AMDG On 12/10/2010 4:52 AM, André Berthe wrote:
I would like to calculate a shear stress based on a shear rate and a viscosity using Boost.Units. Here I have problems using derived dimensions and conversion factors based on a existing system. The error messages of my compiler are in German, so I tried to document my problems with a test code:
<main.cpp> #include <boost/units/systems/si/dynamic_viscosity.hpp> #include "stress.h" #include "shearrate.h"
namespace bu = boost::units; namespace SI = bu::si;
main(){ //Define a quantity of unit stress. Unfortunately there is no //realization of the defined physical dimension stress in si. //So a user defined type from stress.h (see below) is used: bu::quantity<SI::stress> WSS_x;
//Define a shear rate. I did not understand how to use a //inverse time as dimension. Therefore I defined the derived //dimension per_time in shearrate.h. Is it possible to add it to //the si system (as I tried) or do I have to define //my own system and a conversion factor? bu::quantity<SI::per_time> WSR_x(1*per_second);
The inverse of time is frequency.
//There is a existing si type for the dynamic viscosity. //However, there is no defined constant. How is the //correct way to assign a value? Should I define this //constant? bu::quantity<SI::dynamic_viscosity> eta;
You can define the constant or you can combine other units: si::kilograms/(si::meter*si::second).
//Calculation of the shear stress. It seems that this calculation //tries to use a »sizeof« on eta, that seems to have a // incomplete type //(»boost::STATIC_ASSERTION_FAILURE<false>«). And there //seems to be something wrong with the conversion. //Do I have to define a conversion factor? Or is this error //caused by my previous mistakes? WSS_x=WSR_x/eta;
The dimensions of stress are L^-1 M T^-2. Dynamic viscosity is M L^-1 T^-1. Thus, the ratio WSR_x/eta should be (1/T)/(M L^-1 T^-1) = M^-1 L. You get an error because the dimensions don't match. In Christ, Steven Watanabe