Texas Tech University (TTU) Office of the Vice President for Research

Institute for Materials, Manufacturing, and Sustainment (IMMS) Lubbock, Texas

Dy D. Le, Director

Intelligent State Awareness for Intelligent State Awareness for Fatigue-Free Aviation Platforms

Presented to: Rensselaer Polytechnic Institute (RPI)

Troy, NY Dec 5, 2018

!"#$

Presentation Outline •! Why “Fatigue-Free Aviation Platforms” ?

•! Machine versus human “Longevity Perspective” and the search for “Bio-Inspired Living Aerial Platform”

•! Defying “Impossibilities” and envisioning “Discoveries”

-! Finding & catching “Materials Damage Precursors”

-! Cloning “Digital Nanomaterials Architecture (DNA)”

-! Enabling “Reconfigurable & Self-Healing Elements” and “Intelligent Sensing Network”

•! Demonstrating “Health State Awareness” concept of operation for achieving “Fatigue and Maintenance-Free Aircraft”

•! Enabling “Digital Twin” and “Autonomous Component Tracking” to increase aircraft safety and fleet health management

•! Highlighting “Artificial Intelligence” perspectives

•! Presenting “Intelligent State Awareness” hypothesis

•! Summary

1

%$

•! Achieve “zero-maintenance” to reduce sustainment costs

•! Increase safety and availability

Intelligent Health State Awareness Vision for Automated Structural

Health Monitoring and Fatigue-Free Aviation Platforms

Unleashing

Revolutionary Capability

MRO: Maintenance, Repair, and Operation

2

Demand

2015

2024

$79B

$69B $69B

Types Types

!"#$

$69B

2015

2024

$69B $69B Projection

$69B $69B $69B $69B 21%

Source: ICF International

19%

$69B $69B

(RW)

22%

Fixed-Wing Aircraft

Source: Global Fleet & MRO Market Forecast Summary, Oliver Wyman

2017: $76B

2027: $109B 2027: $109B Projection

North America

Western Europe Asia

Pacific

China

Value Proposition Rotary-Wing

Aircraft

2 Rotary-Wing Rotary-Wing Rotary-Wing Rotary-Wing Rotary-Wing Rotary-Wing

Aviation MRO Projections

Risk Assessment

Quantitative Total Risk

Health State Health Treatment Longevity

Sustainment

Genetics Life Style/Stress-Fatigue

Accident/Transmission

Human-Machine Longevity Sustainment

Autonomous & Assisted Environment

Preventive

Autonomous & Assisted

Design Usage/Stress & Fatigue

Environment Accident/Threats

NOT Autonomous BUT Assisted

Preventive

cs.stanford.edu Courtesy of photosearch

Adaptive

NOT Adaptive NOT Autonomous BUT Assisted

cs.stanford.edu

Platform Design Multifunctional Structures Intelligence Next-Gen Platform Intelligence Intelligence Intelligence Next-Gen Platform Intelligence Next-Gen Platform

Bio-inspired Living Aerial Platform (BiLAP)

Living Skin

3

FEEL

Multifunctional Precursor ID

Data Fusion

Diagnostics Info Format Precursor-based Indication

Physics Models

Diagnostics

Prognostics

High

Hazard Mapping

Degree of Risk

medium Low

Negligible

Quantitative Total Risk

Functional Hazard Assessment Fault Tree Analysis

Decompose SSC&I into Components

Probable Occasional

Remote Improbable

Identified Risk Acceptable Risk

Unacceptable Risk Unidentified Risk

Risk Matrix

Types of Risk

Health State Precursors

Adaptive Controls

Safety Significant Subsystems & Interfaces

Indirect Adaptive Controls

BiLAP Prototype

Multi-Scale Modeling Nano-Macro Interfaces Assessment Method

Smart Materials

Living-Skin & Robots

Integration

1 0 1 0 1 0 1 1 0 1 1 0 10110110101 0 1 0 1 0 1 0 1 0 1 1 0 1 0 101010

Prototype Prototype

Distributed Sensing

& Transmission

Direct Adaptive Controls

Assess Remaining Capabilities: Structural Integrity and Survivability

Inform Risk on Demanding Capability ASSESS RECOVER & ADAPT

Quantitative Total Risk

Probable Occasional

Remote Improbable

Identified Risk Acceptable Risk

Unacceptable Risk Unacceptable Risk Unidentified Risk

Risk Matrix

Types of Risk

On-Board Nano-Robots

Fly-by-Feel

High

Hazard Mapping

Degree of Risk

medium Low

Negligible

Functional Hazard Assessment Fault Tree Analysis Fault Tree Analysis

Decompose SSC&I into Components

Health State

Safety Significant Subsystems & Interfaces Safety Significant Subsystems & Interfaces Safety Significant Subsystems & Interfaces

Nano-Macro Interfaces

Quantitative Total Risk Quantitative Total Risk Quantitative Total Risk 1 0 1 0 1 0 1 1 0 1 1 0 10110 1 0 1 0 1 0 1 0 1 1 0 1 0 10101Self-Healing Precursors

Useful Life Remaining

Science & Technology for “Bio-Inspired Living Aerial Platform - BiLAP

4

1 0 1 0 1 0 1 1 0 1 1 0 10110 1 0 1 0 1 0 1 0 1 1 0 1 0 10101

Prototype Prototype 1 0 1 0 1 0 1 1 0 1 1 0 10110 1 0 1 0 1 0 1 0 1 1 0 1 0 10101

Risk Assessment

Adaptive Controls

Indirect Adaptive Controls

BiLAP Prototype

Direct Adaptive Controls

Living-Skin & Robots

Prototype 1 0 1 0 1 0 1 1 0 1 1 0 10110110101 0 1 0 1 0 1 0 1 0 1 1 0 1 0 1010101 0 1 0 1 0 1 1 0 1 1 0 10110 1 0 1 0 1 0 1 0 1 1 0 1 0 10101

Digital Nanomaterial Architecture- Additive Manufacturing

Self-Healing/Advanced Sensing Network

Material Damage Precursor - Failure Correlation

Intelligent Health State Awareness Technology

5

Intelligent Health State Awareness Technology

CBM Tech

•!Benefits to operators: Enable aviation digital health monitoring, effective maintenance “big data” management - Immediate vehicle health state - Moving beyond Condition-based Maintenance (CBM) to informed material state-based awareness - Automated component health tracking

•!Empower operators with “breakthrough” technologies & capabilities to operate vertical lift aircraft with substantial reduction in maintenance costs

Present Future

Cos

ts

• Empower operators & capabilities to operate vertical lift aircraft with

Oxidation

Atomic Force Microscopy Elastic

Modulus Map of High Performance

Microfibers

Embedded Sensing

Corrosion

Rotorcraft Fatigue Crack Growth

Oxidation

Minimum Required

6

Nanorestructuring Prototype Materials “DNA” Cloning Nanorestructuring Materials “DNA” Cloning

Char

acte

rizat

ion

Build

ing

Bloc

ks

Engi

neer

ing

“Maintenance-Free” Aviation

“Fatigue-Free” Structures Materials by Design “Maintenance-Free” Aviation “Maintenance-Free” Aviation “Maintenance-Free”

Char

acte

rizat

ion

Build

ing

Bloc

ks

Build

ing

Bloc

ks C

hara

cter

izatio

n Materials Development Continuum

Materials Database – Computational M/S – Experimental & Digital Data

“Materials Genome Initiative for Global Competitiveness” – June 2011

DNA: Digital Nanomaterial Architecture

Extre

mely

Lig

htwe

ight

, Ad

aptiv

e, Du

rabl

e, an

d Da

mag

e Tol

eran

t Mat

eria

ls G

enom

e

Impr

oved

Reli

abilit

y an

d Lo

ngev

ity

7

•!Embedded microvascular networks within structural materials

•!Continuous transport of healing agents throughout structural lifetime

Can this technology be applied to composites materials with fiber

reinforcement in the resin?

•!Fire ants collectively entangle themselves to form an active structure capable of changing state from liquid to solid when subject to applied loads

Can we dynamically alter interconnections among subsystems to direct the flow of energy and entropy within networks to

achieve desired macroscopic properties?

Potential new process for new types of active, reconfigurable materials for structural morphing & healing, vibration attenuation, and dynamic load mitigation

%&'()

Hea

ling

8

&$&$

&$

Loading Threshold

Optimizing Solution for Sustaining “Fatigue-Free”

Delayed Failure

Loading Threshold

.. !.. !.. !.. !.. !.. !..

!.. !.. !.. !.. !.. !..

.. ..

....

..

..

1

&$

5

&$&$&$&$&$Fatigue Cracks

Optimized Load

Aviation Health State Awareness Concept of Operation

7

3

6

Extended Usage

%&*+&,$-&.&/.,$0$$12343/.&456&$

Assessing Damage & Risk

Incidental Damage Delayed Failure

7

6

Extended Usage Highly Damaging Load

4

735*.&*3*/&)84&&$9:&43;*<$=&45>?$2 &$ 2 Damage Precursor

Executing Adaptive Maneuvers

Optimized Load

9

Platform Health State Awareness Controller Dashboard Health State Visualization Integrated Artificial Intelligence

'$

@A%=B$C!D!E$VISP TN2124

()()()()()()$)()()()()()($()()()()()()$()()()()()()$)()()()()()($()()()()()()$)()()()()()($()(()()())()$()()()()()()$)(()()())()($()(()()())()$)()(()()())($()(()()())()$()(()()())()*+$

010101010101010101010101010101101010101011 101010101010010101 010101010101010101101010 010101010101 010101010101010101010101010101101010101011 101010101010010101 010101010101010101101010 010101010101

01010101 1010101 0101010101010101 0101010101001\01010101011 101010101010010101 010101010101010101101010 010101010

01010101101010101101010 0101101 1010101001\010101001\01010110101 0101010101010 001\010101101 0101010101001\01010101010111010

@&/>4?$73*&FG&4+$0$-H*3I5/+$

13:.F4&$-3I3<&$=4&/F4+>4$

J//&++$K3+&'5*&$

@F*$7L%$7>?&'5*<$

J++&++$$$$@5+M$$$ 1>*N<F4&$J?3:;G&$73*&FG&4$

()()()()()()$)()()()()()($()()()()()()$()()()()()()$)()()()()()($()()()()()()$)()()()()()($()(()()())()$()()()()()()$)(()()())()($()(()()())()$)()(()()())($()(()()())()$()(()()())()*+$

10

– Platform Health Management – !! Digital health monitoring !! Beyond CBM to informed health state-based awareness

– Beyond Intelligent Health State Awareness – !! Enable digital twin technology in near or real-time !! Allow autonomous component tracking

11 11

TTU IMMS-iDVL Collaborations

– Enable Digital Twin Concept – !! Update virtual life-limited components augmented with incoming data in real-time or near RT !! Determine remaining useful life based on actual usage and characterized damage

12

01010101 1010101 0101010101010101 0101010101001\01010101011 101010101010010101

Data uploading

TTU IMMS-iDVL

RFID

13

– Allow Autonomous Component Tracking – !! Aggregate “big data” to highlight patterns and trends from individual aircraft and fleet by serial

numbers – enable autonomous tracking of critical components

Aircraft Fleet at Glance

Zooming & Filtering

Details on Demand

TTU IMMS-iDVL Collaborations

-! Rule-Based AI – !! Good for well-defined problems and system

parameters with good known certainty !! Incapable of training and difficult to address new

hidden states and uncertainty - Statistical Learning AI -

!! Don’t follow exact rules but based on statistical models of certain types of problems – Deal with uncertainty & probability

!! Artificial Neural Network with different computation layers to process data

!! Couldn’t explain informed decision but could tell with level of probability

!! Difficult to train/address new hidden states - Physics-Centric Model Based AI-

!! Construct and/or update models in real environment & address new hidden states

!! Enable self training !! Capable of perceiving, learning, abstracting, and

reasoning

Probability of Outcomes

Machine Learning Decision Tree

Yd = f(Vi) + fe

Capable of perceiving, learning, abstracting, and

14

Cognitive capability with direct feedback

and learning

(b) Rule-based Pattern Recognition & Statistical Learning

(a) Artificial Intelligence Machine Learning (ML)

Recognition & Statistical

(a) Artificial Intelligence

Recognition & Statistical

(c) C

ogni

tive

Cap

abili

ty

(3) Facilitate system behavior change for sustaining longevity or self healing to disrupt failure cascade $

(2) Enable cognitive cueing and human-machine teaming, interaction, & communication in RT

(1) Identify Model of

Models properties,

e.g., ingredients of

longevity or

precursors of onset

of failure

Auto-Feedback Cog

nitiv

e C

apab

ility

(2) Enable cognitive Self-Learning

Achieve “Zero-Maintenance” thru intelligent comprehensive integrated solution

1

3

Goal

2 3 3 (a) Model of Models

(b) System of Systems

15

!"#$

Summary

!!Next-generation aviation platform needs multifunctional and intelligence capabilities to achieve “zero-maintenance” and “fatigue-free” vision

!!Extensive human-manual maintenance labor presents substantial cost burden for aviation stakeholders

!!CBM lack automation capability – and improving reliability - not a total solution

!!Advanced discoveries in materials damage precursor detection and characterization, materials genome, and self-healing are possible to help ease some poor reliability concerns

!!Health state awareness technology also enables Digital Twin concept and automated SHM and component tracking

!! In addition to rule-based and statistic learning, next generation of AI will include physics-models to provide cognitive capability including direct feedback and learning for Bio-inspired Living Living Platform

16

Dy D. Le, Director IMMS, Texas Tech University, Lubbock, TX

dy.d.le@ttu.edu

THANK YOU AND QUESTIONS?