CN111310926A - Fault alarm threshold determination method fusing multivariate information - Google Patents

Fault alarm threshold determination method fusing multivariate information Download PDF

Info

Publication number
CN111310926A
CN111310926A CN202010052024.7A CN202010052024A CN111310926A CN 111310926 A CN111310926 A CN 111310926A CN 202010052024 A CN202010052024 A CN 202010052024A CN 111310926 A CN111310926 A CN 111310926A
Authority
CN
China
Prior art keywords
interval
fault
threshold
threshold value
confidence
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010052024.7A
Other languages
Chinese (zh)
Other versions
CN111310926B (en
Inventor
周志杰
唐帅文
胡昌华
刘涛源
曹友
陈媛
张超丽
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.)
Rocket Force University of Engineering of PLA
Original Assignee
Rocket Force University of Engineering of PLA
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 Rocket Force University of Engineering of PLA filed Critical Rocket Force University of Engineering of PLA
Priority to CN202010052024.7A priority Critical patent/CN111310926B/en
Publication of CN111310926A publication Critical patent/CN111310926A/en
Application granted granted Critical
Publication of CN111310926B publication Critical patent/CN111310926B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/18—Status alarms
    • G08B21/182—Level alarms, e.g. alarms responsive to variables exceeding a threshold

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Mathematical Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Computing Systems (AREA)
  • Computational Mathematics (AREA)
  • Biophysics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Biomedical Technology (AREA)
  • Physiology (AREA)
  • Emergency Management (AREA)
  • Operations Research (AREA)
  • Probability & Statistics with Applications (AREA)
  • Business, Economics & Management (AREA)
  • Algebra (AREA)
  • Databases & Information Systems (AREA)
  • Monitoring And Testing Of Exchanges (AREA)

Abstract

The invention is suitable for the field of state monitoring and information fusion of industrial processes, and aims to reduce the false and missing report rate of fault alarm and improve the fault detection precision. A fault alarm threshold value determining method fusing multivariate information comprises the following steps: converting data into a unified interval confidence structure; constructing a threshold updating model based on interval evidence reasoning; and optimizing a fault alarm threshold value based on the minimum false alarm rate. The method innovatively utilizes an interval evidence reasoning algorithm to fuse the monitoring data and the interval fault threshold value to obtain the overall interval confidence, adaptively updates the alarm threshold value based on the monitoring data and the projection covariance matrix adaptive evolution strategy, and effectively solves various uncertain problems caused by environmental factors. And obtaining the alarm interval threshold with the lowest fault detection false alarm rate through threshold optimization. The invention can effectively reduce the times of false alarm and false alarm of the monitoring system, improve the safety and reliability of the industrial process and has good engineering application value.

Description

Fault alarm threshold determination method fusing multivariate information
Technical Field
The invention is suitable for the field of state monitoring and information fusion of industrial processes, and aims to reduce the false and missing report rate of fault alarm and improve the fault detection precision of a monitoring system. Relates to a method for determining a fault alarm threshold value by fusing multivariate information.
Background
In the state monitoring of the industrial process, a fault alarm threshold value is used as an important criterion for system fault, and the reasonability of the setting directly influences the monitoring precision of the fault. As shown in FIG. 1, the solid blue line represents the actual failure threshold of the system, if the threshold is set too high (e.g., dashed line S)1) Although the monitoring value P exceeds the actual threshold value, the alarm system does not perform fault early warning, so that the fault is not reported, and the personal safety and the system safety of workers are threatened; if the threshold is set too low (e.g., dashed line S)2) When the monitored value Q is lower than the actual threshold value and exceeds S2In time, the alarm system can also frequently give an alarm, so that fault false alarm is caused, and the work order is influenced. Therefore, the accurate and reasonable fault alarm threshold value is set, and the method has great significance for improving the working efficiency and ensuring the safe and reliable operation of the system.
At present, alarm threshold determination methods based on models, statistics, neural networks and fuzzy reasoning can complete threshold determination in corresponding fields. However, in the industrial process, the monitored data is interfered by factors such as noise and the like, and various uncertainties such as data loss and information distortion may exist. In addition, for a complex system, the monitoring indexes are often complex and diverse, the characteristics are different, and quantitative information and qualitative knowledge are mutually interwoven and are called as multivariate information in the invention. The expert system-based method can quantify various uncertainties to the same frame, effectively fuses the multi-element information, and is an effective multi-attribute decision making (MADM) method under the uncertainties. Therefore, in the field of state monitoring and information fusion of industrial processes, the threshold determination method based on expert knowledge has more obvious advantages than other methods.
The industrial environment is complex and changeable, and the accurate numerical threshold value generally cannot directly represent fault information under the interference action; the interval threshold value can be fully combined with expert knowledge and system mechanism analysis, the fault threshold value is controlled within a certain reasonable range, and more reasonable judgment is made on the fault according to real-time monitoring data. It can be seen that the study of alarm thresholds is essentially the MADM problem under one interval uncertainty. Therefore, the invention provides a method for determining a fault alarm threshold value by fusing multivariate information, which expands a numerical alarm threshold value to an interval type, and aims to reduce the false and missed alarm rate of fault detection and improve the performance of a monitoring system.
Disclosure of Invention
Aiming at the problem of determining the alarm threshold in the industrial process state monitoring, the invention provides a fault alarm threshold determining method which integrates multivariate information and aims to effectively integrate expert knowledge and solve various uncertainties such as monitoring data loss, information distortion and the like.
The technical scheme of the invention is as follows:
a failure alarm threshold value determining method fusing multivariate information is characterized by comprising the following steps:
step 1: monitoring data x and interval fault threshold value y1,y2]Converting into the same interval confidence structure;
the invention monitors data x and interval fault threshold value y1,y2]As two pieces of evidence, the IER algorithm cannot directly fuse them because they have different expression forms and numerical units. In view of the advantage of the confidence structure being able to express any form of data, the monitoring data x and the interval fault threshold y are first compared before applying the IER algorithm1,y2]And (4) converting into a unified confidence structure.
The interval confidence structure is as follows:
Figure BDA0002371512640000021
wherein F is an evaluation scale, F { (F)1Normal), (F)2Fault), (F)3Serious fault) }; the selection of the reference value is determined according to historical monitoring data and expert knowledge;
Figure BDA0002371512640000022
expressed as index aiRelative to evaluation grade FnInterval confidence of (2);
monitoring data x and interval fault threshold value y are converted by adopting rule-based input information conversion technology1,y2]Respectively converting to interval confidence structures relative to the reference levels;
converting the monitoring data x into an interval confidence structure, and expressing the following steps:
Figure BDA0002371512640000023
section fault threshold [ y ]1,y2]And converting into an interval confidence structure, and expressing as:
Figure BDA0002371512640000024
step 2: constructing a threshold updating model based on interval evidence reasoning;
firstly, converting an interval confidence structure into interval probability quality by adopting a weighting method;
secondly, sequentially fusing monitoring data x and interval fault threshold value y by adopting an interval evidence reasoning algorithm IER1,y2]Obtaining fused interval confidence by the corresponding interval probability quality;
thirdly, threshold optimization is carried out by adopting an algorithm based on a Projection covariance matrix adaptive evolution strategy (P-CMA-ES), and the confidence coefficient of the overall interval is obtained by updating the monitoring data x
Figure BDA0002371512640000025
Finally, will
Figure BDA0002371512640000026
Converting into an interval threshold;
after the interval confidence structure shown in the formula (1) is obtained, firstly converting the interval confidence structure into interval probability quality;
suppose monitoring data x and interval fault threshold y1,y2]Are respectively weighted as omegaiWherein i is 1, 2; and omega1＝ω2The interval probability mass can be obtained by the following equations (4) to (7):
Figure BDA0002371512640000031
Figure BDA0002371512640000032
Figure BDA0002371512640000033
Figure BDA0002371512640000034
wherein the content of the first and second substances,
Figure BDA0002371512640000035
representing the residual probability mass caused by the immaterial nature of the evidence,
Figure BDA0002371512640000036
representing the residual probability mass caused by the imperfection of the evaluation,
Figure BDA0002371512640000037
since the formula (1) is complete,
Figure BDA0002371512640000038
fusing monitoring data x and interval fault threshold value y by constructing the following nonlinear optimization model and based on IER algorithm1,y2]Obtaining the confidence of the overall interval
Figure BDA0002371512640000039
Figure BDA00023715126400000310
Figure BDA00023715126400000311
Figure BDA00023715126400000312
Figure BDA00023715126400000313
Figure BDA00023715126400000314
Figure BDA00023715126400000315
Figure BDA00023715126400000316
Figure BDA00023715126400000317
Solving the nonlinear optimization model based on P-CMA-ES algorithm, and using the result
Figure BDA00023715126400000318
This means that the updated section threshold is obtained based on equation (3).
And step 3: optimizing a fault alarm threshold based on the minimum false missing report rate;
constructing a threshold optimization model of the minimum false missing report rate;
the threshold optimization model is as follows:
Figure BDA0002371512640000041
wherein s is the false-missing report rate of the fault, w is the false-report rate, l represents the false-missing report rate, N is the influence degree of the false-missing report rate on the system, y1And y2Respectively represent initial interval fault threshold value y1,y2]The upper and lower boundaries of the target fault alarm threshold value y are obtained through optimization1,y2]'L。
The invention has the technical effects that:
the core theory of the invention is interval evidence reasoning, and the monitoring data and the fault threshold value are converted into a uniform interval confidence structure by combining expert knowledge; updating a threshold interval is realized by constructing a pair of nonlinear optimization models; and finally, optimizing the interval threshold by solving a threshold optimization model taking the minimum false alarm rate as a target function.
Drawings
Fig. 1 is a diagram illustrating the setting of a malfunction alarm threshold value according to the present invention.
Fig. 2 is a general technical scheme diagram of the invention.
FIG. 3 is a diagram of a threshold update model based on interval evidence reasoning in step 2 of the present invention.
FIG. 4 is a diagram of a threshold optimization model based on the minimum false negative rate in step 3 of the present invention.
FIG. 5 is a graph of data for an embodiment of the present invention.
Detailed Description
The invention takes the threshold value determination of the key parameter (pull-in time) of a certain type of space relay as an example to illustrate the effectiveness of the method. In an accelerated life experiment of a JRC-7M relay, 5800 groups of pull-in time data shown in figure 5 are selected, the known relay pull-in time is in a fault state after about 3500 times of actions, and 300 groups of data are selected as training data x in a critical state1,x2,...,x300Then 200 groups of data are selected respectively in the fault state and the normal state as the test data c of the fault false and missing report rate1,c2,...,c200And c'1,c′2,...,c′200. The method for determining the fault alarm threshold value of the pull-in time comprises the following steps:
monitoring data x and interval fault threshold value y1,y2]Converting into the same interval confidence structure;
according to the technical specification of the JRC-7M space relay, the initial threshold value is set to be 7.6000,7.7000]Define an evaluation level F { (F)1Normal), (F)2Fault), (F)3Serious failure), the reference values of the levels "normal" and "failure" are f, respectively1＝6.5,f27.52, and "catastrophic failure" reference value f3Given by an expert in combination with historical fault data of the pull-in time of the relay, h37.96. The initial threshold and pull-in time data are converted to an interval confidence structure, as shown in table 1:
TABLE 1 Interval confidence of fusion index relative assessment level
Fusion index F1 F2 F3
[y1，y2]0 [0，0] [0.5909，0.8182] [0.1818，0.4091]
x1 [0.0256，0.0256] [0.9744，0.9744] [0，0]
x2 [0.0192，0.0192] [0.9808，0.9808] [0，0]
... ... ... ...
x300 [0，0] [0.6579，0.6579] [0.3421，0.3421]
Step 2: constructing a threshold updating model based on interval evidence reasoning;
weighting ω1＝ω20.5. First, x is obtained based on the formulae (4) to (7)1Interval confidence of, will ω1、ω2The sum interval confidence coefficient is substituted into the formulas (8) to (11), and an optimization model is solved based on a P-CMA-ES algorithm to obtain the fused interval confidence coefficient [ min β ]n，maxβn]1. Next, data x are used separately2，...，x300Updating to obtain confidence [ min β ] of overall intervaln，maxβn]300Based on formula (3), [ min βn，maxβn]300And converting into an interval threshold value.
And step 3: optimizing a fault alarm threshold based on the minimum false missing report rate;
and respectively using 200 groups of normal state data and 200 groups of fault data to test the interval threshold value, and counting the false alarm rate w and the false failure rate l. According to expert experience, taking N as 2 and giving an initial threshold value y1，y2]0Constraint of y1∈[7.57，7.64]And y2∈[7.65，7.73]. By optimization, when the initial threshold is [7.6207, 7.6824 ]]Then, the minimum false alarm rate min s is 10%, w is 9%, and l is 0.5%. At this time, the updating process of the interval confidence structure is shown in table 2.
TABLE 2 Overall Interval confidence update procedure
Overall interval confidence F1 F2 F3
[minβn，maxβn]1 [0，0] [0.7093，0.7619] [0.2381，0.2907]
[minβn，maxβn]2 [0，0] [0.7078，0.7710] [0.2290，0.2922]
... ... ... ...
[minβn，maxβn]100 [0，0] [0.6842，0.7135] [0.2865，0.3158]
... ... ... ...
[minβn，maxβn]300 [0，0] [0.6547，0.7488] [0.2512，0.3453]
Finally, the optimized overall interval confidence is converted into an interval threshold, and the result is [7.6305, 7.6719 ].
In addition, the initial threshold and the optimized threshold obtained by the mean method, the 3 σ method and the IER method are tested, and the false alarm rate statistics are shown in table 3:
TABLE 3 threshold optimization method comparisons
Optimization method Threshold value False alarm rate Rate of missing reports s
Not optimized [7.6207，7.6824] 15.5％ 0％ 15.5％
3 sigma method [7.6443，7.6596] 7％ 3％ 13％
Neural network method [7.6394，7.6496] 7％ 3.5％ 14％
IER method [7.6305，7.6719] 9％ 0.5％ 10％
.
According to the statistics of the false alarm rate, in the case of accelerating the service life of the aerospace relay, the fault false alarm rate value is the largest when the threshold value is not optimized and is 15.5%; the minimum value after optimization by adopting the IER method is 10%, and the method is obviously superior to other methods, and the effectiveness of the method on the threshold judgment problem is verified.

Claims (3)

1. A failure alarm threshold value determining method fusing multivariate information is characterized by comprising the following steps: the method comprises the following steps:
step 1: monitoring data x and interval fault threshold value y1,y2]Converting into the same interval confidence structure;
step 2: constructing a threshold updating model based on interval evidence reasoning;
firstly, converting an interval confidence structure into interval probability quality by adopting a weighting method;
secondly, sequentially fusing monitoring data x and interval fault threshold value y by adopting an interval evidence reasoning algorithm IER1,y2]Obtaining fused interval confidence by the corresponding interval probability quality;
thirdly, threshold optimization is carried out by adopting a projection covariance matrix-based adaptive evolution strategy P-CMA-ES algorithm, and the confidence coefficient of the overall interval is obtained by updating the monitoring data x
Figure FDA0002371512630000011
Finally, will
Figure FDA0002371512630000012
Converting into an interval threshold;
and step 3: optimizing a fault alarm threshold based on the minimum false missing report rate;
constructing a threshold optimization model of the minimum false missing report rate;
the threshold optimization model is as follows:
Figure FDA0002371512630000013
wherein s is the false-missing report rate of the fault, w is the false-report rate, l represents the false-missing report rate, N is the influence degree of the false-missing report rate on the system, y1And y2Respectively represent initial interval fault threshold value y1,y2]The upper and lower boundaries of the target fault alarm threshold value y are obtained through optimization1,y2]'L。
2. The method for determining a threshold value of a failure alarm based on fusion multivariate information as defined in claim 1, wherein: the interval confidence structure is as follows:
the interval confidence structure is as follows:
Figure FDA0002371512630000014
wherein F is an evaluation scale, F { (F)1Normal), (F)2Fault), (F)3Serious fault) };
Figure FDA0002371512630000015
expressed as index aiRelative to evaluation grade FnInterval confidence of (2);
converting the monitoring data x into an interval confidence structure, and expressing the following steps:
Figure FDA0002371512630000016
section fault threshold [ y ]1,y2]And converting into an interval confidence structure, and expressing as:
Figure FDA0002371512630000021
3. the method for determining a threshold value of a failure alarm based on fusion multivariate information as defined in claim 2, wherein: the specific implementation process of the establishment of the threshold updating model based on the interval evidence reasoning is as follows:
after the interval confidence structure shown in the formula (1) is obtained, firstly converting the interval confidence structure into interval probability quality;
suppose monitoring data x and interval fault threshold y1,y2]Are respectively weighted as omegaiWherein i is 1, 2; and omega1＝ω2The interval probability mass can be obtained by the following equations (4) to (7):
Figure FDA0002371512630000022
Figure FDA0002371512630000023
Figure FDA0002371512630000024
Figure FDA0002371512630000025
wherein the content of the first and second substances,
Figure FDA0002371512630000026
representing the residual probability mass caused by the immaterial nature of the evidence,
Figure FDA0002371512630000027
representing the residual probability mass caused by the imperfection of the evaluation,
Figure FDA0002371512630000028
since the formula (1) is complete,
Figure FDA0002371512630000029
fusing monitoring data x and interval fault threshold value y by constructing the following nonlinear optimization model and based on IER algorithm1,y2]Obtaining the confidence of the overall interval
Figure FDA00023715126300000210
Figure FDA00023715126300000211
Figure FDA00023715126300000212
Figure FDA00023715126300000213
Figure FDA0002371512630000031
Figure FDA0002371512630000032
Figure FDA0002371512630000033
Figure FDA0002371512630000034
Figure FDA0002371512630000035
Solving the nonlinear optimization model based on P-CMA-ES algorithm, and using the result
Figure FDA0002371512630000036
This means that the updated section threshold is obtained based on equation (3).
CN202010052024.7A 2020-01-17 2020-01-17 Fault alarm threshold determination method fusing multivariate information Active CN111310926B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010052024.7A CN111310926B (en) 2020-01-17 2020-01-17 Fault alarm threshold determination method fusing multivariate information

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010052024.7A CN111310926B (en) 2020-01-17 2020-01-17 Fault alarm threshold determination method fusing multivariate information

Publications (2)

Publication Number Publication Date
CN111310926A true CN111310926A (en) 2020-06-19
CN111310926B CN111310926B (en) 2022-10-11

Family

ID=71144899

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010052024.7A Active CN111310926B (en) 2020-01-17 2020-01-17 Fault alarm threshold determination method fusing multivariate information

Country Status (1)

Country Link
CN (1) CN111310926B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112270814A (en) * 2020-12-21 2021-01-26 长沙树根互联技术有限公司 Dynamic alarm method, device, electronic equipment and readable storage medium
CN112950908A (en) * 2021-02-03 2021-06-11 重庆川仪自动化股份有限公司 Data monitoring and early warning method, system, medium and electronic terminal

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109086804A (en) * 2018-07-12 2018-12-25 中石化石油机械股份有限公司 A kind of hydraulic device fault forecast method merged based on multi source status monitoring information and reliability characteristic
CN109946542A (en) * 2019-03-22 2019-06-28 杭州电子科技大学 Inverter DC loop capacitive faults detection method based on the fusion of long short-term information
CN110633450A (en) * 2018-06-06 2019-12-31 中国石油化工股份有限公司 Method for optimizing alarm threshold of hydrogen peroxide device alarm system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110633450A (en) * 2018-06-06 2019-12-31 中国石油化工股份有限公司 Method for optimizing alarm threshold of hydrogen peroxide device alarm system
CN109086804A (en) * 2018-07-12 2018-12-25 中石化石油机械股份有限公司 A kind of hydraulic device fault forecast method merged based on multi source status monitoring information and reliability characteristic
CN109946542A (en) * 2019-03-22 2019-06-28 杭州电子科技大学 Inverter DC loop capacitive faults detection method based on the fusion of long short-term information

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
司小胜等: "基于证据推理的故障预报模型", 《中国科学:信息科学》 *
徐海洋等: "基于置信规则库推理的证据滤波报警器设计", 《山东科技大学学报(自然科学版)》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112270814A (en) * 2020-12-21 2021-01-26 长沙树根互联技术有限公司 Dynamic alarm method, device, electronic equipment and readable storage medium
CN112950908A (en) * 2021-02-03 2021-06-11 重庆川仪自动化股份有限公司 Data monitoring and early warning method, system, medium and electronic terminal
CN112950908B (en) * 2021-02-03 2022-12-30 重庆川仪自动化股份有限公司 Data monitoring and early warning method, system, medium and electronic terminal

Also Published As

Publication number Publication date
CN111310926B (en) 2022-10-11

Similar Documents

Publication Publication Date Title
CN111539515B (en) Complex equipment maintenance decision method based on fault prediction
CN106872657B (en) A kind of multivariable water quality parameter time series data accident detection method
CN109150564B (en) Prediction method and device for cell fault alarm
CN106932670B (en) A kind of distribution power automation terminal method for diagnosing status based on D-S evidence theory
CN111310926B (en) Fault alarm threshold determination method fusing multivariate information
EP3571809A1 (en) Methods and apparatus for analysing performance of a telecommunications network
CN110348752B (en) Large industrial system structure safety assessment method considering environmental interference
CN101718634B (en) Equipment state comprehensive dynamic alarming method based on multivariate probability model
CN110703214A (en) Weather radar state evaluation and fault monitoring method
CN115508672B (en) Power grid main equipment fault tracing reasoning method, system, equipment and medium
CN110865625A (en) Process data anomaly detection method based on time series
CN115063056B (en) Construction behavior safety risk dynamic analysis method based on graph topology analysis improvement
CN112085869A (en) Civil aircraft flight safety analysis method based on flight parameter data
CN114444582A (en) Mechanical equipment fault diagnosis method based on convolutional neural network and Bayesian network
CN117078105A (en) Production quality monitoring method and system based on artificial intelligence
CN116611593A (en) Method, device and medium for predicting failure of air compressor
CN115481726A (en) Industrial robot complete machine health assessment method and system
CN115081647A (en) Industrial intelligent instrument fault pre-diagnosis method based on Bayesian network model
CN111523727A (en) Method for predicting remaining life of battery by considering recovery effect based on uncertain process
CN111126477A (en) Learning and reasoning method of hybrid Bayesian network
CN113610397A (en) Petrochemical enterprise safety evaluation method based on PSO-BP neural network
Smirnov Intelligent decision support system for the control of complex technical systems
Wei et al. A new BRB model for cloud security-state prediction based on the large-scale monitoring data
CN115208773B (en) Network hidden fault monitoring method and device
CN112036653A (en) Fire risk early warning method and system based on Bayesian network

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