CN117640636A - Cloud computing-based dynamic ring monitoring method and system - Google Patents
Cloud computing-based dynamic ring monitoring method and system Download PDFInfo
- Publication number
- CN117640636A CN117640636A CN202311614384.1A CN202311614384A CN117640636A CN 117640636 A CN117640636 A CN 117640636A CN 202311614384 A CN202311614384 A CN 202311614384A CN 117640636 A CN117640636 A CN 117640636A
- Authority
- CN
- China
- Prior art keywords
- data
- humidity
- temperature
- abnormal
- predicted
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 72
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000012545 processing Methods 0.000 claims abstract description 60
- 238000004458 analytical method Methods 0.000 claims abstract description 28
- 238000005516 engineering process Methods 0.000 claims abstract description 22
- 230000007613 environmental effect Effects 0.000 claims abstract description 16
- 230000007246 mechanism Effects 0.000 claims abstract description 11
- 238000003062 neural network model Methods 0.000 claims abstract description 10
- 230000002159 abnormal effect Effects 0.000 claims description 99
- 230000005856 abnormality Effects 0.000 claims description 28
- 238000013136 deep learning model Methods 0.000 claims description 22
- 238000012098 association analyses Methods 0.000 claims description 13
- 230000006399 behavior Effects 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 11
- 238000000605 extraction Methods 0.000 claims description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 5
- 238000013135 deep learning Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 2
- 238000012552 review Methods 0.000 claims description 2
- 230000000007 visual effect Effects 0.000 claims description 2
- 230000004044 response Effects 0.000 abstract description 7
- 238000004891 communication Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000009286 beneficial effect Effects 0.000 description 3
- 238000010219 correlation analysis Methods 0.000 description 3
- 239000003795 chemical substances by application Substances 0.000 description 2
- 239000012141 concentrate Substances 0.000 description 2
- 238000012806 monitoring device Methods 0.000 description 2
- 230000002787 reinforcement Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/54—Extraction of image or video features relating to texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/06—Management of faults, events, alarms or notifications
- H04L41/0631—Management of faults, events, alarms or notifications using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N23/00—Cameras or camera modules comprising electronic image sensors; Control thereof
- H04N23/50—Constructional details
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Signal Processing (AREA)
- Multimedia (AREA)
- General Physics & Mathematics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a cloud computing-based dynamic ring monitoring method and a cloud computing-based dynamic ring monitoring system, which relate to the technical field of dynamic ring monitoring and comprise the steps of acquiring environmental data by using monitoring equipment, processing the environmental data at a gateway end based on an edge computing technology, acquiring first processing data and uploading the first processing data to a cloud; constructing a neural network model based on an attention mechanism, and analyzing and processing the first processing data to obtain second processing data; and constructing a management console based on multi-terminal cooperative control, visualizing the second processing data, analyzing the operation behavior and the system log, and constructing a dynamic ring monitoring system. According to the invention, the local control loop processing is performed at the gateway end by utilizing the edge computing technology, so that the real-time processing of the environmental data can be realized at the local site of the equipment, the cloud burden is reduced, the cost of cloud service is reduced, the response speed of the system is improved, and the attention mechanism is introduced to enable the neural network model to be more focused on key information, so that the analysis efficiency and accuracy are improved.
Description
Technical Field
The invention relates to the technical field of dynamic ring monitoring, in particular to a dynamic ring monitoring method and system based on cloud computing.
Background
In recent years, in the technical field of monitoring systems, a dynamic ring monitoring system is used as a precise and complex system, intelligent video monitoring and analysis are carried out by using equipment such as cameras, remote measurement and remote control are carried out on various parameters distributed on power equipment and machine room environments of each machine room, operation parameters of the equipment are monitored in real time, faults are diagnosed and processed, relevant data are recorded and analyzed, centralized monitoring and centralized maintenance are carried out on the equipment, and great help is provided in the aspect of safety protection. In the prior art, a video acquisition technology is mainly applied, and a high-definition video is acquired by utilizing a high-definition network camera and fish-eye camera equipment; by using video coding and transmission techniques, video data is efficiently transmitted over a network; the video analysis technology can realize intelligent analysis such as face recognition, target tracking, anomaly detection and the like based on technologies such as computer vision, deep learning and the like.
However, the currently common solutions suffer from a number of drawbacks including: the dependence on the network is high, and the image quality is reduced due to insufficient bandwidth; the accuracy of the intelligent analysis algorithm is to be improved, and false alarms can be generated; the video storage cost is high, and sufficient cloud computing resource support is needed; privacy security problems need to be enhanced, and video data may be illegally acquired; these are the current technical difficulties and needs to be further improved for the ring monitoring system.
Disclosure of Invention
The invention is provided in view of the problems that the dependency degree on the network is high, the accuracy of the algorithm needs to be improved, the security of the data is not high and the like in the prior art when the monitoring data obtained by the monitoring system is processed.
Therefore, the invention aims to provide a method for reducing cloud computing pressure, improving response speed and intelligent degree of data processing and guaranteeing system safety.
In order to solve the technical problems, the invention provides the following technical scheme:
in a first aspect, an embodiment of the present invention provides a cloud computing-based dynamic ring monitoring method, which includes acquiring environmental data by using a monitoring device, processing the environmental data at a gateway end based on an edge computing technology, obtaining first processing data, and uploading the first processing data to a cloud; constructing a neural network model based on an attention mechanism, and analyzing and processing the first processing data to obtain second processing data; and constructing a management console based on multi-terminal cooperative control, visualizing the second processing data, analyzing the operation behavior and the system log, and constructing a dynamic ring monitoring system.
As a preferable scheme of the dynamic ring monitoring method based on cloud computing, the invention comprises the following steps: the monitoring equipment comprises a temperature sensor, a humidity sensor and a camera; the environmental data includes temperature, humidity and image information; the processing environment data includes the steps of: denoising and normalizing the acquired temperature data and humidity data; setting a sliding window and calculating data in the window to obtain temperature fluctuation and humidity fluctuation in time T, if the temperature fluctuation in the window is greater than a temperature threshold value, carrying out fluctuation marking, and if the humidity fluctuation in the window is greater than a humidity threshold value, carrying out fluctuation marking; extracting statistical features from temperature data and humidity data, extracting texture features and color features from image data, and performing feature dimension reduction; and judging whether the temperature data and the humidity data are abnormal or not.
As a preferable scheme of the dynamic ring monitoring method based on cloud computing, the invention comprises the following steps: the process for judging whether the temperature data and the humidity data are abnormal is as follows: predicting the temperature and humidity change trend in real time by adopting a time sequence prediction model based on deep learning; if the actual temperature and humidity deviate from the predicted value and exceed the preset error range, carrying out abnormal early warning, wherein the specific steps are as follows: constructing a deep learning model LSTM, inputting temperature data and humidity data in a past period of time, outputting temperature prediction and humidity prediction in m hours in the future, and training network parameters; temperature data are acquired in real time, the latest m-hour data are input every n minutes, a deep learning model LSTM is operated to conduct temperature prediction and humidity prediction, and the calculation formula of the temperature prediction is as follows:
wherein,is the predicted temperature; />Predicting the weight of an output layer for the temperature; h is a t Is in a hidden state; />Predicting the bias of the output layer for the temperature; the calculation formula of the humidity prediction is as follows:
wherein,to predict humidity; />Predicting the weight of an output layer for humidity; />Predicting the bias of the output layer for humidity; when the difference between the predicted temperature and the actual temperature occurs in different error intervals, the processing operation performed when the difference between the predicted humidity and the actual humidity occurs in different error intervals is as follows: if the difference value between the predicted temperature and the actual temperature is smaller than the first threshold value and the absolute value between the predicted humidity and the actual humidity is between the second threshold value, the data are all normal, and the log is recorded for continuous monitoring; if the absolute value of the difference value between the predicted humidity and the actual humidity is not between the second threshold value, the data is abnormal, and abnormal early warning is carried out; if the difference value between the predicted temperature and the actual temperature is between the first threshold value and the second threshold value and the absolute value between the predicted humidity and the actual humidity is between the third threshold value, the data are all normal, and the log is recorded for continuous monitoring; if the absolute value of the difference value between the predicted humidity and the actual humidity is not between the third threshold value, the data is abnormal, and abnormal early warning is carried out; if the difference value between the predicted temperature and the actual temperature is between the second threshold value and the third threshold value, carrying out abnormal early warning and recording the data condition; if the difference value between the predicted temperature and the actual temperature is larger than a third threshold value, abnormal early warning is carried out, and the system is stopped in an emergency.
As a preferable scheme of the dynamic ring monitoring method based on cloud computing, the invention comprises the following steps: analyzing and processing the first processed data comprises the steps of: normalizing temperature and humidity data in the first processing data, and carrying out mean value centering on the image data; calculating attention weights between different features according to the context information of the input data; constructing a deep learning model LSTM based on attention, combining attention weights, automatically focusing key features at different moments, and predicting normal states and abnormal states of input data; and marking and writing the abnormal state data back to the database, and performing manual review.
As a preferable scheme of the dynamic ring monitoring method based on cloud computing, the invention comprises the following steps: the normal state and abnormal state prediction of the input data comprises the following steps: inputting the image data into a deep learning model LSTM to obtain a predicted normal probability and a predicted abnormal probability, wherein the calculation formulas of the predicted abnormal probability and the predicted normal probability are as follows:
P abnormality of ＝sigmoid(W out ×h t +b out )
P Normal state ＝1-P Abnormality of
Wherein P is Abnormality of To predict anomaly probability; w (W) out Weights for the output layer; b out Bias for the output layer; p (P) Normal state To predict a normal probability; comparing the predicted normal probability with a normal probability threshold, and comparing the predicted abnormal probability with an abnormal probability threshold, wherein the comparison result is as follows: if the predicted normal probability is greater than the normal probability threshold, the input data belongs to a normal state; if the predicted abnormality probability is greater than the abnormality probability threshold, the input data belongs to an abnormal state; if the predicted normal probability is not greater than the normal probability threshold and the predicted abnormal probability is not greater than the abnormal probability threshold, namely the model cannot clearly judge normal or abnormal on the current input data, the input data belongs to a fuzzy interval, the system automatically increases the image acquisition frequency from once every n minutes to once every d seconds, continuously inputs newly acquired data into a deep learning model LSTM, repeatedly performs the comparison operation, judges whether the input data is still in the fuzzy interval, judges that the input data is not in the fuzzy interval, restores the image acquisition frequency to the original frequency, continuously maintains the existing image acquisition frequency and performs the comparison operation if the input data is still in the fuzzy interval, simultaneously checks related temperature and humidity data, and outputs the data if the fluctuation of the temperature data and the humidity data in the latest time T is found to be increasedAnd carrying out abnormality prediction and abnormality early warning.
As a preferable scheme of the dynamic ring monitoring method based on cloud computing, the invention comprises the following steps: the analysis of the operational behavior and the system log comprises the following steps: collecting logs from each link and storing the logs into a log server by using a unified log format; developing resolvers for different types of logs, and analyzing log contents; through carrying out association analysis on various logs, finding out abnormal traces; analyzing an operation log, evaluating the matching performance of the operation and the function, and finding out abnormal operation; constructing an abnormality detection model of user behavior and system operation by using a machine learning technology, detecting a new log in real time, and outputting an abnormality level result; and feeding back an optimization rule engine module according to the detection result, and adjusting the association rule weight and the detection model threshold value to realize dynamic update.
As a preferable scheme of the dynamic ring monitoring method based on cloud computing, the invention comprises the following steps: the association analysis of various logs comprises the following steps: collecting various logs stored in the system; defining a regular expression rule according to the log type, and analyzing time, equipment ID and event key fields in the log; identifying keywords frequently appearing in logs of different types by adopting a multi-granularity keyword extraction method; summarizing association rules among different log fields and keywords according to the history log analysis; visually displaying the association strength among different journals through a visual association matrix; real-time analyzing the relevance between newly generated logs by using a relevance rule and a time sequence model algorithm, finding potential fault points and positioning root causes; and feeding back log association analysis results, continuously optimizing a rule base and a model, and improving analysis effects.
In a second aspect, in order to further solve the security problem existing in the dynamic ring monitoring, an embodiment of the present invention provides a dynamic ring monitoring system based on cloud computing, which includes: the data acquisition module is used for acquiring data by using the monitoring equipment and preprocessing the data by using an edge computing technology to obtain first processed data; the analysis module is used for processing and analyzing the first processing data by utilizing the neural network model based on the attention mechanism and judging whether an abnormality occurs or not; and the system construction module is used for constructing a management console according to the multi-terminal cooperative control, analyzing the operation behaviors and the system logs and constructing a dynamic ring monitoring system.
In a third aspect, embodiments of the present invention provide a computer apparatus comprising a memory and a processor, the memory storing a computer program, wherein: the computer program when executed by a processor implements any step of a cloud computing-based method for monitoring a moving ring according to the first aspect of the present invention.
In a fourth aspect, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: the computer program when executed by a processor implements any step of a cloud computing-based method for monitoring a moving ring according to the first aspect of the present invention.
The invention has the beneficial effects that: according to the invention, the local control loop processing is performed at the gateway end by utilizing the edge computing technology, so that the real-time processing of the environmental data can be realized locally on the equipment, the cloud burden is reduced, the cost of cloud service is reduced, and the response speed of the system is improved; the attention mechanism is introduced to enable the neural network model to concentrate on key information more, so that analysis efficiency and accuracy are improved; the multi-granularity keyword extraction method is introduced, and through word-level and phrase-level keyword extraction, more comprehensive and deep log analysis is realized, so that the method is beneficial to accurately finding abnormal traces and potential fault points.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. Wherein:
fig. 1 is a flowchart of a dynamic ring monitoring system based on cloud computing in embodiment 1.
Fig. 2 is a flowchart of the analysis and processing of the first processed data in embodiment 1.
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will become more readily apparent, a more particular description of the invention will be rendered by reference to specific embodiments thereof which are illustrated in the appended drawings.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways other than those described herein, and persons skilled in the art will readily appreciate that the present invention is not limited to the specific embodiments disclosed below.
Further, reference herein to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic can be included in at least one implementation of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.
Example 1
Referring to fig. 1 and 2, in a first embodiment of the present invention, a cloud computing-based method for monitoring a moving ring is provided, including the following steps:
s1: and acquiring environmental data by using monitoring equipment, processing the environmental data at a gateway end based on an edge computing technology, obtaining first processing data and uploading the first processing data to a cloud.
Preferably, as shown in fig. 1, a flowchart of a dynamic ring monitoring system based on cloud computing is shown, new energy equipment is monitored at any time by using monitoring equipment, environmental data is obtained, the data is subjected to preliminary processing and judgment by using a local edge computing technology, the subsequent response speed is accelerated, the processed data is uploaded to a cloud end, the data is further judged by using a deep learning model deployed at the cloud end, abnormal judgment is obtained by integrating temperature, humidity and image data, and finally the safety of the data in the system is ensured by analyzing operation behaviors and system logs stored in the system, so that the safety of the system is further protected, wherein the monitoring equipment comprises a temperature sensor, a humidity sensor and a camera, and the environmental data comprises the temperature, the humidity and the image information of a wind driven generator.
Further, the monitoring device is connected to the edge gateway, and the sampling frequency is set to acquire temperature, humidity and image data.
Preferably, a deep learning model LSTM is deployed on the edge gateway, and the collected data is preprocessed, including cleaning, feature extraction and preliminary judgment, so as to obtain a preprocessing result in the first stage, and the preprocessing result is uploaded to the cloud platform.
Specifically, the processing environment data includes the steps of: and denoising and normalizing the acquired temperature data and humidity data.
Setting a sliding window and calculating data in the window to obtain temperature fluctuation and humidity fluctuation, if the temperature fluctuation in the window is greater than a temperature threshold value, carrying out fluctuation marking, and if the humidity fluctuation in the window is greater than a humidity threshold value, carrying out fluctuation marking.
And (3) extracting statistical features from the temperature data and the humidity data, extracting texture features and color features from the image data, and performing feature dimension reduction.
And judging whether the temperature data and the humidity data are abnormal or not.
Further, judging whether the temperature data and the humidity data are abnormal or not, predicting the temperature and humidity change trend in real time by adopting a time sequence prediction model based on deep learning, and carrying out abnormal early warning once the actual temperature and humidity deviate from the predicted value and exceed a preset error range, wherein the specific steps are as follows: and constructing a deep learning model LSTM, inputting temperature data and humidity data in the past week, outputting temperature prediction and humidity prediction in the future 24 hours, and training network parameters.
Temperature data of a wind turbine cabin are obtained in real time, data of 24 hours are input every 10 minutes, a deep learning model LSTM is operated to conduct temperature prediction and humidity prediction, and a calculation formula of the temperature prediction is as follows:
wherein,is the predicted temperature; />Predicting the weight of an output layer for the temperature; h is a t Is in a hidden state; />The bias of the output layer is predicted for the temperature.
The calculation formula of humidity prediction is as follows:
wherein,to predict humidity; />Predicting the weight of an output layer for humidity; />The bias of the output layer is predicted for humidity.
When the difference between the predicted temperature and the actual temperature occurs in different error intervals, the processing operation performed when the difference between the predicted humidity and the actual humidity occurs in different error intervals is as follows: if the difference between the predicted temperature and the actual temperature is less than 1 ℃ and the difference between the predicted humidity and the actual humidity is within +/-3% RH, all data are normal, log is recorded for continuous monitoring, and if the difference between the predicted humidity and the actual humidity is not within +/-3% RH, data are abnormal, and abnormal early warning is carried out.
If the difference between the predicted temperature and the actual temperature is between 1 and 3 ℃ and the difference between the predicted humidity and the actual humidity is between +/-5% RH, the data are all normal, the log is recorded for continuous monitoring, and if the difference between the predicted humidity and the actual humidity is between +/-5% RH, the data are abnormal, and abnormal early warning is carried out.
If the difference between the predicted temperature and the actual temperature is between 3 and 5 ℃, abnormal early warning is carried out and the data condition is recorded.
If the difference between the predicted temperature and the actual temperature is greater than 5 ℃, abnormal early warning is carried out and the system is stopped in an emergency.
Specifically, the values in the above steps are obtained by collecting related data in the field, processing the related data by using an established LSTM model, and continuously adjusting the related data according to specific conditions represented by corresponding data, if the difference between the predicted data and the actual data selects a smaller range, the tolerance of the system is reduced, the system is easily affected by noise or temporary change, frequent false alarms are caused, and if the difference between the predicted data and the actual data selects a larger range, the system reacts later to abnormal conditions of the actual problem, and some potential problems are missed.
S2: and constructing a neural network model based on the attention mechanism, and analyzing and processing the first processing data to obtain second processing data.
Preferably, a deep learning model LSTM of the attention mechanism is constructed on the cloud server, the model automatically focuses on important features, redundant information is filtered, and the first processing data is subjected to deep analysis to obtain second processing data.
Specifically, as shown in fig. 2, the depth analysis of the first processed data includes the following steps: and normalizing the temperature and humidity data in the first processing data, and carrying out standardization processing on the image data to center the mean value.
And acquiring the context information of the input data by utilizing a bidirectional deep learning model LSTM, calculating the attention weight for different features, and highlighting important features.
An attention-based deep learning model LSTM is constructed, and the network performs model training by automatically focusing on key features.
And predicting the normal state and the abnormal state of the input data by using a network model, marking the corresponding data as suspected abnormal data if the predicted data are abnormal, judging according to three predicted data aspects of an image, temperature and humidity, finally confirming whether the corresponding data are abnormal data, marking and writing the abnormal state data back to a database if the corresponding data are abnormal and the data are abnormal, and manually checking, otherwise, recovering the normal monitoring state of the system.
Preferably, historical image data including data of normal state and abnormal state are collected, a data set is divided into a training set and a test set, a deep learning model is trained by using the training set, the probability that each sample belongs to normal or abnormal is output, and a self-adaptive algorithm is used for updating a normal probability threshold and an abnormal probability threshold in real time according to actual feedback and system state.
Further, the normal state and abnormal state prediction of the input data includes the steps of: inputting the image data into a deep learning model LSTM to obtain a predicted normal probability and a predicted abnormal probability, wherein the calculation formulas of the predicted abnormal probability and the predicted normal probability are as follows:
P abnormality of ＝sigmoid(W out ×h t +b out )
P Normal state ＝1-P Abnormality of
Wherein P is Abnormality of To predict anomaly probability; w (W) out Weights for the output layer; b out Bias for the output layer; p (P) Normal state To predict the normal probability.
Comparing the predicted normal probability with a normal probability threshold, and comparing the predicted abnormal probability with an abnormal probability threshold, wherein the comparison result is as follows: if the predicted normal probability is greater than the normal probability threshold, the input data belongs to a normal state.
If the predicted anomaly probability is greater than the anomaly probability threshold, the input data belongs to an anomaly state.
If the predicted normal probability is not greater than the normal probability threshold and the predicted abnormal probability is not greater than the abnormal probability threshold, namely the model cannot clearly judge normal or abnormal on the current input data, the input data belongs to a fuzzy interval, the system automatically increases the image acquisition frequency from once every 5 minutes to once every 30 seconds, continuously inputs newly acquired data into the deep learning model LSTM, repeatedly performs the comparison operation, judges whether the input data is still in the fuzzy interval, judges that the input data is not in the fuzzy interval, restores the image acquisition frequency to the original frequency, continues to maintain the existing image acquisition frequency and perform the comparison operation if the input data is still in the fuzzy interval, simultaneously checks related temperature and humidity data, and outputs abnormal prediction and performs abnormal early warning if the fluctuation of the temperature data and the humidity data in the last 10 minutes is found to be increased.
S3: and constructing a management console based on multi-terminal cooperative control, visualizing the second processing data, analyzing the operation behavior and the system log, and constructing a dynamic ring monitoring system.
Preferably, the operation log and the system log are collected and uploaded to a unified log platform, log analysis and user behavior analysis technology is applied to find abnormal conditions, and a dynamic monitoring system is built through a rule engine technology.
Specifically, the analysis operation log and the system log include the following steps: logs are collected from various links such as environmental monitoring, edge calculation and deep learning modules, and include operation logs, system logs and alarm logs, and are stored in a log server using a unified log format.
And developing resolvers for different types of logs, and analyzing log contents, wherein the operation log resolves user operation information, and the system log resolves program running states and performance indexes.
Through carrying out association analysis on various logs, abnormal traces are found, wherein the abnormal traces comprise correspondence between functional abnormality and system resource use surge and correspondence between file reading and file integrity check.
And analyzing the operation log, evaluating the matching performance of the operation and the function, and finding out abnormal operation, including frequent restarting of the system and reading abnormal operation of the high-strength file.
And constructing an abnormality detection model of user behavior and system operation by using a machine learning technology, detecting a new log in real time, and outputting an abnormality level result.
And feeding back an optimization rule engine module according to the detection result, and adjusting the association rule weight and the detection model threshold value to realize dynamic update.
Further, the association analysis of various logs comprises the following steps: collecting a server equipment log, a network equipment log, an environment monitoring log and an access control log, defining a regular expression rule according to the log type, and analyzing time, equipment ID and event key fields in the log.
And identifying keywords frequently appearing in logs of different types by adopting a multi-granularity keyword extraction method.
The association rules between different log fields and keywords, such as the time association between "server failure" and "temperature alarm", are summarized according to the history log analysis.
Through the visualized association matrix, the association strength among different journals is visually displayed, association rules and a time sequence model algorithm are applied to analyze the association among newly generated journals in real time, potential fault points are found, and the root cause is located.
And continuously optimizing the rule base and the model according to the feedback log association analysis result, and improving the analysis effect.
Preferably, an initial abnormal probability threshold is set based on historical data, the abnormal probability threshold is optimally learned by interaction of the reinforcement learning agent and the environment, and the reinforcement agent is guided to learn and adjust the threshold by using a reward function so as to optimize the overall system effect, and finally the set threshold is obtained.
Further, the potential fault points are found and judged according to the association analysis result obtained by analyzing the association between newly generated logs, the initial abnormal probability of each event is obtained through the association analysis result, the initial abnormal probability is compared with a set threshold value, and the time sequence model adjustment factors, the dynamic adjustment factors and the factors comprehensively considered by multiple factors are combined for judgment.
Specifically, if the initial abnormal probability is not greater than the threshold value, the system operates everything normally, and the log continues to be recorded; if the initial abnormal probability is greater than the threshold value, marking the event as a potential fault point, and respectively carrying out dynamic adjustment, multi-factor comprehensive consideration and real-time dynamic update processing; if the anomaly probability after the dynamic adjustment is greater than a threshold value, sending anomaly alert and recording a log; if the abnormality probability after the comprehensive consideration of the multiple factors is greater than a threshold value, triggering corresponding emergency operation and notifying related personnel; if the abnormal probability after the real-time dynamic update is greater than the threshold value, stopping the corresponding system component and carrying out abnormal early warning.
The embodiment also provides a cloud computing-based dynamic ring monitoring system, which comprises: the data acquisition module is used for acquiring data by using the monitoring equipment and preprocessing the data by using an edge computing technology to obtain first processed data; the analysis module is used for processing and analyzing the first processing data by utilizing the neural network model based on the attention mechanism and judging whether an abnormality occurs or not; and the system construction module is used for constructing a management console according to the multi-terminal cooperative control, analyzing the operation behaviors and the system logs and constructing a dynamic ring monitoring system.
The embodiment also provides a computer device, which is applicable to the situation of a cloud computing-based dynamic ring monitoring method, and comprises the following steps: a memory and a processor; the memory is used for storing computer executable instructions, and the processor is used for executing the computer executable instructions to realize the dynamic ring monitoring method based on cloud computing, which is provided by the embodiment.
The computer device may be a terminal comprising a processor, a memory, a communication interface, a display screen and input means connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless mode can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
The present embodiment also provides a storage medium, on which a computer program is stored, which when executed by a processor, implements a cloud computing-based moving ring monitoring method as set forth in the above embodiment; the storage medium may be implemented by any type or combination of volatile or nonvolatile Memory devices, such as static random access Memory (Static Random Access Memory, SRAM), electrically erasable Programmable Read-Only Memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), erasable Programmable Read-Only Memory (Erasable Programmable Read Only Memory, EPROM), programmable Read-Only Memory (PROM), read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk, or optical disk.
In conclusion, the invention utilizes the edge computing technology to carry out local control loop processing at the gateway end, can realize real-time processing of environmental data at the equipment locally, lightens cloud burden, reduces the cost of cloud service and improves the response speed of the system; the attention mechanism is introduced to enable the neural network model to concentrate on key information more, so that analysis efficiency and accuracy are improved; the multi-granularity keyword extraction method is introduced, and through word-level and phrase-level keyword extraction, more comprehensive and deep log analysis is realized, so that the method is beneficial to accurately finding abnormal traces and potential fault points.
Example 2
Referring to tables 1 to 7, for the second embodiment of the present invention, which is different from the first embodiment, experimental data of the present invention in practical cases are provided in order to verify the advantageous effects thereof.
Specific data collected during actual experiments of the present invention are shown in tables 1 and 2, wherein table 1 is data collected by a temperature sensor and a humidity sensor, and table 2 is an image feature extraction result.
Table 1 data collected by temperature and humidity sensors
| Time | Temperature (. Degree. C.) | Humidity (% RH) |
| Time 1 | 23.5 | 58 |
| Time 2 | 23.8 | 57 |
| ...... | ...... | ...... |
| Time n | 24.0 | 55 |
TABLE 2 image feature extraction results
| Time | Texture features | Color characterization |
| Time 1 | 0.62，0.73，0.81 | 0.52，0.36，0.64 |
| Time 2 | 0.61，0.72，0.83 | 0.51，0.35，0.63 |
| ...... | ...... | ...... |
| Time n | 0.63，0.74，0.82 | 0.53，0.37，0.65 |
From the 2 tables, the invention can be seen that the real-time monitoring data acquisition function of the temperature and humidity sensor is arranged in the key region of the movable ring, the image information is acquired by using the camera, the important characteristics are extracted by using the image processing technology, and the response speed is improved by preprocessing huge data.
As shown in table 3, the fluctuation analysis is performed based on the window, and the temperature fluctuation and the humidity fluctuation are obtained by setting a sliding window and calculating the data in the window, so as to judge whether the temperature data and the humidity data have large fluctuation, and observe the change trend of the temperature and the humidity.
Table 3 wave analysis based on window
| Time | Temperature fluctuation (. Degree. C.) | Humidity fluctuation (% RH) |
| Time 1 | 0.3 | 1 |
| Time 2 | 0.2 | 2 |
| ...... | ...... | ...... |
| Time n | 0.5 | 3 |
As shown in table 4, the temperature data and the humidity data are determined based on the edge computing technology, the future temperature and humidity are predicted by constructing the deep learning model LSTM, and the corresponding data are compared with the actual data to determine whether the data exceeds the preset threshold value, so as to determine whether the data is abnormal, and the response speed of the data to be uploaded to the cloud for the next determination is improved.
TABLE 4 determination of temperature and humidity data based on edge calculation techniques
As shown in table 5, the prediction of the normal state and the abnormal state is performed, the image data is analyzed by constructing a deep learning model LSTM at the cloud end, the corresponding predicted normal probability and predicted abnormal probability are obtained and compared with the normal probability threshold and the abnormal probability threshold, and meanwhile, the monitoring equipment is further judged whether to be normal or not by combining the corresponding fluctuation data of the temperature data and the humidity data.
TABLE 5 prediction of Normal and abnormal states
| Time | Predicting normal probability | Predicting anomaly probability | Results |
| Time 1 | 0.8 | 0.2 | Normal state |
| Time 2 | 0.6 | 0.4 | Abnormality, triggering abnormality early warning |
The rules of specific log correlation analysis and the corresponding correlation analysis result feedback are shown in tables 6 and 7.
Table 6 rules for log association analysis
TABLE 7 correlation analysis result feedback
As can be seen from the table, when the log analysis is performed, the used rules have different weights, which indicates that the system realizes comprehensive various information through log association analysis, evaluates the abnormal probability and enables the abnormal detection to be more comprehensive.
It should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention may be modified or substituted without departing from the spirit and scope of the technical solution of the present invention, which is intended to be covered in the scope of the claims of the present invention.
Claims (10)
1. A cloud computing-based dynamic ring monitoring method is characterized by comprising the following steps of: comprising the following steps:
acquiring environmental data by using monitoring equipment, processing the environmental data at a gateway end based on an edge computing technology, obtaining first processing data and uploading the first processing data to a cloud;
constructing a neural network model based on an attention mechanism, and analyzing and processing the first processing data to obtain second processing data;
and constructing a management console based on multi-terminal cooperative control, visualizing the second processing data, analyzing the operation behavior and the system log, and constructing a dynamic ring monitoring system.
2. The cloud computing-based dynamic ring monitoring method as claimed in claim 1, wherein: the monitoring equipment comprises a temperature sensor, a humidity sensor and a camera;
the environmental data includes temperature, humidity and image information;
the processing environment data includes the steps of:
denoising and normalizing the acquired temperature data and humidity data;
setting a sliding window and calculating data in the window to obtain temperature fluctuation and humidity fluctuation in time T, if the temperature fluctuation in the window is greater than a temperature threshold value, carrying out fluctuation marking, and if the humidity fluctuation in the window is greater than a humidity threshold value, carrying out fluctuation marking;
extracting statistical features from temperature data and humidity data, extracting texture features and color features from image data, and performing feature dimension reduction;
and judging whether the temperature data and the humidity data are abnormal or not.
3. The cloud computing-based dynamic ring monitoring method as claimed in claim 2, wherein: the process for judging whether the temperature data and the humidity data are abnormal is as follows:
predicting the temperature and humidity change trend in real time by adopting a time sequence prediction model based on deep learning;
if the actual temperature and humidity deviate from the predicted value and exceed the preset error range, carrying out abnormal early warning, wherein the specific steps are as follows:
constructing a deep learning model LSTM, inputting temperature data and humidity data in a past period of time, outputting temperature prediction and humidity prediction in m hours in the future, and training network parameters;
temperature data are acquired in real time, the latest m-hour data are input every n minutes, a deep learning model LSTM is operated to conduct temperature prediction and humidity prediction, and the calculation formula of the temperature prediction is as follows:
wherein,is the predicted temperature; />Predicting the weight of an output layer for the temperature; h is a t Is in a hidden state; />Predicting the bias of the output layer for the temperature;
the calculation formula of the humidity prediction is as follows:
wherein,to predict humidity; />Predicting the weight of an output layer for humidity; />Predicting the bias of the output layer for humidity;
when the difference between the predicted temperature and the actual temperature occurs in different error intervals, the processing operation performed when the difference between the predicted humidity and the actual humidity occurs in different error intervals is as follows:
if the difference value between the predicted temperature and the actual temperature is smaller than the first threshold value and the absolute value between the predicted humidity and the actual humidity is between the second threshold value, the data are all normal, and the log is recorded for continuous monitoring; if the absolute value of the difference value between the predicted humidity and the actual humidity is not between the second threshold value, the data is abnormal, and abnormal early warning is carried out;
if the difference value between the predicted temperature and the actual temperature is between the first threshold value and the second threshold value and the absolute value between the predicted humidity and the actual humidity is between the third threshold value, the data are all normal, and the log is recorded for continuous monitoring; if the absolute value of the difference value between the predicted humidity and the actual humidity is not between the third threshold value, the data is abnormal, and abnormal early warning is carried out;
if the difference value between the predicted temperature and the actual temperature is between the second threshold value and the third threshold value, carrying out abnormal early warning and recording the data condition;
if the difference value between the predicted temperature and the actual temperature is larger than a third threshold value, abnormal early warning is carried out, and the system is stopped in an emergency.
4. The cloud computing-based dynamic ring monitoring method as claimed in claim 3, wherein: analyzing and processing the first processed data comprises the steps of:
normalizing temperature and humidity data in the first processing data, and carrying out mean value centering on the image data;
calculating attention weights between different features according to the context information of the input data;
constructing a deep learning model LSTM based on attention, combining attention weights, automatically focusing key features at different moments, and predicting normal states and abnormal states of input data;
and marking and writing the abnormal state data back to the database, and performing manual review.
5. The cloud computing-based moving ring monitoring method as claimed in claim 4, wherein: the normal state and abnormal state prediction of the input data comprises the following steps:
inputting the image data into a deep learning model LSTM to obtain a predicted abnormal probability and a predicted normal probability, wherein the calculation formulas of the predicted abnormal probability and the predicted normal probability are as follows:
P abnormality of ＝sigmoid(W out ×h t +b out )
P Normal state ＝1-P Abnormality of
Wherein P is Abnormality of To predict anomaly probability; w (W) out Weights for the output layer; b out Bias for the output layer; p (P) Normal state To predict a normal probability;
comparing the predicted normal probability with a normal probability threshold, and comparing the predicted abnormal probability with an abnormal probability threshold, wherein the comparison result is as follows:
if the predicted normal probability is greater than the normal probability threshold, the input data belongs to a normal state;
if the predicted abnormality probability is greater than the abnormality probability threshold, the input data belongs to an abnormal state;
if the predicted normal probability is not greater than the normal probability threshold and the predicted abnormal probability is not greater than the abnormal probability threshold, namely the model cannot clearly judge normal or abnormal on the current input data, the input data belongs to a fuzzy interval, the system automatically increases the image acquisition frequency from once every n minutes to once every d seconds, continuously inputs the newly acquired data into the deep learning model LSTM, repeats the comparison operation, judges whether the input data is still in the fuzzy interval, judges that the input data is not in the fuzzy interval, restores the image acquisition frequency to the original frequency, if the input data is not in the fuzzy interval, continues to maintain the existing image acquisition frequency and execute the comparison operation, simultaneously checks related temperature and humidity data, and if the fluctuation of the temperature data and the humidity data in the latest time T is found to be increased, outputs abnormal prediction and carries out abnormal early warning.
6. The cloud computing-based moving ring monitoring method as claimed in claim 5, wherein: the analysis of the operational behavior and the system log comprises the following steps:
collecting logs from each link and storing the logs into a log server by using a unified log format;
developing resolvers for different types of logs, and analyzing log contents;
through carrying out association analysis on various logs, finding out abnormal traces;
analyzing an operation log, evaluating the matching performance of the operation and the function, and finding out abnormal operation;
constructing an abnormality detection model of user behavior and system operation by using a machine learning technology, detecting a new log in real time, and outputting an abnormality level result;
and feeding back an optimization rule engine module according to the detection result, and adjusting the association rule weight and the detection model threshold value to realize dynamic update.
7. The cloud computing-based moving ring monitoring method as claimed in claim 6, wherein: the association analysis of various logs comprises the following steps:
collecting various logs stored in the system;
defining a regular expression rule according to the log type, and analyzing time, equipment ID and event key fields in the log;
identifying keywords frequently appearing in logs of different types by adopting a multi-granularity keyword extraction method;
summarizing association rules among different log fields and keywords according to the history log analysis;
visually displaying the association strength among different journals through a visual association matrix;
real-time analyzing the relevance between newly generated logs by using a relevance rule and a time sequence model algorithm, finding potential fault points and positioning root causes;
and feeding back log association analysis results, and continuously optimizing a rule base and a model.
8. A cloud computing-based moving ring monitoring system, based on the moving ring monitoring method based on cloud computing as set forth in any one of claims 1 to 7, characterized in that: comprising the steps of (a) a step of,
the data acquisition module is used for acquiring data by using the monitoring equipment and preprocessing the data by using an edge computing technology to obtain first processed data;
the analysis module is used for processing and analyzing the first processing data by utilizing the neural network model based on the attention mechanism and judging whether an abnormality occurs or not;
and the system construction module is used for constructing a management console according to the multi-terminal cooperative control, analyzing the operation behaviors and the system logs and constructing a dynamic ring monitoring system.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that: the steps of the cloud computing-based moving ring monitoring method according to any one of claims 1 to 7 are realized when the processor executes the computer program.
10. A computer-readable storage medium having stored thereon a computer program, characterized by: the computer program when executed by a processor implements the steps of a cloud computing-based moving ring monitoring method as set forth in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311614384.1A CN117640636A (en) | 2023-11-28 | 2023-11-28 | Cloud computing-based dynamic ring monitoring method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311614384.1A CN117640636A (en) | 2023-11-28 | 2023-11-28 | Cloud computing-based dynamic ring monitoring method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117640636A true CN117640636A (en) | 2024-03-01 |
Family
ID=90026522
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311614384.1A Pending CN117640636A (en) | 2023-11-28 | 2023-11-28 | Cloud computing-based dynamic ring monitoring method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117640636A (en) |
-
2023
- 2023-11-28 CN CN202311614384.1A patent/CN117640636A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11120127B2 (en) | Reconstruction-based anomaly detection | |
| JP4856396B2 (en) | Method for creating a unified quality assessment for turbine mechanical systems and the like and providing an automatic fault diagnosis tool | |
| Yan et al. | Fisher’s discriminant ratio based health indicator for locating informative frequency bands for machine performance degradation assessment | |
| US11742934B2 (en) | Method for predictive maintenance of satellites | |
| CN111143173A (en) | Server fault monitoring method and system based on neural network | |
| US20210232104A1 (en) | Method and system for identifying and forecasting the development of faults in equipment | |
| KR20190021560A (en) | Failure prediction system using big data and failure prediction method | |
| CN115059634B (en) | Fault diagnosis method and device for air blowing equipment and electronic equipment | |
| CN113760670A (en) | Cable joint abnormity early warning method and device, electronic equipment and storage medium | |
| CN113670434B (en) | Method and device for identifying sound abnormality of substation equipment and computer equipment | |
| CN115060312A (en) | Building material safety monitoring system based on artificial intelligence | |
| CN111708687B (en) | Equipment abnormality index determination method, device, equipment and storage medium | |
| CN118051822A (en) | Equipment running state abnormality detection method based on voiceprint recognition technology | |
| CN118070199A (en) | Power equipment detection method and detection system | |
| CN115524002B (en) | Operation state early warning method, system and storage medium of power plant rotating equipment | |
| CN118408583B (en) | Encoder fault diagnosis method and system | |
| CN114666117A (en) | Network security situation measuring and predicting method for power internet | |
| CN118035904A (en) | Runout monitoring fault diagnosis method and system based on multi-band data fusion | |
| CN117640636A (en) | Cloud computing-based dynamic ring monitoring method and system | |
| CN118194699A (en) | On-line evaluation method and system for excitation system of thermal power generating unit | |
| CN117930819A (en) | Industrial computer behavior analysis and anomaly detection system based on artificial intelligence | |
| CN115871745B (en) | Intelligent maintenance method and device applied to rail transit | |
| CN117874680A (en) | Operation and maintenance management system for fort machine | |
| US20180087489A1 (en) | Method for windmill farm monitoring | |
| CN116467575A (en) | Real-time data monitoring system and method design of ball screw pair 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 |