CN106155867A - The alarm method of monitoring performance data similarity tolerance and system - Google Patents

The alarm method of monitoring performance data similarity tolerance and system Download PDF

Info

Publication number
CN106155867A
CN106155867A CN201610707419.XA CN201610707419A CN106155867A CN 106155867 A CN106155867 A CN 106155867A CN 201610707419 A CN201610707419 A CN 201610707419A CN 106155867 A CN106155867 A CN 106155867A
Authority
CN
China
Prior art keywords
cycle
time series
similitude
performance data
sequence
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
CN201610707419.XA
Other languages
Chinese (zh)
Inventor
肖飞
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.)
Zhuhai Wiscom Dimensional Mdt Infotech Ltd
Original Assignee
Zhuhai Wiscom Dimensional Mdt Infotech Ltd
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 Zhuhai Wiscom Dimensional Mdt Infotech Ltd filed Critical Zhuhai Wiscom Dimensional Mdt Infotech Ltd
Priority to CN201610707419.XA priority Critical patent/CN106155867A/en
Publication of CN106155867A publication Critical patent/CN106155867A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3065—Monitoring arrangements determined by the means or processing involved in reporting the monitored data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/32—Monitoring with visual or acoustical indication of the functioning of the machine

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Alarm Systems (AREA)

Abstract

The invention provides alarm method and the system of a kind of monitoring performance data similarity tolerance, relate to performance data warning technology field.The method includes: the performance data according to this cycle builds this cycle time series by wavelet transform；Cycle time series on corresponding this cycle seasonal effect in time series of inquiry；Judge whether this cycle time series and upper cycle seasonal effect in time series similitude exceed excursion, continue to build this cycle time series according to the performance data in this cycle by wavelet transform if it is not, return；If so, warning information is triggered.The Time Series Similarity strategy that the present invention provides is on the basis of conventional measures, enrich system exception event detection means, fill up deficiency in system dynamics monitoring for the conventional measures, by the alarm that the i.e. realizability energy time series data behavioral characteristics of simplicity configuration is abnormal.

Description

The alarm method of monitoring performance data similarity tolerance and system
Technical field
The invention belongs to performance data warning technology field, particularly relate to the announcement of a kind of monitoring performance data similarity tolerance Alarm method and system.
Background technology
The monitoring data warning strategies of existing performance monitoring system mainly has with several: the 1st, monitor data variation alarm: prison Control index does not typically change, and triggers as long as there being change to be all probably system exception；2nd, the unchanged alarm of data is monitored: monitor control index It is typically in being continually changing, as long as there being the unchanged for a long time system exception that is all probably to cause；3rd, monitor data threshold to alert: Monitor control index value, in the range of one or several codomains, causes as long as being all probably system exception beyond codomain scope；4th, dynamic Baseline alerts: monitor control index value is in different codomain scope at different time points, and Dynamic Baseline data build a series of according to time point Whether codomain scope, use specific threshold range judge index abnormal at specific time point.Above 1st, the 2nd, 3 warning strategies full Foot administrative staff are to the more stable desired value alarming processing of operation characteristic in system, and for the complicated system of operation characteristic mutation System, above warning strategies tends not to the responsibility of competent performance alarm, the 4th strategy solve the 1st, the 2nd, 3 warning strategies to mutation system The adjustment of system, but brings Dynamic Baseline data and calculates complexity, and baseline threshold arranges troublesome problem.
The problems referred to above are urgently to be resolved hurrily.
Content of the invention
For the complication system for operation characteristic mutation, existing alarm strategy tends not to the responsibility of competent performance alarm Or calculating complexity, arranging troublesome defect, the embodiment of the present invention provides the alarm of a kind of monitoring performance data similarity tolerance Method and system.
The present invention provides the alarm method that a kind of monitoring performance data similarity is measured, comprising:
Performance data according to this cycle builds this cycle time series by wavelet transform；
Cycle time series on corresponding described the cycle seasonal effect in time series of inquiry；
Judge whether described cycle time series and upper cycle seasonal effect in time series similitude exceed excursion, if it is not, Return and continue to build this cycle time series according to the performance data in this cycle by wavelet transform；
If so, warning information is triggered.
Preferably, described build this cycle time series according to the performance data in this cycle by wavelet transform before Including:
The performance data of current point in time is obtained from monitoring system and index performance data buffering area；
Inquire about this cycle to the History Performance Data of current point in time.
Preferably, whether described cycle time series of described judgement and upper cycle seasonal effect in time series similitude be beyond change Include before scope:
Specify section effective time needing to add up in object time sequence；
Extract the sequence all effective time in section described effective time with identical index；
The all effective time extracting described in acquisition the sequence similitude two-by-two between sequence；
Set similitude excursion according to described sequence similitude two-by-two.
Preferably, described sequence similitude two-by-two specifically includes:
Similarity standard difference and similitude mean value.
Preferably, described according to described sequence similitude two-by-two set similitude excursion also include:
Set excursion according to monitoring operation reserve.
The present invention also provides the warning system that a kind of monitoring performance data similarity is measured, comprising:
Build module, for building this cycle time series according to the performance data in this cycle by wavelet transform；
Enquiry module, is used for inquiring about cycle time series on corresponding described cycle seasonal effect in time series；
Whether determination module, be used for judging described cycle time series and upper cycle seasonal effect in time series similitude beyond change Change scope, continues to build this cycle time series according to the performance data in this cycle by wavelet transform if it is not, return；
Alarm module, is used for triggering warning information.
Preferably, described system is additionally operable to:
The performance data of current point in time is obtained from monitoring system and index performance data buffering area；
Inquire about this cycle to the History Performance Data of current point in time.
Preferably, described system is additionally operable to:
Specify section effective time needing to add up in object time sequence；
Extract the sequence all effective time in section described effective time with identical index；
The all effective time extracting described in acquisition the sequence similitude two-by-two between sequence；
Set similitude excursion according to described sequence similitude two-by-two.
Preferably, described sequence similitude two-by-two specifically includes:
Similarity standard difference and similitude mean value.
Preferably, described according to described sequence similitude two-by-two set similitude excursion also include:
Set excursion according to monitoring operation reserve.
Beneficial effect: time series data is represented that algorithm and similarity measurements quantity algorithm are incorporated into monitoring data and accuse by the present invention Alert field, enriches performance data warning strategies, different by the i.e. realizability energy time series data behavioral characteristics of simplicity configuration Normal alarm.Time Series Similarity strategy, on the basis of conventional measures, enriches system exception event detection means, fills out Mend deficiency in system dynamics monitoring for the conventional measures.
Brief description
The alarm method block diagram of the monitoring performance data similarity tolerance that Fig. 1 provides for the embodiment of the present invention；
The alarm method block diagram of the monitoring performance data similarity tolerance that Fig. 2 provides for another embodiment of the present invention；
The alarm method block diagram of the monitoring performance data similarity tolerance that Fig. 3 provides for another embodiment of the present invention；
The time-serial position of certain equipment cpu busy percentage of a week that Fig. 4 provides for another embodiment of the present invention；
The warning system structure chart of the monitoring performance data similarity tolerance that Fig. 5 provides for the embodiment of the present invention.
Detailed description of the invention
In order to make the purpose of the present invention, technical scheme and advantage clearer, below in conjunction with drawings and Examples, right The present invention is further elaborated.It should be appreciated that specific embodiment described herein only in order to explaining the present invention, and It is not used in the restriction present invention.
Monitoring system loading condition often possesses time cycle characteristic, and this feature makes a large amount of monitoring performance time series Data are changed by identical graphic feature in specific period, if performance time sequence and a upper periodic performance time in this cycle Sequence similarity exceed similitude excursion, then show performance time sequence within this cycle not according to intrinsic periodic characteristic Run, simultaneously also with regard to there being anomalous event to occur in this cycle of surface.
The present invention possesses time cycle characteristic according to monitoring system loading condition, as it is shown in figure 1, provide a kind of monitoring performance The alarm method of data similarity tolerance, comprising:
S100, build this cycle time series according to the performance data in this cycle by wavelet transform；
Concrete, similarity measurement is the foundation weighing two sequence similarities, main point of Time Series Similarity tolerance It is that two steps complete: (1) collects the feature measured from time series, is then expressed as certain form；(2) design formula Go to measure two sequences represent between distance.Existing character representation method mainly completes the first step and works, and second step master To be completed by distance metric method.Typical at present time series data character representation form mainly has discrete Fourier transform (DFT), singular value decomposition (SVD), wavelet transform (DWT), Piecewise Linear Representation (PLR, including PLA, PAA, APCA), symbol Number representation (SA) and landmark model (Landmarks) etc..In order to realize that first Similarity Measure needs to select suitable performance Data represent algorithm, the inconsistent feature of each index collection frequency of the time series according to performance data, select discrete wavelet transformer The graphic feature changing (DWT) algorithm to realize performance data extracts, and completes the structure to performance time sequence signature vector.
Have chosen standardized Haar wavelet algorithm, specific algorithm when being embodied as wavelet transform (DWT) algorithm It is described as follows:
If { x1,x2,x3,x4It is a time series.Define the average of it and details:
a 1 , 0 = ( x 1 + x 2 ) / 2 d 1 , 0 = ( x 1 - x 2 ) / 2
Here, a1,0It is original signal the first two value x1、x2Average.It is again low-frequency component, reflect the first two value x1、x2Base Eigen or coarse trend；d1,0Reflect x1、x2Difference, i.e. detailed information, be again radio-frequency component.
Have found x3、x4And a1,1、d1,1Relation.
Equally, a1,1It is latter two value x of original signal3、x4Average, d1,1Reflect x3、x4Details.{ a1,0,a1,1, d1,0,d1,1Regard as to { x1,x2,x3,x4Implement the result of linear transformation.
Conversion can also down be carried out:
a0,0=(a1,0+a1,1)/2
=((x1+x2)/2+(x3+x4)/2)/2
=(x1+x2+x3+x4)/4
a0,0It is that it is the most basic information of original signal to final average of 4 signal elements；d0,0=(a1,0-a1,1)/2。
Through quadratic transformation, obtain former seasonal effect in time series another kind and represented:
{a0,0,d0,0,d1,0,d1,1}
Here it is the time series characteristic vector after wavelet transform, need to select as the case may be in being embodied as Select step-length and number of transitions.
Further, wavelet transform (DWT) algorithm can sufficient expression time sequence in time series feature extraction The graphic feature of row, while reducing distance calculating latitude, remains time enough sequence pattern information.Discrete fourier Conversion (DFT), as the dimension reduction method proposing the earliest, implements also can reach good effect in this programme.
Cycle time series on S200, corresponding described the cycle seasonal effect in time series of inquiry；
S300, judge whether described cycle time series and upper cycle seasonal effect in time series similitude exceed excursion, Continue to build this cycle time series according to the performance data in this cycle by wavelet transform if it is not, return；
S400, if so, triggering warning information.
Preferably, as in figure 2 it is shown, described when building this cycle according to the performance data in this cycle by wavelet transform Between include before sequence:
S400, the performance data obtaining current point in time from monitoring system and index performance data buffering area；
S500, inquire about this cycle to the History Performance Data of current point in time.
Preferably, as it is shown on figure 3, described cycle time series of described judgement and upper cycle seasonal effect in time series similitude are No include beyond before excursion:
S600, appointment object time sequence need section effective time added up；
S700, the sequence data all effective time extracting same monitor control index in section described effective time；
Concrete, the performance time sequence data for same monitor control index is carried out by intrinsic frequency when gathering, and makes With can guarantee that time series data characteristic length is equal, therefore have selected calculate relatively simple again can triangle inequality Euclidean distance is as performance time sequence similarity measure.
1) Euclidean distance:
The seasonal effect in time series characteristic vector of a length of n is regarded as one of n dimension Euclidean space point, and its coordinate value divides It not the value in each moment for the time series, then the Euclidean distance between the time series of two a length of n is exactly n Distance between two points in dimension space.Its mathematical form is described as follows:
Given two time serieses X={x1, x2 ..., xn}, Y={y1, y2 ..., yn}, the euclidean between them Distance definition is:
D ( X , Y ) = Σ i = 1 n | x i - y i | 2 .
Its popularizing form is also referred to as Miknowksi distance, also referred to as Lp distance, and formula is:
As p=1, referred to as manhatton distance；
As p=∞, referred to as ultimate range.
The advantage of Euclidean distance is that calculating is simple, easy to understand, and meets distance triangle inequality, but does not support the time The linear drift of sequence and time shaft stretch；
The all effective time extracting described in acquisition the sequence similitude two-by-two between sequence；
S800, according to described sequence similitude two-by-two set similitude excursion.
Concrete, the similitude statistical information of performance time sequence, it is special that the change of discovery similitude also possesses normal distribution Levy, performance time sequence similarity excursion can be chosen easily based on this principle, and realize time series based on this scope The performance indications of over range are realized monitoring alarm process by the abnormal test of similitude.
Preferably, described sequence similitude two-by-two specifically includes:
Similarity standard difference and similitude mean value.
If the time series data of same index is as follows:
2016-07-11:t
2016-07-11:t2
2016-07-11:t3
2016-07-11:t4
2016-07-11:t5
The characteristic vector obtaining after using wavelet transformation is: p1, p2, p3, p4, p5.
The use calculated similarity measurement of Euclidean distance is:
D1,2, D1,3, DIsosorbide-5-Nitrae, D1,5, D2,3, D2,4, D2,5, D3,4, D3,5, D4,5。
Similarity measurement mean value:
u = D 1 , 2 + D 1 , 3 + D 1 , 4 + D 1 , 5 + D 2 , 3 + D 2 , 4 + D 2 , 5 + D 3 , 4 + D 3 , 5 + D 4 , 5 10 .
Similarity measures difference:
It by distance value again regular following table is:
D1, D2, D3, D4, D5, D6, D7, D8, D9, D10。
Then standard deviation is expressed as:
s = Σ t = 1 10 ( D t - u ) 2 9 2 .
Concrete, monitor system performance data center achieves based on above-mentioned wavelet transform and Euclidean distance calculating side The similitude index of performance time sequence data and inquiry, by the Similarity measures result of the same metric history data of inquiry, Target capabilities time series and history each performance time sequence similarity data can be got, thus obtain performance time sequence Similitude statistical information (similarity standard difference and similitude mean value).
Preferably, described according to described sequence similitude two-by-two set similitude excursion also include:
Set excursion according to monitoring operation reserve.
Actual motion also can simply set empirical value as excursion, in system is run, progressively adjust change Scope is arranged.
As shown in Figure 4, the time-serial position for certain equipment cpu busy percentage of a week.Twice is empirically set during beginning Standard deviation as normality threshold scope, in the face of the situation of the 5th daybreak explict occurrence operation exception, system does not provides warning information, On inspection, system has twice is the state of going offline, and does not collects cpu busy percentage data, the data that other times section collects Consistent with day-to-day operation, after checking this situation, threshold value is changed into one times of standard deviation, test uses fire wall to close target and sets Standby collection port 20 minutes, after reopening continuous 4 to 5 collection period of port, it is different that system sends cpu busy percentage behavioral characteristics Often alarm.
Time series data is represented that algorithm and similarity measurements quantity algorithm are incorporated into prison by the method that the embodiment of the present invention provides Control data alarm field, is enriched performance data warning strategies, is moved by simplicity configuration i.e. realizability energy time series data The alarm of state feature abnormalities.Time Series Similarity strategy, on the basis of conventional measures, enriches the inspection of system exception event Survey means, have filled up deficiency in system dynamics monitoring for the conventional measures.
The present invention also provides the warning system that a kind of monitoring performance data similarity is measured, as shown in Figure 5, comprising:
Build module 100, for building this cycle time of sequence according to the performance data in this cycle by wavelet transform Row；
Enquiry module 200, is used for inquiring about cycle time series on corresponding described cycle seasonal effect in time series；
Determination module 300, is used for judging whether described cycle time series and upper cycle seasonal effect in time series similitude surpass Go out excursion, continue to build this cycle time of sequence according to the performance data in this cycle by wavelet transform if it is not, return Row；
Alarm module 400, is used for triggering warning information.
Preferably, described system is additionally operable to:
The performance data of current point in time is obtained from monitoring system and index performance data buffering area；
Inquire about this cycle to the History Performance Data of current point in time.
Preferably, described system is additionally operable to:
Specify section effective time needing to add up in object time sequence；
Extract the sequence all effective time in section described effective time with identical index；
The all effective time extracting described in acquisition the sequence similitude two-by-two between sequence；
Set similitude excursion according to described sequence similitude two-by-two.
Preferably, described sequence similitude two-by-two specifically includes:
Similarity standard difference and similitude mean value.
Preferably, described according to described sequence similitude two-by-two set similitude excursion also include:
Set excursion according to monitoring operation reserve.
It should be noted that modules in the said system of embodiment of the present invention offer, due to the inventive method in fact Executing example based on same design, its technique effect bringing is identical with the inventive method embodiment, and particular content can be found in the present invention Narration in embodiment of the method, here is omitted.
Time series data is represented that algorithm and similarity measurements quantity algorithm are incorporated into prison by the system that the embodiment of the present invention provides Control data alarm field, is enriched performance data warning strategies, is moved by simplicity configuration i.e. realizability energy time series data The alarm of state feature abnormalities.Time Series Similarity strategy, on the basis of conventional measures, enriches the inspection of system exception event Survey means, have filled up deficiency in system dynamics monitoring for the conventional measures.
The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all essences in the present invention Any modification, equivalent and the improvement etc. made within god and principle, should be included within the scope of the present invention.

Claims (10)

1. the alarm method of a monitoring performance data similarity tolerance, it is characterised in that include:
Performance data according to this cycle builds this cycle time series by wavelet transform；
Cycle time series on corresponding described the cycle seasonal effect in time series of inquiry；
Judge whether described cycle time series and upper cycle seasonal effect in time series similitude exceed excursion, if it is not, return Continue to build this cycle time series according to the performance data in this cycle by wavelet transform；
If so, warning information is triggered.
2. the method for claim 1, it is characterised in that described pass through discrete wavelet transformer according to the performance data in this cycle Include before changing structure this cycle time series:
The performance data of current point in time is obtained from monitoring system and index performance data buffering area；
Inquire about this cycle to the History Performance Data of current point in time.
3. method as claimed in claim 2, it is characterised in that described cycle time series of described judgement and upper cycle time Whether the similitude of sequence includes beyond before excursion:
Specify section effective time needing to add up in object time sequence；
Extract the sequence all effective time in section described effective time with identical index；
The all effective time extracting described in acquisition the sequence similitude two-by-two between sequence；
Set similitude excursion according to described sequence similitude two-by-two.
4. method as claimed in claim 3, it is characterised in that described sequence similitude two-by-two specifically includes:
Similarity standard difference and similitude mean value.
5. method as claimed in claim 3, it is characterised in that described according to the setting similitude change of described sequence similitude two-by-two Change scope also includes:
Set excursion according to monitoring operation reserve.
6. the warning system of a monitoring performance data similarity tolerance, it is characterised in that include:
Build module, for building this cycle time series according to the performance data in this cycle by wavelet transform；
Enquiry module, is used for inquiring about cycle time series on corresponding described cycle seasonal effect in time series；
Whether determination module, be used for judging described cycle time series and upper cycle seasonal effect in time series similitude beyond change model Enclose, continue to build this cycle time series according to the performance data in this cycle by wavelet transform if it is not, return；
Alarm module, is used for triggering warning information.
7. system as claimed in claim 6, it is characterised in that described system is additionally operable to:
The performance data of current point in time is obtained from monitoring system and index performance data buffering area；
Inquire about this cycle to the History Performance Data of current point in time.
8. system as claimed in claim 7, it is characterised in that described system is additionally operable to:
Specify section effective time needing to add up in object time sequence；
Extract the sequence all effective time in section described effective time with identical index；
The all effective time extracting described in acquisition the sequence similitude two-by-two between sequence；
Set similitude excursion according to described sequence similitude two-by-two.
9. system as claimed in claim 8, it is characterised in that described sequence similitude two-by-two specifically includes:
Similarity standard difference and similitude mean value.
10. system as claimed in claim 8, it is characterised in that described according to described sequence similitude two-by-two setting similitude Excursion also includes:
Set excursion according to monitoring operation reserve.
CN201610707419.XA 2016-08-23 2016-08-23 The alarm method of monitoring performance data similarity tolerance and system Pending CN106155867A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610707419.XA CN106155867A (en) 2016-08-23 2016-08-23 The alarm method of monitoring performance data similarity tolerance and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610707419.XA CN106155867A (en) 2016-08-23 2016-08-23 The alarm method of monitoring performance data similarity tolerance and system

Publications (1)

Publication Number Publication Date
CN106155867A true CN106155867A (en) 2016-11-23

Family

ID=57341727

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610707419.XA Pending CN106155867A (en) 2016-08-23 2016-08-23 The alarm method of monitoring performance data similarity tolerance and system

Country Status (1)

Country Link
CN (1) CN106155867A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109614148A (en) * 2018-12-11 2019-04-12 中科驭数（北京）科技有限公司 Mathematical logic operation method, monitoring method and device
CN110337816A (en) * 2017-03-02 2019-10-15 三菱电机株式会社 Monitoring system
CN110647647A (en) * 2019-09-03 2020-01-03 西安外事学院 Closed graph similarity searching method based on time sequence complexity difference
CN111104438A (en) * 2019-11-21 2020-05-05 新浪网技术(中国)有限公司 Method and device for determining periodicity of time sequence and electronic equipment
CN115827384A (en) * 2022-12-08 2023-03-21 国网湖北省电力有限公司信息通信公司 Optimized real-time data center network equipment monitoring method

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102129525A (en) * 2011-03-24 2011-07-20 华北电力大学 Method for searching and analyzing abnormality of signals during vibration and process of steam turbine set
CN104317681A (en) * 2014-09-02 2015-01-28 上海交通大学 Behavioral abnormality automatic detection method and behavioral abnormality automatic detection system aiming at computer system
CN104636999A (en) * 2015-01-04 2015-05-20 江苏联宏自动化系统工程有限公司 Detection method for building abnormal energy consumption data
US20150169654A1 (en) * 2013-12-13 2015-06-18 International Business Machines Corporation Managing time series databases
CN104933175A (en) * 2015-06-30 2015-09-23 深圳市金证科技股份有限公司 Performance data dependency analyzing method and performance monitoring system
CN105205568A (en) * 2015-10-14 2015-12-30 广东电网有限责任公司电力调度控制中心 Alarm transaction extracting method and system

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102129525A (en) * 2011-03-24 2011-07-20 华北电力大学 Method for searching and analyzing abnormality of signals during vibration and process of steam turbine set
US20150169654A1 (en) * 2013-12-13 2015-06-18 International Business Machines Corporation Managing time series databases
CN104317681A (en) * 2014-09-02 2015-01-28 上海交通大学 Behavioral abnormality automatic detection method and behavioral abnormality automatic detection system aiming at computer system
CN104636999A (en) * 2015-01-04 2015-05-20 江苏联宏自动化系统工程有限公司 Detection method for building abnormal energy consumption data
CN104933175A (en) * 2015-06-30 2015-09-23 深圳市金证科技股份有限公司 Performance data dependency analyzing method and performance monitoring system
CN105205568A (en) * 2015-10-14 2015-12-30 广东电网有限责任公司电力调度控制中心 Alarm transaction extracting method and system

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
应海松: "《铁矿石品质波动评估中时间序列预测的应用》", 30 November 2014 *
张海勤 等: "基于小波变换的时间序列相似模式匹配", 《计算机学报》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110337816A (en) * 2017-03-02 2019-10-15 三菱电机株式会社 Monitoring system
CN110337816B (en) * 2017-03-02 2022-02-25 三菱电机株式会社 Monitoring system
CN109614148A (en) * 2018-12-11 2019-04-12 中科驭数（北京）科技有限公司 Mathematical logic operation method, monitoring method and device
CN110647647A (en) * 2019-09-03 2020-01-03 西安外事学院 Closed graph similarity searching method based on time sequence complexity difference
CN110647647B (en) * 2019-09-03 2022-02-08 西安外事学院 Closed graph similarity searching method based on time sequence complexity difference
CN111104438A (en) * 2019-11-21 2020-05-05 新浪网技术(中国)有限公司 Method and device for determining periodicity of time sequence and electronic equipment
CN115827384A (en) * 2022-12-08 2023-03-21 国网湖北省电力有限公司信息通信公司 Optimized real-time data center network equipment monitoring method
CN115827384B (en) * 2022-12-08 2024-02-02 国网湖北省电力有限公司信息通信公司 Monitoring method for network equipment of optimized real-time data center

Similar Documents

Publication Publication Date Title
CN106155867A (en) The alarm method of monitoring performance data similarity tolerance and system
US20160116892A1 (en) Method and system of cause analysis and correction for manufacturing data
WO2008157505A1 (en) Remote monitoring systems and methods
JP2005534192A5 (en)
You et al. Disturbance location determination based on electromechanical wave propagation in FNET/GridEye: a distribution‐level wide‐area measurement system
CN106599487B (en) The dynamic variance Realtime Alerts method regulated and controled based on ARIMA model and user
WO2020008760A1 (en) Abnormality detection system
AU2012209207A1 (en) Method for analyzing and diagnosing large scale process automation control systems
CN104931040B (en) The Big Dipper II generation navigation system electric power tower deformation monitoring equipment installation based on machine learning and adjustment method
CN106681183A (en) Method, apparatus, and system for monitoring manufacturing equipment and computer readable storage media
CN114611778B (en) Reservoir water level early warning method and system based on warehousing flow
CN114450646B (en) System and method for detecting wind turbine operational anomalies using deep learning
CN106663086A (en) Apparatus and method for ensembles of kernel regression models
CN105793789A (en) Computer-implemented method and system for automatically monitoring and determining status of entire process segments in process unit
CN108804703A (en) A kind of data exception detection method and device
US20080255773A1 (en) Machine condition monitoring using pattern rules
CN112069464A (en) Building curtain wall engineering acceptance detection management system based on big data
CN117371339B (en) Transformer operation monitoring system based on Internet of things
Yang et al. Continuous tracking of bridge modal parameters based on subspace correlations
CN109213057A (en) Intelligent diagnostics device and method
CN101251576A (en) Circuit failure diagnosis method based on neural network
CN106598026B (en) Power transmission and transformation equipment state overhauling time detection method and device
KR102582180B1 (en) Small river smart prediction management system using accumulated data and its method
CN117290685A (en) Power plant power equipment expert diagnosis system and method based on historical data
Mousavi et al. Simulation-based real-time performance monitoring (simmon): A platform for manufacturing and healthcare systems

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20161123

RJ01 Rejection of invention patent application after publication