CN103678886A - Satellite Bayesian Network health determination method based on ground test data - Google Patents
Satellite Bayesian Network health determination method based on ground test data Download PDFInfo
- Publication number
- CN103678886A CN103678886A CN201310611919.XA CN201310611919A CN103678886A CN 103678886 A CN103678886 A CN 103678886A CN 201310611919 A CN201310611919 A CN 201310611919A CN 103678886 A CN103678886 A CN 103678886A
- Authority
- CN
- China
- Prior art keywords
- satellite
- data
- node
- bayesian network
- ground test
- 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
Images
Landscapes
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a satellite Bayesian Network health determination method based on ground test data and provides a satellite health status comprehensive and quantitative determination method based on ground test data. From the prospective of historical data, probabilities between a satellite and related parameters as well as between the satellite and onboard equipment are mined; features of data under actual conditions are fully embodied; implicit association between parameters and system health level are mined through multi-parameter conditional probabilities; the system health level is truly indicated. The health status of the satellite during a test phase can be estimated accurately, unified health level of the satellite is quantitated from multi-index telemetry parameters, and the effective method is provided for ground test and analysis of the satellite.
Description
Technical field
The present invention relates to a kind of satellite health and determine method, relate in particular to a kind of healthy definite method of satellite Bayesian network based on ground test data, belong to satellite ground stage prognostic and health management technical field.
Background technology
The aspects such as the condition discrimination of the ground test data of satellite before for satellite launch, transmitting decision-making all have very important effect.In conjunction with the ground test data before satellite launch, ground test personnel and design of satellites personnel can grasp the overall health before satellite launch, thereby further take corresponding ground test safeguard, to guarantee initiatively stage and the in orbit kilter in stage after satellite launch.Satellite ground test period, test data is to embody the whether normal important evidence of satellite current operating state.Satellite health status based on ground test data is determined method, mainly utilizes the ground test data of satellite, and the injection failure condition of combined ground test phase, by the reasoning of network, provides the general health grade of object to be determined.
In current satellite ground test process, main employing is interpretation method at present, by the interpretation of multidimensional telemetry bound, rate of change and variation tendency being determined to the health status of on-board equipment, this method is abnormal to satellite and the detection in subsystem ground test stage, fault etc. played vital role.But because satellite ground amount of test data is large, data dimension is high, not yet according to multi-Dimensional parameters, carry out integral body and quantize, cannot reflect satellite holistic health state.The undue artificial experience knowledge that relies on, is difficult to abnormal occurrence to make the reaction of decision-making fast and accurately simultaneously.
Therefore, need to find a kind of solution, can be for above deficiency, in conjunction with satellite ground testing parameter data, historical ground test data and current ground test data are carried out to complex reasoning, provide in real time current ground test stage satellite health grade aggregate level.
Summary of the invention
The technical matters that the present invention solves is: overcome the deficiencies in the prior art, provide a kind of satellite Bayesian network based on ground test data healthy definite method, can determine more exactly the whole star of satellite and the health status in equipment ground testing stage, for the ground test analysis of satellite provides efficient means.
Technical scheme of the present invention is: a kind of healthy definite method of satellite Bayesian network based on ground test data, comprises the steps:
(1) build Bayesian network: build for determining the Bayesian network of satellite health status to be assessed, wherein the leaf node of Bayesian network is whole star health status, and intermediate node is on-board equipment, and root node is the ground test parameter that on-board equipment is corresponding;
(2) determine training sample: from the ground test historical data of satellite to be assessed, extract the ground test parameter historical data that each root node is corresponding, carry out error code and singular point rejecting, null value filling and mean value computation pre-service; Continuous variable is wherein carried out to discretize, determine the corresponding relation from connection attribute to Category Attributes, the codomain that is about to numerical attribute is divided into some sub-ranges, each interval corresponding discrete value; According to historical evaluation result, determine the value of intermediate node under the different value conditions of root node, and the value of leaf node under the different value conditions of intermediate node; Through the root node data of pre-service, discretize and corresponding intermediate node and leaf node data, form the healthy training sample of determining model of satellite Bayesian network;
(3) the constructed Bayesian network of training sample training step (1) that utilizes step (2) to obtain, obtain conditional probability: the prior probability of adding up root node and each intermediate node in constructed Bayesian network by training sample, and further calculate accordingly the conditional probability that each father node is got child node under different value condition, obtain each root node and get the conditional probability that the conditional probability of intermediate node under different value condition, each intermediate node are got leaf node under different value condition;
(4) determine satellite health status: by satellite to be assessed through error code and singular point reject, null value is filled and mean value computation pre-service after in the constructed Bayesian network of the current ground test parameter input step (1) of each root node of obtaining, the conditional probability that Bayesian network obtains according to step (3) is calculated the health status of the whole star of output.
Extract ground test parameter historical data corresponding to each root node in described step (2), carry out that error code and singular point are rejected, that null value is filled pretreated implementation method is as follows:
Singular point is rejected main serial mean and the variance yields that calculates the ground test parameter historical data that each root node is corresponding that adopt, the mode of adding and subtracting three times of variance yields by mean value determine in test data 95% data should in data area, thereby eliminate the not singular point data within the scope of this, directly this data point is set to sky; Minute average of test data after calculating rejecting error code and singular point; For null value point data, to utilize the method for linear interpolation to carry out benefit value to null value point, thereby form a complete ordered series of numbers, the missing data section of supposing sequence of test data is X
t, X
t+1, L, X
t+M-1, a linear interpolation data benefit point formula is:
Wherein, i=0,1,2, L, M-1,2≤t≤M-1, M is disappearance test data length, and M is more than or equal to 3 natural number, and X is sequence of test data to be detected, X
t, X
t+1, L, X
t+M-1for disappearance test data, X
t-1, X
t+Mbe respectively the previous data of disappearance test data section beginning, a rear telemetry at disappearance test data section end.
The method that each father node of the middle calculating of described step (3) is got the conditional probability of child node under different value condition is as follows:
Wherein, Y={y
1, y
2..., y
n, X={x
1, x
2..., x
k, 1≤i≤K, 1≤j≤N, K, N is natural number; X, Y are the node in constructed Bayesian network, the child node that Y is X, the father node that X is Y; P (Y=y
j, X=x
i) be Y=y
j, X=x
iprior probability, P (Y=y
j) be Y=y
jprior probability; P (Y=y
j| X=x
i) be X=x
iy=y under condition
jconditional probability.
The present invention compared with prior art has following beneficial effect: the satellite health status comprehensive quantification based on ground test data provided by the invention is determined method, from historical data angle, excavated and system under evaluation correlation parameter between probability, fully demonstrated the feature of data under actual condition, the implicit expression incidence relation that has excavated parameter and the entire system general level of the health with the conditional probability of multiparameter, truly reflects system health level.The present invention can realize the whole star of satellite more exactly, the health status in on-board equipment ground test stage is determined, realize the objects such as the whole star of satellite, on-board equipment and be quantified as unified Health Category by many indexs telemetry parameter, for the ground test analysis of satellite provides efficient method.
Accompanying drawing explanation
Fig. 1 is the inventive method process flow diagram;
Fig. 2 is the Bayesian network that comprises battery pack and two on-board equipments of momenttum wheel that utilizes the inventive method to set up;
Fig. 3 is the battery pack Bayesian network that root node only comprises monomer voltage and A group voltage.
Embodiment
In order to solve problems of the prior art, the invention provides a kind of satellite Bayesian network health evaluating method based on ground test data, the core concept of described method is: because satellite is in the ground test stage, working condition difference, the actual general level of the health of satellite set are different, cause its test data to change and differ, are difficult to comprehensively to determine the general level of the health.By system under evaluation, work condition state and test parameter are set up to corresponding Bayes, determine network, utilize historical data to carry out after network training, realize and utilize current data to determine satellite health status.
The principle of the method for the invention is: method of the present invention realizes the problem of detection and the extraction of satellite abnormality for the historical telemetry of utilizing satellite, first, by means of satellite fields ground stage design principle knowledge and ground test engineering experience, be designed for the Bayesian Network Topology Structures of satellite ground test health evaluating; Secondly, the historical ground test data of analyzing and processing, set up Bayesian network training sample in conjunction with expertise experience and historical data; Again, according to training sample, calculate Bayesian network each node prior probability and conditional probability, train the conditional probability between each level node of Bayesian network; Finally, utilize current test data to be assessed to calculate corresponding on-board equipment and whole star health status.
Specifically comprise and build Bayesian network, process historical test data and determine training sample, utilize training sample training Bayesian network to obtain prior probability and conditional probability, determine satellite health status;
(1) build Bayesian network: according to the whole star structure of satellite, form and each parameter index, build for determining the Bayesian network of satellite health status to be assessed, determine the hierarchical relationship between whole star, on-board equipment, ground test parameter; Bayesian network is a directed acyclic graph, and it consists of the directed edge that represents the node of variable and connect these nodes, and directed edge points to child node by father node, without the node of child node, is leaf node, without the node of father node, is root node; The leaf node of Bayesian network is whole star health status, and intermediate node is that the father node of leaf node is on-board equipment, and root node is that the father node of on-board equipment is this equipment ground testing parameter, and ground test parameter, without father node, is root node;
(2) process historical data and determine training sample: from the ground test historical data of satellite to be assessed, extracting the ground test parameter historical data that each root node is corresponding, carry out error code and singular point rejecting, null value filling and mean value computation pre-service; Continuous variable is wherein carried out to discretize, determine the corresponding relation from connection attribute to Category Attributes, the codomain that is about to numerical attribute is divided into some sub-ranges, each interval corresponding discrete value; According to historical evaluation result, determine the Health Category of on-board equipment and whole star under every group of ground test Parameter Conditions, determine the value of intermediate node under the different value conditions of root node, and the value of leaf node under the different value conditions of intermediate node; Through the root node data of pre-service, discretize and corresponding intermediate node and leaf node data, form the healthy training sample of determining model of satellite Bayesian network;
(3) utilize the constructed Bayesian network of training sample training step (1), obtain prior probability and conditional probability: utilize training sample to train constructed Bayesian network, by training sample, add up the prior probability of root node and each intermediate node in built Bayesian network, then calculate successively the conditional probability that each father node is got child node under different value condition, calculate each root node and get the conditional probability of intermediate node under different value condition, each intermediate node is got the conditional probability of leaf node under different value condition, finally draw the conditional probability of whole star health status under on-board equipment Health Category condition,
(4) determine satellite health status: by satellite to be assessed through error code and singular point reject, null value is filled and mean value computation pre-service after in the constructed Bayesian network of the current ground test parameter input step (1) of each root node of obtaining, the conditional probability that Bayesian network obtains according to step (3) is calculated the health status of the whole star of output.
Embodiment:
Describe implementation procedure of the present invention below in detail, for simplifying the analysis, only consider the impact of two on-board equipments of battery pack and momenttum wheel on whole star health status here, set up for determining the Bayesian network of satellite health status, as shown in Figure 2.And to take two test points of monomer voltage and A group voltage be example explanation Bayesian network probability calculation and Bayesian inference process, comprise Bayesian network that monomer voltage and A organize two root nodes of voltage as shown in Figure 3.
(1) set up for determining the Bayesian network of satellite health status to be assessed.
Set up for the Bayesian network of determining satellite health status as shown in Figure 2, the leaf node that whole starlike state is Bayesian network, its father node is that on-board equipment is accumulator and momenttum wheel; The father node of momenttum wheel is its ground test parameter, i.e. momenttum wheel bearing temperature, momenttum wheel current of electric, momenttum wheel motor current and momenttum wheel rotating speed, and these parameter points are root node; The father node of accumulator is its ground test parameter, and A group temperature, B group temperature, monomer voltage, charging current and A organize voltage, and these parameter nodes are root node.
(2) determine the sample set of training Bayesian network.
A, the ground test parameter historical data pre-service corresponding to root node
From the ground test historical data of satellite to be assessed, extract the ground test parameter historical data that each root node is corresponding, carry out error code and singular point rejecting, null value filling and mean value computation pre-service, the per minute average of fetching data here;
Singular point is rejected main serial mean and the variance yields that calculates ground test data that adopt, the mode of adding and subtracting three times of variance yields by mean value determine in test data 95% data should in data area, thereby eliminate the not singular point data within the scope of this, directly this data point is set to sky; Minute average of test data after calculating rejecting error code and singular point; For null value point data, to utilize the method for linear interpolation to carry out benefit value to null value point, thereby form a complete ordered series of numbers, the missing data section of supposing sequence of test data is X
t, X
t+1, L, X
t+M-1, a linear interpolation data benefit point formula is:
Wherein, i=0,1,2, L, M-1,2≤t≤M-1, M is disappearance test data length, and M is more than or equal to 3 natural number, and X is sequence of test data to be detected, X
t, X
t+1, L, X
t+M-1for disappearance test data, X
t-1, X
t+Mbe respectively the previous data of disappearance test data section beginning, a rear telemetry at disappearance test data section end.
B, to the continuous data discretize in ground test parameter corresponding to root node
The supplemental characteristic of each root node by satellite ground test data, detected, continuous variable is wherein carried out to discretize, the codomain that is about to numerical attribute is divided into some sub-ranges, each interval corresponding discrete value, numerical discretization algorithm is exactly that requirement can be determined the corresponding relation from connection attribute to Category Attributes automatically.Take the leaf node monomer voltage of battery pack and A group voltage is example, by actual data analysis being determined to the boundary that each node state is divided is 3 classes.When parameter value is lower than minimum threshold time, be made as the 1st grade, when parameter value is during higher than high threshold, its grade is made as to 3rd level.For the value in normal fluctuation range, it is evenly divided into 2 grades.In accumulator health evaluating Bayesian network, the grade classification scope of parameters is as shown in table 1.
The discrete data table of table 1 ground test data
| Discrete level | 1 | 2 | 3 |
| Monomer voltage | <0.4 | 0.4～1.0 | >1.0 |
| A organizes voltage | <6 | 6～14.75 | >14.75 |
C, intermediate node and leaf node Health Category are divided, and on-board equipment and whole star Health Category are divided
As shown in table 2, accumulator and satellite are all divided into Three Estate: normal, abnormal.When health status result of calculation is greater than 1.5, determine that satellite/equipment state is normal, when health status result of calculation is less than 1.5, determine that satellite/equipment state may exist extremely, need to further analyze, diagnose.
Table 2 accumulator and whole star Health Category are divided table
| Health Category | 1 | 2 |
| Health status is described | Extremely | Normally |
Through the root node data of pre-service, discretize and corresponding intermediate node and leaf node data, form the healthy training sample of determining model of satellite Bayesian network.
(3) utilize the constructed Bayesian network of training sample training step (1), obtain prior probability and conditional probability.
A, calculating prior probability
Through sample data statistical study, obtain root node monomer voltage, the prior probability of A group voltage under different discrete level is as shown in table 3, the prior probability of accumulator is as shown in table 4.
The prior probability table of table 3 ground test discretize data
| Discrete level | 1 | 2 | 3 |
| Monomer voltage | 0.478 | 0.1016 | 0.4086 |
| A organizes voltage | 0.2257 | 0.3645 | 0.4098 |
Table 4 accumulator prior probability table
| Health Category | 1 | 2 |
| Prior probability | <0.102 | 0.898 |
B, design conditions probability
The method that conditional probability is calculated is as follows:
Wherein, Y={y
1, y
2..., y
n, X={x
1, x
2..., x
k, 1≤i≤K, 1≤j≤N, K, N is natural number; X, Y are satellite health and determine the node in Bayesian network, the child node that Y is X, the father node that X is Y; P (Y=y
j, X=x
i) be Y=y
j, X=x
iprior probability, P (Y=y
j) be Y=y
jprior probability; P (Y=y
j| X=x
i) be X=x
iy=y under condition
jconditional probability.
Calculate by statistics, the conditional probability of relative its father node of each node is as shown in table 5, table 6, and wherein XDC represents accumulator, and ZX represents whole star, and ADY represents A group voltage, and DDY represents monomer voltage.
Table 5 accumulator conditional probability
| (DDY,ADY) | XDC=1 | XDC=2 |
| (1,1) | 0.811 | 0.189 |
| (1,2) | 0.372 | 0.628 |
| (1,3) | 0.122 | 0.878 |
| (2,1) | 0.431 | 0.869 |
| (2,2) | 0.129 | 0.871 |
| (2,3) | 0.233 | 0.767 |
| (3,1) | 0.011 | 0.989 |
| (3,2) | 0.061 | 0.939 |
| (3,3) | 0.824 | 0.176 |
The whole star conditional probability of table 6
| XDC | ZX=1 | XDC=2 |
| 1 | 0.361 | 0.639 |
| 2 | 0.197 | 0.803 |
(4) determine satellite health status
In the constructed Bayesian network of step (1), input satellite to be assessed through error code and singular point reject, null value is filled and mean value computation in the current ground test data of each root node of obtaining after pre-service, the conditional probability drawing according to step (3) calculates the health status of whole star.
As, current monomer voltage=0.84, A organize voltage=15.13, and known discretize monomer voltage=2, A organize voltage=3; By in the Bayesian network of the monomer voltage of discretize and A group magnitude of voltage input structure, obtaining whole star health status result of calculation is 1.765, thereby determines that current satellite is in normal condition.
Bayesian network internal calculation process is: by table 5, obtained, the probability of accumulator=1 is 0.233, and the probability of accumulator=2 is 0.767, and obtaining accumulator health status result of calculation is 1.767, is greater than 1.5, determines that accumulator is in normal condition; Associative list 6, the probability that whole star is 1 is 0.233*0.361+0.767*0.197=0.235, the probability that whole star is 2 is 0.233*0.639+0.767*0.803=0.765, obtains whole star health status to be:
0.235*1+0.765*2=1.765>1.5
The current state of determining satellite is normal.
The present invention not detailed description is known to the skilled person technology.
Claims (3)
1. the healthy definite method of the satellite Bayesian network based on ground test data, is characterized in that comprising the steps:
(1) build Bayesian network: build for determining the Bayesian network of satellite health status to be assessed, wherein the leaf node of Bayesian network is whole star health status, and intermediate node is on-board equipment, and root node is the ground test parameter that on-board equipment is corresponding;
(2) determine training sample: from the ground test historical data of satellite to be assessed, extract the ground test parameter historical data that each root node is corresponding, carry out error code and singular point rejecting, null value filling and mean value computation pre-service; Continuous variable is wherein carried out to discretize, determine the corresponding relation from connection attribute to Category Attributes, the codomain that is about to numerical attribute is divided into some sub-ranges, each interval corresponding discrete value; According to historical evaluation result, determine the value of intermediate node under the different value conditions of root node, and the value of leaf node under the different value conditions of intermediate node; Through the root node data of pre-service, discretize and corresponding intermediate node and leaf node data, form the healthy training sample of determining model of satellite Bayesian network;
(3) the constructed Bayesian network of training sample training step (1) that utilizes step (2) to obtain, obtain conditional probability: the prior probability of adding up root node and each intermediate node in constructed Bayesian network by training sample, and further calculate accordingly the conditional probability that each father node is got child node under different value condition, obtain each root node and get the conditional probability that the conditional probability of intermediate node under different value condition, each intermediate node are got leaf node under different value condition;
(4) determine satellite health status: by satellite to be assessed through error code and singular point reject, null value is filled and mean value computation pre-service after in the constructed Bayesian network of the current ground test parameter input step (1) of each root node of obtaining, the conditional probability that Bayesian network obtains according to step (3) is calculated the health status of the whole star of output.
2. a kind of satellite Bayesian network health based on ground test data according to claim 1 is determined method, it is characterized in that: extract ground test parameter historical data corresponding to each root node in described step (2), carry out that error code and singular point are rejected, that null value is filled pretreated implementation method is as follows:
Singular point is rejected main serial mean and the variance yields that calculates the ground test parameter historical data that each root node is corresponding that adopt, the mode of adding and subtracting three times of variance yields by mean value determine in test data 95% data should in data area, thereby eliminate the not singular point data within the scope of this, directly this data point is set to sky; Minute average of test data after calculating rejecting error code and singular point; For null value point data, to utilize the method for linear interpolation to carry out benefit value to null value point, thereby form a complete ordered series of numbers, the missing data section of supposing sequence of test data is X
t, X
t+1, L, X
t+M-1, a linear interpolation data benefit point formula is:
Wherein, i=0,1,2, L, M-1,2≤t≤M-1, M is disappearance test data length, and M is more than or equal to 3 natural number, and X is sequence of test data to be detected, X
t, X
t+1, L, X
t+M-1for disappearance test data, X
t-1, X
t+Mbe respectively the previous data of disappearance test data section beginning, a rear telemetry at disappearance test data section end.
3. a kind of satellite Bayesian network health based on ground test data according to claim 1 is determined method, it is characterized in that: the method that each father node of the middle calculating of described step (3) is got the conditional probability of child node under different value condition is as follows:
Wherein, Y={y
1, y
2..., y
n, X={x
1, x
2..., x
k, 1≤i≤K, 1≤j≤N, K, N is natural number; X, Y are the node in constructed Bayesian network, the child node that Y is X, the father node that X is Y; P (Y=y
j, X=x
i) be Y=y
j, X=x
iprior probability, P (Y=y
j) be Y=y
jprior probability; P (Y=y
j| X=x
i) be X=x
iy=y under condition
jconditional probability.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310611919.XA CN103678886B (en) | 2013-11-26 | 2013-11-26 | A kind of satellite Bayesian network health based on ground test data determines method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310611919.XA CN103678886B (en) | 2013-11-26 | 2013-11-26 | A kind of satellite Bayesian network health based on ground test data determines method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103678886A true CN103678886A (en) | 2014-03-26 |
| CN103678886B CN103678886B (en) | 2017-01-04 |
Family
ID=50316416
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310611919.XA Active CN103678886B (en) | 2013-11-26 | 2013-11-26 | A kind of satellite Bayesian network health based on ground test data determines method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103678886B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104590594A (en) * | 2015-01-27 | 2015-05-06 | 北京空间飞行器总体设计部 | Method for testing and verifying information flow among spacecraft |
| CN105468917A (en) * | 2015-12-01 | 2016-04-06 | 北京无线电计量测试研究所 | Pipeline fault prediction method and apparatus |
| CN105573329A (en) * | 2015-12-16 | 2016-05-11 | 上海卫星工程研究所 | Attitude and orbit control data analysis method based on decision tree |
| CN110046376A (en) * | 2019-02-26 | 2019-07-23 | 中国西安卫星测控中心 | A kind of Satellite Attitude Control System multi-state health evaluating method based on Bayesian network |
| CN110334395A (en) * | 2019-05-28 | 2019-10-15 | 中国地质大学（武汉） | The satellite momentum wheel fault diagnosis method and system of initialization EM algorithm based on JADE |
| CN112668749A (en) * | 2020-11-24 | 2021-04-16 | 江苏中矿安华科技发展有限公司 | Coal mine gas early warning method based on class mark weighting extreme learning machine |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101354757A (en) * | 2008-09-08 | 2009-01-28 | 中国科学院地理科学与资源研究所 | Method for predicting dynamic risk and vulnerability under fine dimension |
| US20110270671A1 (en) * | 2010-05-03 | 2011-11-03 | Hsiu-Kheurn Tang | Predicting number of selections of advertisement using hierarchical Bayesian model |
| CN103218689A (en) * | 2013-04-23 | 2013-07-24 | 湖南工学院 | Analyzing method and analyzing device for operator state assessment reliability |
-
2013
- 2013-11-26 CN CN201310611919.XA patent/CN103678886B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101354757A (en) * | 2008-09-08 | 2009-01-28 | 中国科学院地理科学与资源研究所 | Method for predicting dynamic risk and vulnerability under fine dimension |
| US20110270671A1 (en) * | 2010-05-03 | 2011-11-03 | Hsiu-Kheurn Tang | Predicting number of selections of advertisement using hierarchical Bayesian model |
| CN103218689A (en) * | 2013-04-23 | 2013-07-24 | 湖南工学院 | Analyzing method and analyzing device for operator state assessment reliability |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104590594A (en) * | 2015-01-27 | 2015-05-06 | 北京空间飞行器总体设计部 | Method for testing and verifying information flow among spacecraft |
| CN105468917A (en) * | 2015-12-01 | 2016-04-06 | 北京无线电计量测试研究所 | Pipeline fault prediction method and apparatus |
| CN105468917B (en) * | 2015-12-01 | 2019-03-08 | 北京无线电计量测试研究所 | A kind of pipeline failure prediction technique and device |
| CN105573329A (en) * | 2015-12-16 | 2016-05-11 | 上海卫星工程研究所 | Attitude and orbit control data analysis method based on decision tree |
| CN105573329B (en) * | 2015-12-16 | 2018-05-18 | 上海卫星工程研究所 | Rail control data analysing method based on decision tree |
| CN110046376A (en) * | 2019-02-26 | 2019-07-23 | 中国西安卫星测控中心 | A kind of Satellite Attitude Control System multi-state health evaluating method based on Bayesian network |
| CN110046376B (en) * | 2019-02-26 | 2022-11-22 | 中国西安卫星测控中心 | Multi-working-condition health assessment method for satellite attitude control system based on Bayesian network |
| CN110334395A (en) * | 2019-05-28 | 2019-10-15 | 中国地质大学（武汉） | The satellite momentum wheel fault diagnosis method and system of initialization EM algorithm based on JADE |
| CN112668749A (en) * | 2020-11-24 | 2021-04-16 | 江苏中矿安华科技发展有限公司 | Coal mine gas early warning method based on class mark weighting extreme learning machine |
| CN112668749B (en) * | 2020-11-24 | 2023-07-07 | 江苏中矿安华科技发展有限公司 | Coal mine gas early warning method based on class mark weighting extreme learning machine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103678886B (en) | 2017-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xiang et al. | Condition monitoring and anomaly detection of wind turbine based on cascaded and bidirectional deep learning networks | |
| CN110221225B (en) | Spacecraft lithium ion battery cycle life prediction method | |
| CN106779505B (en) | Power transmission line fault early warning method and system based on big data driving | |
| CN103678886A (en) | Satellite Bayesian Network health determination method based on ground test data | |
| CN102789545B (en) | Based on the Forecasting Methodology of the turbine engine residual life of degradation model coupling | |
| CN102208028B (en) | Fault predicting and diagnosing method suitable for dynamic complex system | |
| CN108960303B (en) | Unmanned aerial vehicle flight data anomaly detection method based on LSTM | |
| CN104064023B (en) | A kind of Dynamic Traffic Flow Prediction method based on space time correlation | |
| CN105300692B (en) | A kind of bearing failure diagnosis and Forecasting Methodology based on expanded Kalman filtration algorithm | |
| CN103646167A (en) | Satellite abnormal condition detection system based on telemeasuring data | |
| CN106484949A (en) | Momenttum wheel fail-safe analysis and method for predicting residual useful life based on degraded data | |
| CN111680875B (en) | Unmanned aerial vehicle state risk fuzzy comprehensive evaluation method based on probability baseline model | |
| CN102879677A (en) | Intelligent fault diagnosis method based on rough Bayesian network classifier | |
| CN107807860B (en) | Power failure analysis method and system based on matrix decomposition | |
| CN104268375A (en) | Ship electric power station fault diagnosing method based on knowledge petri network | |
| Qiu et al. | The early-warning model of equipment chain in gas pipeline based on DNN-HMM | |
| CN107942994A (en) | A kind of satellite temperature control system method for diagnosing faults based on temperature curve feature | |
| CN111126477A (en) | Learning and reasoning method of hybrid Bayesian network | |
| Chen et al. | A novel Bayesian-optimization-based adversarial TCN for RUL prediction of bearings | |
| CN114036647A (en) | Power battery safety risk assessment method based on real vehicle data | |
| Jiang et al. | Remaining useful life prediction of rolling bearings based on Bayesian neural network and uncertainty quantification | |
| Wang et al. | Assessing the performance degradation of lithium-ion batteries using an approach based on fusion of multiple feature parameters | |
| CN105741184A (en) | Transformer state evaluation method and apparatus | |
| CN112232570A (en) | Forward active total electric quantity prediction method and device and readable storage medium | |
| Dang et al. | seq2graph: Discovering dynamic non-linear dependencies from multivariate time series |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |