Reservoirs can sometimes be very sensitive to drilling fluids. Typically, the lower the permeability of the rock, the greater the likelihood of experiencing phase interference effects; as the average diameter of the porous features decreases, the greater the capillary pressure. Many times this phase interference effect significantly reduces well productivity and hence, alternative drilling fluids are used. Moreover, aqueous phase fluids can react with the reservoir rock causing clay swelling, clay flocculation and fines migration.

In an attempt to mitigate the above-mentioned deleterious phenomena, alternative non-aqueous drilling muds are sometimes used. These are designed to be compatible with the reservoir fluids in-situ and because they are hydrocarbon-based, they do not interact with the rock matrix. One of the drawbacks, however, is the mixing that occurs with the oil in-situ. When early-time samples are taken, the oil is contaminated with the drilling fluid. This is a near-wellbore and early-time problem that is self-correcting as more fluids are produced from the well. However, for small volume samples (typically bottom hole samples), contamination can seriously distort the properties measured on the sample. The small-volume samples are often very expensive to procure and many of the reservoirdevelopment decisions are based on the properties of these small samples. The properties need to be accurately measured therefore.

How can the contamination be quantified both in terms of mass or mole percent in the fluid and its impact on the sample properties? This paper describes two technologies developed to achieve this. The first is applicable to synthetic hydrocarbon drilling fluids where the concentration of components is restricted to very few components. The second technique applies to those fluids that contain a more broad distribution of hydrocarbon components.

The results indicate that the resolution of contamination can be achieved to within 1 mass percent accuracy. Using the degree of contamination with Equation of State methods, can they predict properties of the reservoir fluid to within 4% of actual value?


The objective of this work is to develop a technique that will quantify how much contamination is present in the sampled oil and then provide a means by which PVT parameters on reservoir fluid can be approximated from a contaminated sample. The basis for this work is having no "clean" oil available and therefore the uncontaminated properties must be extrapolated from the available samples. The scope of the work consists of:

  • Mathematical development

  • Additional experimental data on selected samples

  • Validation of mathematical approach

  • Estimation of the oil-based mud filtrate contamination on selected samples


Depending on the type of drilling fluid, the determination of the amount of contamination can be facile or it may be somewhat more involved. From the authors' experience, drilling fluids whose compositional distribution is very narrow are the easiest to correct. This type will be discussed in the first section. Those drilling fluids that possess a much broader distribution of components are more difficult to quantify. This type is discussed in Section 2.

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