PORO is a semi-analytical simulator for tracer flux in an oil reservoir that supposes homogeneous and non-isotropic horizontal layers in which several vertical sources and sinks are present. The conceptual model is based on the analytical solutions of Darcy and convention diffusion equations. This simple model is more applicable than a more detailed one because of the high level of uncertainty in the reservoir description, especially after the water flooding in oil secondary recovery projects. The quality of the results was checked by comparing the simulator predictions with experimental results from laboratory and field exercises under different conditions.

How does PORO work?

The PORO simulator allows the evaluation of the principal parameters of the watered layers, by matching the experimental data on the basis of:

• A number of vertical injector wells (with arbitrary coordinates);

• A number of vertical producer wells (with arbitrary coordinates);

• Uniform water flow rates;

• Homogeneous horizontal layers;

• Non-lateral boundaries or sealing faults as boundaries;

• Anisotropy (the Kmax/Kmin ratio and the direction of Kmax must be specified,

K being the permeability).

The PORO simulator works by taking into account:

• That the wells are ‘vertical lines’ sources or sinks (cylindrical geometry);

• Analytical solutions for the velocity field (and the superposition principle);

• The generation of the streamlines between each injector and the connected

producers;

• The solution of the convection diffusion equation on each streamline and

computing the overall concentration in each producer.

For solving the convection diffusion equation, PORO converts the time into frequency (by Fourier transform) and transforms the space into a discrete one along each streamline. Finally, a simple scheme of finite differences is employed.

(77)

Prior to starting the simulation, the following information has to be entered:

• The type (injector or producer) and the coordinates of the wells;

• The water flow rates of each well;

• The layer thickness, porosity and water saturation;

• The layer dispersivity (in oilfield scale, it is in the order of 10% of the distance between wells);

• The anisotropy ratio (Kmax/Kmin ) and the direction of Kmax;

• The sealing faults, if any;

• The time and frequency ranges (between 0.00001 and 1 for the conventional oilfield scale);

• The number of streamlines and the distance between consecutive points in each streamline.

Preparation of a software package for covering the tracer dynamics in all rock reservoir situations is a difficult, but necessary task. The most frequent cases include water stationary fluxes in which a conservative tracer moves in approximately horizontal and homogeneous layers (with simple primary porosity). These cases can be interpreted acceptably by the methods discussed here. Also, absorption, partition in the hydrocarbon phase, radioactive decay and tracers in stationary gas flows can be easily included in the models. However, it is very frequent to find cases in which the tracer moves in rocks with double porosity, particularly along conductive fractures [41-45]. Other anomalous cases are related to the tracer flow in ‘wormholes’, which are typical in unconsolidated sands and heavy oil reservoirs [46, 47]. Finally, it is necessary to include other geometries for covering 3-D flows and 2-D special cases, such as flow in horizontal wells.

The most common beta radioactive tracers for interwell studies are labelled with tritium 3H, 14C or 35S. All of them are usually measured by means of a liquid scintillation counting technique. A small volume of a liquid sample is mixed with a special solution known as a ‘scintillation cocktail’, commonly in a 20 mL light transparent (glass, polypropylene, teflon) vial. Beta particles cause emission of light when passing through and slowing down in the scintillation cocktail. These light pulses are registered by photomultipliers (PMTs) suitable for that particular photon wavelength. The light output in a pulse (light intensity) is proportional to the energy of the beta particle. This process is termed scintillation, and since it happens in liquid media, it is known as liquid scintillation.

The vial is placed inside an instrument, a liquid scintillation counter, which normally has two PMTs operating coincidentally to reduce the background. The liquid scintillation counter analyses the pulses from the PMTs and provides information about the energy of the beta particles and the rate of beta emission (activity) in the sample.

Pulses are sent to an analogue-to-digital converter where they are digitized and stored in an address memory according to their amplitudes, which are proportional to their beta energies (energy spectrum in a multichannel analyser).

In order to reduce further the background coming from natural radiation, a lead shield usually surrounds the PMTs and the vial while the sample is in the measuring position. Modern low background detection equipment also has a so-called active shield. In most cases it consists of a liquid scintillation detector surrounding the PMTs and the counting sample. This shield detector is operated in anticoincidence with the PMTs, such that any event which is registered both in the two PMTs and in the shield (cosmic rays, environmental radiation) detector simultaneously is rejected. In the case of simple non-spectrometric detection equipment (single channel analyser), the contribution of the background to the sample count rate can be further reduced by setting a counting window over only the interesting energy portion of the energy distribution. This is achieved by selecting narrow upper and lower limits. In the case of tritium, the upper gate should, for instance, be set at 19 keV.

Various processes may perturb the beta spectrum obtained in a liquid scintillation process. The most important of these are:

• Chemiluminescence: When different chemicals are mixed in the sample vial together with the scintillation cocktail, chemical processes may start which have relatively slow kinetics and which result in the emission of low energy photons. These photons may contribute to the very low energy end of the beta spectrum. Chemiluminescence may be reduced or completely removed by gentle heating of the vial to 50-60°C for some minutes before counting in order to speed up the chemical process.

• Phospholuminescence: When a sample vial with the scintillation cocktail is exposed to white light (daylight or lamp light), the light energy may be temporarily ‘stored’ and slowly released during sample counting (phosphorescence). Also, this light will contribute to the very low energy end of the beta spectrum. Therefore, counting samples should always be stored in the dark for a few hours before counting starts.

• Colour quenching: A coloured sample liquid may absorb some of the light emitted by the scintillator. Yellowish or brown colours are the heaviest colour quenchers. Hence, attempts should be made to remove such colours during the sample preparation process and before counting.

• Chemical quenching: Some components in the sample may kill the energy transfer process that takes place in the scintillation cocktail and which eventually results in light emission. Such chemicals absorb the energy and release it in the form of heat. Heavy chemical quenchers include, for instance, organic compounds containing oxygen and in particular chlorine.

• Physical quenching: Solid particles or non-transparent emulsions in the sample may prevent light from being detected by the PMTs.

All of these forms of quenching result in a shift of the energy spectrum towards lower channel numbers because the number of photons detected by the PMTs per beta decay is reduced. Figure 21 shows in principle the effect of quenching. Quenching may change from one sample to another. Evaluation of the quenching effect is necessary in order to calculate counting efficiency.

In summary, liquid scintillation counting requires careful sample preparation. Chemical separations are most often involved and when these procedures are optimized, very low detection limits may be obtained, ranging from 2 Bq/L for HTO to <0.02 Bq/L for S14CN-.

HTO (10 Ci) was injected into well 1R8D on 12 July 2006. Ten production wells were chosen for tracer breakthrough monitoring. These are wells 101, 103, 105D, 109D, 202, 209A, 212, 214, 2R2, 2R3D and 2R4D. Wells 105D, 103 and 109D were selected on the basis of their rapid physicochemical response, employing 1R8D as a reinjection well. Wells with considerable steam fraction (enthalpy of 1800-2700 kJ/kg) were selected deliberately to determine the behaviour of tritium in the vapour phase. However, a few watery wells were also included in the programme to determine tritium’s fractionation between the water and vapour phases.

As a basis for selecting samples from ‘watery’ wells, 1,5-naphthalene disulphonate (NDS) was injected on 14 June 2006, one month prior to the injection of tritium in 1R8D. The watery wells that showed NDS breakthroughs were chosen for tritium analysis (i. e. 2R3D, 2R4D, 202, 214). Moreover, the results of NDS tracer were also used to compare and/or confirm the tritium data collected.

Continuous sampling was conducted for a year in the selected monitor wells. However, analysis was terminated for the wells which did not yield positive tritium returns after six months from injection. Tritium was analysed at the Philippine National Research Institute and at CAIRT, Jakarta, Indonesia. The NDS analysis, on the other hand, was done at the LGPF Geoservices Laboratory using high performance liquid chromatography.

Data reduction, both for tritium and NDS, was achieved using the Anduril 2.3 package specifically designed for radioactive tracers. NDS data were processed using the ICEBOX software package (United Nations University Geothermal Training Programme, 1994) and Anduril 2.3. Results from both software packages revealed almost identical values.

For relatively pure water samples (type A), the procedure is as follows:

(1) Filter the sample through a lipophobic filter to remove any traces of dispersed oil droplets and any suspended particles.

(2) The oilfield samples usually contain high salt contents and chemical load. This raises the quenching effect profoundly during sample measurement, which introduces inaccuracy in the results. Though correction factors for chemical quenching can be applied to rectify the quenching effect, it is always desirable to avoid such complicated processes. For this reason, samples are distilled before measurement. All necessary glassware should be thoroughly washed and dried to avoid contamination. Therefore, transfer preferably 250 mL of the water filtrate into a 500 mL round flask.

(3) Assemble the distillation equipment as shown in Fig. 90.

(4) Distil with gentle boiling and collect distilled water in the sidearm of the glass equipment.

(5) Discard the two first fractions of the distilled water because these may contain volatile hydrocarbon components which have been dissolved in the original water sample and which may quench the scintillation process.

(6) Collect the third 8-10 mL sample for analysis. If, for any reason, the water sample cannot be analysed the same day that it is purified and distilled, the samples (vials) should be stored in the refrigerator until they are prepared for counting.

(7) Mix with 10-12 mL of an appropriate scintillation cocktail which is able to accommodate about 50% of water without a detrimental decrease in counting efficiency, for instance Ultima Gold.

(8) Store the counting sample for at least 1 h in the dark in order for any chemiluminescense or phosphorescence to die out before starting the counting sequence.

(9) Three background samples (‘dead’ water) and three standard samples with known activity on a given date are prepared and counted along with the produced water samples.

(10) If variable quenching is suspected in the produced water samples, a correction has to be made for the corresponding varyiable counting efficiency. Most modern liquid scintillation counting equipment provides an instrumental quench correction method. In case this is not the case, the quench correction can most easily be performed by the internal standard method. After having counted the samples, a known activity of a non­quenching tritium compound (standard solution) is added to each vial and counted again.

This gives the counting efficiency directly according to the formula:

where

Rx is the background corrected counting rate of the produced water sample;

Rx+s is the background corrected counting rate of the sample after addition of a known quantity of standard;

Ds is the disintegration rate of the added quantity of standard (Bq).

The activity of tritium in the original sample, Dx (Bq), can then be calculated by:

and the activity concentration, Dxc in Bq/L, in the produced water by:

where Vx is the volume of the water in the counting sample in millilitres.

For samples with a visible layer of oil (type B), remove the oil layer on top of the water with a pipette, then carry out the same procedure as outlined in point 1 above.

For the most difficult samples consisting largely of an oil-water emulsion (type C), the first step is to break the emulsion in order to sample any water dispersed in the oil. There are a variety of commercial emulsion breakers available (a chemical supplier can assist).

Then carry out the same procedure as for point 1 above, with the modification which takes into consideration the volume of water available after emulsion breaking may be limited, i. e. a few millilitres only. In that case, the distillation process may be omitted or performed in miniaturized equipment.

Tracer applications can be found in almost any phase of oil field development. Interwell tracer technology is an important reservoir engineering tool for the secondary and tertiary recovery of oil. Interwell tracer testing is also used in geothermal reservoirs to gain a better understanding of reservoir geology and to optimize production and reinjection programmes. The main purpose of conducting interwell tracer tests in oil and geothermal reservoirs is to monitor, qualitatively and quantitatively, the injected fluid connections between injection and production wells and to map the flow field, reservoir heterogeneities and volumetric sweep (contacted volumes) between wells. Tracer is added into injection fluid via an injection well and observed in the surrounding production wells (Fig. 1). Tracer response is then used to describe the flow pattern and thereby gain a better understanding of the reservoir. This knowledge is important in optimizing oil recovery. Most of the information given by the tracer response curves cannot be obtained by means of other techniques.

Fluid flow in most reservoirs is anisotropic. The reservoir structures are usually layered and frequently contain significant heterogeneities leading to

directional variations in the extent of flow. Hence, the effective fluid movement can be difficult to predict. This is where tracer technology plays an important role, assuming that the movement of the tracer reflects the movement of the injected fluid. Obviously, it is most important to assure that the properties of the tracer meet this requirement as closely as possible; there should be a minimum quantity of undesired loss or delay. The physical and geochemical conditions of the reservoir define the constraints. As a result, tracers found to work properly in one reservoir, may not work satisfactorily in another.

Apart from radioactive and chemical tracers, stable isotopes of the water molecule (2H and 18O) can be employed as effective tracing tools to identify the source (origin) of produced water, both in geothermal as well as oilfield applications. On the basis of stable isotope indices, the relative contribution of different sources of water towards produced water may be estimated. However, in the cases of geothermal reservoirs and high temperature oil reservoirs, the 18O content of injected water is likely to be modified due to 18O exchange between water and host rock. However, 2H is considered as conservative and can safely be used to estimate relative contributions.

A field radiotracer investigation consists, in brief, of the following main

steps:

(1) Design of tracer strategy, involving consultation with reservoir engineers

(2) Selection of applicable tracers

(3) Application to the relevant authorities based on a safety report

(4) Tracer mixture preparation, calibration and quality assurance

(5) Selection/design of tracer injection and sampling procedures

(6) Tracer transportation to injection site

(7) Implementation of radiation safety procedures at the injection site

(8) Tracer injection

(9) Radioactivity contamination survey

(10) Injection equipment decontamination and handling of radioactive waste

(11) Tracer sampling and sample transportation to analytical laboratory

(12) Tracer analysis

(13) Data evaluation and simulation

(14) Reporting of results

The IAEA Coordinated Research Project (CRP) on Validation of Tracers and Software for Interwell Investigations has developed, prepared, tested and validated several tracers, techniques and software packages. The main group activities of the CRP were:

• Laboratory intercomparison on analysis of the tritiated water (HTO) in field samples (brines);

• Intercomparison on evaluation of field data with a simple software package (Anduril);

• Laboratory intercomparison on analysis of mixtures of the two water tracers HTO and 14CH3OH;

• Application of the PORO streamline simulator on field data provided by different companies.

A short summary of the main achievements of the CRP is given below: Tracer preparation, quality assurance and analysis: Synthesis, preparation, analysis and quality control of several tracers both individually and in mixtures: HTO, SCN — (14C or 35S labelled), radiolabelled alcohols, [Co(CN)6]3- (radiolabelled), 125I — (131I ) and gold nanoparticles have been established or validated. Criteria for selecting the more adapted tracers have been investigated. Laboratory intercomparison analysis of mixtures of HTO and 14CH3OH was successfully done. Ions and stable isotopes in produced water have been used as indicators to support the interwell tracer test.

Experimental procedure for tracer tests: Intercomparison of the injection and sampling strategies has been done and rules have been proposed to carry out the tracer experiments. Tracer injection techniques, both bypassing and direct pumping into the well head, have been compared. Well head samplers and procedures for collecting water sample have been developed and tested in the field. Safety procedures have been established and implemented.

Interpretation and modelling: Several models for interwell tracer data interpretation have been tested and compared. Rules and advice have been established to select the more suitable model and/or software packages, depending on the field structure and configuration. The following models (software packages) have been studied: Brigham (home-made code), dispersion (Anduril), streamlines (PORO), chemical engineering (Disproof) and computational fluid dynamics (CFD)(Caste and CONSOL). New possible approaches of compartmental modelling have been suggested for fractured oil or geothermal reservoirs.

The publication represents a form of monograph dedicated to tracer methods as applied to interwell investigations in oil and geothermal reservoirs. It consists of three sections and four appendices.

The first section gives the background and arguments for the use of tracers and presents the general view on tracers and tracer techniques as applied in interwell investigations in oil and geothermal fields. The status of tracer technology worldwide is given as well.

Section two deals with technical steps in the practical application of interwell tracer technology, including planning of field tests (tracer selection, injection and sampling), field related operation and implementation, tracer measurement and data interpretation.

Section three covers new tracer development, including tracer quality control, behaviour of tracers in various environments, and analytical methods for tracer measurement. Finally, the CRP achievements are summarized in a short section.

The four appendices provide the following information:

• Appendix I is allocated to field case studies performed by the various institutions involved.

• Appendix II deals with laboratory intercomparison tests on analysis of HTO and HTO + 14CH3OH in mixtures as well as operation of the tracer interpretation software Anduril on practical cases common for all laboratories.

• Appendix III provides procedures and protocols for measuring tracers in produced water.

• Appendix IV describes two software packages produced and tested during the CRP period: Anduril software for simple data treatment and PORO software for more advanced streamline simulation.

A second analytical step makes use of simple models. The response curves are decomposed in simple functions using a simple mathematical model based on the standard equations for diffusion in porous media.

where

C(x, t) is the tracer concentration as a function of distance and time (Bq/m3); tN is the normalized time;

D1 is the coefficient of dispersion (m2/d); v is the tracer velocity (m/d); x is the distance from the injection point (m);

CREF is the reference tracer concentration (Bq/m3).

Dispersivity may be calculated by multiplying the ratio (D1/vx) by the distance between the injection and production wells. Therefore, the extrapolated curve obtained before may be decomposed in two simple functions using Anduril software to give the curves shown in Fig. 33, where both functions and their sum are shown together with the experimental data.

The following table presents obtained results using Anduril software for the pair of wells CnE241-CnE324.

The decomposition in two curves may be explained by the fact that the tracer follows two different paths in its migration from the injector to the producer, each of them having different permeabilities.

Results of HTO analysis from the laboratories are summarized in Table 22. Figure 83 presents a summary of the results for HTO from all participants and for each sample. The nominal value is presented as the red bar with the error of 5% as an uncertainty of the secondary standard. The errors in results as reported by the authors were presented in error bars as 3o. Figure 84 compares the results for each laboratory with the nominal values for HTO concentration in radar plots with a logarithmic scale.

Discussion

• All HTO analytical data are acceptable in terms of consistency and are relatively close to the nominal values, except for the results from China, which show high values in comparison with the nominal ones. In general, the overall uncertainty of analytical results of tracer laboratories matches the requirement of field works as well.

• The results for HTO reported by China are generally higher than the nominal values. This may possibly be due to contamination of samples.

• The inconsistency of the Indonesian results on samples with low concentration could be due to high background from liquid scintillation counting or the luminescence effect.

IV. 2.1. Data interpretation: Conversion from raw data to information

The importance of interwell tracer testing in oil exploitation is indicated by the great number of tests conduced worldwide over the last 40 years. Normally,
water is injected into injector wells to ‘push’ the oil to the producer wells, from which it is extracted. The end of the secondary recovery process occurs when the water cut increases to the point where water injection becomes economically inefficient to continue further. At this stage, the fluid flow in the reservoir consists mainly of injection water. The oil remaining in the reservoir is largely stagnant (residual saturation) in the swept volume but can also comprise larger untouched oil volumes. This incomplete sweeping of the oil is a consequence of the natural heterogeneity of the reservoirs and the usually unfavourable mobility differences between water and oil. Hence, channelling of water between injector and producer wells is a very common problem that counteracts achieving acceptable sweep efficiencies.

The interwell tracer tests permit detection of this problem and also determination of some parameters of the watered zones, which are necessary for corrective action. Owing to the considerable uncertainty associated with fluid flow knowledge in the water flooding process, there is insufficient basis for assuming a very detailed reservoir model. Therefore, a ‘kit’ of simple methods can provide acceptable information, including moment analysis and Brigham based analysis.

2.2.1. Preparation of a technical safety report

Before implementing a tracer operation in the field, an application for the radiotracer experiment has to be submitted to the national radiological protection authority for evaluation and acceptance. To ensure that persons are protected from harmful effects of radiation, such application must comply with the International Basic Safety Standards [18] or equivalent national regulations. No practical action is taken before permission is granted by the relevant authority.

The safety evaluation report is the basis for this application. Such a report may contain an introductory description of experiments, descriptions of radiotracers, transportation and storage of radiotracers and injection methods, as well as radiation safety measures for all radiotracer related works.

The authorized person or organization will have the prime responsibility to ensure that radioactive material is used safely and in compliance with relevant regulations and standards. Guidance on occupational radiation protection, development of safety assessment plans and safe transport of radioactive material has been published by the IAEA [18-21].

The general procedure was as follows. The anion exchange resin based on a copolymer of styrene and divinylbenzene was immersed in water for 24 h for full swelling before loading onto a column made of polyethylene. The resin was then washed with 10M HCl to convert it into the Cl — form, followed by ion exchanged water until pH 7 is reached in the water eluate.

For experiments where the main purpose was the separation of [Co(CN)6]3-, the column had dimensions 0 = 5 mm and L = 25 mm, corresponding to a resin volume of ~0.5 mL. For the experiments where the main purpose was absorption of SCN-, the column dimensions were 0 = 6 mm and L = 300 mm, corresponding to a resin volume of ~8.5 mL. A sample volume of 100-1000 mL of seawater containing radiotracer was passed through the columns. Flow rates were kept at ~0.5 mL/min.

Absorption yield (%) is expressed as:

cs, feed cs, raf

Vabs = R •i00%

Rcs, feed

where Rcsfeed denotes the net sample counting rate per mL (proportional to activity concentration (Bq/mL) of the feed solution and Rcsraf denotes the net sample counting rate per millilitre of the raffinate.

Elution yield is determined as:

where Rcs elut denotes the net sample counting rate per mL of the eluate.

The separation factor of [Co(CN)6]3- and SCN — in the complete ion exchange procedure (feed plus elution) is then defined as:

All anion exchangers are of the amine type. The main parameters are given in Table 8.

OO

U