Creating a surface:
Data can be entered manually, or copied and pasted, into the data tables if desired. However in most instances data will be loaded from one of the many accepted file formats as shown on the ribbon:
Allows all of the current data to be cleared
Deletes the currently selected row in either table.
Load an existing Surface:
If an existing XPSX file needs to be referenced by a particular phase you can use the Load Surface->XPSX option. This will replace any data on the form and any existing XPSX links for the current phase.
Saving a Surface
Once loaded the data can be viewed and edited within the tables on the form if needed.When OK is clicked the software will perform a triangulation and save the surface file created to an XPSX file that is then referenced by the current phase. This file can be referenced by several different phases within the design to save disk space as well as being reused by other designs if needed.
The XPSX file will be named and stored relative to the Design file itself, if it has already been saved to XPDX. If this is not the case the software will prompt the user for a suitable filename.
It is possible to save the current state of the file to a specific file while editing. This allows control over the filename should this be needed. Click 'Save As' to specify a filename that will then be used to store the surface data, which will then be referenced by the current phase.
Optimising the Surface Data:
Surface Data information can come from a variety of sources and is often in a very raw format based on a survey. This data maybe purchased and supplied for a much larger area than is required, or may contain large flat areas that contain many thousand similar points. The software provides two additional options to help reduce the data over head and therefore reduce files sizes and loading times:
The surface boundary can be defined by using the Trim TIN option to remove information that is not necessary by the design.
Optimisation will be applied automatically if the option is checked when OK is clicked. This will then perform an optimsation of the surface data as outlined below and will request that the surface data is stored to a file.
As LIDAR data becomes more readily available it becomes possible to build terrain models with increasing levels of detail. However, more data is not necessarily better as it may increase file size and degrade application performance without contributing anything to the triangulation.
In the past data sets were thinned by simply removing each second point (for example) or by other methods that were not data aware. Approaches such as this run the risk of removing crucial points such as sinks or ridgelines. The TIN Optimiser examines the data set and selects the most important points with regard to maintaining gradient as this is the single most important consideration for 2D flow analysis. Points are selected based on most significant contribution until either the resulting TIN contains the specified number of points or the vertical separation between the remaining points, which will be discarded, and the resulting TIN surface is within the acceptable tolerance.
This process is computationally expensive and run time can be excessive for very large data sets. Run time can be dramatically reduced by splitting the terrain data into a number of tiles, each of which is optimised individually. For ordered data sets (such as lidar grids) this approach produces very good results. As the data becomes less ordered it is possible to see slight errors being introduced as triangles are formed differently at the borders of the tiles when the data is re-combined (e.g. the maximum vertical error on the final TIN is slightly higher than that requested). This is partly due to the optimiser algorithm and partly due to the ambiguous nature of a Delaunay triangulation. Consider a square with two different levels (A and B) at the corners:
The Delaunay triangulation will split this square into two triangles. The dividing diagonal could be defined as AA or BB, both are mathematically correct.
The resulting level at X is then seen as either A or B. A tile may be considered optimised if this square is split along AA but yields a different answer if BB is selected later when all the points are recombined.
Although using a heavily optimised surface is beneficial whilst the design is ongoing the final model should be tested against as large a dataset as possible.
The tile size and area can be entered under TIN Area.
TIN points can be optimised using preferred maximum points or maximum vertical separation (m).
Viewing the Data:
Once created the Surface File will be shown in the Tree View against the Surface node. The set of Surface related layers are shown below the Surface Node in the Tree. It is possible to adjust the visibility of each layer in the file using the checkbox shown next to the label.
Clearing the Data:
It is possible to remove the Surface File reference by Right-Clicking on the 'Surface Data' node on the Tree View and selecting 'Clear All'.