Pillateunlinux

Hombre, como dice el nombre del programa, CAELINUX es un programa CAE (Computer Aided Engineering), y no CAD (Computer Aided Design). En cristiano, una pieza se diseña en un programa CAD y se obtiene una geometría 2D ó 3D (formato IGES o STL, por ejemplo); después se importa ese fichero en CAELINUX y se pueden hacer cálculos de ingeniería (resistencia estructural, fluidodinámica, etc.)

El problema que yo veo es que, para empezar, no hay un programa CAD libre de calidad. Y en segundo lugar, por mi experiencia con SALOME, no es compatible con los programas más usados en ingeniería y que, creo, se pueden considerar un estándar “de facto”, como pueden ser ANSYS o NASTRAN. Es más, si lo simplificamos todo y nos vamos a un programa CAD “sencillo”, de dibujo en 2D, y ampliamente usado como puede ser AUTOCAD, no conozco ningún software libre que se acerque ni de lejos a sus prestaciones. Y el único que conozco, el QCAD, hay que pagar por la licencia para usarlo profesionalmente. Ergo deja de ser libre.

(Por favor, añádase (R) y (TM) donde corresponda)

Vayamos por partes.

El artículo al que se me remite, si nos paramos a examinarlo, no da mucha información. Sólo glosa que el BRL-CAD es una alternativa para AUTOCAD y otros programas CAD. Si hacemos una búsqueda en Internete, ¿qué encontramos? Las mismas fotos una y otra vez, todo el mundo diciendo que por fin hay una alternativa, etc, etc. Es probable que me equivoque, pero creo que la mayor parte de los que han escrito esas páginas no lo han probado. Así que, a riesgo de equivocarme, creo que hablan “de oídas”.

El BRL-CAD es un programa que está orientado para análisis balísticos (penetración de un proyectil en un objeto, concretamente) y análisis de respuesta al radar.

Para ello, a partir de un sólido 3D, se realizan una serie de hipótesis de densidad del material, propiedades mecánicas (resistencia a la rotura, módulo de Young, límite elástico, coeficiente de Poisson, etc.) y después se lo somete a un análisis del tipo ¿qué pasaría si una ojiva de uranio empobrecido penetrase en este material con esta geometría y estas características? ¿O si una onda de radar de determinada longitud de onda y energía incidiese sobre esta geometría?

BRL-CAD se asemeja a programas como SALOME (abierto) o ANSYS y NASTRAN (propietarios) en el sentido de que están orientados a hacer simulaciones sobre una pieza. Es decir, son programas CAE (Computed Aided Engineering). Tienen un editor gráfico que permite crear sólidos y montar conjunto de sólidos. Pero, normalmente, son editores sencillos. El del BRL-CAD lo es. Porque su objetivo no es crear una serie de planos constructivos para fabricar lo que se diseña en él.

Aunque BRL-CAD puede construir geometría, si vamos a la documentación del proyecto leemos:

http://brlcad.org/w/images/6/66/Converting_Geometry.pdf

There are numerous benefits associated with the use of commercial packages in vulnerability
studies. The Survivability/Vulnerability Information Analysis Center (SURVIAC) identified
some of the most common benefits in its 2002 State-of-the-Art Report (SOAR) on geometric
modeling. They include the following (SURVICE Engineering Company, 2002):
• Reduced Modeling Time and Effort – Manufacturers often spend hundreds of hours
constructing detailed CAD models to streamline their design, production, and assembly
processes (e.g., through computer numerical control equipment). It therefore makes
economic sense—and is consistent with the military’s ongoing commitment to leverage
commercial off-the-shelf technology—to take advantage of existing geometric models and
data where possible and avoid the significant cost and effort of building new models from
scratch.
3
• Increased Funding and Support – Because commercial packages must maintain a larger
user base to stay competitive in the open market, the most popular packages typically
possess ample funding and personnel to continuously develop, support, and improve them.
• Compatibility With Standardized Formats – Most commercial packages possess the
direct or indirect (through third-party vendors) capability to convert to standard or
intermediary geometry formats, making the packages compatible (at least to some degree)
with a wide range of other CAD formats.
• Third-Party Add-On Support – Large commercial packages typically offer a variety of
plug-ins for other packages/utilities.
Of course, commercial CAD packages also have some common liabilities when used for
vulnerability studies. They include the following (SURVICE Engineering Company, 2002):
• Incompatible/Inaccessible File Formats – As discussed in section 2, some CAD formats
use the boundary representation (BREP) approach to solid modeling, which is largely
incompatible with the constructive solid geometry (CSG) approach that BRL-CAD uses.
In addition, although most commercial packages have some capability for data exchange
conversion, data storage is often in a proprietary (and therefore inaccessible) native format.
Moreover, when target descriptions are converted to a format designed for vulnerability
assessment (i.e., BRL-CAD or FASTGEN), they often require manual checking,
adjustment, and additional modeling (see section 7, Postconversion Issues). Typical
problems that must be addressed include the translation of curved and irregular surfaces,
the representation of solids of rotation, tolerancing, and interference handling.
• Too Much Detail – Commercial geometry files often contain too much of a good thing for
vulnerability analysis. That is to say, commercial CAD packages often model geometry all
the way to the “nuts and bolts” level, whereas vulnerability analyses are typically
concerned only with details down to the level of shielding and critical components.
Unfortunately, added detail produces unnecessarily large and complex input files and thus
longer processing times for vulnerability assessments.
• Too Little Detail – In addition to providing too much detail, commercial packages
sometimes provide too little detail for vulnerability studies. Vulnerability analysts and the
applications designed to interrogate geometry rely on geometric measurements and
material properties not always present in commercial CAD formats.
• Package-Specific Naming Conventions – Some organizations and CAD packages use
unique object naming schemes that make geometry difficult to organize and work with
when converted to or from BRL-CAD format.
4
• Relatively Slow Raytracing Capability – Commercial CAD packages typically have
relatively slow raytracing speeds, and raytracing is the primary means of geometry
interrogation in vulnerability assessment.
• Limited Integrated Vulnerability Assessment Support – Commercial CAD packages are
designed for engineering analysis, not ballistic analysis and therefore offer few, if any,
“shotlining” tools and limited integrated vulnerability assessment support.
• High Cost – In addition to facing the typically high cost of commercial CAD software, the
user often faces the decision of whether or not to invest in non-PC hardware (e.g., UNIX
workstations) to obtain maximum performance, especially with large, complex geometry.
In addition, users may be required to pay continuing licensing and maintenance fees, often
on a per-seat basis (although recent developments have offered PC-based implementations
and short-term leasing “seats” to make these packages more affordable).

(Perdón por el rollo patatero)

Básicamente, lo que viene a decir es que los tanques, aviones, satélites, etc. no los diseña ni el Departamento de Defensa ni la NASA, sino empresas particulares que usan otro tipo de software y que, por tanto, como lo que interesa es el análisis balístico o electromagnético, lo más rápido es importar geometrías de esos programas. En los inconvenientes,

Saludos como que no hay CAD en Linux, si lo que deseas es algo al estilo CATIA, NX, PROENGINEER pues les comento que NX y ProEngineer hay para Linux. Y ojo que estos software son netamente superiores a Autocad, Solidworks, Inventor o Solid Edge.