Aircraft Fuel Quantity Indication With Neural Flow Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional aircraft fuel quantity indication systems exhibit an error margin of 1 to 3%, leading to unnecessary fuel weight, which affects environmental sustainability and reduces the capacity for passengers and cargo.
Innovation Solution
A neural network-based fuel quantity indicator system (FQIS) utilizing machine learning to process inputs from fuel tank sensors and fuel flow sensors, providing accurate fuel quantity estimation with an error margin under 1% by integrating fuel flow parameters and benchmark fuel consumption curves.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional fuel quantity indication systems are used, then the system is simple and reliable, but the measurement precision is only 1 to 3% leading to unnecessary fuel weight
Solution Approach 1:
The system segments the fuel quantity measurement problem into multiple independent measurement components: dielectric sensors for volume measurement, densitometers for density measurement, and water detectors for contamination monitoring. Each sensor type independently measures a specific parameter, and their combined data provides high-precision fuel quantity and composition analysis, resolving the contradiction between measurement precision and system complexity.
Solution Approach 2:
The system introduces an intermediary processing unit that integrates data from multiple sensor types (dielectric sensors, densitometers, water detectors) and applies correction algorithms. This intermediary layer processes raw sensor data, applies temperature and density corrections, and synthesizes accurate fuel quantity measurements, enabling high precision without requiring direct complex sensor-fuel interaction.
2Weight of moving object
If conventional fuel quantity systems with 1 to 3% error margin are used, then the system is easy to operate, but unnecessary fuel weight of 675 to 2,025 lbs must be carried
Solution Approach 1:
The system implements continuous feedback measurement where dielectric sensors continuously monitor fuel volume, densitometers continuously measure fuel density, and the processing unit continuously calculates and updates fuel quantity and composition. This real-time feedback loop enables precise tracking of fuel consumption and accurate determination of remaining fuel, allowing the aircraft to carry minimal necessary fuel rather than excessive reserves, directly reducing fuel weight while maintaining high measurement precision.
Solution Approach 2:
The system measures and compensates for changes in fuel parameters including density variations due to temperature changes, compositional changes from fuel blending, and contamination levels. By continuously monitoring and correcting for these parameter changes, the system maintains high measurement precision across varying operating conditions, enabling accurate fuel quantity determination that reduces unnecessary fuel carriage.
3Quantity of substance
If conventional fuel quantity systems are used, then the system structure is simple, but the capacity for passengers and cargo is reduced due to excess fuel weight
Solution Approach 1:
The dielectric sensor system serves multiple functions: it measures fuel volume, detects water contamination, and monitors fuel composition changes. The densitometers provide both density measurements for volume-to-mass conversion and compositional analysis. This multi-functionality allows the sensor system to perform comprehensive fuel monitoring while minimizing the number of separate devices required, thereby reducing overall system complexity while enabling weight reduction for increased passenger and cargo capacity.
Data Source
AI summary
A fuel quantity indicator system (FQIS) for an aircraft or other vehicle includes memory/data storage and a fuel quantity processing unit (FQPU). A neural network trained via machine learning and running on the FQPU receives fuel quantity (FQ) inputs from fuel tank sensors, e.g., tank density, fuel volume, water presence within the tank, at or near a given measurement time. The neural network additionally receives fuel flow (FF) inputs from flow sensors at the measurement time, indicating fuel flow to engines and auxiliary power units (APU) of the vehicle. Based on the FQ inputs and the FF inputs, the neural network calculates an estimated fuel quantity (EFQ) remaining, e.g., across all fuel tanks at or near a particular measurement time.