0 Replies Latest reply on Nov 5, 2011 1:29 PM by Karl Mohr

    Solidworks Simulation Flow Simulation limitations accuracy

    Karl Mohr

      Hi All

       

      Any comments welcome!  There has been talk (under; lift drag l:d) about;

       

      1)What are the Reynolds number limitations of the code, to maintain any degree of accuracy? (Anything from Solidworks on ~6M limitation?)

      2)Boundary location (sizing) if the flow stream is set in the Wizard, will the Wizard's Boundary settings be adequately large?

      3)Boundary spacing of 200 C, being necessary to achieve accuracy, where C is the length or the object in the flow stream or chord of an airfoil/profile! (I do not believe to be correct!)

      4)Advice NOT to use 'Solution Adaptive' mesh refinement!  Seems a good tool to me, (maybe just lazy! ) responding during analysis to 'high shear' areas that can not be accurately predicted!  Seems better than resorting to manual increases in mesh density that must be 'tied' to a model surface!

      5)I noticed that the "criteria" set in the solver by the 'auto' setting, gave VASTLY different CRTERION and DELTA values from run to run!  I blamed the inconsistency of results (for graphing; l:d etc.) on this!  So I recently started setting "criterion" as a "tolerance" (acceptable inaccuracy) on my 'estimated result/objective' (in the case of Aerodynamic LIFT and DRAG, the drag derived from the KNOWN lift objective and my target l:d) I understand that as (for example) LIFT drops/increases, with angle attack or speed (Re) changes, from run to run the "tolerance" will effectively increase! (so does a +-0.1mm as length changes! ) COMMENTS WELCOME

      6)With a default "Mesh 8" selected, there will be a MINIMUM of 3 travels including refinements of the mesh, the refinement defaults to lvl=2 giving 6 sizes of cell usable to the mesher during the refinement! (I use lvl=3 this setting has up to 7 cell sizes available to the mesher!)

      7) With Lift and Drag goals and an l:d equation goal, the l:d line, during the calculation exhibits a STRONG bi-travel (bi-refinement) pattern, this is yielding a VAST difference in the result depending on the number of travels! (+-17.3%) SEEMS this is because; one new travel starts close to the level of the last, one starts with a "stab in the dark" plunge! ANY IDEAS??

      8)delta seems to be some kind of 'deviation' statistic on the MIN/MAX values in the 'goals'! MORE advice appreciated!

       

      Karl