## ENVELOPING

## Overview

REPTILE provides several utilities for filtering its own output (or really any similarly formatted CSV file) to help find the critical load cases. These tools are described below.

## Max/Min Sorting

The simplest way to filter loads data is by finding the maximum and minimum load case for each component. Often times, this is sufficient to find critical load cases, particularly when one component is dominant or highly correlated to other significant components. For more complex loading/analysis, it is recommended to use one of the other methods described on this page.

REPTILE's

The basic command line for

REPTILE's

*--envelope*mode takes a full-output CSV file and outputs only the lines that have a maximum or minimum for some component of load. This is done for each location independently, or can optionally be done across all locations (this option is still in work).The basic command line for

*--envelope*looks like this:reptile --envelope -i loads.csv -o loads_env.csv

That will perform max/min enveloping for every data column (column >= 5) of the CSV file.

If you only want to envelope certain columns, you can use the "-COL" option to specify which columns to process. The desired columns headers are given in a comma-separated list (no spaces allowed).

If you only want to envelope certain columns, you can use the "-COL" option to specify which columns to process. The desired columns headers are given in a comma-separated list (no spaces allowed).

reptile --envelope -COL NXX,NYY -i loads.csv -o loads_env.csv

To envelope automatically after running another REPTILE function, simply add the "-ENV" option to that function, for example:

reptile --node_forces -ENV -COL PX,MY -i input.def -o output.csv file1.xdb

After creating the regular output file, REPTILE will then perform the enveloping and write the output to a file with the same base name as the regular output file plus "_env".

## Potato Plots

A potato plot is a useful tool for identifying critical load cases where there is an interaction between two components of load (e.g. a moment and an axial force). In technical terms, a potato plot is the "2D Convex Hull" of a set of data points. REPTILE uses the Graham Scan algorithm for this application.

REPTILE's

*--potato*mode takes a full-output CSV file and outputs only the lines that are vertices of the potato plot. This is done for each location (element, panel, cut, etc.) independently. The components to be use must be specified with the "-COL" parameter followed by a comma-separated list of the respective column headers (no spaces allowed). For example:reptile --potato -COL PX,MY -i loads.csv -o loads_potato.csv

To make a potato plot automatically after running another REPTILE function, simply add the "-POT" option to that function, for example:

reptile --node_forces -POT -COL PX,MY -i input.def -o output.csv file1.xdb

After creating the regular output file, REPTILE will then perform the potato plotting and write the output to a file with the same base name as the regular output file plus "_pot".

## VodKA Plots

<< FUNCTION STILL IN WORK >>

There are times when more than two load components play a significant role in an analysis, and this can complicate the identification of critical load cases. Fortunately, REPTILE can produce Vodka plots!

Vodka plots are like potato plots, but more filtered and more distilled. This is simply an extension of the 2D potato plot into higher dimensions (i.e. an N-dimensional Convex Hull). REPTILE uses the Quickhull algorithm for this application. The figure below shows a 3-dimensional convex hull, but the concept can be applied to an arbitrary number of dimensions (though the computational run-time increases rapidly).

<< insert vodka plot figure >>

REPTILE's

There are times when more than two load components play a significant role in an analysis, and this can complicate the identification of critical load cases. Fortunately, REPTILE can produce Vodka plots!

Vodka plots are like potato plots, but more filtered and more distilled. This is simply an extension of the 2D potato plot into higher dimensions (i.e. an N-dimensional Convex Hull). REPTILE uses the Quickhull algorithm for this application. The figure below shows a 3-dimensional convex hull, but the concept can be applied to an arbitrary number of dimensions (though the computational run-time increases rapidly).

<< insert vodka plot figure >>

REPTILE's

*--vodka*mode takes a full-output CSV file and outputs only the lines that are vertices of the vodka plot. This is done for each location independently. The components to be use must be specified with the "-COL" parameter followed by a comma-separated list of the respective column headers (no spaces allowed). For example:reptile --vodka -COL PX,MY,MZ -i loads.csv -o loads_vodka.csv

To make a vodka plot automatically after running another REPTILE function, simply add the "-POT" option to that function, for example:

reptile --node_forces -VOD -COL PX,MY,MZ -i input.def -o output.csv file1.xdb

After creating the regular output file, REPTILE will then perform the vodka plotting and write the output to a file with the same base name as the regular output file plus "_vod".