Ships are subject to strict safety regulations. These not only cover the use and maintenance of the vessel but also its construction and the materials of which it is made. Until a few years ago, EU regulations prevented composite materials from being used in the construction of commercial ferries and cruise ships due to safety fears. Now though, technological advances have brought many composite materials up to the same safety levels as steel in terms of non-combustibility and these lightweight materials can be used.
To ensure that composites meet the demands for seaworthiness, they require a variety of tests simulating the effect of the oceans. Current methods are mostly based on trial and error or oversimplification. During the design of ships, the non-linearities of composite structures are often neglected and, in order to compensate, an overestimated safety factor is employed leading to heavier structure than are necessary.
Help is at hand though. Engineers from Bournemouth University and Longitude Consulting Engineers are testing an Artificial Neural Network (ANN) which can accurately monitor loads on ship components in real-time, whilst they are in service on the sea. This will reduce over-design whilst real-time strain measurements can improve safety. As part of their experiments they are using a Windmill strain measurement system
Ship constructed of composite materials
Photo credit: Heiko Schöne und Gunnar Richter, CC BY 3.0.
Integrated Sensors
Some vessels already have integrated sensors, but most do not. As part of their investigations, the researchers - Ramazani, Sewell, Noroozi, Koohgilani and Cripps from Bournemouth University and Longitude Consulting Engineers - asked:
"what is the minimum number of sensors that should be employed to reduce the time to train the system, to reduce costs and to reduce weight?"
Training the Artificial Neural Network
During their tests the researchers attached 20 strain gauge rosettes to a 1 m2 composite panel, which provided 60 strain channels. They applied weights to the panel. The Windmill 751-SG strain monitoring and control data acquisition system, with a resolution of +/- 1 microstrain, captured the strain data.
Collecting Strain Measurements with Windmill
The 751-SG package comprises Windmill Software, a USB (universal serial bus) measurement unit which monitors 16 strain gauges, giving up to 64 strain measurements, through differential inputs at up to 80 samples per second. Eight USB units can be attached to one computer to monitor 128 strain gauges. The researchers developed their own DAQ software in MATLAB which utilised the MATLAB Artificial Neural Network Toolbox capabilities and collected data from Windmill via DDE (dynamic data exchange).
The 751-SG package is currently on special offer for just £475 from our internet shop at https://www.windmillsoft.com/daqshop/
Validating Designs and Improving Safety
The ability to measure the actual load history of a ship in service would enable the designer to validate the load estimation and structural design tools used during the design stage of the craft. The operational safety of the vessel can also be improved by having a real-time load monitoring system that is able to detect any degradation of structural integrity and defects within the structure.
Further Reading and More Information
Sensor Optimisation for in-Service Load Measurement of a Large Composite Panel under Small Displacement MR Ramazani, P Sewell, S Noroozi, M Koohgilani, B Cripps. Applied Mechanics and Materials Volume 248 (2013) Pages 153-161
Determination of the static pressure loads on a marine composite panel from strain measurements utilising artificial neural networks MR Ramazani, S Noroozi, M Koohgilani, B Cripps, P Sewell. Proceedings of the Institution of Mechanical Engineers, Part M: Journal of Engineering for the Maritime Environment Volume 227 (2013) no 1 12-21
Making safer composite materials for ship building, European Commission
Strain Measurement Tutorial, Windmill Software
For more information contact Windmill Software. For more case studies, and tips and tutorials on strain measurement and other data acquisition topics, subscribe to our free, monthly, Monitor newsletter (ISSN 1472-0221)
