© 2015 Real-Time Innovations, Inc. The Communications Platform for the Industrial Internet of Things™ Sumant Tambe Real-Time Innovations, Inc. May, 2015.

© 2015 Real-Time Innovations, Inc.

An Extensible CBM Architecture for Naval Fleet Maintenance Using Open Standards

The Communications Platform for the Industrial Internet of Things™

Sumant TambeReal-Time Innovations, Inc.

May, 2015

Abdel-Moez E. Bayoumi, Alex Cao, Rhea McCaslin, Travis EdwardsCondition-Based Maintenance Center

University of South Carolina, Columbia, SC

Condition-Based Maintenance Center

© 2015 Real-Time Innovations, Inc.2

Outline• Overview

– Condition-Based Maintenance (CBM)– South Carolina CBM Center– Real-Time Innovations (RTI)– Data Distribution Service (DDS)– Open System Architecture for CBM (OSA-CBM)

• Open Architecture using DDS and OSA-CBM Standards

Consider the US industry is spending $260 billion each year on maintenance, of that $85 billion is lost on unneeded maintenance

3

Improper Maintenance Costs

33% is wasted on unnecessary maintenance

4

Maintenance Management Methods

Fix it when it fails

Continuous monitoring of actual conditions

Fixed time, calendar or operating hours

Time-BasedPreventive

Failure-BasedCorrective

Condition-Based· Connect· Analyze· Action

Predictive

Maintenance Practice

5

South Carolina CBM Activities and Objectives

2006

2004

1999

1998

1996

2007

2009

2015

2010

2013

Predictive Maintenance & Health Score

Vibration Control Initiative

Structural/Regime Recognition and Damage Fraction Calculations

University of South Carolina (USC) Helicopter Test Stand

Blade wedges improves accuracy in track adjustments/saves blades

AWR allows track and balance adjustments without test flight

USC Cost Benefit Analysis (Reduced O&S cost, MTF reduction)

Last change to DA PAM 738-751

VMEP development begins on AH-64 A

Tail and Main Rotor Drivetrain Testing

6

Partnerships and Collaborations

7

CBM Areas of Research

DoD Funding 2008$15M & 5 years

DoD Funding2013$15M & 5 years

IM_13_ 7148_finalResearch

Focus

Cost Benefit Analysis

(ROI)Component

Testing

Tribology (Friction,

Lubricants, Wear)

Advanced Signal

Processing

Natural Language

Processing

Predictive Analytics

Diagnosis, Prognosis, &

Life Prediction

USC Research Infrastructure ~ $8M

Interdisciplinary• Medical• Business• Engineering• Computer Science

6 Faculty Members5 Military Personnel2 Post Docs6 PhD Students6 MS Students8 Undergraduates1-2 HS Students

Data Collection

8

Benefits of CBM

Maintenance man hoursDays in maintenanceTime between overhaulDays pending repair

Total parts costPremature part failureMaterial cost avoidanceOperational cost avoidance

PerformanceMoraleSense of safetySense of time savingsConfidence

Direct cost benefits

Intangible benefits

Time savings

9

• ROI: Largest ROI of any Army funded project at 20.2:1

• Economic Impact: Overall Economic Impact exceeds $250M over the past 10 years

• Cost Avoidance: Overall Cost Avoidance for the US Army exceeds $125M over the past 10 years

• Testing Hours: Over 9000 testing hours completed on the USC Test Stands examining multiple faults

• TBO: Time Between Overhauls increased on multiple components by up to 500 hours

15-Year History of CBM at USC

10

Testing Capabilities at USC

Tail Rotor Drive Train (TRDT)/Main Rotor Swashplate (MRSP) on actual

AH-64D

TRDT/MRSPtest stands

at USC

FHB

#4#5

#6

TGB

AHBIGB

#3

Sensing• Vibration• Temperature• Speed• Torque• Metals in oil• Acoustic• Doppler Radar• Strain gauges

TRSP

MRSP

Objective:Demonstrate whether the aircraft could continue to operate with a leaking output seal until the aircraft reached a major maintenance event

11

Tail Rotor Gearbox Output Seal Study

Approach:Discovered that grease was moving freely between the main gear compartment into the static mast

Outcome:‒ Maintenance practice changed from replacing entire TGB

assembly to either replacing the output seal or continue adding grease

‒ Led to reduction in part costs, necessary downtime, and labor requirements

‒ New output seal removal tool was designed and is currently used in the field

U.S. Army Aviation and Missile Command | AMCOM Logistics CenterUNCLASSIFIED / FOUO

Cost-Wise Readiness (CWR) Solution through USC CBM Center Innovation & Maintenance Improvement

Return on Investment (ROI)

Technical Bulletin ReleasedOct 2009

Present Cost Avoided

10/09-12/13$12.68M

Project Cost $627K

ROI 20.22 : 1

Issue: Tail Rotor Gear BoxRemovals Due to Leaking

Required Depot Level Repair

Solution: Repair Developed by University of S. Carolina (USC) CBM

Center & Team Redstone

Followed by IETM Update

Yields within 51 Months Since Implementation:

AH-64 TRGB Output Seal Field RepairIncreasing Readiness & Reducing Need for Spares

Greatest ROI to Date

Avg. Cost Avoided: $2.98M/Year

Metrics Before After

Demands/ 10K FH 10.58 6.71

36.5

5%

324 Fewer TRGBs Demands in 51 Months

Benefits Up 8.84% from Prior Quarter; Significant Contributor to $53.25M Sum Costs Avoided

Objective:Produce a set of tools to be used as an indicator on the success of a system maintained by CBM

Approach:– Accurately simulate a full drivetrain– Use predictive analytics to determine component life and condition– Automate maintenance action reports article replacement– Correlate how maintenance affects supply chain

Outcome:– Longer time on-wing for PPMS monitored components– Extended TBO on articles– Higher morale for the soldiers– Increased sense of safety for entire crew

13

Sustainment Project

14

Preventive Predictive Maintenance (PPM)

Predictive Tools

Useful Life

Health Condition ...Design

Phase II – PPM: PHYSICS MODELING Phase III – PPM:

INTEGRATION & VALIDATION

Phase I – PPM: MEASUREMENT

CORRELATIONS & IMPLEMENTATION

© 2014 QMT, LLC. All Rights Reserved © 2014 QMT, LLC. All Rights Reserved

QMT & CBM@USC Confidential 15

Connecting Components with DDS

DATA BUS

Publisher Publisher

Subscriber

Publisher

Subscriber

Publisher

Subscriber

NI LabVIEW Modeling End User 1 End User 2

Sensor Data

• OSA-CBM specifies a standard architecture and framework for implementing CBM

• There are 4 primary OSA-CBM data classes: DataEvent, Configuration, Explanation and Extensible– DataEvent is the dynamic data

generated by an OSA-CBM module such as measurements

– Our prototype implementation of OSA-CBM uses DataEvent and related classes 16

OSA-CBM

© 2015 Real-Time Innovations, Inc.

Real-Time InnovationsStandards-based Open Architecture forCondition-Based Maintenance

17

© 2015 Real-Time Innovations, Inc.18

Industrial Internet of Things (IIoT)

© 2015 Real-Time Innovations, Inc.19

RTI Provides Connectivity Platform for Industrial IoT Systems• Industrial IoT Systems Interact with the Real World• Must adapt to changing environment• Cannot stop processing the information• Must react at the speed imposed by the environment

© 2015 Real-Time Innovations, Inc.20

DDS: Data Connectivity Standard for the Industrial IoT

© 2015 Real-Time Innovations, Inc.21

RPC over DDS

2014

DDS Security

2012

Family of Specifications

DDSImplementation

Network / TCP / UDP / IP

App

DDSImplementation

App

DDSImplementation

DDS Spec

2004

DDSInteroperablity

2006

UML Profilefor DDS

2008

DDS forLw CCM

2009

DDS X-Types

2010 2010

DDS-STD-C++DDS-JAVA5

App

© 2015 Real-Time Innovations, Inc.22

“Global Data Space” generalizes Subject-Based Addressing• Data objects addressed by Domain ID, Topic and Key• Domains provide a level of isolation • Topic groups homogeneous subjects (same data-type & meaning) • Key is a generalization of subject

Data Writer

Data Writer

Data Writer

Data Reader

Data Reader

Data Reader

Data Writer

Domain

Topic

Instance

Key (subject)

Data-Centric Model

© 2015 Real-Time Innovations, Inc.23

Open System Architecture for Condition-Based Maintenance (OSA-CBM) Standard

OSA-CBM Functional Architecture

• OSA-CBM specifies a standard architecture and framework for implementing CBM

• The functional blocks are often provided by different vendors/teams due to vastly different skillset required to implement each

• An end-to-end CBM system must integrate them together somehow…

– Point-2-point– Bus Architecture

© 2015 Real-Time Innovations, Inc.24

Advisory Generation (AG)

Prognosis Assessment (PA)

Health Assessment (HA)

State Detection (SD)

Data Manipulation (DM)

Data Acquisition (DA)

Inflexible N2 Point-to-Point Communication

• Complex• Inflexible• Tightly-coupled• O(N2) connections• Not scalable• Encourages

custom protocols• Potential vendor

lock-in

© 2015 Real-Time Innovations, Inc.25

Advisory Generation (AG)

Prognosis Assessment (PA)

Health Assessment (HA)

State Detection (SD)

Data Manipulation (DM)

Data Acquisition (DA)

OSA-CBM Data Model in DDS X-Types

Flexible Open Extensible Data Bus Architecture

• Simpler• Flexible• Loose-coupled• O(N) connections• Scalable• No custom

protocols• No vendor lock-in

© 2015 Real-Time Innovations, Inc.26

Example:Types for Data Acquisition (UML)

© 2015 Real-Time Innovations, Inc.27

OSA-CBM Compliant

Data Acquisition

XML Representation

X-Types Representation

DDS DataWriter

OSA-CBM Compliant

Health Assessment and

Prognostics

XML Representation

X-Types Representation

DDS DataReader

OSA-CBM Data Model in DDS X-Types

CBM Dataflow via DDS

Transformation

Write

Reverse Transformation

Publish Subscribe

Read

ProcessAcquire

© 2015 Real-Time Innovations, Inc.28

Modeling Data Acquisition Types using RTI Connext DDS

Option 1 Option 2

1. Manually write IDL for OSA-CBM (lot of it)2. Use RTI code-generator3. Write glue code from source data model

to the IDL-defined data model (again, lot of it)

4. Done!5. Example: Cranfield University [1]

1. Generate code from the OSA-CBM XSD using a third party code generator

2. Describe application-level types in C++ using RTI RefleX [2]

3. Auto-generate glue code from C++ application data model

4. Done!5. Example: RTI Github OSA-CBM [3]

[1] Tarapong Sreenuch, Antonios Tsourdos, Ian Jennions, and Peter Silson, “Application of the Data Distribution Service for Implementing OSA-CBM Standard”, Infotech@Aerospace 2011.[2] http://rticommunity.github.io/rticonnextdds-reflex/ [3] https://github.com/rticommunity/rticonnextdds-reflex/tree/master/test/osacbm

© 2015 Real-Time Innovations, Inc.

Thank You!