Laboratory predictive maintenance method and system based on equipment portrait

By preprocessing energy consumption data of laboratory equipment operating environment and constructing equipment profiles, the problems of difficulty in quantifying equipment degradation symptoms and lag in health status determination in existing technologies have been solved, realizing unified representation of equipment health assessment and risk prediction and closed-loop linkage of maintenance actions.

CN122453379APending Publication Date: 2026-07-24CHINA RAILWAY CONSTRUCTION ENGINEERING GROUP
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
CHINA RAILWAY CONSTRUCTION ENGINEERING GROUP
Filing Date
2026-04-28
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Existing predictive maintenance methods struggle to establish consistent event correlations before and after alarms are triggered in laboratory settings, resulting in difficulties in timely quantification of equipment degradation symptoms and delays in health status assessment.

Method used

By collecting energy consumption data of the operating environment in real time, performing preprocessing, dividing the operating segments according to equipment identification, extracting event anchor points, constructing event response windows to generate window feature vectors, training equipment operation disturbance coupling models, generating equipment profiles, and combining sliding window linear regression to calculate equipment health assessment values ​​and risk prediction models, a set of maintenance actions is generated.

Benefits of technology

It achieves a unified timescale and comparable dimensions across devices and systems, improves the ability to capture short-term electrical parameter distortion, environmental disturbances and mechanical vibration coupling symptoms, forms a quantifiable set of degradation trend parameters, reduces the dependence of maintenance action selection on human experience, and improves the reusability and consistency of maintenance plans.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122453379A_ABST
    Figure CN122453379A_ABST
Patent Text Reader

Abstract

The application discloses a laboratory predictive maintenance method and system based on equipment portraits, and relates to the technical field of equipment operation and maintenance management, comprising the following steps: S1, collecting real-time operation environment energy consumption data, and performing preprocessing on the operation environment energy consumption data; S2, generating a window feature vector by constructing an event response window, constructing and training an equipment operation disturbance coupling model, outputting an operation disturbance coupling index, and generating an equipment portrait; S3, calculating an equipment health evaluation value, constructing and training an equipment risk prediction model, and outputting an equipment risk prediction value; and S4, triggering an alarm and generating a maintenance action set based on the equipment risk prediction value and a risk threshold, and searching for a matching historical treatment record to generate a recommended maintenance action set. The application solves the problems that in the existing predictive maintenance, multi-source operation data is difficult to form consistent event correlation before and after the alarm is triggered, and degradation signs are difficult to be quantified in time and health state determination is lagged.
Need to check novelty before this filing date? Find Prior Art