Arc Sensor Training via Simulated Arcs

Solution Overview

Problem

Existing arc detection methods in electrical installations often fail to accurately identify dangerous arcs, leading to insufficient protection against fires and network unavailability due to false shutdowns.

Innovation Solution

An arc sensor is trained in the electrical environment by simulating and recording arcs, allowing it to distinguish between arc-related and other electrical events, thereby enabling reliable detection and appropriate switching off of affected parts of the installation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If frequency spectrum pattern recognition methods are used to detect arcs, then arc detection capability is provided, but measurement precision deteriorates leading to false identification of harmful events

Engineering Contradiction:

Improvearc detection capabilityVSAvoididentification accuracy

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The system performs preliminary adaptation by recording electrical signals during simulated arc events and non-arc events before actual operation. This pre-training phase allows the evaluation algorithm to learn and store characteristic patterns of actual arcs versus false events, enabling accurate distinction during subsequent monitoring operations.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system incorporates feedback through the adaptation process where recorded signals from simulated events are fed back into the evaluation algorithm for analysis. The algorithm continuously refines its ability to distinguish arc patterns from non-arc patterns based on this feedback, improving measurement precision through iterative learning.

Inventive Principle:

Principle #23Feedback

2Reliability

If arc detection methods are implemented, then protection against arc effects is improved, but network availability deteriorates due to false shutdowns

Engineering Contradiction:

Improveprotection against arc effectsVSAvoidnetwork availability

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system performs preliminary adaptation by recording electrical signals during simulated arc events and non-arc events before actual operation. This pre-training phase allows the evaluation algorithm to learn and store characteristic patterns of actual arcs versus false events, enabling accurate distinction during subsequent monitoring operations.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system incorporates feedback through the adaptation process where recorded signals from simulated events are fed back into the evaluation algorithm for analysis. The algorithm continuously refines its ability to distinguish arc patterns from non-arc patterns based on this feedback, improving measurement precision through iterative learning.

Inventive Principle:

Principle #23Feedback

Data Source

PatentEP2737330B1Method for adapting an arc sensor

Publication Date: 2021.07.21 EATON INTELLIGENT POWER LTD

Figure 1
Figure 2
Figure 3~5

AI summary

In a method for adapting an arc sensor (35) to a position in an electrical installation system, according to the invention a specifiable number of specifiable arcs are simulated and/or produced at least at a first position in the installation system, wherein after each simulated or produced arc, at least one current curve and/or voltage curve is recorded in a measured-value recording unit (2), wherein at least one characteristic of the recorded current curves and/or voltage curves is determined and stored, and the arc sensor (35) is trained for the electrical installation system.