CN-122020194-A - Industrial equipment state intelligent monitoring system based on big data

CN122020194ACN 122020194 ACN122020194 ACN 122020194ACN-122020194-A

Abstract

The invention relates to the technical field of intelligent manufacturing of industrial Internet of things, and discloses an intelligent industrial equipment state monitoring system based on big data. The system comprises a data acquisition module, a feature abstraction module, a knowledge-graph evolution module and an intelligent early warning decision module. The method comprises the steps of firstly, carrying out cleaning and verification and multi-scale feature extraction on raw data collected by a sensor to generate a device state feature abstract, then, autonomously constructing and dynamically updating a device state monitoring knowledge graph through analyzing association rules among features, carrying out projection and similarity matching on real-time operation data in a topological structure of the knowledge graph, and triggering grading early warning according to matching deviation. The method overcomes the defects of poor adaptability and lack of interpretability of the traditional static monitoring model, realizes accurate sensing and self-adaptive judgment on the dynamic evolution of the equipment state, and improves the accuracy and reliability of monitoring and early warning.

Inventors

YANG SEN
WANG ENNING
Ma Qingzhai
ZHANG JIE
CHEN YANJI
ZHENG DAN
CAO HUIHUI

Assignees

中科宝航(苏州)智造科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260113

Claims (10)

1. An intelligent industrial equipment status monitoring system based on big data, characterized in that the system comprises: The industrial equipment operation data acquisition module is used for continuously capturing an original monitoring data stream from a sensor network deployed in the industrial equipment, and carrying out data cleaning and quality check on the original monitoring data stream to form a standardized equipment operation data set; The device state feature abstract module is used for carrying out multi-scale feature transformation on the standardized device operation data set, generating a device state feature image with time sequence association, and carrying out feature reduction on the device state feature image to obtain a device state feature abstract; The monitoring knowledge graph evolution module is used for analyzing the characteristic association rule implicit in the equipment state characteristic abstract, constructing an equipment state characteristic association network according to the characteristic association rule, and enabling the equipment state characteristic association network to be evolved into an industrial equipment state monitoring knowledge graph through a dynamic updating mechanism; And the intelligent early warning decision module is used for projecting the real-time acquired industrial equipment operation data into the industrial equipment state monitoring knowledge graph to perform similarity matching and deviation calculation, and triggering hierarchical early warning decision according to matching and calculation results.
2. The intelligent industrial equipment status monitoring system based on big data according to claim 1, wherein the raw monitoring data stream comprises vibration signals, temperature readings, pressure values, current-voltage waveforms, equipment control parameters, and environmental temperature and humidity data.
3. The intelligent monitoring system for industrial equipment status based on big data according to claim 1, wherein the performing data cleaning and quality checking on the original monitoring data stream to form a standardized equipment operation data set specifically comprises: performing outlier rejection and missing value interpolation on the original monitoring data stream to obtain primarily processed monitoring data; performing dimension unification and numerical normalization on the primarily processed monitoring data to obtain standardized monitoring data; And aligning and integrating the standardized monitoring data according to the equipment numbers and the time stamps to form a standardized equipment operation data set.
4. The intelligent monitoring system for industrial equipment status based on big data according to claim 1, wherein the performing a multi-scale feature transformation on the standardized set of equipment operation data to generate an equipment status feature image with time sequence association specifically comprises: respectively carrying out time domain statistical feature extraction, frequency domain energy feature decomposition and time-frequency domain joint feature analysis on the standardized equipment operation data set to obtain a multi-scale feature set; And performing sliding splicing on various features in the multi-scale feature set according to time windows to construct a device state feature image.
5. The intelligent monitoring system for industrial equipment status based on big data according to claim 1, wherein the feature reduction of the equipment status feature image to obtain the equipment status feature abstract specifically comprises: calculating the association strength between each feature dimension in the equipment state feature image and the equipment health state; sorting the feature dimensions according to the association strength, and selecting feature dimensions with the association strength exceeding a preset threshold; and recombining the selected feature dimensions to generate a device state feature abstract with reduced dimensions.
6. The intelligent industrial equipment state monitoring system based on big data according to claim 1, wherein the analyzing the characteristic association rule implicit in the equipment state characteristic abstract, and constructing the equipment state characteristic association network according to the characteristic association rule specifically comprises: analyzing the equipment state feature abstract by adopting an association rule mining method, and finding a feature combination mode of frequent co-occurrence; Taking the characteristics as nodes and the co-occurrence intensity among the characteristics as edges, and constructing an initial equipment state characteristic association network; And performing redundant edge trimming and core community discovery on the initial equipment state characteristic association network to obtain a simplified equipment state characteristic association network.
7. The intelligent industrial equipment state monitoring system based on big data according to claim 1, wherein the enabling the equipment state feature association network to be evolved into the industrial equipment state monitoring knowledge graph through a dynamic updating mechanism specifically comprises: continuously injecting a new device state feature abstract into the device state feature association network; Adjusting the weight of the network node and the edge according to the newly injected data, and identifying the newly added characteristic association relation; After iterative updating, an industrial equipment state monitoring knowledge graph containing an equipment state evolution rule is formed.
8. The intelligent industrial equipment state monitoring system based on big data according to claim 1, wherein the projecting the industrial equipment operation data acquired in real time into the industrial equipment state monitoring knowledge graph to perform similarity matching and deviation calculation specifically comprises: carrying out the same characteristic abstract processing on the real-time industrial equipment operation data as the historical data to obtain a real-time equipment state characteristic abstract; calculating the similarity of the characteristic vector of each typical state mode in the real-time equipment state characteristic abstract and the industrial equipment state monitoring knowledge graph; and calculating the characteristic deviation degree of the state characteristic abstract of the real-time equipment relative to the normal state reference in the knowledge graph.
9. The intelligent industrial equipment state monitoring system based on big data according to claim 1, wherein the triggering of the hierarchical early warning decision according to the matching and calculation result specifically comprises: Obtaining a matching calculation result of a real-time equipment state feature abstract and an industrial equipment state monitoring knowledge graph, wherein the matching calculation result comprises a feature vector similarity and a feature deviation degree; The similarity of the feature vectors and the degree of deviation of the features are synthesized into a comprehensive abnormal index by adopting a weighted fusion method; step-by-step comparing the comprehensive abnormal index with a plurality of preset level thresholds, and dividing the state level of the equipment; inquiring an early warning rule base according to the equipment state level to obtain corresponding early warning action configuration; and executing early warning action configuration, generating an early warning signal and distributing the early warning signal to the monitoring terminal.
10. The intelligent monitoring system for industrial equipment status based on big data according to claim 5, wherein the calculating the association strength between each feature dimension in the equipment status feature map and the equipment health status specifically comprises: Collecting historical equipment operation data and labeling equipment health state labels to form a training data set; extracting a numerical sequence of each feature dimension in the training data set for each feature dimension in the equipment state feature map; calculating the statistical correlation measure of the numerical value sequence of each characteristic dimension and the equipment health state label sequence; determining the association strength value of each feature dimension according to the absolute value of the statistical correlation measure; And normalizing the association strength value to obtain a normalized association strength weight.

Description

Industrial equipment state intelligent monitoring system based on big data Technical Field The invention relates to the technical field of intelligent manufacturing of industrial Internet of things, in particular to an intelligent industrial equipment state monitoring system based on big data. Background Current industrial equipment status monitoring relies primarily on alarm or orphan data analysis models based on fixed thresholds. The fixed threshold method is insensitive to early performance degradation of the device and to progressive failure caused by complex factors. However, the conventional prediction model mostly regards the monitoring parameters as independent or simply linearly related variables, and is difficult to capture complex nonlinear dynamic coupling relations among multidimensional parameters such as vibration, temperature, current and the like, so that false alarm and false alarm frequently occur. Knowledge representation and decision mechanisms of the prior art are typically static. After the system is deployed, internal rules or model parameters are fixed, and self-adaptive evolution capability is lacked. The baseline of the normal operation state of the equipment can drift due to abrasion, aging and working condition change in long-term operation, a static model cannot adapt to the dynamic process, early warning precision attenuation is necessarily generated, and even a new normal working condition is misjudged as abnormal. Traditional early warning decisions rely on threshold decisions on individual data points or abnormal scores of isolated models, lacking macroscopic insight into the overall operational situation of the device. The method can not place real-time data in a knowledge network rich in context association for semantic comparison, the decision process is isolated, spanning from anomaly detection to root cause analysis is difficult to realize, and early warning results are poor in interpretation. The invention aims to solve the core defects of insufficient adaptability of the static model and lack of context associated with a decision mechanism. Disclosure of Invention The invention aims to provide an industrial equipment state intelligent monitoring system based on big data so as to solve the problems in the background technology. In order to achieve the above object, the present invention provides an intelligent monitoring system for status of industrial equipment based on big data, the system comprising: The industrial equipment operation data acquisition module is used for continuously capturing an original monitoring data stream from a sensor network deployed in the industrial equipment, and carrying out data cleaning and quality check on the original monitoring data stream to form a standardized equipment operation data set; The device state feature abstract module is used for carrying out multi-scale feature transformation on the standardized device operation data set, generating a device state feature image with time sequence association, and carrying out feature reduction on the device state feature image to obtain a device state feature abstract; The monitoring knowledge graph evolution module is used for analyzing the characteristic association rule implicit in the equipment state characteristic abstract, constructing an equipment state characteristic association network according to the characteristic association rule, and enabling the equipment state characteristic association network to be evolved into an industrial equipment state monitoring knowledge graph through a dynamic updating mechanism; And the intelligent early warning decision module is used for projecting the real-time acquired industrial equipment operation data into the industrial equipment state monitoring knowledge graph to perform similarity matching and deviation calculation, and triggering hierarchical early warning decision according to matching and calculation results. Preferably, the primary monitoring data stream includes vibration signals, temperature readings, pressure values, current-voltage waveforms, equipment control parameters, and environmental temperature and humidity data. Preferably, the performing data cleaning and quality checking on the original monitoring data stream to form a standardized equipment operation data set specifically includes: performing outlier rejection and missing value interpolation on the original monitoring data stream to obtain primarily processed monitoring data; performing dimension unification and numerical normalization on the primarily processed monitoring data to obtain standardized monitoring data; And aligning and integrating the standardized monitoring data according to the equipment numbers and the time stamps to form a standardized equipment operation data set. Preferably, the performing multi-scale feature transformation on the standardized device operation data set, and generating the device state feature image with time sequence associ