The invention concerns a method of determining the produc- tion rate of each of the phases in a well stream and of the type defined in the introductory portion of claim 1.

A method of the above-mentioned type has been described in two papers (by Flemming Jensen). One of these "A Substi- tute for Test Separators" was presented at "The SPE Euro¬ pean Petroleum Computer Conference held in Stavanger, Nor¬ way, 25-27 May 1992" and the other "Production Testing without Test Separator" at the "Petrotech Workshop on the Importance of Quality Fluid Data in Fluid Evaluation and Process Optimisation" held in Haugesund, Norway, 24-25 November 1992.

Both papers are incorporated in the present patent appli¬ cation by way of reference.

In this known approach, called SATS below, which stands for "Simulation of a Test Separator", the production rate of gas, oil, condensate and water in the well stream from a well in a hydrocarbon field is calculated by means of pressure and temperature measurements at the bottom of the well as well as at the tubing head and downstream of the choke. When, during production, the well stream flows upward through the well, the pressure and temperature changes. These changes depend upon parameters such as composition and the production rate of the well stream, the geometrical conditions of the well and heat exchange with the surroundings.

The known approach applies both sensors for measurement of pressure and temperatures, and a software model which, on

the basis of compositional analyses of gas, oil or conden¬ sate present in the reservoir, is capable of calculating the physical properties of the mixtures, such as density, phase split, viscosity, surface tension, etc., at a given pressure and temperature. The software model contains a module which simulates the flow of gas, oil and water in the well under stationary conditions by means of informa¬ tion on the geometrical conditions of the well, the ele¬ ments used in the well completion such as the choke, and finally a software module for optimization capable of com¬ paring measured and calculated results. In this approach the software is utilized to calculate the measured physi¬ cal quantities on the basis of the estimated production rates of gas, oil and water in the well stream. The cal- culated and measured quantities are compared, and the estimated production rates of gas, oil, condensate and water in the well stream are adjusted until the difference between the calculated and measured physical quantities is minimized.

Each phase of gas, oil, condensate or water which flows into the well from the reservoir represents a degree of freedom in the system of equations determining the re¬ sults, and at least the corresponding number of con- straints on the system is therefore to be introduced through measurements. If there are more constraints than degrees of freedom, the solution can be optimized to pro¬ vide a more reliable determination of the desired produc¬ tion rates. The known approach uses temperature differen- ces and pressure differences, but it is moreover proposed - but not yet implemented - to measure other physical parameters, such as density, relative permittivity, mass flow, phase split, total density, etc., by means of sen¬ sors at the well head.

The problem of this known approach is that the number of measurements of pressure and temperature available in practice cannot provide the desired accuracy in the deter¬ mination of the amount of oil, condensate, gas and water in the well stream while the well is flowing. The use of additional sensors, - which are not temperature or pres¬ sure sensors - proposed for the method, but not tested in this connection, is expensive and/or inaccurate and/or re¬ quires significant maintenance. Furthermore, practical tests of this known approach show that measurement of tem¬ perature differences only gives poor information on the system, and in practice it will often be necessary to as¬ sume the composition of the well stream to be known, and then, on this assumption, using the approach for calculat- ion of the total rate of produced well stream, which greatly reduces the applicability of the approach.

The object of the invention is to improve the above- mentioned known approach in a manner such that the stated drawbacks can be minimized or avoided completely. Further, the accuracy and additional costs will allow this approach to be used as a full or partial substitute for test sepa¬ rators in connection with the production from hydrocarbon reservoirs.

This is obtained by means of the new and unique features of the invention which are defined in the characterizing portion of claim 1.

An additional improvement of the method is obtained by means of the subject-matter stated in the dependent claims 2-5.

The invention will be explained more fully by the follow- ing description of a design of the method, which just serves as an example, with reference to the drawings, in

which

fig. 1 is a schematic view of a hydrocarbon reservoir with a fraction of a well,

fig. 2 is a likewise schematic view of a well,

fig. 3 is a diagram showing the density of four gas/oil mixtures in response to the pressure,

fig. 4 is a diagram showing the pressure difference be¬ tween two vertically different positions in the well in response to the pressure at the well bottom for the mix¬ tures shown in fig. 3,

fig. 5 is a diagram showing experimental data from an ac¬ tual field in response to the time,

fig. 6 is an enlarged view of the same,

fig. 7 is a diagram showing the pressure difference be¬ tween two vertically different positions in the well in response to the time, and

fig. 8 is a view of the same, but in response to the pres¬ sure at the well bottom.

Fig. 1 is a schematic view of a hydrocarbon reservoir, de¬ signated 1. The reservoir is depleted by means of a well designated 2 via one or more openings designated 3 in it. In the case shown, the reservoir contains the phases gas, oil and water, which are distributed in this order, seen from above and downwardly, due to the differences in den¬ sity.

Production from such a reservoir will normally result in changes in the gas oil ratio (GOR = Gas Oil Ratio) and in the water cut (WC = Water Cut). The problem is outlined in fig. 1, which indicates the possibility of drawing water and gas (coning) into the tubing. The WC and GOR normally ' change as a consequence of production from the field. In addition, the use of secondary recovery methods (EOR = Enhanced Oil Recovery) like gas and water injection could cause a significant change of these parameters. For a re- servoir to be utilized optimally, it is important to have information available concerning the actual GOR and WC of the produced well stream.

When a hydrocarbon reservoir is discovered in an explora- tion drilling, an analysis of the composition and physical properties of oil, gas and water in the reservoir is per¬ formed in connection with the testing of the reservoir. These results are summarized in a PVT report, where P stands for pressure, V for volume and T for temperature. This report is then used in connection with the operation of the reservoir (reservoir management). Information from the PVT report is combined with information on recorded pressures and temperatures in the reservoir as well as measured production of gas, oil and water. This informa- tion constitutes an important element in the optimization of the future production from the reservoir.

On a processing platform, the produced well stream from the various wells is mixed together and thereafter sepa- rated into oil, gas and water in a production separator (typical size: D=3-5 m, L=10-20 m). A somewhat smaller test separator (typical size: D=2-3 m, L=5-8 m) is uti¬ lized to measure the production rates of gas, oil and water from the individual wells by alternately producing each well to said separator. The test separator separates the produced well stream into gas, water and oil.

Installation of a test separator on a satellite platform requires a significant investment and subsequently in¬ volves high operational costs. Instead, the wells are pro¬ duced directly in a multi-phase pipeline connected with a centre platform where the processing takes place. It is difficult, however, to utilize the test separator on the centre platform, because the well streams from the various wells on the satellite platform are mixed in the multi¬ phase pipe. Test of wells on a satellite platform without a test separator therefore presently has to be carried out in one of the following ways.

1. One or more wells are closed, and then the flow through the multi-phase pipe is tested in the test separator on the centre platform. Owing to the lost/ postponed production this is an expensive approach, which takes significant time and is inaccurate be¬ cause of the length and dimension of the multi-phase pipe.

2. The wells are tested together (commingled test), and the measured oil, gas and water production is distri¬ buted between the various wells in view of the know¬ ledge on reservoir, recorded pressures and previous tests. This approach is highly uncertain.

3. Installation of an additional multi-phase pipeline between the satellite platform and the process plat¬ form. This is an expensive solution which often in- volves operational difficulties (increased back pres¬ sure to the wells, slugging, etc. ) which decrease the accuracy of the test results obtained.

Development of marginal fields in deep water by means of subsea completion is also possible, and if several of these wells are connected to a common multi-phase pipe-

line, fundamentally the same problem is faced with respect to well testing as on satellite platforms without a test separator.

A more inexpensive and/or faster approach to replace test separators will therefore be of signficant value to an oil company. Such a method is described below, viz. the al¬ ready mentioned SATS approach.

The idea of the SATS method is to calculate production rates of gas, oil and water from measurements of pressure and temperature, which are very inexpensive and are fre¬ quently already available. When during production a well fluid flows up through the well, temperature and pressure changes. These changes depend on parameters like the com¬ position and the production rates, the geometrical condi¬ tions of the well and heat exchanged with the surround¬ ings. In principle, all parameters, except composition and production rates of the well stream, can be estimated ex- plicitely on the basis of information already known. The hydrostatic head in the well is a function of the average density of the well fluid present in the well, and this in turn depends on parameters like the composition of gas and oil as well as the flow conditions of the well (the resi- dence time of the individual phases in the well). The dy¬ namic pressure loss depends on the friction in the well and is therefore an indirect measure of the velocity of the flow.

The SATS method therefore consists of the following basic elements:

1. A software based thermodynamic module calculating physical properties (characterization, phase split, heat conductivity, viscosity, surface tension, enthalpy, etc. ).

2. A software based flow module simulating the behaviour of the well during production.

3. A software based numeric optimization module perform- ing the iterative calculation procedure outlined below.

4. Sensors. These may be positioned e.g. at the bottom hole, at the tubing head and after the choke. In the tests published, pressure and temperature sensors are used. It has moreover been proposed to possibly sup¬ plement the concept in future tests with sensors cap¬ able of measuring flow rate, phase split, total den¬ sity or relative permittivity.

Fig. 2 schematically shows the well 2 shown in fig. 1. A pressure and/or temperature sensor 4 is arranged at the tubing head, and another pressure and/or temperature sen¬ sor 6 downstream of a choke 5. A third pressure and/or temperature sensor 7 is moreover arranged at the bottom hole. In addition to these sensors, other instruments 8 for measuring other parameters, e.g. as suggested above, may be arranged, as needed.

On the basis of the results presented in the PVT report of the well and the latest available information on produc¬ tion rates of gas, oil and water production, the SATS method makes an initial estimate of the actual rates and composition of the produced well stream. Based on this, and on the measured bottom hole pressure and bottom hole temperatures the stationary flow conditions in the well are simulated. The changes in temperature, pressure and optionally other physical properties calculated hereby are compared with the values measured in the well by means of the instruments. The initial estimates of the rate and composition of the produced well stream are then optimized

by minimizing the difference between the measured and calculated pressures, temperatures, etc.

Depending upon the information available from the measure- ents, a more or less accurate determination of composi¬ tion and rates can be obtained. Each phase (gas, oil or water) flowing into the well adds a degree of freedom to the system. Correspondingly, at least the same number of constraints (information) must be added through measure- ments. There may be temperature difference, pressure dif¬ ference, relative permittivity, phase split, total density or the like. If there are more measurements than degrees of freedom in the system, the above-mentioned iterative calculation procedure is merely a minimization.

Results of a test of the SATS method are given in the pre¬ viously mentioned two papers by Flemming Jensen. It is concluded that the concept is useful as a basic idea, but that a higher accuracy is required for the concept to qualify for commercial use. In some of the performed tests it was necessary to assume the GOR to be known to match the number of degrees of freedom. The temperature measure¬ ments were found to contain very little information, and their usefulness as estimation parameters is questioned. Further, sensitivity analyses of the system showed that the pressure profiles at relatively small flow rates were relatively flat, which resulted in significant inaccuracy of the GOR and WC determinations. Finally, it is concluded in the above-mentioned paper entitled "Production Testing without Test Separator" that: "High accuracy of the pre¬ dictions of the flow rates requires a more accurate flow model and/or more experimental measurements".

The basic idea of the invention, which constitutes an im- provement of the SATS method, is to calculate the composi¬ tion of the well stream by including an analysis of the

measurements of the transient behaviour of the bottom hole pressure and the tubing head pressure after closing of the well. The information already available is supplemented with a new and different kind of information, and all the information is thus used in an improved method.

The manner in which the density of oil or gas changes with the pressure at constant temperature, depends upon the components included in the sample concerned. This appears from fig. 3, in which the density of four hydrocarbon mix¬ tures with different GOR is shown versus the pressure. In this connection the pressure range from 50 to 150 bar is interesting, and the figure shows a great difference in the density sequence for the four mixtures.

The difference between the bottom hole pressure (BHP) and the tubing head pressure (THP) when the well is at rest, is an expression of the average density at the current temperature of the condensate/gas/oil/water mixture which is present between the two pressure measuring points.

In fig. 4 the difference between the BHP and the THP cal¬ culated for the four gas/oil mixtures al, a2, a3 and a4 is shown versus the BHP. The vertical depth of the well is about 1800 m, and experimental data are also shown in the figure (data from a North Sea well). In the calculation of the pressure profiles in fig. 4 no allowance has been made for the relative movements of the phases under the statio¬ nary flow conditions immediately before the choke is closed. The experimental data should preferably agree with the a4 simulation, and considering the relatively rough assumptions a satisfactory, good consistency is found, indicating the usefulness of the method.

When both gas, oil and water in a specific ratio flow into a well from the reservoir, the same ratio can be found

using a test separator on the platform. However, the ave¬ rage ratio of oil, gas and water in the actual well bore will be displaced toward the heavy phases (more oil and even more water). This is due to the mutual slip of the phases during flow (which means that the phases flow at different rates). The size of the slip depends upon a number of parameters, such as flow rate, surface tension, density, viscosity, pressure, temperature and composition of the phases. Information on these parameters can be ob- tained from the flow simulations, undertaken in the SATS method.

An example of experimental data from a North Sea well is shown in fig. 5, and the use of these is outlined below. The results are in this figure presented versus the time. The sequences interesting in this connection are marked with a circle in fig. 5, where the bottom hole and the tubing head pressures increase abruptly owing to the clos¬ ing of the well. The sequence is shown on an enlarged scale in fig. 6.

In fig. 7 the BHP-THP is shown versus the time. Initially, there is a fairly constant pressure drop of about 1870 psi. As mentioned before, this pressure drop is mainly caused by the static head (height of the liquid column) and the dynamic pressure loss (friction pressure loss). In this figure the well is closed after a little less than 24 hours, and the contribution from the friction to the total difference between the BHP and the THP is essentially now just the static head in the well (just under 1720 psi), and under the given conditions this may be ascribed to the average density of fluid present between the points of pressure measurements. Then the BHP-THP increases within about 5 hours to just under 1880 psi. This build-up of pressure is caused by inflow from the reservoir.

In fig. 8 the difference between the BHP and the THP is shown versus the BHP. As expected, the curve reflects the dependence of the density on the pressure, as mentioned before. The experimental data shown in fig. 8 are also shown in fig. 4.

The invention is based on the above relations being uti¬ lized as follows:

The original model is expanded with a supplement serving to establish a relation between the composition of the well stream under the given conditions (i.a. the flow velocity before the well is shut in) and the pressure at various locations in the well after the well has been closed. Other information required for this supplement such as gas and/or oil and/or condensate and/or water is already available.

The well is closed, and the pressure is measured at least at one point in time in at least two vertically different positions.

The expanded mathematical model, which also describes the behaviour of the well after the closing, is solved with respect to the composition of the well stream which means the production rate of each of the present phases. This is done by minimizing the deviation between the measured and corresponding calculated data which, for the expanded mo¬ del, include the pressure differences between the selected vertically different positions measured after the well has been closed.

The measurements of the pressures typically take place at the top and at the bottom of the well, where the necessary measuring sensors equipment is usually already available. In this case no additional measurement equipment will be

required to implement the new method.

The method has been described above on the assumption that the well was closed completely. However, within the scope of the invention there is nothing to prevent the method from being applied even if the well is just partly closed, which enables production to go on, although to a limited degree, while the measurements take place.

The accuracy of the method can moreover be improved by measuring, in addition to the pressures, the temperatures at corresponding or other locations in the well.

The kind of mathematical model used by the SATS method is a stationary flow model. Instead, a dynamic flow simulator may advantageously be used, by which further information on the actual change in measured pressures after closing the well can be utilized.

The proposed improvement of the existing SATS method dif¬ fers from this by collecting and using data in a fundamen¬ tally new way.

1. The SATS method uses measurements performed when the well flows under stationary conditions. The method proposed here includes measurements of pressure dif¬ ferences after the well has been closed in. In this case, the friction pressure loss vanishes, and the total average density of the contents of the actual well can be calculated. Due to the density dependency on pressure it is posible to gain additional informa¬ tion on the system by including measurements of pres¬ sure differences at several points in time (i.e. at different BHP).

2. Installation of a total density meter, as proposed by Flemming Jensen, is relatively expensive and also rather inaccurate in operation, because the indivi¬ dual phases in an oil/gas well only rarely flow in a fairly homogeneous mixture. In the new method an average density of the fluid is calculated for the entire well bore, which gives a result, in practice independent of slugging

The invented method provides the following advantages:

1. The well needs only to be closed quite briefly.

2. Existing pressure and temperature sensors are used.

3. With the normally occurring number of degrees of freedom as well as the number of available measuring sensors, the SATS concept in its basic form is fre¬ quently not capable of predicting changes in the com- position of the produced well stream (see the refe¬ rence paper "A Substitute for Test Separators"). The new method adds at least one additional constraint to the system of equations.