• info@accreteltd.com
• +234 908 555 0001, +44 122 428 4386

# Posts

## PVT Modelling Fundamentals – Part 2

### BLACK OIL PVT MODELIING

Black Oil (BO) PVT modelling is a 2-component description of the hydrocarbon fluids as ”surface gas” and “stock tank oil”. This PVT modelling method is widely used due to its simplicity and limited input data requirement. This method relies on empirical correlations derived from different oil samples. Some correlations focused on crude samples within a geographical region (e.g., Glaso for North Sea crudes) while some focused-on crudes with specific properties (e.g., De Ghetto et al. for heavy and extra heavy crudes). As a result, different PVT correlations sometimes yield significantly different results from the same input.

Generally, the inputs into BO correlations are properties of the surface gas and oil i.e., API gravity of the stock tank oil, specific gravity of the surface gas and the gas oil ratio (GOR). With these inputs, several other PVT properties like formation volume factors, viscosities, densities etc. can be calculated over different pressure and temperature ranges.

With the same set of inputs (API, GOR, Gas Gravity), different BO correlations will give different results thereby presenting a challenge on which correlation to select.

### Selection of BO PVT Correlations

Ideally, PVT data are determined from laboratory experiments performed on representative fluid samples collected from wellhead, surface separators, or downhole. However, these experiments are either conducted at limited pressure-temperature conditions or only consider a few PVT properties. Hence, the need for models that can cover wider pressure-temperature conditions and applicable to any required PVT properties.

Therefore, in selecting an appropriate BO correlation, laboratory experimental data (if available and reliable) is often used as a guide. The approach is to compare the results from the different BO correlations with the measured laboratory data and then select the correlation closest to the measured data. This can be done using a single data point (recommended at the saturation pressure) or for a range of datapoints if available.

There is usually a higher level of confidence when comparing with multiple data points. The figure below illustrates the selection process for a single point and for multiple datapoints using Rs as an example.

When there is no laboratory data, which is sometimes the case, other methods can be adopted in selecting a suitable BO correlation. Some of them include:

• geographical location: In the absence of laboratory data, many engineers choose BO correlations that took crude samples within similar geographical regions as the fluid they are modelling. For instance, using Standing BO correlation for a Californian reservoir considering Standing took his samples from the same region. Although it is logical to assume that crudes in different reservoirs in a particular geographical location are of similar origin (source rock), it is erroneous to assume that this approach is always reliable. The samples considered in developing the BO correlation may have been taken from significantly different depths compared to the reservoir fluid in question. This may then result to major difference in properties. It is also possible that there are different source rocks with different properties within a geographical location.
• Selection based on fluid types: De Ghetto et al. correlation for instance was developed for heavy and extra heavy crude oils. It is therefore recommended to use this correlation for heavy crudes, and it becomes unreliable when used for other types of crude.
• Selection based on Experience and Analogues: In mature fields with extensive production and history, it may have been observed that certain BO correlations have performed reliable well over time. So, when trying to choose a correlation for a reservoir within the same region, without lab data, it can be assumed that the same correlation will perform reliable well. This assumption is based on the premise that the crude in question have similar properties/origins as the other crudes in the mature filed/region. It is however erroneous to assume that this approach is always reliable.

Several other selection methods/criteria have been developed, some of them incorporating multiple considerations much more complex than outlined here. The aim is to have a systematic and logical approach to choosing BO correlations while making considerations for the associated limitations of each approach.

While BO correlations are extensively used due to their simplicity, there are other situations when BO correlations in general become unreliable:

• When fluid properties change rapidly for a given change in pressure-temperature, it becomes difficult to capture this change with BO correlations. This is usually the case for gas condensates or volatile oil. In general, as the reservoir conditions get closer to the critical conditions, the more difficult it is to capture the fluid behavior with BO correlations.
• Some modeling objectives are better achieved using detailed EoS PVT modelling e.g., flow assurance studies, miscible gas injections, tracking the evolution of a particular component etc.

It therefore becomes necessary to consider other modelling options in situations where BO correlations are generally unreliable or when the modelling objectives cannot be achieved with BO correlations. One of the alternatives is to modify the BO correlations to account for oil that comes out of gas at surface conditions i.e. Modified Black Oil (MBO) correlations or to model each component of the reservoir fluid in detail using Equations of States (EoS).