Abstract
The apparent rocks permeability, measured under cyclic loading, tends to constantly decrease with each subsequent loading/unloading cycle, the main reason for this, the researchers consider irreversible deformations. However, this understanding is not complete, as attention is not paid to the method of determining the permeability. Under cyclic loading, permeability is measured by fluid flow, which is able to carry colloids and they can clog pore throats and reduce permeability, but this is not taken into account in the study of porous rocks. Colloidal migration is also not taken into account in routine core tests, although colloids can significantly reduce the apparent permeability of rocks during migration. Also it is impossible to determine exactly which of the factors (mechanical compaction or colloid migration) contributed to the change in apparent permeability. The purpose of this research is to provide experimental evidence that under cyclic loading, permeability decreases not only due to mechanical compaction, but also due to colloid migration within the porous medium. For this task, a methodology has been developed, the feature of which is that between loading/unloading cycles, the core is blowning with a large pressure gradient. It has been established that under cyclic loading, the apparent permeability of rock samples changes as a result of colloid migration up to 20 %. The dynamics of permeability becomes unpredictable and weakly depends on the confining pressure when changing the direction and regime of the gas flow. At the end of the multi-cycle confining pressure test, the apparent permeability of 3 out of four core samples, despite the decrease during injection, recovered to the initial value which also confirms that the apparent permeability has affected by colloid migration. A mechanism for changing the apparent permeability during gas injection and cyclic confining pressure is proposed. The outcomes indicate that colloid migration should be taken into account during core tests not only in cyclic loading but also in routine coreflooding tests.