Development and experimental validation of a computational fluid dynamics-discrete element method sand production model

Sand production is a dynamic process that occurs during the production of oil or natural gas, costing the petroleum industry billions of dollars per year to control. The main aim of this research is to develop a numerical model combining the discrete element method with computational fluid dynamics and the using particle flow code to simulate the sand production process, so as to obtain a deeper insight into its mechanisms. The numerical model was validated using sand production data from a specially designed sand production cell. The effects of various variables affecting the sand production process including fluid pressure, sand particle size distribution, and fluid type were investigated. Changing the random number will not influence the macroscopic properties of the sample but affect the microstructure. Selecting the same random number is necessary in each simulation if one wants to keep both the macro and micro properties of samples the same. Running one numerical model with different random numbers can eliminate the haphazard of test. Two main failure modes resulted in sand production: the collapse of thin inner layers of stable sand arches and the thorough collapse of the sand body. When the drawdown pressure was lower than the critical drawdown pressure, the collapse of thin inner layers of stable sand arches prevailed as a failure mode. An unfavourable increase in sand production by thorough collapse of the sand body took place when the drawdown pressure was greater than the critical drawdown pressure. Therefore, determination of the critical drawdown pressure is very important to prevent catastrophic sand production. The closer the drawdown pressure and critical drawdown pressure are, the more difficult it is to form a stable sand arch. The formation of a sand arch was random and haphazard. The proposed numerical model is a promising method for studying the sand production mechanisms.