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

CN116149295A - Double-redundancy position sensor fault diagnosis method based on dynamic time warping - Google Patents

Double-redundancy position sensor fault diagnosis method based on dynamic time warping Download PDF

Info

Publication number
CN116149295A
CN116149295A CN202211611312.7A CN202211611312A CN116149295A CN 116149295 A CN116149295 A CN 116149295A CN 202211611312 A CN202211611312 A CN 202211611312A CN 116149295 A CN116149295 A CN 116149295A
Authority
CN
China
Prior art keywords
position sensor
fault
paths
consistency
dynamic time
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.)
Pending
Application number
CN202211611312.7A
Other languages
Chinese (zh)
Inventor
李款
包旭光
王厚浩
唐德佳
潘卫东
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.)
Shanghai Aerospace Control Technology Institute
Original Assignee
Shanghai Aerospace Control Technology Institute
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 Shanghai Aerospace Control Technology Institute filed Critical Shanghai Aerospace Control Technology Institute
Priority to CN202211611312.7A priority Critical patent/CN116149295A/en
Publication of CN116149295A publication Critical patent/CN116149295A/en
Pending legal-status Critical Current

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
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B21/00—Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant
    • 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
    • G05B23/0208—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the configuration of the monitoring system
    • G05B23/0213—Modular or universal configuration of the monitoring system, e.g. monitoring system having modules that may be combined to build monitoring program; monitoring system that can be applied to legacy systems; adaptable monitoring system; using different communication protocols

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Transmission And Conversion Of Sensor Element Output (AREA)

Abstract

The invention provides a dual-redundancy position sensor fault diagnosis method based on dynamic time warping, which is suitable for being integrated into a light-weight miniaturized servo system to realize the detection and positioning of the position sensor fault. The invention designs a method for calculating the consistency distance of a low-pass filtered position sensor signal based on a dynamic time warping algorithm, setting a non-consistency detection threshold value by using a random algorithm, realizing consistency discrimination of the position sensor, constructing a virtual position sensor by using the angle of a rotary transformer, and completing positioning of the fault position sensor and discrimination of an extreme value zero fault mode. The method has certain diagnosis robustness and good fault compatibility.

Description

Double-redundancy position sensor fault diagnosis method based on dynamic time warping
Technical Field
The invention relates to the technical field of servo system fault diagnosis, in particular to a double-redundancy position sensor fault diagnosis method based on dynamic time warping.
Background
The servo system is used as a key actuating mechanism of the aircraft, and can quickly follow a position command signal to realize pose adjustment of a follow-up control surface by means of position-rotating speed-current three-ring control, so that the triaxial motion pose of the aircraft can be stably adjusted in real time.
As an important component of the servo system, a healthy and reliable position sensor is critical to the safe operation of the servo system. However, the position sensor is a fragile measuring device, and is susceptible to failure caused by various environmental factors such as short-time overtemperature, electromagnetic interference, abnormal impact and the like which are possible in long-endurance flight. To improve the reliability of the servo system, the position sensor is usually designed in a redundant manner, so as to utilize the increased redundancy to assist in fault diagnosis and fault-tolerant control. The traditional position sensor redundancy generally adopts a triple redundancy design mode, and a voting decision mode of 'three taking two' (3 oo 2) is adopted to judge the fault position sensor. Currently, an on-board servo system is developing towards the application direction of light miniaturization, and the design scheme of the dual-redundancy position sensor is widely focused from the perspective of reasonably reducing design redundancy. How to perform fault diagnosis on the dual redundant position sensor so as to ensure the healthy operation of the servo system is a problem to be solved urgently at present.
Disclosure of Invention
The technical solution of the invention is as follows: the fault diagnosis method for the double-redundancy position sensor based on dynamic time warping is provided for overcoming the defects of the prior art.
The technical scheme of the invention is as follows:
a fault diagnosis method of a dual redundant position sensor based on dynamic time warping is characterized by comprising the following steps:
measuring the consistency distance r of the outputs of the two paths of position sensors after low-pass filtering based on a dynamic time warping algorithm;
setting a non-uniform detection threshold value by using a random algorithm, and realizing uniformity discrimination of the position sensor;
on the basis of consistency discrimination, a virtual position sensor is constructed by utilizing the angle of the rotary transformer, and the detection, the positioning and the fault mode discrimination of the position sensor are assisted.
Preferably, the method for measuring the consistency distance of the outputs of the two paths of position sensors after low-pass filtering based on the dynamic time warping algorithm is as follows:
step A1: two paths of servo system in running state are collected in real timePosition sensor signal pos 1 ，pos 2 Dynamically intercepting two paths of position sensor signals with length of l from the two paths of position sensor signals respectively
Figure BDA0003997250390000021
And->
Figure BDA0003997250390000022
Wherein the method comprises the steps of
Figure BDA0003997250390000023
k is the current sampling time;
step A2: initializing an accumulated distance matrix D, D epsilon R l×l ；
Step A3: setting the length w of a matching window, and updating the accumulated distance matrix by adopting a dynamic time warping algorithm, wherein the updating formula of an ith row and jth column element D (i, j) of the accumulated distance matrix D is as follows:
D(i，j)＝d(i，j)+min(D(i-1，j)，D(i-1，j-1)，D(i，j-1))
wherein I i-j is less than or equal to w,
Figure BDA0003997250390000024
step A4: and calculating the consistency distance r of the two paths of position sensor signals.
Preferably, the consistency distance r of the two paths of position sensor signals satisfies the following conditions:
Figure BDA0003997250390000025
wherein dist () is a consistent distance function of two paths of position sensor signals.
Preferably, a random algorithm is used for setting a non-uniform detection threshold value, so that the uniformity judgment of the position sensor is realized, and the specific implementation mode is as follows:
step B1: collecting position sensor signals in a fault-free state, and constructing a consistency distance training set
Figure BDA0003997250390000026
Figure BDA0003997250390000027
Wherein Nsamp is a sufficiently large constant;
step B2: given epsilon (0, 1), delta epsilon (0, 1), selecting random sample number N, satisfying
Figure BDA0003997250390000028
Step B3: setting an allowable fault false alarm rate FAR and an allowable increment delta of single-step iteration;
step B4: initializing a non-uniformity detection threshold
Figure BDA0003997250390000029
Let->
Figure BDA00039972503900000210
Wherein J is 0 Is a sufficiently small real number;
step B5: from a collection
Figure BDA00039972503900000212
N groups of training samples are randomly extracted to construct a random training set +.>
Figure BDA00039972503900000211
Step B6: traversing random training set
Figure BDA0003997250390000031
Calculating that the element in the set is greater than the current +.>
Figure BDA0003997250390000032
The number of samples n, the return duty cycle
Figure BDA0003997250390000033
Step B7: iterative calculation of non-uniformity detection threshold according to the following criteria
Figure BDA0003997250390000034
If it is
Figure BDA0003997250390000035
Return to the current +.>
Figure BDA0003997250390000036
Step B8 is entered;
if it is
Figure BDA0003997250390000037
Let->
Figure BDA0003997250390000038
Returning to the step B5;
step B8: judging the consistency of the signals of the two paths of position sensors, if
Figure BDA0003997250390000039
Judging that the signals of the two paths of position sensors are consistent; if->
Figure BDA00039972503900000310
And judging that the signals of the two paths of position sensors are inconsistent.
Preferably, nsamp satisfies Nsamp > N.
Preferably, on the basis of consistency discrimination, the virtual position sensor is constructed by utilizing the angle of the rotary transformer, and the detection, the positioning and the fault mode discrimination of the fault of the position sensor are assisted, and the realization modes are as follows:
step C1: constructing a virtual position sensor by utilizing a rotary transformer angle signal according to the transmission relation between the rotary motion and the linear motion;
step C2: dynamically intercepting virtual position sensor signal with length of l
Figure BDA00039972503900000311
Step C3: determining two-way position sensor signal
Figure BDA00039972503900000312
And virtual position sensor signal->
Figure BDA00039972503900000313
A non-uniformity detection threshold value between +.>
Figure BDA00039972503900000314
And->
Figure BDA00039972503900000315
Step C4: after the virtual position sensor is introduced, comparing the consistency of the position sensor signals in pairs, and if the two paths of position sensor signals are consistent, the two paths of position sensors have no faults, and the position feedback is selected from the position feedback with smaller consistency distance with the virtual position sensor; if the two paths of position sensor signals are inconsistent, and one path of position sensor signals are consistent with the virtual position sensor signals, judging that the other path of position sensor has a fault; if the comparison results are inconsistent, judging that the two paths of position sensors have faults, and selecting a virtual position sensor for position feedback.
Step C5: setting an extreme fault detection threshold pos max Comparing the maximum value of the absolute value of the fault position sensing signal in the time window with the length of l with the extremum fault detection threshold value, and carrying out extremum fault mode identification;
step C6: setting a zero-value fault detection threshold pos zero And comparing the maximum value of the absolute value of the fault position sensing signal in the time window with the length of l with a zero-value fault detection threshold value, and carrying out zero-value fault identification.
Preferably, in the step C1, pos 3 θ×ratio, where pos 3 And (3) reconstructing a virtual position sensor signal, wherein θ is an angle output by the rotary transformer, and ratio is a transmission ratio.
Preferably, the implementation manner of the step C4 is as follows:
if it is
Figure BDA0003997250390000041
The two paths of position sensors have no faults, and the position feedback is selected to be the position with smaller consistency distance with the virtual position sensor;
if it is
Figure BDA0003997250390000042
Judging that the first path of position sensor fails, and selecting a second path of position sensor for position feedback;
if it is
Figure BDA0003997250390000043
Judging that the second path of position sensor fails, and selecting the first path of position sensor for position feedback;
if it is
Figure BDA0003997250390000044
And judging that the two paths of position sensors are failed, wherein the position feedback is a virtual position sensor.
Preferably, in the step C5, if the fault location sensor is
Figure BDA0003997250390000045
i= {1,2}, it is determined that the jth path position sensor has an extremum fault.
Preferably, in the step C5, if the fault location sensor is
Figure BDA0003997250390000046
i= {1,2}, it is determined that the jth path position sensor has zero value fault.
Compared with the prior art, the invention has the following advantages:
(1) the invention realizes the consistency measurement of the position signal sequence in the time window by utilizing the dynamic time warping algorithm when the consistency of the two paths of position signals is judged, and has certain robustness to the signal wild value and the asynchronous communication delay compared with the point-to-point consistency measurement mode.
(2) The invention designs the non-uniform detection threshold by adopting a random algorithm, has a certain guiding significance compared with manual experience selection, and can avoid the problem of missed detection caused by overlarge threshold selection to a certain extent.
(3) The invention introduces a soft redundancy design of the position sensor, utilizes the angle signal output by the rotary transformer to construct the virtual position sensor, so that the double redundancy diagnosis problem can be converted into a triple redundancy diagnosis problem, the diagnosis logic is simpler, and when two paths of position sensors have faults, fault tolerance exists, the fault compatibility is good, and the reliability is high.
Drawings
FIG. 1 is a schematic diagram of a servo system dual redundant position sensor fault diagnosis;
FIG. 2 is a simplified schematic diagram of a specific embodiment technique;
FIG. 3 is a graph of virtual sensor output effects;
FIG. 4 is a diagram of the effect of position sensor fault detection;
fig. 5 is a diagram of the effect of the second path position sensor zero-value fault discrimination.
Detailed Description
The invention aims to solve the problem of fault diagnosis of double redundant position sensors, measures the consistency distance of the outputs of two paths of position sensors after low-pass filtering based on a dynamic time warping algorithm, and utilizes a random algorithm to set a non-consistency detection threshold value so as to realize consistency judgment of the position sensors; based on consistency discrimination, the virtual position sensor constructed by utilizing the angle of the rotary transformer is combined to realize the detection, positioning and fault mode discrimination of the fault of the position sensor. The fault diagnosis principle of the dual redundant position sensor of the servo system is shown in fig. 1, and key links of the fault diagnosis design comprise: (1) a consistency distance measure; (2) non-uniform detection threshold design; (3) fault diagnosis logic.
The flow chart of the invention is shown in fig. 2, and is mainly realized by the following technical scheme:
(1) consistency distance metric
According to the collected two paths of position sensor signals, consistency distance measurement is carried out based on a dynamic time warping algorithm, and the method specifically comprises the following steps:
step A1: real-time acquisition servoPosition sensor signal pos in system operation state 1 ，pos 2 Dynamically intercepting two paths of position sensing signals with length of l
Figure BDA0003997250390000051
And->
Figure BDA0003997250390000052
Figure BDA0003997250390000053
Where k is the current sampling instant.
Step A2: initializing an accumulated distance matrix D E R l×l The initial value of the matrix element D (i, j) is set to infinity.
Step A3: setting the length w of the matching window, updating the accumulated distance matrix by adopting a dynamic time warping algorithm,
D(i，j)＝d(i，j)+min(D(i-1，j)，D(i-1，j-1)，D(i，j-1))
wherein I i-j is less than or equal to w,
Figure BDA0003997250390000054
step A4: a consistent distance measure r of the position sensor signal is output,
Figure BDA0003997250390000061
wherein D (l, l) is
Figure BDA0003997250390000062
And->
Figure BDA0003997250390000063
Minimum matching distance of two sections of position signal sequences.
(2) Non-uniform detection threshold design
Based on the consistency distance measurement, a random algorithm is utilized to design a non-consistency detection threshold value of the two paths of position sensing signals. The method comprises the following specific steps:
step B1: collecting enough position sensing signals in a fault-free state, and constructing a consistency distance measurement training set
Figure BDA0003997250390000064
Figure BDA0003997250390000065
Wherein Nsamp is a sufficiently large constant.
Step B2: given epsilon (0, 1), delta epsilon (0, 1), selecting random sample number N, satisfying
Figure BDA0003997250390000066
Step B3: setting the allowable fault false alarm rate FAR and the allowable increment delta of single-step iteration.
Step B4: initializing a non-uniformity detection threshold
Figure BDA0003997250390000067
Order the
Figure BDA0003997250390000068
Wherein J is 0 Is a sufficiently small real number.
Step B5: from a collection
Figure BDA0003997250390000069
N groups of training samples are randomly extracted to construct a random training set +.>
Figure BDA00039972503900000610
Step B6: traversing random training set
Figure BDA00039972503900000611
Computing elements in the collection greater than the current J th The number of samples n, the return duty cycle
Figure BDA00039972503900000612
Step B7: iterative calculation of non-uniformity detection threshold
Figure BDA00039972503900000613
The following operations are performed:
if it is
Figure BDA00039972503900000614
Return to the current +.>
Figure BDA00039972503900000615
Exiting the algorithm;
if it is
Figure BDA00039972503900000616
Let->
Figure BDA00039972503900000617
And returning to the step B5.
Step B8: judging the signal consistency of the double-position sensor if
Figure BDA00039972503900000618
Judging that the signals of the two paths of position sensors are consistent; if->
Figure BDA00039972503900000619
And judging that the signals of the two paths of position sensors are inconsistent.
(3) Fault diagnosis
On the basis of consistency discrimination, detecting the occurrence of a fault of the position sensor and positioning the fault position sensor, wherein the specific steps comprise:
step C1: according to the transmission relation between the rotary motion and the linear motion, a virtual position sensor is designed by utilizing the angle signal of the rotary transformer,
pos 3 ＝θ×ratio
wherein pos 3 And (3) reconstructing a virtual position signal, wherein θ is an angle of the output of the rotary transformer, and ratio is a transmission ratio.
Step C2: dynamically intercepting a position sensing signal with the length of l in a mode of step A1
Figure BDA0003997250390000071
Figure BDA0003997250390000072
Step C3: determining two paths of position sensing signals according to the steps B1-B7
Figure BDA0003997250390000073
And virtual position sense signal->
Figure BDA0003997250390000074
A non-uniformity detection threshold value between +.>
Figure BDA0003997250390000075
And->
Figure BDA0003997250390000076
Step C4: position sensor fault detection. After the virtual position sensor is introduced, if
Figure BDA0003997250390000077
The two paths of position sensors have no faults, and the position feedback is selected to be the position with smaller consistency distance with the virtual position sensor; if it is
Figure BDA0003997250390000078
Judging that the first path of position sensor fails, and selecting a second path of position sensor for position feedback; if->
Figure BDA0003997250390000079
Figure BDA00039972503900000710
Judging that the second path of position sensor fails, and selecting the first path of position sensor for position feedback; if->
Figure BDA00039972503900000711
Figure BDA00039972503900000712
And judging that the two paths of position sensors are failed, wherein the position feedback is a virtual position sensor.
Step C5: and (5) identifying an extremum fault mode. Setting an extreme fault detection threshold pos max Comparing the maximum value of the absolute value of the fault position sensing signal in the time window with the length of l with the extreme fault detection threshold value, if
Figure BDA00039972503900000713
The ith path position sensor is determined to have an extremum fault.
Step C6: zero-value fault pattern discrimination. Setting a zero-value fault detection threshold pos zero Comparing the maximum value of the absolute value of the fault location sensing signal in the time window with the length of l with a zero-value fault detection threshold value, if
Figure BDA0003997250390000081
It is determined that the i-th path position sensor has failed by a zero value.
Examples:
the embodiment of the invention is a differential type speed comprehensive dual-redundancy servo system, the position adjustable range is [ -35mm,35mm ], and the transmission ratio ratio=0.0783.
Given a position command signal of 35mm amplitude, 0.05hz, the consistency distance metric parameter sets: dynamically intercepting window length l=50, and matching window length w=2; non-coincidence detection threshold design parameter setting: epsilon=0.01, delta=0.01, the allowed false positive rate far=0.02, the number of random training set samples n=20000, the initial non-coincidence detection threshold J 0 Allowed increment of single step iteration delta=0.0001, =0.0001; failure mode discriminationParameter setting: extreme fault detection threshold pos max =35 mm, zero-value fault detection threshold pos zero ＝0.08mm。
Collecting enough position sensing data under a fault-free state, calculating the consistency distance between three groups of position signals (two paths of position sensing signals and one path of virtual position sensing signals) in a pairwise mode according to the consistency distance measurement mode, and constructing three groups of consistency distance measurement training sets
Figure BDA0003997250390000082
According to the non-coincidence detection threshold design mode, three groups of non-coincidence detection threshold designs are calculated as +.>
Figure BDA0003997250390000083
Figure BDA0003997250390000084
According to the fault diagnosis logic, the detection, the positioning and the mode identification of the fault of the position sensor are carried out, and the fact that the differential type speed comprehensive dual-redundancy structure design is adopted requires the angle theta of the dual-channel rotary transformer 1 And theta 2 The addition is performed to obtain the angle θ for calculating the virtual position signal.
Fig. 3 shows the position output of the virtual sensor. In order to facilitate the observation effect, a partial amplification diagram of the position output is intercepted, and under the fault-free state, the output of the virtual position sensor is well matched with the output of the two paths of position sensors, so that the validity of the designed virtual position sensor is verified.
FIG. 4 shows the results of dual redundant position sensor fault diagnosis, in which a zero value fault is injected into the second path position sensor at 50 seconds in the experiment, and the consistency distance between the first path and the second path sensor exceeds a threshold value at the moment of fault occurrence
Figure BDA0003997250390000085
The two paths of signals are inconsistent; the consistency distance between the first path and the virtual sensor does not exceed a threshold value +.>
Figure BDA0003997250390000086
The two paths of signals are consistent; the coincidence distance of the second path from the virtual sensor exceeds a threshold value +.>
Figure BDA0003997250390000087
The two signals are not identical. Thus, it can be determined that a failure occurred at 50 th second, and the second path position sensor failed. In the case of no fault, the phenomenon that the threshold value is occasionally exceeded is that the threshold value is designed under the allowable fault false alarm rate, and the threshold value design mode can avoid missing detection caused by too large threshold value selection, and in this example, if 10 sampling points do not exceed the threshold value, no fault can be determined. Furthermore, +_of the second path failure position sensor>
Figure BDA0003997250390000091
As shown in fig. 5, it may be determined that a zero-value fault has occurred.
The invention is suitable for being integrated into a light-weight miniaturized servo system, and realizes the detection and the positioning of the fault of the position sensor. The invention designs a method for calculating the consistency distance of a low-pass filtered position sensor signal based on a dynamic time warping algorithm, setting a non-consistency detection threshold value by using a random algorithm, realizing consistency discrimination of the position sensor, constructing a virtual position sensor by using the angle of a rotary transformer, and completing positioning of the fault position sensor and discrimination of an extreme value zero fault mode. The method has certain diagnosis robustness and good fault compatibility.
What is not described in detail in the present specification belongs to the known technology of those skilled in the art.

Claims (10)

1. A fault diagnosis method of a dual redundant position sensor based on dynamic time warping is characterized by comprising the following steps:
measuring the consistency distance r of the outputs of the two paths of position sensors after low-pass filtering based on a dynamic time warping algorithm;
setting a non-uniform detection threshold value by using a random algorithm, and realizing uniformity discrimination of the position sensor;
on the basis of consistency discrimination, a virtual position sensor is constructed by utilizing the angle of the rotary transformer, and the detection, the positioning and the fault mode discrimination of the position sensor are assisted.
2. The method for diagnosing faults of the dual redundant position sensor based on dynamic time warping as claimed in claim 1, wherein the method for measuring the consistency distance of the outputs of the two paths of position sensors after low-pass filtering based on the dynamic time warping algorithm is realized by the following steps:
step A1: two paths of position sensor signals pos under running state of real-time acquisition servo system 1 ，pos 2 Dynamically intercepting two paths of position sensor signals with length of l from the two paths of position sensor signals respectively
Figure FDA0003997250380000011
And->
Figure FDA0003997250380000012
Wherein->
Figure FDA0003997250380000013
Figure FDA0003997250380000014
k is the current sampling time;
step A2: initializing an accumulated distance matrix D, D epsilon R l×l ；
Step A3: setting the length w of a matching window, and updating the accumulated distance matrix by adopting a dynamic time warping algorithm, wherein the updating formula of an ith row and jth column element D (i, j) of the accumulated distance matrix D is as follows:
D(i，j)＝d(i，j)+min(D(i-1，j)，D(i-1，j-1)，D(i，j-1))
wherein I i-j is less than or equal to w,
Figure FDA0003997250380000015
step A4: and calculating the consistency distance r of the two paths of position sensor signals.
3. The method for diagnosing a dual redundant position sensor fault based on dynamic time warping of claim 2, wherein the consistency distance r of the two paths of position sensor signals is as follows:
Figure FDA0003997250380000016
wherein dist () is a consistent distance function of two paths of position sensor signals.
4. The method for diagnosing faults of the dual redundant position sensor based on dynamic time warping as claimed in claim 1, wherein a random algorithm is used for setting a non-uniform detection threshold value to realize the uniformity judgment of the position sensor, and the specific implementation mode is as follows:
step B1: collecting position sensor signals in a fault-free state, and constructing a consistency distance training set
Figure FDA0003997250380000021
Figure FDA0003997250380000022
Wherein Nsamp is a sufficiently large constant;
step B2: given epsilon (0, 1), delta epsilon (0, 1), selecting random sample number N, satisfying
Figure FDA0003997250380000023
Step B3: setting an allowable fault false alarm rate FAR and an allowable increment delta of single-step iteration;
step B4: initializing a non-uniformity detection threshold
Figure FDA0003997250380000024
Let->
Figure FDA0003997250380000025
Wherein J is 0 Is a sufficiently small real number;
step B5: from a collection
Figure FDA0003997250380000026
N groups of training samples are randomly extracted to construct a random training set +.>
Figure FDA0003997250380000027
Step B6: traversing random training set
Figure FDA0003997250380000028
Calculating that the element in the set is greater than the current +.>
Figure FDA0003997250380000029
Is the number n of samples of (1) the return duty cycle +.>
Figure FDA00039972503800000210
Step B7: iterative calculation of non-uniformity detection threshold according to the following criteria
Figure FDA00039972503800000211
If it is
Figure FDA00039972503800000212
Return to the current +.>
Figure FDA00039972503800000213
Step B8 is entered;
if it is
Figure FDA00039972503800000214
Let->
Figure FDA00039972503800000215
Returning to the step B5; />
Step B8: judging the consistency of the signals of the two paths of position sensors, if
Figure FDA00039972503800000216
Judging that the signals of the two paths of position sensors are consistent; if->
Figure FDA00039972503800000217
And judging that the signals of the two paths of position sensors are inconsistent.
5. The method for dynamic time warping based dual redundant position sensor fault diagnosis according to claim 4, wherein Nsamp satisfies Nsamp > N.
6. The method for diagnosing faults of the double redundant position sensor based on dynamic time warping as claimed in claim 1, wherein the virtual position sensor is constructed by utilizing the angle of the rotary transformer on the basis of consistency discrimination, and the fault detection, positioning and fault mode discrimination of the position sensor are assisted, and the implementation modes are as follows:
step C1: constructing a virtual position sensor by utilizing a rotary transformer angle signal according to the transmission relation between the rotary motion and the linear motion;
step C2: dynamically intercepting virtual position sensor signal with length of l
Figure FDA00039972503800000218
Step C3: determining two-way position sensor signal
Figure FDA00039972503800000219
And virtual position sensor signal->
Figure FDA00039972503800000220
A non-uniformity detection threshold value between +.>
Figure FDA00039972503800000221
And->
Figure FDA00039972503800000222
Step C4: after the virtual position sensor is introduced, comparing the consistency of the position sensor signals in pairs, and if the two paths of position sensor signals are consistent, the two paths of position sensors have no faults, and the position feedback is selected from the position feedback with smaller consistency distance with the virtual position sensor; if the two paths of position sensor signals are inconsistent, and one path of position sensor signals are consistent with the virtual position sensor signals, judging that the other path of position sensor has a fault; if the comparison results are inconsistent, judging that the two paths of position sensors have faults, and selecting a virtual position sensor for position feedback.
Step C5: setting an extreme fault detection threshold pos max Comparing the maximum value of the absolute value of the fault position sensing signal in the time window with the length of l with the extremum fault detection threshold value, and carrying out extremum fault mode identification;
step C6: setting a zero-value fault detection threshold pos zero And comparing the maximum value of the absolute value of the fault position sensing signal in the time window with the length of l with a zero-value fault detection threshold value, and carrying out zero-value fault identification.
7. The method for diagnosing a dual redundant position sensor fault based on dynamic time warping as claimed in claim 6, wherein in said step C1, pos 3 θ×ratio, where pos 3 And (3) reconstructing a virtual position sensor signal, wherein θ is an angle output by the rotary transformer, and ratio is a transmission ratio.
8. The method for diagnosing a dual redundant position sensor fault based on dynamic time warping of claim 6, wherein said step C4 is implemented as follows:
if it is
Figure FDA0003997250380000031
The two paths of position sensors have no faults, and the position feedback is selected to be the position with smaller consistency distance with the virtual position sensor;
if it is
Figure FDA0003997250380000032
Judging that the first path of position sensor fails, and selecting a second path of position sensor for position feedback;
if it is
Figure FDA0003997250380000033
Judging that the second path of position sensor fails, and selecting the first path of position sensor for position feedback;
if it is
Figure FDA0003997250380000034
And judging that the two paths of position sensors are failed, wherein the position feedback is a virtual position sensor.
9. The method for diagnosing a dual redundant position sensor fault based on dynamic time warping as recited in claim 6, wherein said step C5 is performed for a faulty position sensor if
Figure FDA0003997250380000035
The ith path position sensor is determined to have an extremum fault. />
10. The method for diagnosing a dual redundant position sensor fault based on dynamic time warping as recited in claim 6, wherein said step C5 is performed for a faulty position sensor if
Figure FDA0003997250380000036
It is determined that the i-th path position sensor has failed by a zero value. />
CN202211611312.7A 2022-12-13 2022-12-13 Double-redundancy position sensor fault diagnosis method based on dynamic time warping Pending CN116149295A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211611312.7A CN116149295A (en) 2022-12-13 2022-12-13 Double-redundancy position sensor fault diagnosis method based on dynamic time warping

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211611312.7A CN116149295A (en) 2022-12-13 2022-12-13 Double-redundancy position sensor fault diagnosis method based on dynamic time warping

Publications (1)

Publication Number Publication Date
CN116149295A true CN116149295A (en) 2023-05-23

Family

ID=86339891

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211611312.7A Pending CN116149295A (en) 2022-12-13 2022-12-13 Double-redundancy position sensor fault diagnosis method based on dynamic time warping

Country Status (1)

Country Link
CN (1) CN116149295A (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103140814A (en) * 2010-10-11 2013-06-05 通用电气公司 Systems, methods, and apparatus for detecting shifts in redundant sensor signals
CN104834210A (en) * 2015-03-24 2015-08-12 上海新跃仪表厂 Redundancy control method based on double position sensors
CN105467975A (en) * 2015-12-29 2016-04-06 山东鲁能软件技术有限公司 Equipment fault diagnosis method
KR20180025447A (en) * 2016-08-31 2018-03-09 고려대학교 산학협력단 Method and device for fault detection of manufacturing process based on dynamic time warping and exponentioal penaltu(dtwep)
CN110073301A (en) * 2017-08-02 2019-07-30 强力物联网投资组合2016有限公司 The detection method and system under data collection environment in industrial Internet of Things with large data sets

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103140814A (en) * 2010-10-11 2013-06-05 通用电气公司 Systems, methods, and apparatus for detecting shifts in redundant sensor signals
CN104834210A (en) * 2015-03-24 2015-08-12 上海新跃仪表厂 Redundancy control method based on double position sensors
CN105467975A (en) * 2015-12-29 2016-04-06 山东鲁能软件技术有限公司 Equipment fault diagnosis method
KR20180025447A (en) * 2016-08-31 2018-03-09 고려대학교 산학협력단 Method and device for fault detection of manufacturing process based on dynamic time warping and exponentioal penaltu(dtwep)
CN110073301A (en) * 2017-08-02 2019-07-30 强力物联网投资组合2016有限公司 The detection method and system under data collection environment in industrial Internet of Things with large data sets

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
程序，关永刚，张文鹏，唐诚: "《基于因子分析和支持向量机算法的高压断路器机械故障诊断方法》", 电工技术学报, 27 July 2014 (2014-07-27), pages 209 - 215 *

Similar Documents

Publication Publication Date Title
KR101967339B1 (en) System and Method for Diagnosing Fault and Backup of ADAS Sensors based on Deep Learning
Goel et al. Fault detection and identification in a mobile robot using multiple model estimation and neural network
CN102322825B (en) Optical measuring system and method for coaxiality of extra-long-hole part
CN103776654B (en) The method for diagnosing faults of multi-sensor information fusion
Jiang Sensor Fault Detection and Isolation Using System Dynamics Identification Techniques.
CN103344963B (en) A kind of robust reckoning method based on laser radar
CN108490923A (en) The design method of small fault detection and positioning for electric traction system
CN106594000B (en) A kind of Fault Diagnosis Method of Electro-hydraulic
CN113071509B (en) Fault detection method for wheel angle sensor of intelligent vehicle
CN111474919A (en) Aeroengine control system sensor fault diagnosis method based on AANN network group
CN113916306B (en) Pipeline defect detection and positioning method based on multi-sensing information fusion
Li et al. A data driven fault detection and isolation scheme for UAV flight control system
Xiao et al. Prognosis of electric scooter with intermittent faults: Dual degradation processes approach
CN111506049A (en) Multiple fault diagnosis method for aero-engine control system based on AANN network system
CN116071051B (en) Method, system, electronic equipment and medium for determining optimal replacement time of inertial navigation system
CN116149295A (en) Double-redundancy position sensor fault diagnosis method based on dynamic time warping
CN109657329A (en) Based on the sensor fault and Structural Damage Identification for improving adaptive particle filter
CN112801267A (en) Multiple fault diagnosis device for aircraft engine with dynamic threshold value
CN112651087A (en) Train motor fault detection method based on distributed estimation
US20080190166A1 (en) Device and Method for Error-Protected Detection of Measured Values in a Control Unit
Wang et al. UAV actuator fault detection using maximal information coefficient and 1-D convolutional neural network
CN116501183A (en) Mouse displacement regulation and control method and system based on multi-sensor fusion
CN113778891B (en) Embedded software interface failure mode automatic identification and analysis method
CN108763925A (en) A kind of sensor attack detection method measured based on fusion interval and history
Shi et al. Research on Fault Diagnosis of Three Degrees of Freedom Gyroscope Redundant System

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