A water temperature change diagnosis system based on deep reinforcement learning

CN122333170APending Publication Date: 2026-07-03QINGDAO YIHE XINSHUI TECHNOLOGY CO LTD
View PDF -1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
QINGDAO YIHE XINSHUI TECHNOLOGY CO LTD
Filing Date
2026-04-08
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

Existing water temperature monitoring systems rely on static thresholds, resulting in insufficient adaptability, low diagnostic accuracy, lack of physical interpretability, and inability to trace the root cause of anomalies, thus failing to achieve intelligent and robust development under complex operating conditions.

Method used

A water temperature change diagnostic system based on deep reinforcement learning is adopted. Through multi-dimensional data collaborative acquisition, deep evolution of time-series features, reinforcement learning diagnostic decision-making, dynamic reward feedback evaluation, and abnormal root cause tracing execution module, combined with thermodynamic and physical consistency verification, it realizes autonomous learning and physically interpretable multi-level structured diagnosis.

Benefits of technology

It significantly improves the accuracy of water temperature fluctuation detection under complex coupled operating conditions, reduces false alarm and false alarm rates, provides physically interpretable diagnostic results, and supports adaptive adjustment and efficient equipment maintenance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122333170A_ABST
    Figure CN122333170A_ABST
Patent Text Reader

Abstract

The present application relates to a kind of water temperature change diagnosis systems based on deep reinforcement learning, belong to the technical field of automated monitoring and diagnosis, the system includes: multidimensional data collaborative acquisition module;Time series feature deep evolution module;Reinforcement learning diagnosis decision module;Dynamic reward feedback evaluation module;Abnormal root cause tracing execution module;Thermodynamic physics consistency check unit;Safety constraint layer;Self-adapting working condition identification function unit;Historical trend comparison unit;Meta-learning unit;Digital twin verification unit.The present application builds the physical guided reinforcement learning diagnosis framework and man-machine collaborative closed-loop incremental learning architecture by root cause contribution feedback closed loop, physical consistency forced check and confidence entropy trigger artificial intervention and priority experience feedback mechanism, significantly improves the capture accuracy of water temperature fluctuation under complex working condition and diagnosis real-time, and provides clear physical basis for diagnosis result.
Need to check novelty before this filing date? Find Prior Art