CN112711850A - Unit online monitoring method based on big data - Google Patents
Unit online monitoring method based on big data Download PDFInfo
- Publication number
- CN112711850A CN112711850A CN202011609564.7A CN202011609564A CN112711850A CN 112711850 A CN112711850 A CN 112711850A CN 202011609564 A CN202011609564 A CN 202011609564A CN 112711850 A CN112711850 A CN 112711850A
- Authority
- CN
- China
- Prior art keywords
- data
- real
- model
- noise reduction
- abnormal
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2457—Query processing with adaptation to user needs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2465—Query processing support for facilitating data mining operations in structured databases
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Probability & Statistics with Applications (AREA)
- Fuzzy Systems (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- Quality & Reliability (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a big data-based unit online monitoring method, wherein when monitoring is carried out, real-time data is directly accessed into a single-point evaluation model, the running state of equipment or whether the data is noise can be rapidly detected through the single-point evaluation model, or the single-point evaluation model judges that the equipment runs and the real-time data is non-noise data, the real-time data can be accessed into an abnormal state data model, and the abnormal state data model is established based on abnormal state historical data, so that when the real-time data and the abnormal state data model are successfully matched, a system can rapidly and accurately send out an abnormal signal of the equipment, a worker can conveniently investigate and overhaul the abnormal state, and the monitoring efficiency and the accuracy of a monitoring system on faults are improved.
Description
Technical Field
The invention relates to the field of industrial application of generator sets, in particular to a generator set online monitoring method based on big data.
Background
China develops wind power generation rapidly, but the wind power generation method faces the challenge of multiple faults in the primary stage. The wind power industry expects to shift from extensive operation to intensive operation, and needs to improve installed capacity and realize high yield and high efficiency. The fan has low reliability, and small faults cannot be found and maintained due to the lack of a fault early warning function, so that serious safety and equipment accidents are developed, the shutdown loss is caused, and the maintenance cost is high. The report that a certain fan manufacturer fails to pay due to the fact that the fan manufacturer is not laid out because of the frequent product failure of the fan manufacturer is maintained has appeared abroad. The wind power industry must move out of such a dilemma.
The health state monitoring system is based on a big data analysis technology, and provides a technical condition for the wind power industry to break away from the above-mentioned embarrassment. Specifically, the health state monitoring technology takes historical big data of equipment operation as a modeling basis, carries out a series of screening, noise reduction and data removal when the operation is abnormal on the collected historical data, models and optimizes the processed data to obtain a health state monitoring model, and then accesses the real-time data into the health state monitoring model to realize dynamic monitoring on the equipment. The health state monitoring system can enable a user to find out faults in time and give an alarm to prompt maintenance. The health state monitoring model in the prior art is established based on single normal state historical data, then real-time data is processed and accessed into the health state monitoring model, in order to match the real-time data with the health state monitoring model, the preprocessing of the real-time data usually takes longer time, the calculated amount of a system detection chamber is increased, the sensitivity of the monitoring system to interference data is reduced, meanwhile, because the monitoring system is only provided with the health state monitoring model, during monitoring, the preprocessed real-time data is matched with the health state monitoring model, if the matching is successful, the equipment is normal, if the matching is unsuccessful, the system directly adopts an elimination method to immediately judge the equipment is abnormal, the detection method is not beneficial to the system to quickly early warn abnormal real-time data on one hand, and on the other hand, the elimination method is adopted to judge the equipment is abnormal, the accuracy of the health status monitoring model is reduced.
Disclosure of Invention
The invention mainly aims to solve the problem that when real-time data are not matched with a model due to the fact that only a health state monitoring model is integrated in a monitoring system, the system judges that equipment is abnormal through an elimination method, so that the sensitivity and the accuracy of the health state monitoring system of a wind generating set are influenced, and provides a large data-based set online monitoring method.
To achieve these objects and other advantages in accordance with the purpose of the invention, there is provided a big data based on-line monitoring method for a unit, comprising the steps of:
step one, establishing a single-measuring-point evaluation model, a normal state data model and an abnormal state data model;
the single-measuring-point evaluation model is established based on a noise reduction historical data set of typical measuring points, a to-be-evaluated interval corresponding to the typical measuring points is arranged in the single-measuring-point evaluation model, the to-be-evaluated interval comprises all operating parameters of the typical measuring points under the operating state of the wind generating set, and the single-measuring-point evaluation model is used for predicting the operating state of the wind generating set;
the normal state data model is provided with normal ranges of all the typical measuring points when the wind generating set operates normally;
the abnormal state data model is provided with abnormal ranges of all the typical measuring points when the wind generating set operates abnormally;
step two, collecting real-time data of the typical measuring points, and carrying out noise reduction processing on the real-time data to obtain a plurality of noise reduction real-time data sets, wherein the typical measuring points correspond to the noise reduction real-time data sets one by one;
step three, the noise reduction real-time data set in the step two is accessed into the single-measuring-point evaluation model for online fault detection;
step four, selectively accessing the noise reduction real-time data set in the step three into the abnormal state data model for online fault detection;
combining every two noise reduction real-time data sets to be subjected to health degree evaluation in the step four to obtain a plurality of real-time grey correlation degree combinations, and carrying out non-dimensionalization processing on the real-time grey correlation degree combinations to obtain correlation real-time data sequences;
the associated real-time data sequence is accessed into the normal state data model, and if the associated real-time data sequence is located in the normal range of the normal state data model, the wind generating set is judged to normally operate; if the associated real-time data sequence is located outside the normal range of the normal state data model, judging that the wind generating set is abnormal in operation;
and step six, optimizing the noise reduction real-time data corresponding to the abnormal operation of the wind generating set into the abnormal state data model based on the judgment result in the step five.
Preferably, the noise reduction historical data set is obtained by performing noise reduction on historical data of a plurality of monitoring measuring points, and the historical data covers operating parameters of the wind generating set within 1-3 years of operation.
Preferably, the modeling method of the single-point evaluation model comprises the following steps:
and performing trend analysis on the noise reduction historical data set of each monitoring measuring point, determining the typical measuring point by combining a wind generating set use condition record table, and setting the to-be-evaluated interval corresponding to the typical measuring point.
Preferably, the online fault detection method of the single-point evaluation model in step three is as follows:
extracting a typical measuring point measured value corresponding to the typical measuring point from the noise reduction real-time data set, comparing the typical measuring point measured value with the interval to be evaluated, if the typical measuring point measured value is located outside the interval to be evaluated, determining that the wind generating set does not run, finishing the evaluation, if the typical measuring point measured value is located within the interval to be evaluated, determining that the wind generating set runs, and sending the corresponding noise reduction real-time data set to the abnormal state data model.
Preferably, the online fault detection method of the abnormal state data model in the fourth step is as follows:
and receiving the noise reduction real-time data set on the single-point evaluation model, if the noise reduction real-time data set is located in the abnormal range of the abnormal state data model, determining that the wind generating set is abnormal in operation, and finishing evaluation, and if the noise reduction real-time data set is located outside the abnormal range of the abnormal state data model, storing the noise reduction real-time data set, and sending the noise reduction real-time data set to the normal state data model for health degree evaluation.
Preferably, the normal state historical data sets are combined pairwise to obtain a plurality of normal state ash association degree combinations, an association degree analysis method is adopted to calculate the ash association degree value of the normal state ash association degree combinations, redundant measuring points are eliminated according to the ash association degree value, and the reserved normal state ash association degree combinations are subjected to dimensionless processing to obtain associated normal state data sequences.
Preferably, the modeling method of the normal state data model includes:
establishing the normal state data model based on the associated normal state data sequences, mapping all the associated normal state data sequences to a normal state space, constructing a plurality of second external planes based on boundary points of the normal state space, equally dividing each second external plane into a plurality of plane models, enabling the plane models to correspond to the normal state gray correlation degree combinations one by one, and fitting the maximum value and the minimum value on the associated normal state data sequences to the plane models to obtain the normal range of the normal state data model corresponding to the normal state gray correlation degree combinations.
The invention at least comprises the following beneficial effects:
1. the method and the device utilize the historical data of the operation of the wind generating set as the basis of system modeling, carry out noise reduction processing on the historical data through trend analysis to obtain a noise reduction historical data set, can determine a typical measuring point and a to-be-evaluated interval corresponding to the typical measuring point when the wind generating set is in the operation state through the trend analysis of the noise reduction historical data set and combining with a using condition recording table of the wind generating set, and the typical measuring point can reflect the operation state of the device, so that the operation state of the device can be rapidly judged through the to-be-evaluated interval after the monitoring system accesses the real-time data, and the online monitoring process of the operation state of the wind.
2. When the monitoring is carried out, the real-time data is directly accessed into the single-point evaluation model, the running state of the equipment or whether the data is noise can be quickly detected through the single-point evaluation model, or the single-point evaluation model judges that the equipment runs and the real-time data is non-noise data, the real-time data can be accessed into the abnormal state data model, and the abnormal state data model is established based on the abnormal state historical data, so that when the real-time data is successfully matched with the abnormal state data model, the system can quickly and accurately send out an abnormal signal of the equipment, a worker can conveniently examine and repair the abnormal state, and the monitoring efficiency and the accuracy of the monitoring system on faults are improved.
3. The abnormal data model is established based on the abnormal data in the historical data, so that when the real-time data is subjected to abnormal verification, if the real-time data is not detected when the abnormal data model is accessed and the normal data model is also not detected, the real-time data is judged to be the abnormal data, and the real-time abnormal data when a new fault occurs is supplemented into the original abnormal data model, so that the abnormal data model can realize dynamic optimization along with the use and monitoring of equipment, and the accuracy of a monitoring system is further improved.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention.
Drawings
FIG. 1 is a typical test point trend graph for a wind turbine generator system;
FIG. 2 is a comparison graph of correlation between reference measuring points and associated measuring points;
FIG. 3 is a modeling flow diagram of an online monitoring system;
FIG. 4 is a flow chart of a method of a human-computer Internet of things system based on motion gesture recognition.
Detailed Description
The present invention is further described in detail below with reference to the attached drawings so that those skilled in the art can implement the invention by referring to the description text.
The invention provides a big data-based unit online monitoring method, and figures 1-4 show an implementation form according to the invention, which comprises the following steps:
step one, establishing a single-measuring-point evaluation model, a normal state data model and an abnormal state data model, wherein,
the single-measuring-point evaluation model is established based on a noise reduction historical data set of typical measuring points, a to-be-evaluated interval corresponding to the typical measuring points is arranged in the single-measuring-point evaluation model, the to-be-evaluated interval comprises all operation parameters of the typical measuring points in the operation state of the wind generating set, the single-measuring-point evaluation model is used for predicting the operation state of the wind generating set, the noise reduction historical data set is obtained by performing noise reduction on historical data of a plurality of monitoring measuring points, and the historical data covers the operation parameters of the wind generating set in 1-3 years of operation.
The modeling method of the single-point evaluation model comprises the following steps: and performing trend analysis on the noise reduction historical data set of each monitoring measuring point, determining typical measuring points by combining with a use condition recording table of the wind generating set, and setting an interval to be evaluated corresponding to the typical measuring points, wherein a trend graph of the typical measuring points is shown in figure 1, and the interval to be evaluated of the typical measuring points can be quickly determined from the trend graph.
The normal state data model is provided with normal ranges of all typical measuring points when the wind generating set normally operates.
The modeling method of the normal state data model comprises the following steps: selecting any measuring point from typical measuring points as a reference measuring point, using a plurality of measuring points which are not selected as associated measuring points, combining a noise reduction historical data set corresponding to the reference measuring point with a noise reduction historical data set of each associated measuring point to obtain a plurality of data sets to be separated, performing association relation comparison analysis on the data sets to be separated, obtaining a normal state linear regression curve and an abnormal state linear regression curve after the association relation comparison analysis on each data set to be separated, and extracting data on the normal state linear regression curve to obtain a normal state historical data set.
Combining the normal state historical data sets pairwise to obtain a plurality of normal state gray association degree combinations, calculating gray association degree values of the normal state gray association degree combinations by adopting an association degree analysis method, eliminating redundant measuring points according to the gray association degree values, and carrying out dimensionless processing on the reserved normal state gray association degree combinations to obtain associated normal state data sequences.
Specifically, the method for excluding the redundant measurement points comprises the following steps: setting a correlation threshold, if the grey correlation values of all normal state grey correlation combination related to the typical measuring point are smaller than the correlation threshold, judging the measuring point as a redundant measuring point, and removing the normal state grey correlation combination related to the measuring point; and if the grey correlation value of any one or more of the normal state grey correlation degree combinations related to the typical measuring point is larger than the correlation degree threshold value, keeping the normal state grey correlation degree combination related to the measuring point.
Establishing a normal state data model based on the associated normal state data sequences, mapping all the associated normal state data sequences to a normal state space, constructing a plurality of second external planes based on boundary points of the normal state space, equally dividing each second external plane into a plurality of plane models, enabling the plane models to correspond to normal state gray association degree combinations one by one, and fitting the maximum value and the minimum value on the associated normal state data sequences into the plane models to obtain the normal range of the normal state data model corresponding to the normal state gray association degree combinations.
Generally speaking, the normal operation time in the wind generating set is longer than the abnormal operation time, so that the number of normal state historical data sets obtained after the noise reduction historical data is subjected to correlation comparison and analysis is more than that of abnormal state historical data sets, therefore, dimension reduction and optimization processing are required to be carried out on the normal state historical data sets, the normal state historical data sets corresponding to the measuring points are combined pairwise to obtain a plurality of normal state gray association degree combinations, the normal state gray association degree combinations are calculated and analyzed by adopting an association degree analysis calculation method to obtain the gray association degree corresponding to the normal state gray association degree combinations, setting a correlation threshold, comparing the gray correlation of the normal gray correlation combination with the correlation threshold, removing redundant measuring points, and carrying out non-dimensionalization treatment on the reserved normal state ash association degree combination to obtain an associated normal state data sequence, and finally establishing a normal state data model on the basis of the associated normal state data sequence. Based on actual conditions, the invention mainly analyzes the normal state historical data set with larger data volume and more complexity, removes redundant measuring points in the normal state historical data set by adopting an association degree analysis method, reduces the dimensionality of the normal state historical data set by a dimensionless processing method, accesses the normal state historical data set into a normal state data model, continuously optimizes and adjusts an association degree threshold value according to an alarm signal of the system, and finally ensures the high sensitivity and high accuracy of the normal state data model in a monitoring system.
The method for verifying and optimizing the normal state data model comprises the following steps: simulating the normal operation state of the wind generating set, performing correlation analysis on the simulation data of the wind generating set during normal operation to obtain a plurality of normal state simulation correlation data sequences, accessing the normal state simulation correlation data sequences into a normal state data model as a first verification to simulate the abnormal operation state of the wind generating set, performing correlation analysis on the simulation data of the wind generating set during abnormal operation to obtain a plurality of abnormal state simulation correlation data sequences, and accessing the abnormal state simulation correlation data sequences into the normal state data model as a second verification;
if the alarm does not appear in the first verification, the alarm appears in the second verification, and the normal state data model is judged to be available;
if the alarm does not appear in the first verification and the alarm does not appear in the second verification, increasing the correlation threshold, and determining and eliminating a new redundant measuring point;
if the alarm appears in the first verification, the alarm also appears in the second verification, the threshold value of the correlation degree is reduced, and the excluded redundant measuring points are added again;
and if the alarm appears in the first verification and the alarm does not appear in the second verification, judging that the normal state data model is unavailable, and reestablishing the normal state data model.
The abnormal state data model is provided with abnormal ranges of all typical measuring points when the wind generating set operates abnormally.
The modeling method of the abnormal state data model comprises the following steps: and separating the noise reduction historical data sets, specifically, selecting any measuring point from the measuring points as a reference measuring point, using a plurality of measuring points which are not selected as associated measuring points, combining the noise reduction historical data sets corresponding to the reference measuring points with the noise reduction historical data sets of each associated measuring point to obtain a plurality of data sets to be separated, performing association relation comparison analysis on the data sets to be separated, and obtaining a normal state linear regression curve and an abnormal state linear regression curve after the association relation comparison analysis on each data set to be separated.
Extracting data on the abnormal state linear regression curve to obtain an abnormal state historical data set; and establishing an abnormal state data model based on the abnormal state historical data set.
The abnormal state historical data set is mapped into an abnormal state space, a plurality of first external planes are constructed based on boundary points of the abnormal state space, the first external planes correspond to the typical measuring points one by one, and the maximum value and the minimum value on the abnormal state historical data set are fitted into the first external planes to obtain the abnormal range of the abnormal state data model corresponding to the typical measuring points.
For the abnormal state historical data set, the corresponding typical measuring points are all related to the running state of the wind driven generator, and the data volume is small, so that the abnormal state historical data set can be directly modeled without performing relevance analysis calculation. When accessing real-time data, firstly performing single-test evaluation, if the evaluation is not finished, accessing the real-time data into the abnormal state data model, if the real-time data is judged to be in the abnormal range of the abnormal state data model, judging that the equipment is abnormal in operation, if the real-time data is not in the abnormal range, immediately judging that the equipment is normal in operation, storing corresponding real-time data, processing the real-time data and then accessing the real-time data into the normal state data model, and if the processed real-time data is in the normal range of the normal state data model, judging that the equipment is normal in operation, otherwise, judging that the equipment is abnormal in operation. Compared with the monitoring system in the prior art, the method adopts a plurality of data models to monitor the real-time data, and adopts a method of mutual verification of the plurality of data models to improve the accuracy of the state detection of the wind generating set.
And step two, acquiring real-time data of typical measuring points, and performing noise reduction processing on the real-time data to obtain a plurality of noise reduction real-time data sets, wherein the typical measuring points correspond to the noise reduction real-time data sets one to one.
And step three, accessing the noise reduction real-time data set in the step two into a single-measuring-point evaluation model, extracting a typical measuring point actual measurement value corresponding to a typical measuring point in the noise reduction real-time data set, comparing the typical measuring point actual measurement value with an interval to be evaluated, if the typical measuring point actual measurement value is positioned outside the interval to be evaluated, judging that the wind generating set does not run, finishing evaluation, if the typical measuring point actual measurement value is positioned in the interval to be evaluated, judging that the wind generating set runs, and accessing the corresponding noise reduction real-time data set into an abnormal data model.
The method and the device utilize the historical data of the operation of the wind generating set as the basis of system modeling, carry out noise reduction processing on the historical data through trend analysis to obtain a noise reduction historical data set, can determine a typical measuring point and a to-be-evaluated interval corresponding to the typical measuring point when the wind generating set is in the operation state through the trend analysis of the noise reduction historical data set and combining with a using condition recording table of the wind generating set, and the typical measuring point can reflect the operation state of the device, so that the operation state of the device can be rapidly judged through the to-be-evaluated interval after the monitoring system accesses the real-time data, and the online monitoring process of the operation state of the wind.
And step four, fault evaluation is carried out on the noise reduction real-time data set accessed into the abnormal state data model in the step three, if the noise reduction real-time data set is located in the abnormal range of the abnormal state data model, the wind generating set is judged to be abnormal in operation, evaluation is finished, and if the noise reduction real-time data set is located outside the abnormal range of the abnormal state data model, the noise reduction real-time data set is stored, and health degree evaluation is carried out.
When the monitoring is carried out, the real-time data is directly accessed into the single-point evaluation model, the running state of the equipment or whether the data is noise can be quickly detected through the single-point evaluation model, or the single-point evaluation model judges that the equipment runs and the real-time data is non-noise data, the real-time data can be accessed into the abnormal state data model, and the abnormal state data model is established based on the abnormal state historical data, so that when the real-time data is successfully matched with the abnormal state data model, the system can quickly and accurately send out an abnormal signal of the equipment, a worker can conveniently examine and repair the abnormal state, and the monitoring efficiency and the accuracy of the monitoring system on faults are improved.
And step five, combining every two noise reduction real-time data sets to be subjected to health degree evaluation in the step four to obtain a plurality of real-time gray correlation degree combinations, and carrying out dimensionless processing on the real-time gray correlation degree combinations to obtain a correlation real-time data sequence.
The associated real-time data sequence is accessed into the normal state data model, and if the associated real-time data sequence is located in the normal range of the normal state data model, the wind generating set is judged to normally operate; and if the associated real-time data sequence is located outside the normal range of the normal state data model, judging that the wind generating set is abnormal in operation.
And step six, optimizing the noise reduction real-time data corresponding to the abnormal operation of the wind generating set into an abnormal state data model based on the judgment result in the step five. The abnormal data model is established based on the abnormal data in the historical data, so that when the real-time data is subjected to abnormal verification, if the real-time data is not detected when the abnormal data model is accessed and the normal data model is also not detected, the real-time data is judged to be the abnormal data, and the real-time abnormal data when a new fault occurs is supplemented into the original abnormal data model, so that the abnormal data model can realize dynamic optimization along with the use and monitoring of equipment, and the accuracy of a monitoring system is further improved.
When the monitoring is carried out, the real-time data is directly accessed into the single-point evaluation model, the running state of the equipment or whether the data is noise can be quickly detected through the single-point evaluation model, or the single-point evaluation model judges that the equipment runs and the real-time data is non-noise data, the real-time data can be accessed into the abnormal state data model, and the abnormal state data model is established based on the abnormal state historical data, so that when the real-time data is successfully matched with the abnormal state data model, the system can quickly and accurately send out an abnormal signal of the equipment, a worker can conveniently examine and repair the abnormal state, and the monitoring efficiency and the accuracy of the monitoring system on faults are improved. Meanwhile, because the abnormal data model is established based on the abnormal data in the historical data, when the real-time data is subjected to abnormal verification, if the real-time data is not detected when the abnormal data model is accessed and the normal data model is accessed, the real-time data is judged to be the abnormal data, and the real-time abnormal data when a new fault occurs is supplemented into the original abnormal data model, so that the abnormal data model can realize dynamic optimization along with the use and monitoring of equipment, and the accuracy of the monitoring system is further improved.
While embodiments of the invention have been disclosed above, it is not intended to be limited to the uses set forth in the specification and examples. It can be applied to all kinds of fields suitable for the present invention. Additional modifications will readily occur to those skilled in the art. It is therefore intended that the invention not be limited to the exact details and illustrations described and illustrated herein, but fall within the scope of the appended claims and equivalents thereof.
Claims (7)
1. A unit online monitoring method based on big data is characterized by comprising the following steps:
step one, establishing a single-measuring-point evaluation model, a normal state data model and an abnormal state data model;
the single-measuring-point evaluation model is established based on a noise reduction historical data set of typical measuring points, a to-be-evaluated interval corresponding to the typical measuring points is arranged in the single-measuring-point evaluation model, the to-be-evaluated interval comprises all operating parameters of the typical measuring points under the operating state of the wind generating set, and the single-measuring-point evaluation model is used for predicting the operating state of the wind generating set;
the normal state data model is provided with normal ranges of all the typical measuring points when the wind generating set operates normally;
the abnormal state data model is provided with abnormal ranges of all the typical measuring points when the wind generating set operates abnormally;
step two, collecting real-time data of the typical measuring points, and carrying out noise reduction processing on the real-time data to obtain a plurality of noise reduction real-time data sets, wherein the typical measuring points correspond to the noise reduction real-time data sets one by one;
step three, the noise reduction real-time data set in the step two is accessed into the single-measuring-point evaluation model for online fault detection;
step four, selectively accessing the noise reduction real-time data set in the step three into the abnormal state data model for online fault detection;
combining every two noise reduction real-time data sets to be subjected to health degree evaluation in the step four to obtain a plurality of real-time grey correlation degree combinations, and carrying out non-dimensionalization processing on the real-time grey correlation degree combinations to obtain correlation real-time data sequences;
the associated real-time data sequence is accessed into the normal state data model, and if the associated real-time data sequence is located in the normal range of the normal state data model, the wind generating set is judged to normally operate; if the associated real-time data sequence is located outside the normal range of the normal state data model, judging that the wind generating set is abnormal in operation;
and step six, optimizing the noise reduction real-time data corresponding to the abnormal operation of the wind generating set into the abnormal state data model based on the judgment result in the step five.
2. The online monitoring method for the unit based on the big data as claimed in claim 1, wherein the noise reduction historical data set is obtained by noise reduction processing of historical data of a plurality of monitoring points, and the historical data covers operating parameters of the wind generating set within 1-3 years of operation.
3. The big data-based unit online monitoring method according to claim 1, wherein the modeling method of the single-point evaluation model is as follows:
and performing trend analysis on the noise reduction historical data set of each monitoring measuring point, determining the typical measuring point by combining a wind generating set use condition record table, and setting the to-be-evaluated interval corresponding to the typical measuring point.
4. The online monitoring method for the large data-based unit as claimed in claim 1, wherein the online fault detection method for the single-measurement-point evaluation model in step three is as follows:
extracting a typical measuring point measured value corresponding to the typical measuring point from the noise reduction real-time data set, comparing the typical measuring point measured value with the interval to be evaluated, if the typical measuring point measured value is located outside the interval to be evaluated, determining that the wind generating set does not run, finishing the evaluation, if the typical measuring point measured value is located within the interval to be evaluated, determining that the wind generating set runs, and sending the corresponding noise reduction real-time data set to the abnormal state data model.
5. The big data-based unit online monitoring method according to claim 4, wherein the online fault detection method of the abnormal state data model in step four comprises:
and receiving the noise reduction real-time data set on the single-point evaluation model, if the noise reduction real-time data set is located in the abnormal range of the abnormal state data model, determining that the wind generating set is abnormal in operation, and finishing evaluation, and if the noise reduction real-time data set is located outside the abnormal range of the abnormal state data model, storing the noise reduction real-time data set, and sending the noise reduction real-time data set to the normal state data model for health degree evaluation.
6. The big-data-based online monitoring method for the unit according to claim 5, wherein the normal-state historical data sets are combined pairwise to obtain a plurality of normal-state gray association degree combinations, a gray association degree analysis method is used to calculate gray association degree values of the normal-state gray association degree combinations, redundant measuring points are excluded according to the gray association degree values, and the reserved normal-state gray association degree combinations are subjected to dimensionless processing to obtain associated normal-state data sequences.
7. The big data-based unit online monitoring method according to claim 5, wherein the modeling method of the normal state data model is as follows:
establishing the normal state data model based on the associated normal state data sequences, mapping all the associated normal state data sequences to a normal state space, constructing a plurality of second external planes based on boundary points of the normal state space, equally dividing each second external plane into a plurality of plane models, enabling the plane models to correspond to the normal state gray correlation degree combinations one by one, and fitting the maximum value and the minimum value on the associated normal state data sequences to the plane models to obtain the normal range of the normal state data model corresponding to the normal state gray correlation degree combinations.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011609564.7A CN112711850A (en) | 2020-12-30 | 2020-12-30 | Unit online monitoring method based on big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011609564.7A CN112711850A (en) | 2020-12-30 | 2020-12-30 | Unit online monitoring method based on big data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112711850A true CN112711850A (en) | 2021-04-27 |
Family
ID=75547221
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011609564.7A Pending CN112711850A (en) | 2020-12-30 | 2020-12-30 | Unit online monitoring method based on big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112711850A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116418421A (en) * | 2023-06-09 | 2023-07-11 | 北京神州明达高科技有限公司 | Communication equipment detection method based on frequency reception |
| CN117192269A (en) * | 2023-09-20 | 2023-12-08 | 珠海高新区铭越科技有限公司 | Big data monitoring and early warning system and method for electric room environment control box |
| CN117470320A (en) * | 2023-12-26 | 2024-01-30 | 中融固成建设科技有限公司 | On-site environment monitoring system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103226651A (en) * | 2013-03-23 | 2013-07-31 | 中国水利电力物资有限公司 | Wind turbine state evaluation and early-warning method and system based on similarity statistics |
| US20140067327A1 (en) * | 2011-05-03 | 2014-03-06 | China Real-Time Technology Co., Ltd. | Similarity curve-based equipment fault early detection and operation optimization methodology and system |
| CN103674234A (en) * | 2013-12-23 | 2014-03-26 | 北京天源科创风电技术有限责任公司 | State early warning method and system for abnormal vibration of wind generating set |
| CN109185211A (en) * | 2018-10-31 | 2019-01-11 | 扬州大学 | A kind of pumping plant unit pressure pulsation prediction method instantaneously calculated based on Fluid Mechanics Computation |
| CN111008363A (en) * | 2019-11-21 | 2020-04-14 | 西安交通大学 | Multivariable causal-driven complex electromechanical system service safety situation evaluation method |
-
2020
- 2020-12-30 CN CN202011609564.7A patent/CN112711850A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140067327A1 (en) * | 2011-05-03 | 2014-03-06 | China Real-Time Technology Co., Ltd. | Similarity curve-based equipment fault early detection and operation optimization methodology and system |
| CN103226651A (en) * | 2013-03-23 | 2013-07-31 | 中国水利电力物资有限公司 | Wind turbine state evaluation and early-warning method and system based on similarity statistics |
| CN103674234A (en) * | 2013-12-23 | 2014-03-26 | 北京天源科创风电技术有限责任公司 | State early warning method and system for abnormal vibration of wind generating set |
| CN109185211A (en) * | 2018-10-31 | 2019-01-11 | 扬州大学 | A kind of pumping plant unit pressure pulsation prediction method instantaneously calculated based on Fluid Mechanics Computation |
| CN111008363A (en) * | 2019-11-21 | 2020-04-14 | 西安交通大学 | Multivariable causal-driven complex electromechanical system service safety situation evaluation method |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116418421A (en) * | 2023-06-09 | 2023-07-11 | 北京神州明达高科技有限公司 | Communication equipment detection method based on frequency reception |
| CN117192269A (en) * | 2023-09-20 | 2023-12-08 | 珠海高新区铭越科技有限公司 | Big data monitoring and early warning system and method for electric room environment control box |
| CN117192269B (en) * | 2023-09-20 | 2024-04-05 | 珠海高新区铭越科技有限公司 | Big data monitoring and early warning system and method for electric room environment control box |
| CN117470320A (en) * | 2023-12-26 | 2024-01-30 | 中融固成建设科技有限公司 | On-site environment monitoring system |
| CN117470320B (en) * | 2023-12-26 | 2024-03-12 | 中融固成建设科技有限公司 | On-site environment monitoring system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109524139B (en) | Real-time equipment performance monitoring method based on equipment working condition change | |
| CN112711850A (en) | Unit online monitoring method based on big data | |
| US5602761A (en) | Machine performance monitoring and fault classification using an exponentially weighted moving average scheme | |
| CN110823576B (en) | Mechanical anomaly detection method based on generation of countermeasure network | |
| Hayashi et al. | Study of machine fault diagnosis system using neural networks | |
| CN108181105B (en) | Rolling bearing fault pre-diagnosis method and system based on logistic regression and J divergence | |
| JPH07168619A (en) | Method and system for equipment/facility diagnosis | |
| CN112881014B (en) | Offline NVH (noise, vibration and harshness) testing system and method of transmission assembly | |
| CN112393906B (en) | Method for diagnosing, classifying and evaluating health of weak signal fault of bogie bearing of metro vehicle | |
| GB2476246A (en) | Diagnosing an operation mode of a machine | |
| CN114004306A (en) | Equipment fault evaluation system and method based on multi-dimensional data of Internet of things | |
| JP2020071624A (en) | Abnormality diagnosing apparatus, abnormality diagnosing method and program | |
| CN110553789A (en) | state detection method and device of piezoresistive pressure sensor and brake system | |
| CN116625683A (en) | Wind turbine generator system bearing fault identification method, system and device and electronic equipment | |
| Bejaoui et al. | A data-driven prognostics technique and rul prediction of rotating machines using an exponential degradation model | |
| CN112664410B (en) | Big data-based modeling method for unit online monitoring system | |
| CN117786461A (en) | Water pump fault diagnosis method, control device and storage medium thereof | |
| KR102108975B1 (en) | Apparatus and method for condition based maintenance support of naval ship equipment | |
| CN114577470A (en) | Fault diagnosis method and system for fan main bearing | |
| CN109872511B (en) | Self-adaptive two-stage alarm method for monitoring axial displacement sudden change | |
| CN111456915A (en) | Fault diagnosis device and method for internal components of fan engine room | |
| KR20210006832A (en) | Method and apparatus for machine fault diagnosis | |
| US11339763B2 (en) | Method for windmill farm monitoring | |
| KR20240003469A (en) | Predictive diagnosis method and system of nuclear power plant equipment | |
| CN116802471A (en) | Method and system for comprehensively diagnosing defects of rotary machine |
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 |