# 3 Simple Steps to Best Match Your PVT Experimental Data

Updated on June 2022

You may know why it’s vital to obtain the most representative PVT properties of your reservoir fluids. The best practice to do so is to carry out a PVT study in the laboratory.

Almost all black oil properties are measured in a standard PVT experimental study. Once you receive the study results, you would need to match the measured PVT properties to specific correlations. But why?

In most reservoir fluid PVT experimental studies, black oil properties are measured at a limited number of pressure points, ranges from 15 to 30 pressure points only. In petroleum engineering applications, you would need to calculate the black oil properties at almost any value of pressure, and that’s why matching is required.

The purpose of the PVT matching is to predict the black oil properties at pressures not used in the measurement. To do so, you need to obtain the best-fit correlation model to the measured data. And here is how to achieve this:

## Plot lab data and correlations

It will help if you can visualize how the calculated properties are biased from the laboratory-measured data. In other words, you would need to identify the mismatch between model data and actual properties. And here is how to do so:

1. Calculate all the measured black oil properties using multiple PVT correlations from the petroleum literature.
2. On the same chart, plot the measured property, obtained from the laboratory study, and the calculated properties, obtained from various PVT correlations, for each black oil property versus pressure.

Bear in mind that all calculated properties must be obtained at the same reservoir temperature and the same pressure points as measured in the PVT study. Moreover, you have to input the measured field parameters (gas gravity, oil gravity, and solution gas-oil ratio at the bubble point pressure) into the correlations you would use.

## Fit all correlations to lab data

So far, you have built multiple calculated models for each black oil property; it’s time to fit all calculated properties to the lab data. The best practice is to manipulate all used correlations with two modifiers, multiplier (a) and shift (b), as per the equation:

Y = AX + B, where:

• Y indicates the matched, or corrected, correlation;
• X indicates the unmatched, or original, correlation;
• A indicates the multiplier to the original correlation; and
• B indicates the shift to the original correlation.

And here’s how to do so:

1. At each pressure point, calculate the PVT property in interest using the original correlation, X, and you have already done this in step 1.
2. Using non-linear regression, calculate both matching modifiers, a and b.
3. At each pressure point, calculate the PVT property in interest using the matched correlation, Y, as per the above equation.

## Select the best-fit correlation

Now it’s time to select only one PVT correlation for each black oil property. The best fit-correlation is the one with the minimum mismatch to the lab data. And here is how to select the best-fit correlation to your lab data:

1. Calculate the average absolute relative error, AARE, between lab data and all original correlations.
2. Sort all correlations based on the minimum AARE.
3. The correlation with minimum AARE is the most representative model for your lab data.

## Conclusion

This post explains how to match your experimental PVT data to several empirical correlations. The objective is to obtain the most representative model of your reservoir fluids that you can confidently use in almost any petroleum engineering applications.

PVT matching can help you predict the black oil properties at pressures not measured in the experimental study. It’s also may help in validating PVT study results as well.