Method elliptical differential equations

Recently see "Thermal numerical fluid" (Shuji Yang, 2012) are some typical stresses solving partial differential equations, the difference calculation method according to which, using the Java language and Python language is solved formula 2-8. Mathematical model is as follows:

 

When the difference calculation to determine the accuracy of convergence is very important, it must be determined in accordance with the convergence accuracy of the scale grid,
That is, the smaller the mesh size, the convergence accuracy is required to be higher, or the correction value is ignored, a very large difference in the end result may be the case. Accuracy is determined as: while (delta> = 0.01 / n / n):
java faster calculations, but I believe that may be necessary to adjust the calculation of bigdata prevent irregular curves appear in FIG, of course, it may also be due to the drawing. After commissioning, the reason is because the accuracy of the csv export of only two, into five on the right
BigDecimal when the division calculations, the accuracy must be controlled, or slow
BigDecimal ap=aw.add(ae).add(an).add(as).divide(new BigDecimal(4),15,BigDecimal.ROUND_HALF_UP);

Python results

Java results







Published 34 original articles · won praise 9 · views 90000 +

Guess you like

Origin blog.csdn.net/tianyatest/article/details/53641584