Wax precipitation is one of the most challenging flow assurance problems in the oil and gas industry, as it can lead to flow issues, equipment failures, and production loss. This is a complex process that involves both phase separation and solidification.

It is critical to be able to accurately predict the amount of precipitated wax for each well in order to avoid problems during intervention into the well. Our team developed Perturbed Chain Statistical Associating Fluid Theory to accurately predict temperature (WAT) and amount of precipitated wax without the use of costly PNA analysis. The advantage of PC-SAFT is mainly its accuracy in the estimation of the frugalities of heavy components in vapor and liquid mixtures.

We know that the composition of wax is different across the field, which makes it more difficult to predict wax precipitation if wax is represented as one component. We used a different approach by dividing the reservoir into sectors with similar wax properties, calibrating each sector separately, and obtaining one set of PC-SAFT parameters for every sector. If prediction of a new well is required in the future, then the sector to which this well belongs needs to be determined, and this calibrated set of parameters for that part can be used.

Feed-Forward Neural Network trained on Distributed Temperature Sensing (DTS) data was used to predict the temperature inside the wellbore based on the production data for each well. Usually, the temperature of the oil in the flow center is hotter than DTS. Therefore, the heat transfer equation was used to model the radial temperature gradient and therefore determine the temperature of the oil flowing inside the tubing. Applying the heat transfer equation increases the accuracy of the PC-SAFT model and enables company engineers to compare and identify all wells with a high risk of wax precipitation and, thus, to know about the risk of wax precipitation before intervention into the well.

With this product, we developed the underlying theory of how solution-solid and multi-solid models are used together with the PC-SAFT Equation of State in the three-phase flash calculation algorithm.

Fig. 1. (a) Solution solid model, (b) Multi-solid model.

A developed thermodynamic model used sensitivity analysis in order to identify what affects wax precipitation the most, such as gas/oil ratio (GOR), pressure, and temperature. According to sensitivity analysis, temperature is a dominant parameter. Depending on the temperature drop, it may increase the amount of precipitated wax by 2–4 times.