Condition monitoring in an on-ship environment Mike Knowles and David Baglee Institute for Automotive and Manufacturing Advanced Practice (AMAP) University.

Condition monitoring in an on-ship environment

Mike Knowles and David BagleeInstitute for Automotive and Manufacturing

Advanced Practice (AMAP) University of Sunderland

Who we are - AMAP

• AMAP is part of the Faculty of Applied Sciences within the University of Sunderland

• AMAP has been involved in a number of projects in:– Low Carbon Vehicle Manufacturing– Digital Manufacturing– Reliability and Condition Monitoring(Posseidon)– Industrial Maintenance and Efficiency

Facilities and Projects

• Projects– Dynamic Decisions in Maintenance (DYNAMITE)– Intelligent Energy and Maintenance Management– Digital Factory

• Digital Manufacturing– CAD– CNC – Rapid Prototyping– Dynamometer– Driving Simulators

Posseidon Project - Background

• Progressive Oil Sensor System for Extended Identification ON-Line

• Failures in marine diesel engines can be costly and can cause extreme inconvenience

• Current approaches to oil-based condition monitoring involve samples being sent for land based testing.

Impact of failures

• Engine failures can prove to be costly due to delays, time to repair and, in certain cases, environmental costs dues to ships running aground

• Thus onboard Condition Monitoring was borne out of need.

Posseidon

• The Posseidon projects seeks to address these problems by providing a means to monitor the condition of engine lubricating oil

Partners• Fundación Tekniker

• BP Marine

• OelCheck

• Martechnic

• IMM

• Rina

• IB Krates

• University of Sunderland

Diesel Engine Fault ModesFault Symptoms visible in oil properties

Corrosive Wear High increase in wear metals,

A strongly decreased TBN compared to the fresh oil.

Abrasive (mechanical) wear Incorrect viscosity.

Wear particles can be detected optically

Magnetic testing can reveal the presence of Iron.

Deposits The TBN of the drip oil can become slightly decreased compared to the fresh oil

Additionally the calcium content of the drip oil is decreased compared to the fresh oil.

Adhesive (mechanical) wear A strong loss of the viscosity compared to the fresh oil.

Magnetic testing can reveal the presence of large amounts of Iron

Severe sliding particles are visible optically

Soot Contamination Detection of soot particles by IR methods

Increase in Viscosity

Oxidation Increase in Viscosity

Mixture with another oil type Change in Viscosity

Water Contamination Detection of Water by IR methods

Nitration/Sulfation from Blow by gases

Change in base number

Oil Analysis

• Oil analysis at land based laboratories makes advanced analysis possible. Measurements taken include:– Measurement of water content using Karl Fisher titration– Measurement of TBN – Particle counting using optical techniques to detect wear

particles– Infrared spectroscopy techniques for measuring oil condition and

contaminants. – Magnetic PQ index testing to measure iron particle content– Density– Viscosity– Viscosity Index– Fuel Content– Flash Point

Sensor selection

Sensor Output

IR sensor Water concentration

Soot concentration

TBN

Viscosity sensor Viscosity

FTIR sensor TBN

Water content

Insoluble content

Optical particle detector Particles

IR Sensor

• Developed by IMM

• Monitors water concentration, soot concentration and TBN

Viscosity Sensor

• Developed by IMM

• Functions on vibrating pin principle

thread M30

Pin

thermocouple

housing(coils)

Optical Particle Detector

• Developed by Tekniker

• The smallest particles which can be identified are around 0.1 micron

Role of software

There are two levels of functionality for the system, at the most basic level:– Log the data– Display the data– Give simple assessments of oil condition and

potential faults– Offer simple guidance messages to the operator.

While the more advanced requirements are:– Exploit the multivariate nature of fault conditions– Detect both immediate, fast developing faults and

longer-term, incipient fault

Technologies used

• Java– Platform independence

• XML– Data can be read by spreadsheets etc– Configuration and condition monitoring limits

can easily be edited

Configuration – Design for Extensibility<config>

<datalogConfig><retrievalIntervalShort>0</retrievalIntervalShort><retrievalIntervalLong>3000</retrievalIntervalLong><xmlfile>\xmldata\sensorReadings.xml</xmlfile>

</datalogConfig>

<main><title>Posseidon Software Version 2</title><limitfile>\xmldata\CMLimits.xml</limitfile><messagefile>\xmldata\messages.xml</messagefile>

</main>

<BN><HKBFile>\BayesianNetwork\DieselEngine.hkb</HKBFile>

</BN>

<sensorConfig><sensor>

<name>Water</name> <id>N</id> <units>%</units></sensor><sensor>

<name>Visosity</name> <id>V</id> <units>cSt</units></sensor>

</sensorConfig></config>

Bayesian Network• An artificial intelligence module was developed

based on a Bayesian network to evaluate the probabilities of various faults and component failures

Screenshot

Testing

Posseidon Acheivements

• The need for the product has been demonstrated

• The viability of the system has been proved by the development of the prototype system

Future Development

• Hardware and Miniaturisation

• Display technologies

• Extensibility and Sensor Selection

• On-board/Off-board connectivity

• Design Issues

Hardware and miniaturisation

• Progress has already been made on miniaturising the individual sensors.

• Bespoke design is now required to produce a reliable and robust unit

Display Technologies

• Robust display technologies exist which support marine communication standards and which offer the desired level of robustness.

Extensibility

• Future Sensor additions – beyond oil– Vibration– Temperature– Thermal Imaging– Exhaust Emissions

Onboard/Offboard Connectivity

• Onboard– NMEA 2000 – Supported by proposed display

units– Inter-sensor connectivity – WSNs?

• Ground to shore connectivity – Cost– Update rate

Design issues

• What info is displayed?– Use of software ‘mock-ups’ to obtain feedback

from engineering personnel

• Resilience– Use of bespoke test rigs to simulate vibration,

thermal conditions etc.

Proposed Development Plan

• Create a consortium of interested parties who can support development

• Produce refined prototype– Smaller Sensors– No Laptop– Refined Software developed in collaboration

with industry

• Support needed:– Direct input from Shipping operators– Sensor/instrumentation companies.

Acknowledgements

• This work was supported by the EU Framework Programme 6 under the Posseidon project.

Thank you for listening

Questions?