Sensor optimisation for in-service load measurement of a large composite panel under small displacement
Authors: Ramazani, M.R., Sewell, P., Noroozi, S., Koohgilani, M. and Cripps, B.
Journal: Applied Mechanics and Materials
Volume: 248
Pages: 153-161
eISSN: 1662-7482
ISSN: 1660-9336
DOI: 10.4028/www.scientific.net/AMM.248.153
Abstract:Current methods to estimate the behaviour of composite structures are based on trial and error or oversimplification. Normally the nonlinearities in these structures are neglected and in order to cover this inadequacy in design of composite structures, an overestimate safety factor is used. These methods are often conservative and leading to the heavier structures. A novel technique employs Artificial Neural Network (ANN) as an inverse problem approach to estimate the pressure loads experienced by marine structures applied on a composite marine panel from the strain measurements. This can be used in real-time to provide an accurate load history for a marine structure without requiring knowledge of the material properties or component geometry. It is proposed that the ANN methodology, with further research and development, could be utilised for the quantification of in-service, transient loads in real-time acting on the craft from the craft's structural response (strain response to load). However, to fully evaluate this methodology for load monitoring of marine structures further research and development is required such as sensor optimisation. The number of sensors should be minimised to reduce the time to train the system, cost and weight. This study investigates the number of sensors required for accurate load estimation by optimising the method. © (2013) Trans Tech Publications, Switzerland.
Source: Scopus
Sensor Optimisation for In-service Load Measurement of a Large Composite Panel Under Small Displacement
Authors: Reza Ramazami, M., Sewell, P., Noroozi, S., Koohgilani, M. and Cripps, B.
Journal: Applied Mechanics and Materials
Volume: 248
Pages: 153-161
Abstract:Current methods to estimate the behaviour of composite structures are based on trial and error or oversimplification. Normally the nonlinearities in these structures are neglected and in order to cover this inadequacy in design of composite structures, an overestimate safety factor is used. These methods are often conservative and leading to the heavier structures. A novel technique employs Artificial Neural Network (ANN) as an inverse problem approach to estimate the pressure loads experienced by marine structures applied on a composite marine panel from the strain measurements. This can be used in real-time to provide an accurate load history for a marine structure without requiring knowledge of the material properties or component geometry. It is proposed that the ANN methodology, with further research and development, could be utilised for the quantification of in-service, transient loads in real-time acting on the craft from the craft’s structural response (strain response to load). However, to fully evaluate this methodology for load monitoring of marine structures further research and development is required such as sensor optimisation. The number of sensors should be minimised to reduce the time to train the system, cost and weight. This study investigates the number of sensors required for accurate load estimation by optimising the method.
Source: Manual
Preferred by: Mehran Koohgilani and Philip Sewell