Two phase flow applications in petroleum industry are so widespread. It is a fact that UBD precise bottomhole pressure maintenance ascertains UBD success. UBD hydraulics design, especially for inclined trajectories, is a real challenge. This is greatly dependent on the pressure drop in the annulus. Two phase flow through annulus is an ambiguous area of study to evaluate the bottomhole pressure. Two phase flow correlations on which most of UBD simulators based on over predict and also make extrapolation risky. Although Mechanistic approaches increase the frequency for designing two phase flow systems in pipes, modeling them through annulus by using the hydraulic diameter concept is not so successful. For this reason, their corresponding errors are not small. Therefore, in this paper, Artificial Neural Network is made use of to evaluate BHP in the inclined annulus using two major Iranian Oil Fields. To compare BHP found by neural network, Naseri et al mechanistic model which is a popular mechanistic model for these two fields is applied. ANN shows to perform much better than Naseri et al mechanistic model. The results show that neural network can estimate bottomhole pressure with an error of less than 20%. This proves that in case of existence of measured BHP while under balanced drilling, it is worth to use ANN to simulate BHP rather than mechanistic modeling or correlations. ANN is highly shown to be useful for solving the non-straightforward problem of two phase flow in annulus. Few jobs have been done to prove the superiority of ANN to mechanistic modeling and correlations in terms of pressure prediction especially in under balanced drilling.

Summary Core analysis is one of the most important steps in formation evaluation. The availability of routine–core–analysis (RCAL) and special–core–analysis (SCAL) data results in a better characterization of reservoirs and prediction of their behaviors. Unfortunately, the process of running core experiments, in both RCAL and SCAL phases, is very time consuming. Because all plug samples should be cleaned during the RCAL phase, finding a solvent that can speed up this process is desirable. The cleanliness of a core sample during Soxhlet extraction is usually determined by monitoring the color of solvents qualitatively. The main contribution of this study is to propose a methodology during RCAL to determine the best solvent during the Soxhlet–cleaning experiments. By introducing a novel quantitative method, the cleaning time of different solvents (i.e., tetrachloroethene, acetone, toluene, chloroform, xylene, and n–hexane) is investigated. This quantitative method is based on turbidity measurement of the solvent that siphons periodically from the Soxhlet extractor. Moreover, the wettability alteration of the implemented solvents is monitored by contact–angle measurements. To perform the analysis, two crude–oil samples (a heavy oil and a light oil with different asphaltene/resin fractions) and carbonate rocks from two Iranian formations are implemented. The results show that the polar solvents can speed up the cleaning process while altering the wettability of the carbonate samples toward more–water–wet conditions. The introduced methodology of measuring the cleaning time can be implemented as a routine screening tool in RCAL projects to determine the proper solvent that can reduce the Soxhlet–cleaning time.

The difficulty of maintaining the wellbore pressure should be overcome by precisely predicting the wellbore pressure. The empirical models used widely by UBD simulators give a large error and may not be extendable to other UBD cases. For this reason, Mechanistic modeling has been applied to two phase flow with a number of assumed values for the existent coefficients varying from one UBD case to another. Therefore, in this paper, a new comprehensive mechanistic model is developed mathematically for Iranian Oil fields to match the characteristics of the corresponding UBD jobs in Iran by modifying the constant values of Lage et al, Perez Telez at al., Mousavi et al. and Ansari et al. mechanist models. It is compared with Ansari et al, the specifically developed mechanistic model for use in South of Iran.Since the mechanistic models use mathematical relations with assumptions for the values of the constants rather than pure empirical study of two phase flow, the mechanistic models developed and validated with data outside Iran are not good enough to be used. In this paper different constant values have been tested and the ones which have led to the least average absolute error have been picked up to present a specific mechanistic model for South of Iran.The empirical models which have been developed outside Iran cause erroneous results. The results illustrate that the new mechanistic model predicts BHCP with an error much less than Ansari et al. Mechanistic Model.

No abstract

Knowledge of permeability is critical for developing an effective reservoir description. Permeability data may calculated from well tests, cores and logs. Normally, using well log data to derive estimates of permeability is the lowest cost method. In the last years, the concept of hydraulic flow units (HFU) has been used in the petroleum industry to improve prediction of permeability in un-cored intervals/wells. This concept is strongly related to the flow zone indicator (FZI) which is a function of the reservoir quality index (RQI). Both measures are based on porosity and permeability of cores. It is assumed that samples with similar FZI values belong to the same HFU. This paper will focus on the evaluation of formation permeability in un-cored intervals for a carbonate reservoir in Iran from core and well log data. We used Flow Zone Index method for rock type identification and Artificial Neural Networks for permeability estimation. Identifying the hydraulic flow units that is the first step of predicting permeability always takes lots of time and lack appropriate accuracy. We have developed a new clustering technique that is more precise, easy to apply and taking much less time. IntroductionReservoir rock typing is a process for the classification of reservoir rocks into distinct units. If the rocks are properly classified and defined, the real dynamic characteristics of the reservoir will be provided in the reservoir simulation model. Of course, direct measurement of rock properties using cores is the ideal method to determine HFU's. However, because the costs to cut and analyze cores are so high, few core measurements are routinely available. Several investigators have noted the inadequacy of classical approach and have proposed alternative models for relating porosity to permeability. From the classical approach it can be concluded that for any given rock type, the different porosity/permeability relationships are evidence of the existence of different hydraulic units. Rock type identification and prediction of porosity/permeability relationships is a process that has been done in the literature with various rock-typing techniques but the question we have to deal with is to find the FZI for each log data. Hence, it is crucial to extend the flow unit determination to the un-cored intervals and wells. The relationship between core flow units and well log data was established by artificial neural networks, and then was used as a tool to extend the flow units prediction to un-cored intervals and wells.Petrophysical properties are controlled by both depositional characteristics, such as grain-size and sorting, and by diagenetic features, such as the amount and type of cement or clay minerals. Thus each HFU involves more than a genetic facies of a depositional system. Generally, the variability of petrophysical properties is large among the HFUs and low within them. Permeability calculation by HFU approach offers an improved estimation over traditional regression-based averaged relationships by incorporating geol...