ow

Acknowledgements

The author wishes to acknowledge the NASAInnovative Partnerships Program (IPP) forsupporting the development of the technologiesthat enable Integrated System HealthManagement (ISHM) capability. Special thanksto Dr. Ramona Travis, Manager of the IPP atNASA Stennis, and to John Bailey, who as Chiefof the Science and Technology Division, hasbeen instrumental in making possibledevelopment of the ISHM area.

Outline• Motivation• Concepts and Approaches

— ISHM: Background/Definition— ISHM Model of a system— Detection of anomaly indicators.— Determination and confirmation of anomalies.— Diagnostic of causes and determination of effects.— Consistency checking cycle.— Management of health information— User Interfaces

• Implementation• Conclusions

WE-I

A-1

Support rocket engine testmission with highly reliable,

accurate measurements;reduced costs; etc.

B- I/B-2

K :^ R^ ,w^ 3 A-2

M-1 L

JQMI

Requirements Driving ISHMThrough comprehensive and continuous vigilance

• Improve quality— By more accurately understanding the state of a system.

• Minimize costs— Of configuration— Of repair and calibration— Of operations

• Avoid downtime— By predicting impending failures— By timely intervention— By faster diagnosis and recovery

• Increase safety (protect people and assets)

ISHM Objectives

• Use available data, information, andknowledge to—Identify system state— Detect anomalies— Determine anomaly causes— Predict system impacts— Predict future anomalies— Recommend timely mitigation steps— Evolve to incorporate new knowledge

ISHM implementation is a problem of "management" ofdata, information, and knowledge (DlaK) focused onachieving the objectives of ISHM

Concepts and Approach

OH

aQU

W

O

n

ISHM is Being Done Now ....But

International Space Station Rocket Engine Test Stand

azdN

W

WW

..---•-•-----•- - - - - -: ------ ---.-•-•-•-•-•- - - --

Layer 2 - - — -Astronaut/ "` R

Test JA

Conductor

Layer 3ControlRoom

-------------------- •---------•--------E_-- ---- ------------------

Layer 4 ^s^

BackControlRoom

Layer 1Vehicle/

Test Stand

Signalthreshold

violationdetection

AddedDIaK fromon-boardusers.

AddedDIaK frombroadgroup ofexperts.

AddedDIaKresources

from largercommunity

IntelligentElement

ProcessesIntelligent

ControlProcesses

Data, Information, and Knowledge ManagementArchitecture for ISHM (Information Architecture)

IntelligentSubsystemProcessess

IntelligentSystem Processes

ActuatorIntelligent

Sensor IntelligentProcesses om_oon en is

Classic architecture describinghow systems are built

System of1 Systems1 1

------ -- - - - - - - - - - - - ----- -----•------ 1

1 Pump Storage II System 1 1I II I I

^ I II I pI ^ II I II I ^ ^ 11 I I

11 1 Pump Motor Bearing Tanks Valves Sensor II Sub- Sub- Sub- I 111,II system ystem system

1 ^ I^ I I' ^ ISample System_ _ _ _ _ _ _ _ _ _ ^^ ,Generic System_

to

K

Intelligent SystemProcess

0

OxidizerSubsystemProcesses

IIntelligent Subsystem IProcess

Intelligent SubsystemProcess Over-Pressurization

LeakingPressure collapse

k

Intelligent SensorProcesses

ValveProcesses

OpeningClosingleaking

19c;Intelligent Components

Correspondence between elements in the ISHMInformation Architecture and processes taking

place in a system

Intelligent I J I IntelligentProcess Process

IntelligentProcess

Tray r:YY a oFl^f W Gal'4Q WgGp. m0

nY-,lGx w

Gbvltl

t— rc-rhwo

V.

V I ^^YV-IliSFA

'^ doad

._ YV.I I]HU]

- tlp., eA^A ^ ! .^.

' vrY-t1t74p -.

'rr.l@0649

YY.t 176IIY I Y4., ITFQ.I

.1131

G4wo

a! opn

YY-116}GY

rrrla

pN-iGCRC

MTTPGM

AnomalyDetection:Leaks, etc.

Intelligent

Sensors: IEEE

Standard+Health

`_ eena

YY11GbGY AFF 11afl0 rc.I IIIi:N R1,&OY -,^,g.:^u

t ,^, P4111FgY

dDpGo To Part View Part —

1.:1

n

SSC Integrated System Health Management (ISHM)Capabilities

Health Anomaly Database:_TU'

ISHM Models (Embedded Data, Health Electronic Data SheetsInformation, and Knowledge): Repository of anomaliesMTTP Implementation

Embedding o wol Integrated Awareness:Predictive Models RootCause 3-D Health Visualization ofAnalysis MTTP

i

Leak Pressure Leak Decreasing Pressure

I1^& a-subcomponent-of II& encompassing ^ II^

Is _process-equipment is2pressure-subsystem pressure_sensor

GENERAL ATOMICSELECTRO11WAGNErIC SYSTEMS

RowanoUniversity

Pratt & WhitneA United Technologies Company

MOBITRUMBrings you mobility""

PWI^NVOCONJINalwe

NVE^NVE CORPORATION

Rowan

COLLEGE OFVITHAIN&oor

OxidizerSubsystemProcessesIntelligent System

Process

temIntelligent SubsystemProcess Over- Pres ation

. LeakingPressure collapse

Detection and Confirmation of AnomaliesConsistency Checking Cycle

Activated Model

res

Activated ModelIntelligent I J Activated Model

Process

Health Ilealtn Healin

intelligent nsorProcesses

Health Health Health Health

Intelligent Components

'

ValveProcesses

OpeningClosing

MTTP Embedded DIaK

Root-Cause Tree

Leak

Gas Detection Alarmurithin-the-active-range-^f

isE_p roc-ess-equipment Jas-sensor

Leek

—a-sid component-of Lead Abnormally changing pressure Abnormally changing pressure^^, II _ _ eilGOrflpaSsIr1CJ ^^ ^I

Is? process-equipment

erg-isolation- -of isZ_flow-s^ubsystem isZ flow-subsystem pressure sensorLeak through

Is2 valve

ISHM Enabling Technologies-.Root Cause Analysis

Test Time sq rqTine of Oey Oete 22 Jun 2667 9.51.2 359 am.TO Marker 6.6 ^ ^ PacketeJS ecmd Ok

prea auiixable-subsystem preaaurix'b -euheyatem

?n:a=mricdrF: ?n±an^irindak^ presaurixe6le-subsystemRV-1182-G0 PE-1140-G0 closed

I^RV-1183-GM

claaetl claaetl ctoaedM V•1168•GN

I= PE-1134-GO PE-1143-GO151 -2218.371 MV-1138-GO PE-11¶71-GM PE-1 184-Gh1 PE-11 ^i98-GM

ID

VV I I 1^ RO-1166GM II /ice, claaetl— ^ r'^i 4. closed ^^O-11i -GO i closed — —I .'^ closed — ^I_

MV-1135-GO ^J I ^ M V-11 E6 G MLVy VP V-1139 -G 0 UP V•1176•G M

TT-1208-GO TT-1;5-GM-2218'371 iclosedclosed ^, close .closed

MV-1136-GN MV-1176-GN MV 11 W-GN MV•1139•GN

GOXPct

T+1G CH4

GN

GOX-1G M•1

vclosed closedGN

IGN-CURRENT —

I8 V-1134-G M I ^8V-11 95-GO

GN-1

GCtt9

GDR-2 GM-2

GN

n

GN-2

Virtual Intelligent Sensors• Provides benefits of ISHM capabilities to existing

data acquisition systems by adding VirtualIntelligent Sensor capability. I ISHM

To Control Room & NCAPexisting applications

Sensor1

Sensor 2• DAS

Sensor N

VIRTUAL INTELLIGENT SENSORSmart

•TEDS-NCAP NCAP

Intelligent-Other EDS-Health Algorithms

0

z

Sensor INTELLIGENT SENSOR I NCAP

ISHM Implementations

OSA•EAIOpen S;ilkl irds 0,vd.tw wre forEnterprise Applica!ror Iraegration

i

Site-wideISHM

USA-EAI Element

Data Pepository ISHIM

Spacecraft ProcessingMobile LaunchLaunch Pact

o % Solid Rwket Processing

Generic Architecture to implement ISHM capability forsystems with conventional equipment, with option toincorporate advanced smart/intelligent sensors and

actuators.6

OSA-CBM and OSA-EAI Standards on Ethernet l]^^^----- ^- -I----4---- -- ------ - ---- -----------

^r y

u------------------------'o^ COMponunt Heijllh f

Heallh (OEM) Node i }I I ^ System I I

System ISHM/GOHM Direct Connection I o Supervisor I IISHIM -- — I I I

Computer — I I II System Bus i I

II = Iit

I - I II '~ PLC

-I I

System Health Operator Control and I I1 and Status Display Display 1 DAQ I I1 1451.1 and CSSA-CBM Standards on Ethernet I I I

I i^ Control Control PC1 Supervisor

I II I I

I1455 1451 OSA-CBM OSA-CBM

I 'IN, I

Sensor 1 Sensor M Senor Senor 1 p

Node Noce 1 I I

1 I DAQ PC I IEU Da-uVRaw Umnfs a I

.^ Conventional System I

IS!>k^1 Swrso sEEE 1451.1

GSA-CBM I - Data RepositoryIEEE

'Open S°.are^fds I __ I

^I Ychireuurb — — — —^^ — —°^ — _ —^^^ — — — —

Virtual Intelligent standard for r4CAP for Condi5w-Based -- ii Conventional Control and Data Acquisition System 1 1Cnmrm,,icalia^ klainlerancflSensor Environment - — — — — — — — — — — — — — — — — — — — -- — — — f I1---------------------------------------------------

a — — — — — — — a — ^ Facility ISHM 11

PURGE

Top level view of the 1S& model of theLaunch Complex 20 Facility at NASA Kennedy

1 1 = Space Center

lox•tlevrar-tsw

PURGE

Sensor anomaly indicators detected by an intelligentsensor during a pump test using the LC-20 facility at

NASA Kennedy Space Center

L01_MTAT001_ TEMP_F(Ambient Temperature)

1000

500

a^C 0950 1000 1050 1100 1150 1200 1250

Noise Detection Resultw 1.5

Ms 1U 0.5

0^- -0.5

950 1000 1050 1100 1150 1200 1250Disconnect Detection Result

w 1.5r

^' 1LL 0.5^ 0

~ -0.5950 1000 1050 1100 1150 1200 1250

out-of-Range Detection Resultw 1.505 1

0.5

3 0

-0.5950 1000 1050 1100 1150 1200 1250

Time (sec)

4CV Screenshot of the ISHM model of the LC-20 facility atKSC showing detection of a valve leak created by

opening the valve manually

fie Edt Yew layout Siff Prolea Wirkspate Tools WYIdoW Help

=_, ? c c °. ,, •'I T - our ^,.• f, .) ^ a1 E p ®i ; X --' ®0 UP i 8 _^ too-/. User Mode Devebper Go ToEDS Explorer 4 x

v

Real-Time Plot of GNPT104 'a VJR Workspace

EDS List 35 GNM V 109O GNPT104 ,Hy GNPT103

® Processes 30 LJGNPT104ff ® CEDS II

Lit ® HEDS 25 pg 7qIV^I

- ®TEDS

® Pressure Calibration Maintenance R 20 _q_E 0:a IIII l GNC P104Pressure Transducer-EDS GNPG104 fr YY

© Basic TEDS 75G

GNM V179< > 10 GNCP103EDS Attributes

O Basic TEDS 5-GNRGlO3 GNRG104

Template ID UTemplate Control READ-ONLY 0

20:1 za zo 20:17:25 20:77:30 q ^_ -q .Manufacturer ID SOR 8 Tue A r 2008 Currant Tim HH:MM:SSModel Number ISGTIKAA GNOR108Version LetterVersion NumberSerial Number

GNRV105 GNMV11

^loclose,

T ONIP101l^

Package Browser o x

o U Loaded PackagesO 4h VISE Interface lassesGE 4h EDS Classesto 4L Class Browser Classes

Alarrns 4 x : Root Causes

Val s4j-Z:JAJl tj^-Jeje v

Target Event \ ame Va... — StatusC At= 11121

(- e l m P I: 1^`I . • . I 1 1 121 1Target Event Same Value ^

T.lunction-o4^-Pressure-Su... Higher Than Expected Pressure false

_ m

Gnptso Higher Than Expected Pressure false

Ali

41 n

_Expanded causal-directed graph generated by the_detection of a leak in the subsystem where a valve was

1 opened manually (injected leak)

P Fes' z#z

Leal.

oIFa•aes \,

LeaK

^^LCIII-el Tn>F= \'IUI1 .7t-ju nt.twn •O^ recce re-

Fai {r Tn4n C. C rCEI&Sl n:

I I1-JUnCllon.-047-preS6nre-GU",

Lc PI:

on cplQ4

Y.-^1] PAC1f4n - 0^- F'rO6SV re- 6 V b6 y61em

Hqn< Tnan Ei;:.ecli7 PrG66llre

RUEgn r l l -04

T - Pi &NUre - GUt}6y6^Qn

' Than MXP*CGa6PfO9Cur4

04

A"OrMally t:nangng pres6ufeR &x-jun tic.n }ik3-prat cu m- cu Gc y. clean

•i2+^ermaUy Ch nn9r^8 P re a€wre

1-li nclic•n -047- pne L6ure- Ell n6y6Gern

4nan^^ng P nC Efvn6

gnptl 04

An p1184

Causal directed graph dynamicallyLeakI^ generated from events detectedP 0

during the simulated leak atLeak.

I ' GNCP104ppa• #3

pn mvs $ 8

Window Hey

(ra Get NI Sen- TEDS :1 Clew NI Sensor TEDS

CSGIPA System IPA-RUNTANK

LOXLOX n IJ

^

^^ I

H2ORzo

I^ t

LOX System t\ .

CSGWater System

File Edit _ Layout Go Project Workspace _rook

.tl Connect W TEDS Database a DsconneR from TEDS Database

DS Explorer 4 x

EDS Listp O PIE-14A41o1-IPA

Cpx TEDS-® Basic TEDS

1....0 Bridge Sensor

^^ .3t, =ii

Pilot ISHM Implementation:,hemical Steam Generator (CS GI

slue Line.

B LU E LBV E-DE M 0-S E N 50 R

dins. UP

i0

Bridge Element Impedance (Se; 351.3 ^ —Bridge Type FullCalibration Datc 2008-07-31 n

u

Calihration Period

Calibrators urinals365

TS @03232009

Electronic Datesheet NameFull Scale Electrical Value Percician

Bridge SensormV/V

,,

10^,,^

-

IEEE 1451.4 Template D

Mapping Method31Linear

Real-Time Plot of PE-14A4101-IPA1200

Maximum Electrical Output (V(V) 3.034Maximum Excitation Level (V) 15 1000 - G^lif<dule

Maximum PhyiscalValue 2000Measurement Location iD 9101 800

Minimum Electrical Output (VJV) 0.031Minimum Excitation Level (V) 10 600

Minimum Phoical Value a18:39:40 18:39:47 78:39:42

31 Mon Aug 2009 Current Time HH:MM:SS FGKE2AG11WG1(]GGI CGGTm32MGGW

_] L-101 g ^ llkI:^-I I I@IeIaI EI®Tacet I EVant Name — I Va

r

Foot Caves 4-Xi Repair—

J ^^Jej g op'Ik-1XIBIPPI^^? !110

J 001 lb^-NTercet I Event Name ( Value S'

Pilot ISHM ImplementationChemical Steam Generator (CSG

lit'st-t ®C 'l –L"JJ

^ File Flit View layout Go Project 1^/orkspace 1.hWindow Helpj3 _ y O—ct to TEDS Database Disconnect from TEDS Database 4 Get All S—r TEDS [?. Clear All Sensor TEDS

ES Explorer '+ x1.J/

EDS ListO ®. TE D14A4101-EPA

O-s TEES

-® Basic TEDS® --^ Bridge Sensor – LOX OLLL`aal,' LOX ( M, iJ

1H2O

©}H2O O

^^ I

K Telewi... --J y

untitled..

® I

^ Red Lines

G2

I

.R* 'R* n

Bridge Element Impedance (4) 351.3Bridge Type Full II

Calibration Date 2008-07-31

Calibration Period 365

Calibrator's Initials TSElectronic Datasheet Name Bridge Sensor

Full Scale Electrical Value Percision mV/VIEEE 1451.4 Template ID 33Mapping Method LinearMaximum Electrical Output (V/V) 3.034

Maximum Excitation Level (V) 15

Maximum Phyiscal Value 2000Measurement Location ID 4101Minimum Electrical Output (V/V) 0.031 –Minimum Excitation Level (V) 10

@03232009

Event Name EYentTACHIEVE] 2009108(2] ]1:03:20:8]5

NOT-ACHIEVED 200910612] 11:04:49:030

ACHIEVE] 200910812] 11:04:49:905` NOT-ACHIEVED 200910812711:04:50:453

ACHIEVED 200910812711:04:50:890NOT-ACHIEVED 2009/08/27 11:04:51:328

D11111p Minimum Physical Value 0Nominal Excitation Level (V) 10 J I

AlCFOSDtt ORct

NIC Settings Alarms 4 x RoaC— 4 x Repair Actions 4 x

J ! • 1 ^ *Px LJ-)-1 w 1 n 1 a A mu J tua 11L 0k 1-, 1-k1w1A1F0,, 1 N® J 11® I.gz

34 ITargel ( Event Name (Y. Target EVgd Name S Target Acb-Name Smtus

1 31120 1 II . II I I I08/31/2009

@03232009 '

II --^

II'

1111111 U11 Parameters

Enabled SetErgs StartMme End Mme CandNPntv l ^ nn T_ 7700 ? T_ PE-.4A4184GN HIGH Red U11

gave DAete I r" F___3 Psi Tii -s.000 ; TD F__ PE77,4184GN L01^/ ed Line

Pilot ISHM ImplementationChemical Steam Generator (CSG)

i OleEdt yreW Layout Q, Sro1ea werksPace Dds and— tklp

Jf Connect hr TEOS Database :y Discmn—f.—DSJatabase ^{; Ge[All Senmr Te05 [^Clex All Sensor TEDS

JCSG

r ' 1 IPA-RU NTANK IPA System

LOX o0 LOX L

^ry o - CSGLOX System

H2O

OH2O O CSG 1

Water System 1

41PE-11A1109-IPAPE 14A411IPA

Telewi... PE 14A4117

IP0.APE 1404]]9 IPly untitled... T^^^ PE 14A413i3 ---;

lg untitled... I _P^ 14:4111

DeElI.p

NIc—It 0t111

NIC Setting:

f®^ JiAMomi t-® rA^`3

'Z03Malloy

0>U31/d009

J JJJReal-Time Plot of TC-14A4275-S

5860-

5840

5820

5800

5780

5760

r9A3:45 79:03:50 19:0355 19:04:00 : 4:05

31 Mon Aug 21109 Current Time HH:MM:SS ^^

IWADI xi

Real-Time Plot of PT-14A41

26.10950M I _N

26.10944M

26.109J2M

26.10940M

r 19:03:45 19:03:50 19:03:55 19:04:00 19:04:05

31 MOn Aug 2009 Current Tlme HH:MM:SI

JJJReal-Time Plot of PE-14A4262-LO

Tcn_

740

720

'g:03:4'19:03:50 19:03:55 19:04:00 19:04:05

31 ...A 2009 Curtent Time HH.MM:55

Jt ]Real-Time Plot of PE-14A46-LO

620

600

580

560

19:03:45 19:03:50 19:03:55 19:04:00 19:04:0531 Mon Aug 2009 Current Time HH:MBI:SS

I

Conclusions• A sound basis to guide the community in the conception and implementation of ISHM

capability in operational systems was provided.• The concept of "ISHM Model of a System" and a related architecture defined as a

unique Data, Information, and Knowledge (DIaK) architecture were described. TheISHM architecture is independent of the typical system architecture, which is basedon grouping physical elements that are assembled to make up a subsystem, andsubsystems combine to form systems, etc.

• It was emphasized that ISHM capability needs to be implemented first at a lowfunctional capability level (FCL), or limited ability to detect anomalies, diagnose,determine consequences, etc. As algorithms and tools to augment or improve theFCL are identified, they should be incorporated into the system. This means that thearchitecture, DIaK management, and software, must be modular and standards-based, in order to enable systematic augmentation of FCL (no ad-hoc modifications).

• A set of technologies (and tools) needed to implement ISHM were described. Oneessential tool is a software environment to create the ISHM Model. The softwareenvironment encapsulates DIaK, and an infrastructure to focus DIaK on determininghealth (detect anomalies, determine causes, determine effects, and provideintegrated awareness of the system to the operator). The environment includesgateways to communicate in accordance to standards, specially the IEEE 1451.1Standard for Smart Sensors and Actuators

Backup Slides

K,

A

(5-Im J..w 11100

IND" at *NMN

Kp[lift"

QY

MTTP End-to-End SystemMethane Thruster Testbed Project

E* I. w" 0 Low 0~ " ff"p- "

IT-M fl-

=W4

Mr

OQfI

Al and J2-X ISHM MOD Lr

GN

GN2-PURGEy CIRCUIT a„a 4 c,;; CIRCUIT

LHeCIRCUIT

LOXTAN KV102

Lox-aCIRCUIT L

wx tx: :°_Horn rwilx4xr4uY

LOX-RUNCIRCUIT

LOX-1CIRCUIT

Lax-2CIRCUIT

44 QQ'

--^ IIJ

Runtime Predictive ModelingxEDS Model

Coefficientsr

Sensor Predictive j'r Prediction-

Data Model L MeasurementMismatch

Nleasuremeni Values100r

_ 700=

ID 600 — ---- - -500

a 400300—200

a 100

0

Lt

toga10 20 30 40 50 60 70

Time is)Predictive Values

800 —_ 700

GOO —20 500

d -000300

a 200—^aa

0I I I

-7000 1,0 20 30 d0 50 60 TO

1

Thneis)

Prediction- Measurement MismatchBoo700'-

t^ 600a 500

d 40a300

u 200-a` 700

•700 I ! I^ 1 I 1^ t

0 10 20 30 40 50 60 70Time is)

Checking for Pressure Leaks

No

Wait for ValueState Change

/ Do Closed `,Elements Form a\Boundary?,/

DefineYes Pressurizable

Subsystem

Pressurizable Subsystems

For Each PS

Mark All Elements ,-^ Do Sensors ^

Check Allof PS SUSPECT

Indicate a Change

Pressure

for Leak Anomaly -, in Pressure? - Sensors

For Each Element

Change HealthParameters in LeakProcess Model to

SUSPECT

Root-Cause-Analysis Root Cause

Electronic Datasheets

• Electronic Data Sheets (EDS)— Transducer Electronic Data Sheets (TEDS)

• Calibration— Health Electronic Data Sheet (REDS)

• Codified fault conditions and system phases• Key detection algorithms w/ parameters

— Component EDS (CEDS)• Manufacturing details• Engineering data• Traceability

— Other EDS

W

Intelligent Sensors haveembedded ISHMfunctionality and supportSmart Sensor standards

Intelligent Sensors0

Is

0

Smart sensor– NCAP (Go Active, Announce)– Publish data– Set/Get TEDS

Intelligent sensor– Set/Get H EDS– Publish health

Detect classes of anomalies using:– Using statistical measures

• Mean• Standard deviation• RMS

– Polynomial fits– Derivatives (1st 2nd)

– Filtering—e.g., Butterworth HP– FFT—e.g., 64-point– Algorithms for

— Flat— Impulsive ("spike") noise— White noise

– Other (ANN, etc.)