Remote equipment health prediction method based on machine learning and edge computing

A technology of edge computing and remote equipment, applied in the direction of instruments, electrical testing/monitoring, testing/monitoring control systems, etc., can solve problems such as reduced management efficiency, impact on production activities, property loss, etc., to achieve the effect of reducing impact

Inactive Publication Date: 2020-10-02

NANJING INTELLIGENT MFG RES INC

0 Cites 13 Cited by

AI-Extracted Technical Summary

Problems solved by technology

With the increase in the number of equipment deployed, the maintenance cost increases sharply, and the management efficiency is greatly reduced

In addition, due to the lack of remote management methods, abnormalities in equipment operation cannot be detected early, and sudden failures often have a major impact on production activities, causing major property losses and production safety accidents

[0003] The current remote maintenance methods mainly use remote network technology to monitor the current operating status of the eq...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Method used

The present invention provides a kind of high-efficiency intelligent prediction model based on machine learning, finds the method for associated feature from a large amount of highly nonlinear, high-noise industrial state sensor data, by screening, cleaning and other operations to a large amount of sensor data, Realize the efficient mining of data, discover the correlation change characteristics between different sensor data hidden in the data, and establish the correlation feature model. Through the distribution of model feature sets, the discovery and analysis of abnormal features can be realized, so as to establish ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Abstract

The invention discloses a remote equipment health prediction method based on machine learning and edge computing, and the method comprises the following steps: deploying an edge computing terminal integrating computing, storage, network and application core capabilities on the edge side close to equipment or a data source, realizing connection with production equipment in combination with a communication module, and acquiring operation element data of the production equipment in real time; providing intelligent data analysis services nearby; employing the intelligent analysis model based on machine learning for completing cleaning and preliminary analysis processing of a large amount of real-time data, triggering possible analysis service response according to a deployed prediction mode strategy, and uploading an analysis result to the cloud; and then completing comprehensive analysis and prediction of the data through the cloud intelligent model. The cloud management architecture realizes efficient management of the unattended equipment terminal, and the distributed attribute of the edge computing effectively reduces the data processing load of the cloud platform and ensures the data security at the same time. Operation health state management and fault prediction of remote equipment are effectively realized.

Application Domain

Electric testing/monitoring

Technology Topic

Data processingData analysis +11

Image

Examples

Experimental program(1)

Example Embodiment

[0026] The preferred embodiments of the present invention will be described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described here are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0027] Please refer to Figure 1-2 As shown, a remote device health prediction method based on machine learning and edge computing. This method deploys edge computing terminals that integrate computing, storage, network, and application core capabilities on the edge side close to the device or data source. Edge computing terminals Combine the communication module to realize the connection with the production equipment, and collect the operation element data of the production equipment in real time. The edge computing terminal includes an edge gateway, a network, a cloud data platform and an application terminal; the specific prediction steps are:

[0028] S1: Collect real-time operating data of production equipment through edge computing terminals;

[0029] S2: Store and collect data on the edge side to form formatted data to meet the engineering needs of the system, and the edge computing terminal provides basic data support;

[0030] S3: Adapt network, computing, storage and other resources according to the data model, automatically encapsulate the response strategy, complete data screening and security review, package the rule storage unit, request the cloud and perform asymmetric encrypted communication;

[0031] S4: Extract the associated features of multi-dimensional data from the data set, remove invalid features, mine effective features to comprehensively manage the features, form processing logs, trigger external API rules, and respond to external communication mechanisms;

[0032] S5①: Perform unbalanced data processing on multi-dimensional data features, fill in missing data features, and construct a self-learning model to input basic components;

[0033] S5②: Split multi-dimensional complex data features, classify abnormal features and normal data features, and build basic model components;

[0034] S5③: Clean the original data, eliminate idle complex and invalid data, establish an effective database, and perform data forwarding, using Fourier transform and time series wave decomposition, output the associated sensor data map, and update it to the data model simultaneously;

[0035] S6: Use the associated feature unit of the basic component library to create a prediction model, efficiently import the data set, and output the model prediction map database;

[0036] S7: The model training unit synchronizes the new data in real time, uses the machine learning framework to identify normal and abnormal data features according to strategies, combines the associated data atlas to train the data prediction model, and continuously optimizes the maturity of the feature model;

[0037] S8: Use data outlier graphs, time series distribution graphs and abnormal fault identification database to form a comprehensive analysis strategy, correlate business data, and predict business abnormal data situation;

[0038] S9: After the occurrence of abnormal features, automatically locate the associated data sensor, identify failure precursors, output failure warning prompts, and automatically draw a health life value curve to evaluate equipment health;

[0039] S10: According to the input business conditions and the positioning sensor ID, identify the equipment health warning value, output a fault warning reminder, and synchronize the warning business data to the edge side and the cloud.

[0040] The edge computing gateway is deployed between the cloud data service platform and the managed device, effectively reducing cloud workloads, improving computing performance, and reducing application data latency. With the development of the Internet of Things, converged edge services that integrate network, storage, computing and applications are adopted to respond to user needs faster, so that user needs can be solved at the edge, and the cost of user processing problems can be minimized.

[0041] The system application framework deployment is divided into edge layer, network layer and data computing layer.

[0042] Edge layer: The edge layer is close to the equipment terminal or data source side of the production line. It collects equipment data, integrates storage, network, and computing core capabilities, cleans basic data as required, and calculates the results according to the initial planning model. The core capability of edge computing is to migrate computing tasks from cloud data platforms to edge gateways. The cloud platform comprehensively calculates complex application data according to the needs of the edge. It is precisely because of the delay and elasticity of edge computing that the autonomous decision-making of edge computing does not depend on the characteristics of the cloud data platform, and it has a greater application advantage in IoT applications.

[0043] Network layer: The network layer realizes device connection and data interconnection through wireless or wired means. It also includes network connection and management, processing edge computing results, and ensuring the survival of services locally. The network guarantees the security of data transmission, supports protocol conversion functions, supports a wide range of Internet of Things communication protocols, and ensures the stability of data transmission. Different protocols need to be effectively converted under the cooperation of edge gateways and networks to carry data uniformly on the network. Transfer on.

[0044] Data calculation layer: The cloud data platform collects and uploads data at the edge layer through the network, calculates and analyzes complex data, links and associates edge gateways to achieve comprehensive data management and application, and delivers the calculation results of complex linkage models to the corresponding edge gateways in time Effectively solve customer problems and provide real-time feedback on the edge. Cloud collaboration maximizes the integration advantages of network, computing, storage and applications, and provides agile and effective services for the final application.

[0045] The edge computing terminal of the present invention provides intelligent data analysis services nearby, uses the intelligent analysis model based on machine learning to complete the cleaning and preliminary analysis of a large amount of real-time data, and triggers possible analysis service responses according to the deployed prediction mode strategy, and analyzes The results are uploaded to the cloud, and the comprehensive analysis and prediction of the data is completed by the cloud intelligent model. The architecture of cloud management is applied to realize the efficient management of massive unattended equipment terminals. The distributed nature of edge computing effectively reduces the massive data processing load of the cloud platform while ensuring data security. Through the collaboration between the cloud and the edge, the operational health status management and fault prediction of remote devices can be effectively realized.

[0046] The present invention provides a high-efficiency intelligent prediction model based on machine learning, and a method for discovering associated features from a large number of highly nonlinear and high-noise industrial state sensor data. By filtering and cleaning a large number of sensor data, data recovery is achieved. Efficient mining, discover the correlation change characteristics between different sensor data hidden in the data, and establish the correlation feature model. Through the distribution of the model feature set, the discovery and discrimination of abnormal features are realized, so as to establish the abnormal degradation curve of the data association state and realize the purpose of data prediction.

[0047] The embodiments of the present invention are not limited to this. According to the content of the above-mentioned embodiments of the present invention, using conventional technical knowledge and conventional methods in the field, without departing from the above-mentioned basic technical ideas of the present invention, the above preferred embodiments can also make other Various modifications, substitutions, or combinations, and other obtained embodiments fall within the protection scope of the present invention.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to view more

PUM

Description & Claims & Application Information

We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to view more

Classification and recommendation of technical efficacy words

Reduce the impact

Remote equipment health prediction method based on machine learning and edge computing

AI-Extracted Technical Summary

Problems solved by technology

Method used

Abstract

Image

Examples

Example Embodiment

PUM

Description & Claims & Application Information

Similar technology patents

Assembly method of section box girder

Floating keyboard

Non-blasting digging method for tunnel and underground project

Preparation method of super-hydrophobic/lipophilic oil-water separation meshed membrane

High-precision temporary punching type injecting transonic speed wind tunnel flow field control structure

Classification and recommendation of technical efficacy words

Monitoring, predicting and treating clinical episodes

Moire-Free Touch Screen with Tilted or Curved ITO Pattern

Back side deposition apparatus and applications

Telecommunication networks

Knee brace and method for securing the same