This paper is in two parts. The first part describes the experimental investigation enabling the precipitate to be identified as a single complex-structure solid. The second part describes a numerical model for predicting solid deposits.

In constructing the model, consideration was given to the experimental observation given above. Likewise, the concept of an equilibrium constant applied to each sensitive ion was used in the place of solubility products, which had become inadequate to describe an equilibrium in which the solid phase has a complex composition.

Comparisons were made between deposited masses, determined by computing and by experimental measurements for several water pairs. Quite satisfactory agreement was obtained in each case.

Several measured families of equilibrium constants were introduced into the model, corresponding to water pairs that are often encountered in reservoirs. With this database, the model is able to provide a mass deposit curve for a water pair having any composition. An example of prediction is given. It has to do with an industrial case applied to production.


Among the problems raised by waterflooding in oil fields, the physicochemical incompatibility between the injected water and the reservoir water is incontestably the least mastered problem at present from the production standpoint. This situation stems from the fact that each operation is practically a separate case in itself.

The deposits and corrosions generated by different incompatibility reactions between the waters lead to irreversible degradations of the reservoir rock (reduction of permeability) in wells and surface installations (loss of production). In all cases, they result in an appreciable increase in the production cost.

To help prevent such damage, this paper describes a numerical model for predicting solid deposits, based on knowing the specific structure of the mineral species that are liable to be encountered.

This content is only available via PDF.
You can access this article if you purchase or spend a download.