Generally, EOR processes are divided into four categories: thermal, gas, chemical, and other (see Table 1) [4] .

As aforementioned, higher oil prices and concerns about future oil supply are leading to increased interest in Enhanced Oil Recovery (EOR) around the world. So, extensive researches have been conducted to develop various EOR methods, evaluate their applicability and optimize operation conditions [5] [6] . But, since EOR projects are generally more expensive and involve higher front end costs than conventional secondary projects, one of the principal areas is to develop an effective tool for selection of a suitable EOR method according to oil field characteristics [7] . Due to the high importance of this tool, a large number of studies have been conducted to help the petroleum engineer select efficient EOR methods with limited field information [8] - [15] . In one of

the first efforts, screening criteria were briefly presented in a series of tables and simple graphs [16] . Then, EOR methods were classified based on field data and EOR mechanisms [17] [18] and were updated in 2011 [19] . The main tool for screening of EOR techniques is generally based on the criteria presented in a variety of tables and graphs given in the literature. The most efficient tool was proposed by Taber et al. [16] - [18] and Goodlett et al. [20] established the technical screening criteria of each EOR method. In recent years, simulation methods, artificial intelligence and neural networks have improved the EOR screening methods. Zerafat et al. [7] solved the problem of selecting appropriate EOR methods by Bayesyan network. So, in the current investigation, the screening study was performed to find which EOR method is the most proper method for tertiary oil recovery. For this purpose, intelligent screening method based on Bayesian network (based on Zerafat et al. method [7] ) was used to find the proper EOR methods.

A full description of the used method is given elsewhere [7] . In brief, a data bank has been gathered from worldwide EOR/IOR techniques and analyzed using data mining procedure which is then combined with extracted data from previously published screening tables. Bayesian network quantitative learning technique was applied to different data combinations from the data bank to train the network which is to serve as the expert system. The produced expert system is also applied to the gathered data pertaining to 10 Iranian southwest reservoirs. This expert system is based on Bayesian network analysis in order to sort the proper EOR techniques. By inserting several inputs including viscosity, permeability, specific API and reservoir temperature, and depth of reservoir, the used intelligent method demonstrates several possible EOR methods along to the percent of the EOR method efficiency.

In general, surface samples are obtained at the separator require collection of high stage separator gas and liquid which must be subsequently recombined in a ratio that corresponds to the relative amounts of gas and liquid produced, as the reservoir fluid travels up through the wellbore and on through the surface separation facilities [21] . The recombination of surface separator samples is achieved by recombining the gas and fluid either to match the measured separator GOR or to match a specified saturation pressure at the reservoir temperature. In cases where the reservoir fluid is known to be highly under saturated, the target saturation pressure may be significantly lower than the actual reservoir pressure and therefore the separator GOR may be a better reservoir fluid characteristic to attempt to match [22] .

As shown schematically in Figure 1, firstly the hydrocarbon fluid sample (oil or gas) metered into a visual PVT cell at reservoir temperature and at a pressure in excess of the initial reservoir pressure. Then, the CCE test was started by lowering the pressure in small increments and recording of changes in volume at each pressure. The procedure is repeated until the cell pressure was reduced to a pressure that is considerably lower than the saturation pressure. The original composition of the fluid in the cell does not change at any time during the test because no material is removed from the cell. The fluid may be either oil or a gas with condensate. If the fluid is oil, the saturation pressure is the bubble point pressure. Finally, the volume of the hydrocarbon system as a function of the cell pressure is reported as the ratio of the reference volume [21] .

In general, in the differential liberation process, live oil sample is placed in a PVT cell at reservoir temperature and at a pressure above the bubble point pressure. Then, the solution gas that is liberated from an oil sample during a decline in pressure is continuously removed.

In brief, as shown schematically in Figure 2, the pressure is reduced in steps, usually 10 to 15 pressure levels, and all the liberated gas is removed and its volume is measured at standard conditions. The volume of oil remaining V_L is also measured at each pressure level [21] .

This test is usually conducted for reservoirs to be depleted under gas injection, miscible type displacement or

dry gas cycling. In this test, a gas of known composition, usually similar to the proposed injection gas, is injected in a series of slugs into a reservoir oil sample, as shown in Figure 3. The injection of the first gas slug starts at the bubble point pressure of the reservoir fluid sample and continues until a considerable amount of gas is injected. After each gas addition, the cell is pressured up until only one phase is present [23] .

Generally, it is not possible to consider the best EOR method based on one or two characteristics of the reservoir. The available software capable one to find the proper EOR method using viscosity, permeability, specific API˚ and reservoir temperature, and depth of reservoir at three different modes. The results listed in Table 2 revealed that based on the three different modes, three different methods including miscible CO₂, miscible nitrogen and miscible hydrocarbon injection are the most proper EOR methods. In the light of these findings, CO₂ and N₂ injection on the live crude oil was investigated in the remained stages of this study to reliably check which gas is more efficient for enhanced oil recovery. In addition, the results obtained from the screening method must be corrected using availability of gases and economic and environmental issues.

In this work, recombination of provided gas and oil from separator (P = 334 psi and T = 100˚F) using GOR of 601 scf/STB was performed. To examine if the recombination was correctly carried out, the bubble point pressure of the prepared live-crude oil was measured using visual PVT (CCE test) equipment. The obtained results revealed that the bubble point pressure of the recombined live crude oil was about 1727 psi (see Figure 4). In other words, the average absolute relative deviation percent (AARD%) of the bubble point pressure obtained by prepared crude oil compared to actual sample of live crude oil was about 8% means the recombined crude oil is well close to reservoir sample.

The differential oil formation volume factors B_o from differential liberation (DL) test (commonly called the relative oil volume factors) at all the various pressure levels are calculated by dividing the recorded oil volumes V_L by the volume of residual oil V_sc, or:

The B_o at pressure greater that bubble point pressure could be obtained by simultaneously using of CCE and DL tests (see Figure 5):

The differential solution gas-oil ratio R_sd is also calculated by dividing the volume of gas in solution by the residual oil volume. As illustrated in Figure 5 and Figure 6, the differential relative oil volume (B_o) and the differential gas-oil ratio (R_s) at the bubble-point pressure are 1.47 rbbl/STB and 600 scf/STB, respectively.

Also, the gas formation volume factor B_g (see Figure 7) can be described by the following equation [21] :

where B_g is the gas formation volume factor in ft³/scf, T is the temperature in R, P is the cell pressure in psia and z is the gas deviation. The z-factor was calculated from the recorded gas volume measurements as follows:

where V_R is the volume of the liberated gas in the PVT cell at p and T, V_sc is the volume of the removed gas at standard, T_sc is the standard temperature in R and P_sc is the standard pressure in psia.

Finally, the total live crude oil formation factor, B_t (see Figure 8) and coefficient of isothermal compressibility, c₀ (see Figure 9) were successfully calculated from the recorded data points (see Equations (5) to (7)) [21] .

The live-crude oil was initially metered into visual PVT cell at bubble point pressure of 1727 psi bar and reservoir temperature of 212˚F. After that, injection of CO₂ and N₂ was begun. As it is demonstrated in Figure 10,

injection of CO₂ can increase the relative volume (expansion) to 1.35 at high CO₂ saturation pressure while the injection of N₂ cannot significantly increase the swelling factor even at high saturation pressures. Generally, it is obvious that not only CO₂ leads to the lower bubble point pressure compared with N₂, but also leads to higher swelling factor (see Figure 10). Totally, it can be concluded that it is possible to have higher oil recovery by injection of CO₂ instead of N₂. It should be mentioned that, the N₂ injection leads to a significant increase in the bubble point pressure of the crude oil which is not desired in the gas injection process for EOR methods. In addition, the fluid most commonly used for miscible displacement is CO₂ and N₂ because it reduces the oil viscosity and is less expensive than liquefied petroleum gas [24] . Since the miscibility conditions of N₂ is very difficult to be achieve in most of the cases, CO₂ injection has been gain an increasingly attention. On the other hand, due to the environmental concern of greenhouse gas effect on the global warming CO₂ capture and sequestration is being considered as one of the most attractive research area during the past decade [25] .