#20 Dealing with Mud Contaminated Samples in EoS Modeling

Taking a representative bottom hole fluid sample is a tedious and therefore also an expensive operation. Before a clean fluid sample can be taken, the well must be flooded for some time to remove the oil-based mud (OBM) in the near well bore area. Subsequently, a stable flow must be established from the cleaned well.

The cleaning operation is not always successful and many bottom hole fluid samples are OBM contaminated. In cases when a clean sample does not exist, the question arises as to whether it is possible to develop an EoS model for the clean reservoir fluid using PVT data for a contaminated fluid sample. PVT laboratories do not have the ability to physically remove the contaminant, so any measured PVT data will be for the contaminated fluid.

This TechTalk addresses the following topics:

  • What does the mud contamination consist of?
  • How does OBM contamination affect the phase properties of oils and gas condensates?
  • How to find out whether a fluid sample is mud contaminated?
  • How to develop an EoS model for a reservoir fluid when only PVT data exists for a contaminated fluid sample?

The mud contaminate consists paraffinic components in the range from ~C12 – C24. A typical OBM composition is shown in Table 1. It is dominated by C14 and C16. The mud components are rich in n-paraffins and their densities are lower than of the fractions of the same carbon number in a reservoir fluid.

ComponentMol %Mol WtDensity g/cm³
C120.997158.70.751
C130.173172.50.764
C1470.005187.20.775
C151.235203.00.780
C1627.06218.80.787
C170.269233.60.783
C180.139247.40.786
C190.035259.20.795
C200.023271.00.800
C210.015286.80.804
C220.009300.60.809
C230.008313.40.813
C240.007326.20.816
C250.006340.00.820
C260.006353.80.824
C270.005368.60.827
C280.004382.40.831
C290.005396.20.834
Table 1 Typical OBM composition.

For a clean reservoir fluid composition, ln(mol%) versus carbon number will form an approximately straight line for the components from C10 and onwards. If a fluid sample is OBM contaminated, the contamination will show up as peaks on a plot of ln(mole%) versus carbon number as is illustrated in Figure 1.

Figure 1 ln(mole%) versus C10+ carbon number for a clean and an OBM contaminated reservoir fluid sample.

Figure 2 shows the phase envelopes of an oil mixture and of the same fluid with 10 and 20 weight% OBM contamination of the STO oil  (the oil phase from a flash of the reservoir fluid sample to standard conditions). The saturation pressures of the contaminated fluids are lower than of the clean fluid.

Figure 2 Phase envelope for a clean reservoir oil composition and for the oil contaminated by respectively 10 and 20 weight% OBM of the STO oil.

Figure 3 shows what impact 10 and 20 weight% OBM contamination (% of STO liquid) has on the phase envelope for a light gas condensate. At temperatures above 50 oC, the saturation pressures of the contaminated fluids are slightly higher than of the clean fluid, but the impact on the saturation pressure is not as high as for the oil mixture in Figure 2.

Figure 3 Phase envelope for a clean light gas condensate composition and for the gas condensate contaminated by respectively 10 and 20 weight% OBM of the STO liquid.

The phase envelopes in Figure 3 do however not give the full picture of the impact of the OBM contamination on the phase properties of a light gas condensate. As can be seen from Figure 4, the OBM contamination has a major impact on the Constant Volume Depletion (CVD) liquid dropout.  Each 10 weight% OBM contaminate will increase the maximum liquid dropout by around 20%.

Figure 4 CVD liquid dropout curves for clean light gas condensate composition and for the same fluid contaminated by respectively 10 and 20 weight% OBM contaminate of the STO liquid.

Figure 5 shows the phase envelope of a rich gas condensate and of the same fluid with 20 and 40 weight% OBM contamination (% of STO liquid). The fluid is from a reservoir with a temperature of 143 oC. A PVT lab would classify the clean fluid sample and the one with 20 weight% OBM contaminate as gas condensates as the saturation pressures at the reservoir temperature are dew points. The sample with 40 weight% OBM contaminate would be classified as an oil as the mixture critical temperature is higher than the reservoir temperature and the saturation pressure at the reservoir temperature therefore a bubble point. This is an example of an OBM contamination masking the true reservoir fluid type.

Figure 5 Phase envelope of clean rich gas condensate and of the fluid  contaminated by respectively 20 and 40 weight% OBM of the STO liquid. The red dots show the locations of the mixture critical points.

One of the checks performed by the QC module in PVTsim is whether a fluid composition is contaminated by oil-based drilling mud. The QC module checks whether the ln(mole%’s) versus carbon number for the C10+ fractions follow a straight line or peaks exist as shown in Figure 1, signaling a mud contamination.

In the latter case, the user is directed to PVTsim’s Clean for Mud module, which will divide the sampled composition into reservoir fluid components and OBM components. In the example of Figure 1, the clean reservoir fluid composition is obtained by drawing a straight line through the ln(mole%) versus carbon number for the uncontaminated C10+ fractions. The area above the straight line represents the mud contaminate. When assigning EoS parameters to the reservoir fluid and to the mud, it is taken into account that the mud components are more paraffinic than the reservoir fluid components and therefore to be assigned lower Tc’s and Pc’s than the same carbon number fractions of the reservoir fluid.

Development of an EoS model on the basis of a mud-contaminated sample presents the challenge that the EoS model must be valid for the clean reservoir fluid, while PVT data only exists for the contaminated fluid sample.

Figure 6 illustrates the regression logic used for contaminated reservoir fluid samples in PVTsim Nova. The regression parameters are Tc and Pc of C7+ fractions in the cleaned fluid. In each iterative step the clean fluid is mixed with OBM contaminate to mimic the fluid used in the PVT experiments. The final EoS model is the one for the clean fluid when the PVT data for the Clean Fluid + OBM is matched (illustrated by green liquid dropout curve in Figure 6).

Figure 6 Regression logic used un PVTsim Nova for OBM contaminated samples.

Often, it is not sufficient to characterize a single possibly mud-contaminated fluid. There may be a need to characterize several fluids to a common EOS, of which one or more fluids may be OBM-contaminated. Characterization of multiple fluids for one common EoS model will be covered in a later TechTalk and the procedure to be presented is also applicable when some or all fluids are OBM contaminated.

Reference

Pedersen, K. S., Christensen, P.L and Shaikh, J.A. , “Phase Behavior of Petroleum Reservoir Fluids”, Taylor & Francis, 2015, Chapter 2.6.