ADAPTIVE GRID TECHNIQUES FOR USE IN COMPUTATIONAL FLUID DYNAMICS

THIS PROJECT FOCUSES ON THE TWO MOST IMPORTANT ISSUES OF ALL OF COMPUTATIONAL FLUID DYNAMICS: 1) HOW GOOD ARE THE NUMERICAL ANSWERS OBTAINED IN A COMPUTATION? AND 2) HOW CAN THE BEST POSSIBLE ANSWERS BE OBTAINED FOR A FIXED AND LIMITED COMPUTATIONAL EFFORT, BUDGET, OR COMPUTING CAPABILITY? IN GENERAL TERMS THE ANSWERS TO THESE QUESTIONS ARE 1) TO DETERMINE RELIABLE A-POSTERIORI ESTIMATES OF ERROR BASED ON FLOW RESIDUALS AND 2) TO DEVELOP SELF-ADAPTIVE SCHEMES TO REDUCE LOW ERRORS. THE PROJECT EXPLORES THESE DIFFICULT QUESTIONS AND CONCENTRATES ON THE DEVELOPMENT OF UNIQUE ALGORITHMS FOR RESOLVING THEM. ATTENTION IS GIVEN TO COMPLEX FLOW PHENOMENA IN REACTING FLUIDS AND EMPHASIS ON MODELS OF LASER FLOW FIELDS.