Lift Sensing and Tracking System (concept note)

Introduction

Sensors such as the pitot-static system, angle of attack sensor, lift reserve indicator, gyroscope and accelerometer that provide pilots and automated control systems with inputs needed for effective aerodynamic control of aircraft. The failure of such sensors, while rare, can be catastrophic as in the case of Ethiopian flight 312 (2019), Lion Air flight 610 (2018) and Air France flight 447 (2009). Over the last 15 years, an estimated 2500 air crashes are attributed to stalls or malfunctions of anti-stall systems. Sensor failures were a factor in a majority of these crashes.

The proposed Lift Sensing and Tracking System (LSAT) would measure aerodynamic lift by using an array of strain gauges affixed to the load bearing members of the airframe. LSAT would work in tandem with the air data reference (ADR) and inertial reference(IR) components.

Mechanism

In fixed wing aircraft, the fuselage supports the wings while the aircraft is on the ground. The wings hang off the fuselage creating deformations along the wingspan and on the fuselage. Once airborne, the wings support the fuselage and deformations are caused in approximately the opposite direction.

LSAT would measure, during flight, the relative force/torque within and between the fuselage and the wings using an array of strain gauges placed along potential points of deformation. Airframe designs specify the expected deformations due to various forces allowing for straightforward determination of proper installation points. Commercially available strain gauges, like thin film, semiconductor or diffused semiconductor may be employed after considering frequency response characteristics, lifecycles, maintenance levels, size etc.

The data from the sensors would then feed into the air data computers for processing and then into flight control systems, anti stall computers and autopilot as well as to flight displays.

Thus, elements of the system would be:

Strain gauges placed along anticipated points of deformation
Electrical wiring from the strain gauges to the air data computer or other control systems
Updates to stall algorithms in stall prevention computers, autopilot and flight control computers to utilize the new data
Updates to flight displays

Justification

Given its criticality in sustaining flight, the potential or actual aerodynamic lift needs to be measured. Currently this is done by measuring air pressure differentials on the top and bottom surfaces, air speed, angle of attack etc, which provide the inputs to compute predicted lift. So pitot-static system measures static and dynamic pressures, which in turn is used to compute air speed. Angle of attack detection uses either the pitot static system or an alpha vane. Stall prevention systems use these computed lift to decide whether to take corrective action or not.

In comparison to the ADR, the LSAT uses a different type of input and a circuitry, and the physical location of the sensors is different as well. Apart from providing redundancy, LSAT would be less error prone than the ADR unit.

Supporting arguments to measure the lift force(the effect), instead of measuring the causative phenomenon (inputs) and then computing the effect.

Errors due to Modeling Complexity and Simplification: Aircraft and the air around it form a complex system, whose operation is commonly understood using simplified models employing classical physics - Newton’s Laws, Bernoulli’s equation. However the full physics of a real aerodynamic flight contains a host of complex interactions between the elements - and in different states of temperature, pressure, altitude and so on, such that it is difficult to model and precisely predict due to the impracticality of measuring all the parameters that are in play.
Errors due to Sensor damage and Inherent Errors: The external sensors like the pitot-static tube and the alpha vane, though statistically highly reliable, have inherent vulnerabilities due to their placement outside of the aircraft. They can be affected by ice, moisture, insects and bird hits. There are also various classes of inherent errors that these sensors are subject to.
Sensor Latency: For instruments that rely on air flowing through substantial spans of tubes, latency issues exist
Small sample size and assumptions of homogeneity: The air envelope is non-homogenous, yet current sensor data only picks a limited number of samples. To pick up larger number of samples, the cost burden is substantial. Errors arising out of this will affect the accuracy of all downstream computations.
Quicker troubleshooting: Lift can be adversely affected by - AoA, airspeed, turbulence, icing, altitude/air density, damage to flight surfaces, dirt, bird hit etc. If the actual lift produced and its change is known, it is easier to triangulate on a problem when it arises. For eg: icing of the wings and other flight surfaces is a known hazard. Any lift reduction due to icing would be detected early by LSAT which provides certainly which can then trigger an analysis of other sensor data to find the reason and then mitigation can be done
Special conditions: There are a host of conditions where air pressure measurement will not work effectively like during turns, low speed flight etc.
Handling disagreements in sensors: Typically there are 2 to 3 AOA sensors. When they disagree, the common approach is to average out the values after removing the outlier - and this practice has caused crashes where the presumed outlier was actually providing the correct reading. In the case of strain gauges, it is possible to deploy a larger number of sensors (maybe 20 or 100) and given that there is an anticipated pattern for force distribution on the airframe, it should be possible to eliminate incorrect reading in a simpler fashion.

Additional Applications

Catastrophic failure of control surfaces especially during turbulence and compounded by pilot actions have caused crashes. The strain gauges, if deployed to control surfaces and other potential failure points, allows early intervention as the critical members are exposed to loads approaching design limits.

Additional Components

A digital twin display showing the 3D model if the airplane, and the following sets of parameters overlaid on it with adjustable look-forward(projected values) and look-backward (historical data showing rates of change):
1. Position of various flight & control surfaces
2. Forces detected on the structures & surfaces due to aerodynamic effects
3. Actual Pilot inputs, and forces detected on the structures & surfaces due to those
4. Actual Automated System inputs, and forces detected on the structures & surfaces due to those
5. Warning when forces approach design safe limits
6. Aircraft response to pilot and automated system inputs

Document Properties

Authors	Reuben Jacob;
Date	April 2, 2019; April 6, 2019;
Topics	Aeronautics; Aircraft sensors; Aerodynamic lift; Measurement of lift in aircraft;
Status	Draft; In Review;
Revision	3;