Adaptive Shift Detection in Power Supply Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing power supply systems face challenges in accurately predicting operating performance due to inherent system-to-system variations and changes over time, leading to unreliable detection of unexpected behavior that could cause damage.
Innovation Solution
A shift detection system that generates an adaptive model based on real-time monitoring of operating parameters, using a controller to group data points into bins and calculate nodes, determining an output function, and employing a change detection algorithm to identify significant shifts in behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If pre-calibrated models are used to monitor power supply system performance, then the monitoring system can operate with simple comparison logic, but the detection accuracy deteriorates due to system-to-system variations and changes over time
Solution Approach 1:
The patent transforms the static pre-calibrated model into a dynamic adaptive model that continuously learns from operational data. The monitoring system evolves over time by updating its parameters based on actual system behavior, allowing it to adapt to system-to-system variations and temporal changes while maintaining simple operational logic.
Solution Approach 2:
The patent changes the parameters of the monitoring model from fixed pre-calibrated values to adaptive parameters that are continuously updated based on operational data. This allows the model to adjust to variations between different power supply systems and changes over time, improving detection accuracy without significantly increasing system complexity.
2Productivity
If static pre-calibrated models are used, then the monitoring approach is simple and fast, but it cannot reliably detect unexpected behavior due to inherent variations in power supply systems
Solution Approach 1:
The patent performs preliminary learning during a setup phase where the system collects operational data and builds an adaptive model before actual monitoring begins. This preliminary action allows the system to capture the specific characteristics of each power supply system, ensuring reliable detection from the start without compromising monitoring speed during operation.
Solution Approach 2:
The patent implements a feedback mechanism where the monitoring system continuously compares actual operational parameters against the adaptive model and uses the deviations to update the model. This feedback loop enables the system to maintain high detection reliability by adapting to actual system behavior while preserving fast monitoring performance through efficient comparison logic.
3Adaptability or versatility
If universal pre-calibrated models are applied to all power supply systems, then the monitoring system can be standardized, but it fails to account for individual system variations and wear over time
Solution Approach 1:
The patent enables the monitoring system to self-configure by automatically collecting operational data and generating its own adaptive model during a preliminary learning phase. This self-service approach eliminates the need for manual customization for each power supply system while still capturing individual system characteristics, balancing adaptability with ease of deployment.
Solution Approach 2:
The patent creates a universal adaptive modeling framework that can be applied to any power supply system type. The same basic methodology and algorithms work across different systems, providing standardization while the data-driven nature of the model allows it to adapt to specific system characteristics, wear patterns, and operational conditions of each individual system.
Data Source
AI summary
A shift detection system and method may monitor at least first and second operating parameters of a power supply system over time, and generate an adaptive model based on values of the first and second operating parameters. The adaptive model may include data points defined by the values of the first and second operating parameters. The data points may be grouped into bins according to designated ranges of the first operating parameter, and nodes may be calculated for individual bins based on the data points within the bins. An output function may be determined based on the nodes, and a shift incident may be detected based at least in part on an offset between the output function and recent values of the second operating parameter. A control signal may be generated in response to the shift incident to control the power supply system and/or notify an operator.