The optimal state estimation method. A tool to integrate full scale shock trial measurement data and numerical models

In the joint US-NL research program DYCOSS (an acronym for Dynamic Behavior of Composite Ship Structures), a data analysis tool has been developed. The tool contains a mathematical method called the Optimal State Estimation method (OSE) and a Graphical User Interface (GUI). The OSE-method utilizes a reduced model of the ship consisting of a set of eigenmodes. The eigenmodes include the effect of the surrounding water, resulting in the system boundary being at the wet hull of the ship. Together with this numerical model the OSE-method utilizes the measured accelerations and velocities, either raw or manipulated. The OSE-method uses these eigenmodes to obtain a best fit with the measurement data. In this process the external shock loads on the wet hull of the ship are estimated. In this way the difficult topic of fluid structure interaction is avoided. The quality of the solution and the model and measurement data is indicated by the residuals.
The GUI/OSE package allows the assessment of both the global and the local shock responses. It allows stretching of the measured data set, such that mathematically based predictions for the responses at non-measured areas can be obtained. The estimated shock loads on the wet surface of the ship allow to some extent for extrapolation of the results for varying shock geometries. The use of weighting factors makes it possible to distinguish between more and less reliable variables and measurements. Simultaneous visualization of both estimated and measured responses is one of the features enhancing the user's insight into large scale dynamic response problems involving comprehensive data sets. The user has the possibility to control the solution quality by varying the amount of eigenmodes and selecting subsets of measurement data and eigenmodes