106919048B (en) | A kind of AUV propeller state-detection and thrust curve modification method | |
| Sun et al. | Fault diagnosis method of autonomous underwater vehicle based on deep learning | |
| Gao et al. | Physical model embedding-based generative adversarial networks for unsupervised fault detection of underwater thrusters | |
| Perera et al. | Digitalization of seagoing vessels under high dimensional data driven models | |
| CN117250970B (en) | Method for realizing AUV fault detection based on model embedding generation countermeasure network | |
| Abbasi Hoseini et al. | Multivariate time series data mining in ship monitoring database | |
| CN115056948B (en) | Unmanned underwater vehicle stern rudder position autonomous switching control method and system | |
| Pereira et al. | A 3-D Lightweight Convolutional Neural Network for Detecting Docking Structures in Cluttered Environments | |
| CN116992338B (en) | Method and device for detecting faults of IMU (inertial measurement unit) sensor of aircraft based on typical variable analysis | |
| Mou et al. | Fault diagnosis method of the USV based on improved Res2Net multi-scale feature fusion | |
| CN118410306B (en) | Rolling bearing fault cross-equipment depth domain self-adaptive migration diagnosis method |
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 |