The water produced from petroleum reservoirs must be cleaned before its disposal or re-injection. Re-injection of the produced water is complicated by the fact that this water, after all the cleaning procedures, may still contain some very small solid particles and oil droplets. Upon re-injection, these inclusions are filtered by the rock and may cause injectivity decline and formation damage. This is a long-term process involving different time scales (from milliseconds to years) and different space scales (from micrometers to hundred meters). Similar problems arise under the filtration of water (contamination of the filters).
This project will concentrate on the collective behavior and filtering of the liquid micro- and nanodroplets contained in the injected water. Filtration of particles has been considered previously in multiple studies, including those of the applicant. In particular, software (SNY) was produced, to predict the formation damage by filtration. Meanwhile, the filtration of droplets is still an unexplored area, with several distinctive physical mechanisms.
Due to the different droplet and pore sizes, they may penetrate at different distances inside the reservoir rock. The droplets may interact with each other and with the particles forming agglomerates, layers, and sticking to the rock surface. The droplets may plug narrow throats and stop the flow, due to the so-called Jamin effect. But they may also squeeze through the capillary, if, due to the agglomeration, they have grown to larger sizes. The oil droplets may join the residual oil phase, expanding it and making it slightly more mobile. These and other effects have not been sufficiently described in the literature. Recent microfluidic studies carried out at the DTU Offshore in collaboration with KT, uncover a multiplicity of phenomena occurring with the liquid droplets in the flow through thin capillaries.
All these processes are affected by the production chemicals present in the water: demulsifiers, corrosion inhibitors, scavengers, and others. Recent microfluidics studies carried out by the DOTC (where the applicant has participated) indicate that the chemicals may change the capability of droplets to agglomerate: coalescence frequency, droplet sizes, and others. Water salinity and the collective action of several chemicals may also affect the picture.
The rates of droplet motion and capturing and the depths of penetration depend not only on the droplet and pore sizes but on the size distributions. Hence, probabilistic modeling of the process is necessary, to do not only determine the injectivity decline but also, how deeply it affects the rock around the well. Advanced methods for such modeling have been developed during the last two decades. However, they have mainly addressed solid particles. Liquid droplets behave differently in a porous medium, as described above.
The questions of potential industrial interest are: How rapidly the near-well zone will be contaminated? How deeply the impurities will advance? How does it depend on the methods for produced water cleaning and added/removed chemicals? And what methods for eventual cleaning of the wells will be effective in the removal of the filter cakes and increase of the injectivity?
Answering these questions requires advanced analysis of the existing data on the composition/filtration of the produced water; and the building of a stochastic large-scale model for prediction of the formation damage. Such a model should involve the effect of dispersed droplets. This effect is planned to study in the current research proposal.