CN113592308B - Monitoring data alarm threshold extraction method based on normal model - Google Patents
Monitoring data alarm threshold extraction method based on normal model Download PDFInfo
- Publication number
- CN113592308B CN113592308B CN202110881132.XA CN202110881132A CN113592308B CN 113592308 B CN113592308 B CN 113592308B CN 202110881132 A CN202110881132 A CN 202110881132A CN 113592308 B CN113592308 B CN 113592308B
- Authority
- CN
- China
- Prior art keywords
- monitoring data
- probability
- periodic
- distribution
- extrapolation
- 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.)
- Active
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 104
- 238000000605 extraction Methods 0.000 title claims abstract description 8
- 238000009826 distribution Methods 0.000 claims abstract description 76
- 230000000737 periodic effect Effects 0.000 claims abstract description 76
- 238000000034 method Methods 0.000 claims abstract description 50
- 230000001186 cumulative effect Effects 0.000 claims abstract description 42
- 238000013213 extrapolation Methods 0.000 claims abstract description 30
- 230000007774 longterm Effects 0.000 claims abstract description 27
- 230000008859 change Effects 0.000 claims abstract description 19
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 14
- 238000010183 spectrum analysis Methods 0.000 claims abstract description 10
- 238000007689 inspection Methods 0.000 claims abstract description 5
- 238000000926 separation method Methods 0.000 claims abstract description 4
- 238000012360 testing method Methods 0.000 claims description 9
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 230000008030 elimination Effects 0.000 claims description 2
- 238000003379 elimination reaction Methods 0.000 claims description 2
- 230000007613 environmental effect Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000004321 preservation Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Educational Administration (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a monitoring data alarm threshold extraction method based on a normal state model, which comprises the following steps: a significant periodic set in the monitoring data is obtained by using a spectrum analysis method, and a periodic component set is separated by using a time sequence decomposition method; fitting the residual monitoring data after the separation period component set into a long-term trend by using a straight line fitting method; fitting the first monitoring data and the significant periodic set by adopting a corresponding cumulative probability fitting function, carrying out K-S goodness-of-fit inspection on the cumulative probability distribution model, and obtaining an optimal cumulative probability distribution model based on an inspection result; and respectively constructing an extrapolation probability interval of the monitoring data according to the obtained monitoring data, the period component set and the minimum probability value of the long-term trend, selecting a small probability value of the extrapolation interval, and constructing an alarm threshold value based on the small probability value. The method can accurately alarm the long-term change trend and change period of the monitored data of the cultural relics.
Description
Technical Field
The invention belongs to the field of data mining, and particularly relates to a monitoring data alarm threshold extraction method based on a normal model.
Background
The immovable cultural relics refer to cultural relic entities which have important scientific values in history, culture, building and art and actually exist, include traditional historical historic and historic sites such as ancient buildings and settlement, and cover many aspects such as politics, military affairs, production activities, sacrifice, art, society, building and education.
The protection of the immovable cultural relics refers to a process of preventing the cultural relics with important historical, artistic and scientific values from being damaged by adopting various means, in recent years, China develops a large amount of targeted research and implementation work, obtains a certain result in the aspect of protection of the immovable cultural relics, and obviously restrains and relieves the weathering conditions of a batch of important cultural relics.
However, immovable cultural relics are influenced most by the environment, gradual weathering caused by natural and artificial environmental factors is still unconsciously carried out, and the damage range and the damage speed are continuously increased along with the current environmental change and the aggravation of pollution, so that the immovable cultural relics are a crucial factor influencing the safe preservation of the immovable cultural relics. Most immovable cultural relics have diseases, and some cultural relics are even in endangered states. Therefore, the efficiency of the whole cultural relic protection system is improved by improving the technological content of the immovable cultural relic protection means, and the problem which needs to be solved is now urgent.
In the traditional cultural relic protection, as a channel for collecting cultural relic information in real time is lacked, the related information of the comprehensive and sufficient cultural relics cannot be obtained, and the problems of fuzzy relation between the cultural relic entity and the preservation environment, untimely response to abnormal conditions and the like are caused. In addition, because many cultural relics are located in a remote geographical location and a severe environment, the existing scientific and technological means cannot be completely applied to cultural relics protection under the condition of lacking infrastructure such as electric power, network and the like.
In recent years, with the development of the internet of things, the monitoring work of immovable cultural relics is increasing, and the monitoring work becomes an important means for analyzing the preservation status of the cultural relics and evaluating risk factors. At present, domestic unmovable cultural relic monitoring practice has made more breakthroughs in data acquisition and storage, but a short board still exists in systematization and rationalization of data utilization. Because research foundations such as quantitative traceability prediction of a degradation mechanism, preventive protection of a universal environmental parameter threshold and the like are not mature, people do not know which abnormal data are in the face of massive sensor monitoring data, the data are difficult to give play to guidance, and the timeliness of current protection practice is assisted, so that research on a sensor monitoring data real-time grading alarm system construction method based on normal conditions is urgently needed to be developed in the face of practical application.
An existing cultural relic sensor monitoring data alarm threshold extraction method is a threshold determination and alarm system construction method based on combination of a normal distribution 3 sigma criterion and an extrapolation method. This approach has two problems: firstly, the method is established under the assumption that the sensor monitoring data obeys normal distribution under the normal condition, however, in reality, most of the sensor monitoring data do not accord with the normal distribution; and secondly, only alarming is carried out aiming at the monitoring data of the sensor, and deep analysis and alarming are not carried out on the long-term trend and the change period of the sensor.
Disclosure of Invention
The invention provides a monitoring data alarm threshold extraction method based on a normal model, which can accurately alarm the long-term change trend and change period of the monitoring data of cultural relics.
A monitoring data alarm threshold extraction method based on a normal state model comprises the following steps:
s1: a significant periodic set in the first monitoring data is obtained by using a spectrum analysis method, and a periodic component set is separated from the first monitoring data by using a time sequence decomposition method based on the significant periodic set;
s2: fitting the residual monitoring data after the separation period component set in the step S1 into a long-term trend by using a straight line fitting method;
s3: fitting the first monitoring data and the period component set by adopting a corresponding cumulative probability fitting function based on various cumulative probability distributions, obtaining various cumulative probability distribution models based on a fitting result, carrying out K-S fitting goodness test on the cumulative probability distribution models, and obtaining an optimal cumulative probability distribution model based on a test result;
s4: the method comprises the steps of obtaining the minimum probability values of first monitoring data and a periodic component set based on an optimal cumulative probability distribution model, respectively constructing a first extrapolation probability interval of the first monitoring data and a second extrapolation probability interval of the periodic component set according to the minimum probability values of the first monitoring data and the periodic component set, constructing a third extrapolation probability interval of a long-term trend based on a long-term trend slope, respectively selecting first, second and third small probability values of the first, second and third extrapolation intervals as required, and constructing an alarm threshold based on the first, second and third small probability values.
Firstly, a spectrum analysis method is utilized to obtain a significant periodic set in monitoring data, namely a change period of the monitoring data. And then decomposing the monitoring data by using a time series decomposition method according to the significant periodic set to obtain a periodic component set and a long-term trend. And then constructing an optimal cumulative probability distribution model based on the obtained periodic component set, long-term trend and monitoring data, and obtaining an alarm threshold value by extrapolating a probability interval. Based on the alarm threshold and the optimal cumulative probability distribution model, the long-term trend and the change period of the monitored data of the cultural relics can be accurately alarmed.
The method for obtaining the significant periodic set in the first monitoring data by using the spectrum analysis method comprises the following steps:
fourier transform is carried out on the obtained first monitoring data by using a spectrum analysis method to obtain first to Nth significance periods which are arranged from short to long time, and the first to Nth significance periods construct a significance period set.
Before Fourier transformation is carried out on the first monitoring data, a part of significance period is determined in advance based on priori knowledge, the priori knowledge is temperature, humidity, tourist visiting behavior, scenic spot opening behavior, and monitoring data influence factors with stable change period are arranged in a production plan of a production line, the stable change period is used as a part of significance period of the first monitoring data, and subsequent Fourier transformation can verify results obtained by the priori knowledge so as to achieve the purpose of quickly and accurately separating the first monitoring data to obtain a significance period set.
The separating the periodic component set from the first monitoring data comprises:
a moving average method in a time series method is adopted, a plurality of periodic components are separated from first monitoring data based on a significant periodic set, the periodic components are denoised by a Gaussian noise averaging elimination method, and a periodic component set is constructed by the obtained denoising periodic components.
The separating the plurality of periodic components from the first monitoring data includes:
s11: separating a first periodic component from the first monitoring data by using a time series decomposition method by taking the first significant period as a unit;
s12: removing the first periodic component from the first monitoring data to obtain second monitoring data, taking the second saliency period as a unit for the second monitoring data, and separating the second periodic component from the second monitoring data by using a time series decomposition method again;
s13: sequentially performing the steps according to the step S12 until an N-1 periodic component is removed from the N-1 monitoring data to obtain an N monitoring data, taking the N significant period as a unit of the N monitoring data, and separating the N periodic component from the N monitoring data by using a time sequence decomposition method again; the plurality of periodic components are the first to nth periodic components.
Fitting the first monitoring data and the first periodic component by utilizing a plurality of cumulative probability distributions to obtain parameters of the plurality of cumulative probability distributions, and constructing a plurality of cumulative probability distribution models through the parameters, wherein the plurality of cumulative probability distributions comprise: normal distribution, Gumbel distribution, Weibull distribution, Frechet distribution.
The obtaining of the optimal cumulative probability distribution model based on the test result comprises:
and performing K-S goodness-of-fit inspection on the plurality of cumulative probability distribution models to obtain the probability that the cumulative probability distribution models are similar to the fitted data, and taking the cumulative probability distribution model with the maximum similar probability as the optimal cumulative probability distribution model.
The first extrapolation probability interval is as follows:
the second extrapolation probability interval is as follows:
the third extrapolation probability interval is as follows:
wherein the content of the first and second substances,
is the minimum probability value of the first monitored data,
is the first of a periodic component setiThe minimum probability value of the periodic component,
is the slope of the long-term trend.
The first extrapolation probability interval is used for determining the upper limit distribution probability and the lower limit distribution probability of the first monitoring data distribution, and aims to select a first small probability value in a range outside the upper limit distribution probability and the lower limit distribution probability, regard the condition exceeding the first small probability value as abnormal and early warn the first monitoring data; the second extrapolation probability interval is used for determining the upper limit value distribution probability of each period component, selecting a second small probability value from a range outside the upper limit value distribution probability and early warning the change period; and the third extrapolation probability interval is used for determining the change upper limit distribution probability of the long-term trend, selecting a third small probability value from a range outside the upper limit distribution probability and early warning the long-term trend.
The alarm threshold value is calculated and constructed based on the first, second and third small probability values, and the method comprises the following steps:
and calculating first, second and third parameter values of the optimal cumulative probability distribution model corresponding to the first, second and third small probability values, and constructing an alarm threshold value by the first, second and third parameter values.
Compared with the prior art, the invention has the following beneficial effects:
1. a probability fitting method is used for obtaining a normal model so as to obtain an alarm threshold value, a proper distribution model can be selected according to actual normal distribution of sensor monitoring data/periodic components/long-term trends, and the alarm threshold value can be adjusted according to different given probabilities.
2. The time series decomposition method is used for decomposing the sensor monitoring data, so that the sensor monitoring data can be alarmed, and the long-term change trend and the periodic change fluctuation of the sensor monitoring data can be alarmed.
3. The method is suitable for monitoring data of various sensors and has strong mobility.
Drawings
Fig. 1 is a flowchart of a method for extracting alarm threshold of monitoring data based on a normal state model according to an embodiment of the present invention;
FIG. 2 is a flowchart of a method for obtaining a periodic component according to an embodiment of the present invention;
FIG. 3 is a flowchart illustrating a process of denoising periodic components according to an embodiment of the present invention;
FIG. 4 is a flow chart of selecting a probability distribution model according to an embodiment of the present invention;
FIG. 5 is a diagram of a spectral analysis provided by an embodiment of the present invention;
FIG. 6 is a decomposition composite picture provided by an embodiment of the present invention;
fig. 7 is a probability fit picture provided by the embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the detailed description and specific examples, while indicating the scope of the invention, are intended for purposes of illustration only and are not intended to limit the scope of the invention.
As shown in fig. 1, the method for extracting alarm threshold of sensor monitoring data based on normal state model according to the embodiment of the present invention includes the following steps:
first, a periodic component is separated, as shown in fig. 2, and the specific steps are as follows:
the method specifically uses a spectrum analysis method to find out the significant period in the sensor monitoring data. According to the prior knowledge, namely the daily temperature change rule, the change rule of the first monitoring data with the daily period can be predetermined, the frequency of the first monitoring data is one day, and the position of a certain period close to one month has obvious peaks, and further analysis results show that the 26-day period has the most significance, so that a significance period set with the day and the 26 days is obtained. Separating the significant periodic factors of the sensor monitoring data by using a time sequence decomposition method, decomposing the sensor monitoring data by taking a day as a first significant period, and separating a day period component; and decomposing the remaining data of the separation day period again by taking 26 days as a second significance period, and separating out 26-day period components, wherein the day period components, the 26-day period components and the long-term trend decomposition comprehensive graph are shown in fig. 6, and the influence of different period components on the sensor monitoring data can be more clearly seen from the graph, and the change trend of the sensor monitoring data can be reflected.
Meanwhile, the noise of the periodic component is removed, and as shown in fig. 3, the sensor monitoring data is divided into two parts by using a moving average method according to the period of significance, wherein one part comprises the periodic component and the noise component, and the other part is the residual data. Then, removing the noise component of the part containing the periodic component and the noise component by using a Gaussian noise averaging and eliminating method, namely separating a denoised daily periodic component and a 26-day periodic component, and constructing a periodic component set by the denoised daily periodic component and the 26-day periodic component;
then, probability fitting is performed to obtain a normal state model, as shown in fig. 4, the specific steps are as follows:
and fitting all items of the normal model of the first monitoring data and the periodic component set by using various cumulative probability distribution models, and carrying out K-S goodness-of-fit inspection and comparison on the fitting model. It can be seen from fig. 7 that the fitting curves of the sensor monitoring data in the normal distribution, the Gumbel distribution and the Weibull distribution models are shown, the K-S goodness-of-fit test result is shown in table 1, and the probability that the Gumbel distribution passes the K-S goodness-of-fit test is the highest, so that the Gumbel distribution model is selected as the optimal cumulative probability distribution model of the sensor monitoring data.
TABLE 1 test results of goodness of K-S fit
Respectively constructing a first extrapolation probability interval of the monitoring data, a second extrapolation probability interval of the periodic component set and a third extrapolation probability interval of the long-term trend according to the monitoring data, the periodic component set and the minimum probability value of the long-term trend obtained by the optimal cumulative probability distribution model, and respectively selecting first, second and third small probability values of the first, second and third extrapolation intervals according to needs, wherein the first extrapolation probability interval is as follows:
the second extrapolation probability interval is as follows:
the third extrapolation probability interval is as follows:
wherein the content of the first and second substances,
is the minimum probability value of the first monitored data,
is the first of a set of periodic componentsiThe minimum probability value of the periodic component,
is the slope of the long-term trend.
And finally, calculating the optimal cumulative probability distribution model parameter value under the condition of small probability. And respectively calculating an early warning threshold value of the sensor monitoring data under the probability of 0.0001 and an alarm threshold value under the probability of 0.00005 according to the optimal cumulative probability distribution model parameter values, and referring to table 2. Similarly, early warning/alarm thresholds for periodic components and long-term trends at a given probability can be derived.
TABLE 20.0001 probability and 0.00005 probability early warning/alarm thresholds for sensor monitoring data
Claims (4)
1. A monitoring data alarm threshold value extraction method based on a normal state model is characterized by comprising the following steps:
s1: a significant periodic set in first monitoring data is obtained by using a spectrum analysis method, the first monitoring data are sensor monitoring data, and a periodic component set is separated from the first monitoring data by using a time sequence decomposition method based on the significant periodic set;
the method for obtaining the significant periodic set in the first monitoring data by using the spectrum analysis method comprises the following steps:
carrying out Fourier transform on the obtained first monitoring data by using a spectrum analysis method to obtain first to Nth significance cycles which are ordered from short to long, wherein the first to Nth significance cycles construct a significance cycle set;
the separating the periodic component set from the first monitoring data comprises:
separating a plurality of periodic components from the first monitoring data by adopting a moving average method of time series data based on a significant periodic set, denoising the plurality of periodic components by a Gaussian noise averaging elimination method, and constructing a periodic component set by the obtained plurality of denoised periodic components; wherein said separating the plurality of periodic components from the first monitored data comprises:
s11: separating a first periodic component from the first monitoring data by using a time series decomposition method by taking the first significant period as a unit;
s12: removing the first periodic component from the first monitoring data to obtain second monitoring data, taking the second saliency period as a unit for the second monitoring data, and separating the second periodic component from the second monitoring data by using a time series decomposition method again;
s13: sequentially performing the steps according to the step S12 until an N-1 periodic component is removed from the N-1 monitoring data to obtain an N monitoring data, taking the N significant period as a unit of the N monitoring data, and separating the N periodic component from the N monitoring data by using a time sequence decomposition method again; the plurality of periodic components are the first to Nth periodic components;
s2: fitting the remaining monitoring data of the separation periodic component set in the step S1 to a long-term trend by using a straight line fitting method;
s3: fitting the first monitoring data and the period component set by adopting a corresponding cumulative probability fitting function based on various cumulative probability distributions, obtaining various cumulative probability distribution models based on a fitting result, carrying out K-S fitting goodness test on the cumulative probability distribution models, and obtaining an optimal cumulative probability distribution model based on a test result;
s4: based on the optimal cumulative probability distribution model, obtaining the minimum probability values of first monitoring data and a periodic component set, respectively constructing a first extrapolation probability interval of the first monitoring data according to the minimum probability values of the first monitoring data and the periodic component set, constructing a second extrapolation probability interval of the periodic component set, constructing a third extrapolation probability interval of a long-term trend based on a long-term trend slope, respectively selecting the first, second and third small probability values of the first, second and third extrapolation intervals as required, calculating the parameter values of the optimal cumulative probability distribution model corresponding to the first, second and third small probability values, and constructing an alarm threshold according to the parameter values of the optimal cumulative probability distribution model;
the first extrapolation probability interval is used for determining the upper limit distribution probability and the lower limit distribution probability of the first monitoring data distribution, and aims to select a first small probability value in a range outside the upper limit distribution probability and the lower limit distribution probability, regard the condition exceeding the first small probability value as abnormal and early warn the first monitoring data; the second extrapolation probability interval is used for determining the upper limit value distribution probability of each period component, selecting a second small probability value from the range outside the upper limit value distribution probability and carrying out early warning on the change period; and the third extrapolation probability interval is used for determining the change upper limit distribution probability of the long-term trend, selecting a third small probability value from a range outside the upper limit distribution probability and early warning the long-term trend.
2. The method as claimed in claim 1, wherein the method for extracting the alarm threshold of the monitoring data based on the normal state model is characterized in that a plurality of cumulative probability distributions are used to fit the periodic component set of the first monitoring data to obtain parameters of the plurality of cumulative probability distributions, and the plurality of cumulative probability distributions are constructed by the parameters, and include: normal distribution, Gumbel distribution, Weibull distribution, Frechet distribution.
3. The method for extracting a threshold of alarm for monitoring data based on a normal state model according to claim 1 or 2, wherein the obtaining of the optimal cumulative probability distribution model based on the test result comprises:
and performing K-S goodness-of-fit inspection on the plurality of cumulative probability distribution models to obtain the probability that the cumulative probability distribution models are similar to the fitted data, and taking the cumulative probability distribution model with the maximum similar probability as the optimal cumulative probability distribution model.
4. The method for extracting alarm threshold of monitoring data based on normal state model as claimed in claim 1, wherein the first extrapolation probability interval is:
the second extrapolation probability interval is as follows:
the third extrapolation probability interval is as follows:
wherein the content of the first and second substances,
is the minimum probability value of the first monitored data,
is the first of a set of periodic componentsiThe minimum probability value of the periodic component,
is the slope of the long-term trend.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110881132.XA CN113592308B (en) | 2021-08-02 | 2021-08-02 | Monitoring data alarm threshold extraction method based on normal model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110881132.XA CN113592308B (en) | 2021-08-02 | 2021-08-02 | Monitoring data alarm threshold extraction method based on normal model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113592308A CN113592308A (en) | 2021-11-02 |
| CN113592308B true CN113592308B (en) | 2022-07-12 |
Family
ID=78253736
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110881132.XA Active CN113592308B (en) | 2021-08-02 | 2021-08-02 | Monitoring data alarm threshold extraction method based on normal model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113592308B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115098561B (en) * | 2022-06-28 | 2024-03-26 | 浙江大学 | Method for extracting alarm threshold of cultural relic ontology monitoring data |
| CN116304996B (en) * | 2023-04-26 | 2024-03-22 | 浙江大学 | Method and device for extracting body trend of stone relic monitoring data |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110381083A (en) * | 2019-08-07 | 2019-10-25 | 浙江双成电气有限公司 | A kind of smart grid communication abnormality detection method based on time series |
| CN111148045A (en) * | 2019-12-30 | 2020-05-12 | 中国联合网络通信集团有限公司 | User behavior cycle extraction method and device |
| CN111508216A (en) * | 2020-04-28 | 2020-08-07 | 水利部交通运输部国家能源局南京水利科学研究院 | Intelligent early warning method for dam safety monitoring data |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| GB0902730D0 (en) * | 2009-02-18 | 2009-04-01 | Oxford Biosignals Ltd | Method and apparatus for monitoring and analyzing vibrations in rotary machines |
| CN103745272B (en) * | 2014-01-06 | 2017-02-08 | 国家电网公司 | Power short-term load predicating method based on fast periodic component extraction |
| EP3432177B1 (en) * | 2017-07-17 | 2023-04-26 | Roche Diagnostics GmbH | Method and device for analyzing a dataset |
| CN110889554A (en) * | 2019-11-27 | 2020-03-17 | 东南大学 | Power load fluctuation analysis and risk early warning method based on recurrence time interval analysis method |
| CN111120348A (en) * | 2019-12-25 | 2020-05-08 | 中国石化销售股份有限公司华南分公司 | Centrifugal pump fault early warning method based on support vector machine probability density estimation |
| CN112464169A (en) * | 2020-12-03 | 2021-03-09 | 上海卫星工程研究所 | Big data single-parameter threshold mining and correcting method and system based on statistical distribution |
| CN112819293B (en) * | 2021-01-14 | 2023-01-06 | 中国长江三峡集团有限公司 | Failure early warning analysis method for water reservoir scheduling rule under influence of climate change |
-
2021
- 2021-08-02 CN CN202110881132.XA patent/CN113592308B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110381083A (en) * | 2019-08-07 | 2019-10-25 | 浙江双成电气有限公司 | A kind of smart grid communication abnormality detection method based on time series |
| CN111148045A (en) * | 2019-12-30 | 2020-05-12 | 中国联合网络通信集团有限公司 | User behavior cycle extraction method and device |
| CN111508216A (en) * | 2020-04-28 | 2020-08-07 | 水利部交通运输部国家能源局南京水利科学研究院 | Intelligent early warning method for dam safety monitoring data |
Non-Patent Citations (2)
| Title |
|---|
| 基于测点数据变化趋势告警判定的大型立磨故障诊断专家系统;杜文辽等;《矿山机械》;20170510(第05期);全文 * |
| 预应力混凝土连续梁桥应变监测预警阈值研究;包龙生等;《中外公路》;20180102(第06期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113592308A (en) | 2021-11-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113592308B (en) | Monitoring data alarm threshold extraction method based on normal model | |
| CN115578015B (en) | Sewage treatment whole process supervision method, system and storage medium based on Internet of things | |
| CN109977916B (en) | Online monitoring system for dissolved gas in transformer oil based on embedded platform | |
| CN106873571B (en) | early warning method based on data and model fusion | |
| CN110889088A (en) | Enterprise pollution discharge supervision method assisted by electric model | |
| CN110260774B (en) | GNSS deformation information inspection and early warning method based on Pettitt algorithm | |
| CN111367959B (en) | Zero-time-lag nonlinear expansion Granger causal analysis method | |
| CN111881594B (en) | Non-stationary signal state monitoring method and system for nuclear power equipment | |
| CN115499185A (en) | Method and system for analyzing abnormal behavior of network security object of power monitoring system | |
| CN112633781A (en) | Vehicle energy consumption evaluation method based on Internet of vehicles big data | |
| CN115470973A (en) | ARIMA-based enterprise daily electricity quantity abnormity detection method | |
| CN116359285A (en) | Oil gas concentration intelligent detection system and method based on big data | |
| CN114330120A (en) | 24-hour PM prediction based on deep neural network2.5Method of concentration | |
| CN107147143B (en) | Method for establishing early warning model of fan interlocking off-line fault | |
| CN117034197A (en) | Enterprise power consumption typical mode analysis method based on multidimensional Isolate-detection multi-point detection | |
| CN116862109A (en) | Regional carbon emission situation awareness early warning method | |
| CN114943189B (en) | XGboost-based acoustic velocity profile inversion method and system | |
| CN116910655A (en) | Intelligent ammeter fault prediction method based on device measurement data | |
| CN116756825A (en) | Group structural performance prediction system for middle-small span bridge | |
| CN113821895B (en) | Method and device for constructing power transmission line icing thickness prediction model and storage medium | |
| CN110208876B (en) | Characterization method for cooperative change of radial positions of subtropical zone torrent and polar front torrent | |
| CN113408383A (en) | Audible noise invalid data judgment method based on bounded beta (g, h) distribution and MWKPCA | |
| CN112486096A (en) | Machine tool operation state monitoring method | |
| CN112734123A (en) | Industrial waste gas emission prediction method based on ARIMA model | |
| CN111625525A (en) | Environmental data repairing/filling method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |