In cooperation with MACSEA Ltd., the Univer 

ables,  interfaces  with  DEXTER  (an  agent  based

sity of New Orleans, College of Engineering (UNO

monitoring  system)  and  includes  various  failure

COE)  developed  software  for  automated  machin 

models for fault recognition and diagnostics.  This

ery  diagnostic  reasoning.    This  Condition Based

software system improves engine performance and

Maintenance (CBM) philosophy is a proven strat 

reliability through prudent and robust monitoring.

egy that supports minimum maintenance and in 

The  knowledge  base  of  the  system  is  extended

creases machinery availability.  CBM typically in 

through the continued use of the system due to its

volves  increasing  levels  of  machinery  plant

learning capability.  It reduces user workload for

automation.  The objective was to design an agent

monitoring tasks as well as provides data for fu 

capable of continuous real time machine learning

ture system design improvements.

by  using  an  Artificial  Neural  Network  (ANN),

known as the cerebellar model articulation control 

ler.  An engine simulator that can model both nor 

Wave  Software System

mal and faulty engine operations was used to de 

velop  the  learning  system  controller  in  a  flexible

The University of New Orleans, College of Engi 

and cost efficient manner.

neering collaborated with industry in a prosperous

Eliminating unnecessary maintenance tasks can

manner, which lead to a technology transfer of soft 

save both human and maintenance resources, which

ware development initially designed for Naval ap 

reduces the total ownership cost of ship machinery

plications.  This Wave  software package will al 

plants.  As automation levels increase, the amount

low  industry  to  take  advantage  of  research  and

of machinery data that maintenance personnel must

development efforts performed for the Navy.

track and assimilate can become extensive.  Intelli 

gent,  diagnostic,  and  prognostic  software  agents

The University of New Orleans, College of Engi 

assist  people  in  monitoring  and  troubleshooting

neering (UNO COE) developed software that was

complex machinery processes.  Agents perform te 

designed to store, monitor, and investigate failures

dious, repetitive, and analytically complex diagnos 

experienced from initial factory testing through ship

tic  tasks,  reporting  only  when  exceptions  are  de 

delivery.  This software was developed for Northrop

tected.   They  are  deployed  to  identify  machinery

Grumman Ship Systems in its construction of LPD

conditions that should trigger maintenance activi 

17 (USS San Antonio) for the U.S. Navy.  UNO COE

ties before equipment failures occur.

saw a need in industry for this software to be used

Diagnostic  inference  involves  sensing  perfor 

outside of the military envelope.  The commercial 

mance abnormalities by comparing measured ma 

ized  version  of  the  software  is  called  Wave .

chinery performance to a known baseline.  This com 

Wave   is  the  only  Reliability,  Availability,  and

parison yields symptoms that characterize faults.

Maintainability (RAM) software package designed

The accuracy of the baseline performance has a di 

specifically for the maritime industry.  Its unique

rect impact on the robustness of the diagnostic sys 

features allow maritime operators to lower main 

tem, which in turn can impact the effectiveness of

tenance and repair costs, increase equipment avail 

maintenance decisions.   A real time implementa 

ability, and reduce procurement costs.

tion of an ANN model forms the basis of the learn 

Resurgence Software, headquartered in the Uni 

ing system.  A real time engine simulation code was

versity  of  New  Orleans   Research  &  Technology

used to efficiently develop the learning system by

Park,  delivers  innovative  software  solutions  and

generating  real time  data  streams  in  a  series  of

services that help improve the bottom line for ship

learning  experiments.    While  previous  research

owners and operators. Their product, Wave  Sys 

shows  that  the  ANN  is  suitable  for  the  learning

tem, is an equipment reliability analysis system that

engine, real time issues require careful design and

enables ship owners and operators to optimize the

development considerations.

reliability and financial performance of the fleets

By using the engine simulator, SELENDIA, which

by maximizing vessel uptime, minimizing mainte 

is jointly developed by UNO and Ecole Centrale de

nance  costs,  and  reducing  the  risk  of  equipment

Nantes,  they  were  able  to  compare  these  results

failure.   Resurgence has marketing alliances with

with  actual  measured  engine  characteristics.

other  companies  to  market  the Wave   Software

SELENDIA has a Microsoft Windows graphical in 

System. Most important is the alliance with the clas 

terface, allows real time simulation of engine vari 

sification society, Lloyd s Register.  The agreement,

15