Power Protection Logic Filtering for Faster, More Reliable Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing protection, monitoring, and control systems face challenges in balancing speed and dependability, often leading to false or missed alarms due to sensitivity trade-offs, and require complex parameter setting for decision-making logic.
Innovation Solution
Implementing a machine learning model, particularly with recurrent neural network layers like LSTM or GRU, to process decision-making logic outputs, acting as a nonlinear low-pass filter, thereby balancing speed and dependability without increasing storage or memory requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If the sensitivity of the decision-making logic is increased to ensure quick reaction, then the speed of fault detection is improved, but the system becomes more prone to false alarms due to noise or transients
Solution Approach 1:
The patent applies preliminary action by implementing a counter mechanism that counts consecutive samples exceeding the threshold before triggering an alarm. This preliminary counting action filters out transient noise while maintaining sensitivity to genuine faults, resolving the contradiction between quick detection and false alarm reduction
Solution Approach 2:
The counter serves as an intermediary element between the threshold comparison and the final alarm decision. It mediates the conflict by requiring a predetermined number of consecutive threshold exceedances, thus smoothing out transient fluctuations while preserving genuine fault signals
2Reliability
If the sensitivity is reduced to avoid false alarms, then the reliability is improved, but the reaction speed decreases leading to potential missed alarms
Solution Approach 1:
The counter mechanism performs preliminary accumulation of threshold exceedance events before triggering an alarm. This allows the system to maintain high sensitivity thresholds while requiring multiple consecutive confirmations, thus avoiding both false alarms from noise and missed alarms from genuine faults
3Reliability
If a counter is used to count consecutive samples exceeding the threshold, then the reliability of protection decisions is increased, but the reaction time is slowed down
Solution Approach 1:
The patent applies partial action by requiring only a predetermined number of consecutive samples (not all samples) to exceed the threshold before triggering an alarm. This partial requirement strikes a balance between filtering noise and maintaining fast reaction time, resolving the contradiction between reliability and speed
Data Source
AI summary
An asset protection, monitoring, or control device is operative to execute a decision-making logic to process inputs and generate a decision-making logic output that comprises one or more time series, process the decision-making logic output using a machine learning model, and cause an action to be performed responsive to a machine learning model output.


