Flexcom > Theory > Analysis > Restart Analyses > Data Specification in a Restart

As you can appreciate, the data required in a restart analysis is substantially the same as that in the analysis from which it is restarted. In Flexcom (with a few minor exceptions) you are required to specify only new or altered inputs in setting up the restart run. In fact, a large amount of data (in particular the finite element mesh) cannot be altered in a restart, and the program will object if you attempt to invoke any of the options appropriate to specifying this data. This section categorises the Flexcom input data in terms of those inputs that can and cannot be altered in a restart. The categories represent a logical division of the data and the reason for a particular input being in a particular category is normally self-evident.

Flexcom in fact groups the input data into three categories for this purpose. The first category consists of data that must be specified in the very first of a series of consecutive analyses. This data is carried through to all subsequent restarts and may not be changed. This category naturally includes the finite element discretisation, structural and hydrodynamic properties, plus any other inputs which characterise the initial model configuration (e.g. initial vessel position, seabed properties, ocean depth etc.) – basically any information which cannot logically change from run to run are contained in Category 1.

The second category consists of parameters that may be entered in any run of a series of runs. This data is automatically carried through to all subsequent runs, but may of course be subsequently altered again. This category naturally includes environmental parameters such as current and waves, boundary conditions of various kinds, internal fluid loading, and the analysis and solution type.

Strictly speaking, the majority of parameters which relate specifically to frequency domain dynamic analysis (e.g. harmonic loads), or those which also relate to the time domain but are being discussed in the context of the frequency domain (e.g. wave loading) do not fall into this category. The reason being that all such analyses must restart from converged static solutions, so it is not entirely true to say that the parameters that may be entered in any run of a series of runs. Furthermore, it is not possible to perform another restart analysis subsequent to a frequency domain dynamic analysis. However, such parameters cannot be accurately classified by categories one or three either, so it is logical to include them in this category, along with inputs of similar purpose in the time domain.

The final category consists of parameters which may be specified in any run of a series of runs but which do not carry through to subsequent runs. Rather they are specific to that particular run. The most obvious of these are the solution variables (such as time, damping and convergence parameters for example) – these must be specifically input for each analysis stage. Other parameters would include data storage requests (i.e. database and timetrace), stress properties etc.