AI-Based Series Arcing Detection in Hybrid-Electric Propulsion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hybrid-electric and electric propulsion systems in aircraft often suffer from series arcing, which leads to overheating, component damage, and performance issues, making it difficult to diagnose due to its intermittent nature.
Innovation Solution
A hybrid-electric propulsion system incorporating a thermal engine, electric motor, battery, battery management unit, electrical distribution bus, and an AI model with a monitoring system controller to detect series arcing by comparing actual and predicted operational parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional monitoring methods are used to detect series arcing, then the system structure remains simple, but the detection precision and reliability are insufficient due to the intermittent nature of arcing
Solution Approach 1:
The system performs preliminary actions by continuously collecting operational parameters (current, voltage, temperature, vibration) and training an AI model with historical data before arcing occurs. This pre-positioning of detection capabilities enables the system to identify subtle precursors to series arcing that traditional methods miss, improving detection precision while the AI model automates the complexity of analyzing multiple parameters simultaneously
Solution Approach 2:
An AI model serves as an intermediary between raw operational parameters and arcing detection decisions. The model processes multiple input parameters (current, voltage, temperature, vibration) and translates them into reliable arcing predictions, bridging the gap between simple parameter monitoring and complex diagnostic conclusions, thereby improving detection precision without requiring direct complex interaction between sensors and decision-making components
2Reliability
If AI models are implemented to predict operational parameters, then the detection reliability improves, but the computational energy consumption and processing time increase
Solution Approach 1:
The system applies partial action by selectively focusing AI processing on critical operational parameters and only performing full predictive analysis when anomalies are detected. During normal operation, the system monitors with lower computational intensity, reducing energy consumption while maintaining reliability through continuous but lighter-weight AI inference that can escalate to full analysis when needed
Solution Approach 2:
The AI model is trained offline using historical data, allowing it to self-learn detection patterns without requiring intensive real-time computational resources during operation. Once trained, the model performs lightweight inference that consumes minimal energy while maintaining high reliability, as the complex learning process has already been completed during the self-service training phase
3Measurement precision
If continuous monitoring of operational parameters is performed, then the detection capability improves, but the system complexity and data processing requirements increase
Solution Approach 1:
The system merges multiple monitoring functions (current sensing, voltage sensing, temperature monitoring, vibration detection) into a unified AI-driven analysis platform. By combining these separate measurement streams and processing them through a single neural network model, the system achieves comprehensive detection precision while reducing overall system complexity compared to maintaining separate analysis pipelines for each parameter
Solution Approach 2:
The AI model serves multiple functions simultaneously: it analyzes current parameters, predicts future operational states, detects series arcing, and identifies potential failure modes. This multi-functionality allows the system to maintain high measurement precision across multiple parameters while using a single versatile processing engine rather than requiring separate specialized systems for each function
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
A hybrid-electric propulsion system includes a thermal engine, an electric motor (34), a battery (32), a battery management unit (BMU) (50), an electrical distribution bus (30), an artificial intelligence (Al) model (48), and a monitoring system (MS) controller (46). The electrical distribution bus (30) electrically connects the electric motor (34) and the BMU (50). The MS controller (46) is in communication with the AI model (48), the electric motor (34), the BMU (50), and a memory storing instructions. The instructions when executed cause the MS controller (46) to: receive a first operational parameter from the BMU (50); control the AI model (48) to produce a predicted first operational parameter utilizing at least one electric motor operating parameter; and determine the presence of series arcing within the electrical distribution bus (30) using the first operational parameter and the predicted first operational parameter.