Predictive Failure Warning From Inspection Parameter Trends

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for monitoring and predicting the operational state of critical devices, such as those in power plants, lack effective early warning systems for potential failures, relying on periodic inspections and lacking predictive capabilities based on historical data.

Innovation Solution

A method and apparatus that includes an inspection module to obtain current device parameters, a prediction module to forecast future parameter values using historical records, and an early warning module to alert for potential failures based on predefined rules, integrating knowledge graph technology for accurate prediction and rule construction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If periodic inspection is used to monitor device state, then device operation continuity is maintained, but early warning capability for potential failures is lost

Engineering Contradiction:
Improveearly warning capabilityVSAvoidresponse time to potential failure
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary actions by predicting future device parameter values before actual failures occur. The prediction module uses historical inspection records to forecast parameter trends and issues early warnings when predicted values approach threshold values, enabling preventive maintenance before critical failures happen.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements feedback by continuously comparing predicted parameter values against preset threshold values. When predicted values approach or exceed thresholds, the system generates early warnings that feed back to maintenance personnel, creating a closed-loop monitoring system that improves reliability through timely responses.

Inventive Principle:
Principle #23Feedback

2Reliability

If historical inspection records are analyzed to predict future values, then predictive capability is improved, but system complexity increases

Engineering Contradiction:
Improvepredictive capabilityVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system applies partial action by focusing prediction efforts only on key parameters that have historical inspection records and are critical for failure prediction. Rather than analyzing all possible device parameters, the system selectively predicts parameters where historical data exists and where prediction provides meaningful early warning value.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The system uses copying by creating a virtual model of device parameter evolution based on historical records. The prediction module generates predicted future values that replicate the expected parameter behavior patterns observed in historical data, allowing the system to forecast device state without physically monitoring every possible parameter in real-time.

Inventive Principle:
Principle #26Copying

3Reliability

If early warning system is implemented based on predicted parameter values, then device safety is improved, but additional monitoring and processing resources are required

Engineering Contradiction:
Improvedevice safetyVSAvoidmonitoring resource consumption
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary prediction computations using historical inspection records that have already been collected and stored. By analyzing existing historical data to forecast future parameter values, the system avoids the need for continuous real-time monitoring of all parameters, reducing additional resource consumption while maintaining device safety through predictive capabilities.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4194984A1Method and apparatus for early warning of failure
Publication Date: 2023.06.14 BEIJING BAIDU NETCOM SCI & TECH CO LTD
  • EP4194984A1 patent drawingFigure 1~2
  • EP4194984A1 patent drawingFigure 3~4
  • EP4194984A1 patent drawingFigure 5

AI summary

Provided are a method and apparatus for early warning of a failure, an electronic device, and a non-transitory computer-readable storage medium, and computer program product, which relate to the field of artificial intelligence, particularly, knowledge graph and natural language processing technologies. The implementation scheme includes that: a target device is inspected (S101) in a current inspection period to obtain a current value of a target device parameter; a new value of the target device parameter in a new inspection period is predicted (S102) according to historical inspection records of the target device in historical inspection periods and the current value of the target device parameter; and early warning of a failure is performed (S103) according to a preset early warning rule and the new value of the target device parameter.