A GEOMETRY-INDEPENDENT FINE-MESH-BASED MONTE-CARLO IMPORTANCE GENERATOR

A new importance map approach for Monte Carlo simulation that can be u sed in an adaptive fashion has been identified and developed. It is ba sed on using a mesh-based system of weight windows that are independen t of any physical geometric cells. It consists of an importance map ge nerator and a splitting and Russian roulette algorithm for a mesh-base d weight windows game that is used in an iterative fashion to obtain i ncreasingly efficient results. The general purpose Monte Carlo code MC NP is modified to incorporate this new mesh-based importance map gener ator and matching weight window technique for variance reduction. Two nuclear well logging problems-one for neutrons and the other for gamma rays-are used to test the new importance map generator. Results show that the new generator is able to produce four to six times larger fig ures of merit than MCNP's physical geometry cell-based importance map generator. More importantly, the superior user friendliness of this ne w mesh-based generator makes variance reduction easy to accomplish.