Machinery Health Monitoring System
The system consists of signal sensors, I/O boards for data collection
and data processing - and software applications for data analyses,
alarm management, and data storage which are found in the Node
Processors placed in the thruster rooms.
The sensors include vibration accelerometers which register the
vibrations on the thrusters, sensors for registering the oil
temperature and pressure sensors to register oil pressure on the
thruster’s pitch, sensors to measure motor load - and sensor for
water saturation monitoring in the thrusters oil system.
The measured signals will be collected and analysed through hardware
(HW) instrumentation delivered as part of the system; where after
measured data will be further processed both for direct dynamic alarm
control and database storage.
This data processing takes place in local Node Processors in the
thruster rooms, containing data acquisition boards, hardware and
software for data acquisition, data processing, alarm- and event
management and storage to database. The system will conduct a series
of quality controls every sequence taking into account the actual
permitted motor load range and requirements to the motor load
stability.
All vibration data will be processed through FFT algorithms to
generate frequency analysis spectrums and further classified into a
number of separate frequency bands – whereby narrow band values will
be statistically extenuated over a shorter time period and compared
up against each individual dynamic alarm limit conformed each
thruster at any given time or place. In other words the system
operates with individual dynamic alarm limits adjusted accordingly to
the applicable area with a bandwidth of 10 % in each motor load band.
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Schematic Overview
The Machinery Health Monitoring System contains several functionality
modules with logical relations to obtain the operational consistency
required for such a system. The data flow throughout the system and
relations to all the major functionality modules are described in
schematic diagrams.
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Graphical User Interface
Very user friendly GUI`s has been implemented on the Host PC which is
the communication link between the operator and the Machinery Health
Monitoring System. Remote access, configuration, offline data
analyses, alarm & event handling, status screens and reports are
accessible from the Host GUI. All kind of communication with Node
processors and www. takes place from the GUI as well.
NORMAL VIEW - THRUSTER NO.1 RUNNING IN NORMAL CONDITION
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Condition Based Classification
Principally the thruster’s condition and need for service will be
classified and analysed in accordance to acquired and processed
historical data and information which communicates something about
the thruster’s mechanical condition and status.
Parameters conveying something about the condition are called
condition parameters. Such parameters include both current amplitude
and trend status for vibration, oil temperature, and oil pressure and
water saturation. Shipping companies and oilrig operators are given
the opportunity to take advantages of condition based classification.
Generally speaking the wear and tear and dynamic changes will be
mirrored by the vibration status and history of the thruster itself,
therefore attempts are made to detect any sign and changes which may
indicate potential danger or damages in progress through different
kind of analyses as statistical and frequency analyses of the
vibration data. Gear wheel and bearing damages will - as an example,
always result in a change in vibration pattern which can be detected
through such analyses.
¾ Mechanical conditions are classified by means of data analyses
techniques of normalised historical data and other accessible
and relevant information
¾ Service needs are established after considering the total
relevant and accessible information
¾ Classification societies offer customers such classification
through special rules, regulations and regimes
¾ This will give tremendous economic saving for the customers
¾ Thrusters will be classified based on real condition instead of
time – with help from oil analysis and vibration analysis
parameters, as well as operative time and history
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Node Processor with DAQ and Dynamic Alarm System
Node Processors can be placed locally. These will contain among other
things, the instrument part of the system with multi-channel I/O
boards used for data acquisition, filtering and data processing.
Measured data from all the sensors are all collected in sequences and
processed locally on the Nodes. Statistically extenuated condition
parameters are processed and controlled up against individual dynamic
alarm limits and stored in the Host database for history - for post
analyses and reporting. The advanced dynamic alarm system will be
placed locally on each Node.
Event logic will be activated between the alarm system and the data
acquisition which results in; when an alarm limit has been exceeded,
a new set of vibration data automatically are collected from all
channels for that particular thruster, this will stretch over an
interval capturing both the time laps before and after the alarm was
activated. This ensures extra analyses and troubleshooting
possibilities at the time of the alarm. Extra alarm delay logic is
also implemented.
¾ Node Processors situated in thruster rooms which sequent and
automatically collects different kind of condition data from the
thrusters
¾ Contains hardware and software for data collection from sensors,
data processing, alarm handling and data communication with the
Host PC
¾ sequent and automatically executes dynamic alarm control
regarding the thrusters motor load
¾ Contains event logic regarding the alarm system
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Host PC
Host PC will be placed in the vessels central control room, along
with the software for offline analyses of condition related data
obtained through the database logs. Such as data trends and
corresponding prognoses, frequency analyses and alarm history can be
shown graphically, showing and disclosing the thruster’s mechanical
condition and progression over time. Host PC will be connected to
each Node Processor through Ethernet communication.
One will be able to access software and hardware settings and
configurations on the Node Processors from the Host PC through the
remote configuration module. This junction permits amongst other,
alarm acknowledgement, system configuration of alarm limits, SW and
HW logic etc.
¾ Situated in the Central Control Room and communicates with the
Node Processors through Ethernet and communicates also through
www. to different locations
¾ SW running on Windows XP or 2000 OS and MS database
¾ Contains Graphical User Interface (GUI) for communication
between the operator and the MHM-system
¾ Employs System configuration and communication with Nodes and
networks
¾ Alarm management and acknowledgement
¾ Data retrieval and offline condition analyses
¾ Manually and automatically generated Reporting tools
¾ Statistical trend analyses and prognosis with “time to alarm”
for both linear and nonlinear parameter trends etc.
¾ Alarm and event historical analyses
¾ Filtered vector analysis plots of vibration data
¾ Advanced Filter functionality (Bessel and Butterworth)
¾ Advanced 2D and 3D Frequency analyses of vibration data
¾ Orbit analysis plots and polar plots
¾ Transient analysis of captured run-up vibration data
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Quality Control to ensure reliable signals and analysis parameters
Consists of controlling the degree of signal reliability and
stability - and also statistically based control methods are used to
ensure high and reliable quality regarding stability, data
acquisition, data processing as well as alarm handling and storage.
The dynamic alarm system ensures that all individual alarm limits are
correctly tuned to each load range – classifying existing operating
areas for alarm control and processing independent of load and
running condition. This giving the benefit of reduced possibilities
for false alarms. High sampling rates and the advanced alarm control
are crucial to the systems ability and capacity to detect any
abnormal situations and signs which can indicate mechanical/
technical problems in the thrusters.
¾ Signal stability control by using statistical methods
¾ Flexible multiple load ranges for dynamic alarm limit control
¾ Dynamic alarm limit version control
¾ Data reduction algorithms based on signal variation
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Alarm Control Logic
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Dynamic Alarm Limits
Thrusters often operate within large operative variations where the
load will vary greatly, therefore emphasis have been made by the
alarm system to take into consideration these variations. The
individual parameters needs to over time, stabilize at a certain
level above the corresponding alarm limit before an alarm is
activated, so as to minimize the amount of false and unwanted alarms.
The system will collect load parameters for each thruster with the
same sampling rate as it does for other variables. Measured data will
at all times be individually compared against its individual and load
dependent dynamic alarm limit for that particular area for that
particular thruster.
These alarm limits are established individually based on where the
relevant condition parameters have it’s natural amplitude area whiles
the thruster is in normal operation, with no abnormalities, this will
be a form of learning approach for the system itself. This will
result in a much more precise and effective alarm system and a
condition control system compared to more conventional systems within
the same Health Monitoring categories.
¾ Semi automatic statistical based Alarm limit establishment
¾ Load ranges with Individual dynamic alarm limits
¾ Functionality for minimising false and unwanted alarms
¾ Alarm system will slowly “learn” to differentiate between normal
and abnormal conditions and situations
¾ Increased alarm system quality
¾ Automatic Alarm & Event Control sequence logic
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Dynamic Alarm Limit Control to ensure correct alarm limits
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Alarm and Event Log
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Trend Analyses and Prognoses
When historical data is analysed collectively and graphically one
will be able to view tendencies and trends visually. On top of this
all is included a statistic prognoses function conveying expected
trend development for the future – based on the historic data used.
Statistical prognosis can tell WHEN a defined parameter can be
expected to exceed its alarm limit. It will also be capable of
telling what a parameter level can be expected in an elective period
X days into the future. These are statistic methods - used
reasonably, are very effective in regards to condition analyses of
machinery equipment.
¾ Narrow Band trending of vibration signals
¾ Condition Parameter trending within “Specific Load Windows”
¾ Load and slowly changing parameter trending
¾ Statistical “time to alarm-level” predictions
¾ Statistical “specific date/ time level” predictions
¾ Linear and non-linear trend analysis options
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Vibration Analyses
There are many advanced and useful analyses methods used in the
system, making it possible to convert vibration data to useful
information concerning the thruster’s mechanical condition and need
for service. There are possibilities for filtering and classifying
vibrations which then are compared to individual dynamic alarm
limits. One can execute extremely advanced frequency analyses of both
individual vibration measurements and the comparison between new and
reference frequency analysed vibration data. Zoom function is also
included here for a detailed view of frequency analysis.
¾ Narrow Band (NB) generated condition parameters
¾ NB parameters will be compared against individual load dependent
dynamic alarm limits
¾ Advanced time vector filtering functionality
¾ Sets of different Fast Fourier Transform algorithms
¾ Zoom functionality regarding frequency spectrums
¾ Reference time vectors and frequency spectrums are established
regarding each individual load window for comparison and
diagnostic troubleshooting
Revealing Rotor Resonances by use of Waterfall Plot (Time-Frequency
distribution)
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Vibration Analyses Tools
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Orbit Plot Analysis
System allows plotting measured signals on an orbit graph. In order
to do this, one or several sensors groups have to be defined. Each
group have to include signals for X and Y axes and a tacho signal
key-phasor) for synchronization. Using key-phasor the system is able
to determine current rotating speed, initial angle and extract only
one period of measured data for displaying an orbit.
Configuration
To define a sensor group, a system configuration utility
(Configurator) will be used. All sensor groups defined for particular
machines have to be an orbit type. For each group it is possible to
define some name string, select two already configured channels for X
and Y axes respectively and select synchronization channel. Each
group is referenced by its name string elsewhere in the system.
Online Orbit Analysis
Online orbit analysis is available as a special option in the Host
application. In order to get an orbit plot for selected sensor group
user has to press an “Update” button on a “Machine View” screen. As a
result of this action system returns raw signals for this group and
build an orbit plot. It is possible to plot raw orbit or extract
orbits of another orders (2X, 3X …) with a band pass filter.
Following filtering options are available: start and stop frequency
settings, filter order. Together with the orbit operator can plot and
see raw signals on a separate graph.
Offline Orbit Analysis
In the offline analysis orbits can be plotted for record files. In
the offline all the online orbit analysis options available as well
together with multiply and reference orbit plots. In the multiply
orbits mode tree orbits can be displayed simultaneously for three
different signals group in the same display see Figure 1. Multiply
Orbits View.
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Figure 1. Multiply Orbits View.
In the reference orbit mode user can display current and reference
orbit on the same plot see Figure 2. Reference Orbit. It is possible
to select a reference group from the same or another record file as
well. In this mode user can use an on graph cursor to precisely
determine vertical and horizontal coordinates of each point as well
as a point’s angle.
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Figure 2. Reference Orbit
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Ethernet
I/O Connector
Signals Flowchart
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Vibration Low-speed
Signals Signals
Mean and RMS
Antialiasing Filter
Calculation
Ax+B Ax+B
Other Signals
(scaling) (scaling)
Load Signal
Stability and
FFT
Power Range Alarm Detection
(widowed)
calc.
Integration
(optional)
Power in Band
calculation
Averaging
No Stable? Yes
External or
Alarm Detection Periodic
Request
Alarm? Yes Recording
Update Status
HMSS
Update Data
Information flowchart
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
System Program Parts
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Health Monitoring System Run-Time and Support Components
No. Component Producer Comment
1 LabVIEW Run-Time National Allows running any
Instruments compiled LabVIEW code.
2 DSC Module Run- National Provides data storage
Time Instruments functionality.
3 NI-DAQ driver National Measurement hardware
(traditional) Instruments driver
4 SQL Server Microsoft Installing as a part
Desktop Engine of the LabVIEW DSC
(MSDE) Module.
5 .Net Framework Microsoft Components used: XML
parser, ADO .NET
HMS Development Tools
LabVIEW 7.1
It is a main development environment for the Health Monitoring
System. Also following additional libraries have been used:
1. Sound and Vibration Toolkit for LabVIEW (National
Instruments)
2. LabVIEW Data Supervisory and Control Module (National
Instruments)
3. LabVIEW NI-DAQ support library (National Instruments)
Sound and Vibration toolkit is used only for the Offline Data
Analysis program.
NI-DAQ support library is used for the Node program only.
LabVIEW DSC Module is used for developing the HMSS program.
Microsoft Visual Studio .Net
It is additional development tool that is using for developing
database related functionality. The Records database is based on the
MSDE (reduced version of the Microsoft SQL server). Because of the
LabVIEW does not have build-in functionality suitable for working
with the SQL Server, and because the Database Connectivity Toolkit
for LabVIEW works slowly and requires run-time to be installed for
every system instance, the MS Visual Studio (VS) has been selected
for developing a bridge between LabVIEW code and the database server.
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
Within the VS a custom .Net assembly has been developed. Because
of the LabVIEW 7.1 directly supports .Net technology the integration
of this assembly to LabVIEW code is simply and more reliable than
integration of ordinary dll. The assembly uses ADO .Net technology
for opening connection to database and for transactions.
Any changes in the structure of the Record Database require
changes and rebuilding the assembly. Also Visual Studio .Net has
additional tools that helps develop and debug database applications
such as Server Explorer, query builder and others.
Figure 1. A route for all system transactions to the Records Database
Breivika Industrivei 63 · Box 8013 Spjelkavik · N-6022 Ålesund · NORWAY · [Link]
