Sixto L. Vazquez

Battery health management system for electric UAVs

This paper presents a novel battery health management system for electric UAVs (unmanned aerial vehicles) based on a Bayesian inference driven prognostic framework. The aim is to be able to predict the end-of-discharge (EOD) event that indicates that the battery pack has run out of charge for any given flight of an electric UAV platform. The amount of usable charge of a battery for a given discharge profile is not only dependent on the starting state-of-charge (SOC), but also other factors like battery health and the discharge or load profile imposed. This problem is more pronounced in battery powered electric UAVs since different flight regimes like takeoff/landing and cruise have different power requirements and a dead stick condition (battery shut off in flight) can have catastrophic consequences. Since UAVs deployments are relatively new, there is a lack of statistically significant flight data to motivate data-driven approaches. Consequently, we have developed a detailed discharge model for the batteries used and used it in a Bayesian inference based filtering (Particle Filtering) technique to generate remaining useful life (RUL) distributions for a given discharge. The results section presents the validation of this approach in hardware-in-the-loop tests.12

Structural Anomaly Detection Using Fiber Optic Sensors and Inverse Finite Element Method

NASA Langley Research Center is investigating a variety of techniques for mitigating aircraft accidents due to structural component failure. One technique under consideration combines distributed fiber optic strain sensing with an inverse finite element method for detecting and characterizing structural anomalies - anomalies that may provide early indication of airframe structure degradation. The technique identifies structural anomalies that result in observable changes in localized strain but do not impact the overall surface shape. Surface shape information is provided by an Inverse Finite Element Method that computes full-field displacements and internal loads using strain data from in-situ fiber- optic sensors. This paper describes a prototype of such a system and reports results from a series of laboratory tests conducted on a test coupon subjected to increasing levels of damage.

Predicting Battery Life for Electric UAVs

This paper presents a novel battery health management technology for the new generation of electric unmanned aerial vehicles powered by long-life, high-density, scalable power sources. Current reliability based techniques are insufficient to manage the use of such batteries when they are an active power source with frequently varying loads in uncertain environments. The technique presented here encodes the basic electrochemical processes of a Lithium-polymer battery in an advanced Bayesian inference framework to simultaneously track battery state-of-charge as well as tune the battery model to make accurate predictions of remaining useful life. Results from ground tests with emulated flight profiles are presented with discussions on the use of such prognostics results for decision making.

Automated anomaly detection for orbiter high-temperature reusable surface insulation

The description, analysis, and experimental results of a method for identifying possible defects on high temperature reusable surface insulation (HRSI) of the Orbiter thermal protection system (TPS) is presented. Currently, a visual postflight inspection of Orbiter TPS is conducted to detect and classify defects as part of the Orbiter maintenance flow. The objective of the method is to automate the detection of defects by identifying anomalies between preflight and postflight images of TPS components. The initial version is intended to detect and label gross (greater than 0.1 inches in the smallest dimension) anomalies on HRSI components for subsequent classification by a human inspector. The approach is a modified Golden Template technique where the preflight image of a tile serves as the template against which the postflight image of the tile is compared. Candidate anomalies are selected as a result of the comparison and processed to identify true anomalies. The processing methods are developed and discussed, and the results of testing on actual and simulated tile images are presented. Solutions to the problems of brightness and spatial normalization, timely execution, and minimization of false positives are also discussed.