BRPI0819322A2 - training evaluation method, and cleaning monitoring and real-time forecasting method. - Google Patents

Abstract

METHOD OF EVALUATION OF TRAINING, AND METHOD OF MONITORING CLEANING AND FORECAST IN REAL TIME This is the cleaning monitoring and forecast in real time, aiming at estimating the pumping volume versus final contamination, including the detection of the formation fluid outbreak for a sampling tool and detection of the cleaning regime transition, from a predominantly circumferential cleaning regime for a predominantly vertical cleaning regime. A similar workflow can be used to estimate contamination at the end of cleaning production for a given pumping volume.

Description

METHOD OF EVALUATION OF TRAINING, AND METHOD OF MQNITORING

REAL-TIME CLEANING AND FORECASTING Field of the Invention The present invention relates to techniques for cleaning contaminated supply fluid during sampling operations. In particular, the present invention relates to the prediction of cleaning parameters during sampling operations.

Background of the Invention Sampling of training fluid by Cable Training Tester (WFT) during the drilling operation represents an important component of the training evaluation system established by the oil industry, especially when it comes to high profile and marine wells. It is known that errors in the estimates of the properties of the formation fluid can lead to significant calculation errors in the design and performance forecast of the flow guarantee, well construction, and production facilities. The main challenge in obtaining representative samples of formation fluid by WFT is related to the invasion of mud filtrate during drilling. After a few hours of drilling, the well is generally surrounded by the saturated invasion zone with predominance of mud filtrate. and, therefore, any sampling operation launched during the interruption of the

2 .

drilling must start from the cleaning production, which continues, until the contamination tolerance target is reached, or the time allocated for the sampling operation is exhausted- The great challenge of monitoring 5 of the cleaning production represents the process of drilling with oil-based mud (OBN), due to the miscibility of the OBM filtrate with the formation hydrocarbons and low resistivity contrast.

The variation in the properties of the formation fluid during the cleaning production, however, can be reliably detected by the optical system. Existing Optical Fluid Analyzers (OFA) can measure the optical density (OD) of mixtures produced in a wide spectrum of invisible light, with wavelengths in the range of 400 nm at 2200 nm. In the presence of the initial DO contrast between the OBM filtrate and the formation fluid, the high sensitivity of the optical measurements to the composition of the produced fluid can be observed. This sensitivity, however, does not allow quantification of the contamination of the produced fluid, since the OD of the virgern formation fluid is unknown beforehand. To compensate for the lack of information during the monitoring of the cleaning production, the contamination transport forecast has to be involved to achieve the closure of the optical monitoring model.

Although advanced sampling operations with optical contamination monitoring have been used by the industry for more than ten years, contamination transport modeling for the cleaning prediction is still in its rudimentary stage. The approach to transporting

Contamination during the production of cleaning, which is currently considered to be the state of the art in the industry, is based on the empirical model for the evolution of contamination over time during the final cleaning phase. This model states that contamination, defined as the concentration of the mud filtrate in the fluid produced, should decrease over time as q_5 / 12. This behavior contradicts the analytical model, which is based on the pseudo-spherical flow and the transport of contamination, predicting the faster decay of contamination rj with time, that is, · as t "2/3 0 Recently, numerical siniulations have confirmed the empirical law, q - t_5 / l2, but did not explain the contradiction between the theory and the field observations.

Summary of the Invention Urri a new concept for predicting cleanliness during sampling is proposed. This concept shifts the focus to the first cleaning phase, allowing the initial prediction of the pumping tempQ or volume produced, through "profiles of optical fluid analysis versus certain contamination targets. The first cleaning phase QfeFece potentially a great quantity, of iriformations to be used in the prediction of cleaning, because this phase is

V strongly affected by the local flow pattern and eontaniination transport, and handles a wider range of optical density variation.

5 A new approach to cleaning prediction is based on a truly three-dimensional (3D) model · of flow and transport of contamination to the probe production area on the well wall covered with dry Iama.

This model captures the initial cleaning phase better than the conventional spherical flow model, which incorporates contamination transport on the axis of symmetry to a small production sphere located in the shaft of the well.

The new model provides for the signature of the 3D contamination transport in the cleaning dynamics, which is controlled by the relationship between the depth of invasion and the radius of the well. The analysis of solutions to new problems reveals new details of the evolution of cleaning. In particular, the transition from a predominantly circumferential cleaning regime to a predominantly vertical cleaning has an unmistakable assertion, which can be used to monitor the progress of cleaning and reconstruction and the depth of initial invasion.

Examples of data processing for the sampling work, which support the new concept, are provided.

They indicate that adequate estimates of pumping volumes can be obtained 3 to 5 times before using the new approach.

Brief Description of the Drawings The current disclosure is better understood, starting from the following detailed description, when I deal with the figures that accompany it. It is noteworthy that, in accordance with industry standard practice, · various resources are not mapped on a scale. In fact, the dimensions of the various resources can be arbitrarily increased or reduced for reasons of clarity of the discussion.

Fig. 1 shows an example of the drilling system used to drill a well in an underground formation, in accordance with one or more aspects of the present invention.

Figure 2 illustrates an example of optical density measurements, used for cleaning follow-up, according to one or more aspects of the present invention.

Figure 3 is a contamination transport scheme.

Figures 4A and 4B show details of the model for an imitation of cleaning production, according to one or more aspects of the present invention.

Figures 5 and 6 illustrate the results of the cleaning production simulation using the model of

Figures 4A and 4B, according to one or more aspects of the present invention.

Figure 7 illustrates the evolution of contamination with the volume produced, according to one or more aspects of the present invention.

Figure 8 illustrates the transition of the cleaning regime in relation to the depth of invasion, according to one or more aspects of the present invention.

Figure 9 illustrates an example of a map of cleaning regimes versus pumping volume and depth of invasion, according to one or more aspects of the present invention.

Figure 10 illustrates volume

Inrush P versus depth of invasion, according to one or more aspects of the present invention.

Figures I1A and 11B illustrate an example of optical profiles, according to one or more aspects of the present invention.

The spinner 12 illustrates a method according to one or more aspects of the present invention.

Detailed Description It is necessary to understand that the following disclosure provides many different modalities, or examples, to implement different characteristics of various modalities. Specific examples of components and mechanisms are described below to simplify

This disclosure. These are, of course, only examples and are not intended to be a limitation. In addition, the present disclosure may repeat reference numbers and / or letters in several examples. This repetition has the purpose of simplicity and clarity, and not to dictate in itself a relationship between the different modalities and / or configurations discussed. In addition, the formation of a first resource on a second resource in the description that follows may include modalities, in which the first and second resources

W are formed in direct contact, and may also include modalities, in which additional resources may be formed by interposing the first and second resources, in such a way that the first and second resources may not be in direct contact.

Figure 1 illustrates a drilling system 10 used to drill a well in underground formations, usually shown in 11. A drilling rig 12 on surface 13 is used to rotate a drill column 14, which includes a drill bit 15 in your

W lower end. When the drill bit 15 is being rotated, a "Iaiua" 16 pump is used to pump drilling fluid, commonly referred to as "mud" or "drilling mud", down through the drill column 14, towards the arrow 17 for the drill bit 15. The mud, which is used to cool and lubricate the drill bit, leaves the drill bit 14 through the holes (not shown) in the drill bit 15. The mud then carries gravel drilling away from the bottom of the well 18, as it returns to the surface 13, as indicated by the arrow 19, through the annular space 21 between the drilling column 14 and the formation: l1. Although a drilling column 14 is shown in Figure 1, it will be important to note that this disclosure is also applicable to production columns, tubular columns, flexible tubing, and cable driven tools, among others. On surface 13, the return sludge is filtered and returned to a sludge tank 22 for reuse. The lower end of the drill column 14 includes a downhole assembly (BHA) 23, which includes the drill bit 15, as well as a number of drill commands 24, 25, which may include various instruments, such as profiling during drilling (LWD), or measurement during drilling (MWD), and telemetry equipment. An assessment instrument during the drilling of the 2Q formation may, for example, still include, or be arranged within, a centralizer or stabilizer 26.

The probe may be located on the stabilizer 26 to make incorrect contact with the well wall, or it may be located on a probe module, as part of the BHA.

Alternatively, Qbturadores can be used. People versed in the art will realize that a training probe could be arranged in other Jocals, without departing from the scope of this disclosure. At least portions of the current approach may be particularly relevant for MWD or LWD sampling, as this type of tool 5 usually finds a shallow depth of invasion of the filtrate fluid. In particular, formulas and models for calculating contamination levels in an environment or cable tool may be different from formulas and models for calculating contamination levels in an MWD and LWD environment or tool, since the depth of the invasion fluid filtered in an environment that is very significant (due to long periods of exposure).

It will become more evident from the below disclosure that an initial estimate of contamination levels, important in an LWD or MDW environment, can be made.

The current approach for monitoring contamination during. cleaning production is based on measurements of the optical density of multiple channels of fluid produced. In the presence of the contrast of optical density between c) virgin forming fluid and the OBM filtrate, penetrated into the formation during "drilling, the variation in optical density provides an indication of progress in displacement by the forming fluid in the mud filtrate.

It is not easy, however, to quantify this information, converting it into contamination of the fluid produced. The main difficulty lies in the fact that the composition of the virgin foraging fluid is previously unknown, and cannot be determined with the technology currently available during the cleaning process, making it impossible to determine the amount of Iama filtrate produced with the 5 reservoir fluid.

The quantification of measurements of the optimum density during the production of 1 = clean, however, becomes possible, if the information about the flow pattern and the contamination transport in the reservoir are involved in the interpretation. Unfortunately, the exact solutions for cleaning production through a probe or a sealing gap are not readily available and, for this reason, approximate solutions are used alternatively. These solutions are based on simplifications of the geometry of the I! 1flow pattern, fluidQ properties, and fluid displacement mechanisms. They generally do not capture the evolution of cleaning in the "short term, but provide some insight into their long-term scenario, which is determined by transporting contamination to a distant field. The spherical flow model for a small sphere of production, in the absence of The viscosity contrast between the forming fluid and Iania's filtrate is an example of such an approximate solution.

The information required for the closure of the optical model can be reduced to a single function, which is the variation of rj quantification with time t or q volume 0 produced V. For the spherical displacement flow model similar to piston without contrast of viscosity , the evolution of co-contamination during the final cleaning phase is described by the power law with two 5 parameters: q (t) = Bt "", t ~ w .. + ...... +.-... ( l) Parameter B depends on the rate of production and the depth of the invasion, while the exponent a is determined by the geometry of the distant flow pattern. For spherical flow, we have c = 2/3. The different value a = 5/12 is claimed, to be more consistent with the carnpo data, and is used in groups

The existing interpretation support to monitor and visualize optical data during sampling operations.

The relationship between optical density OD and contamination n can be obtained from an empirical relationship between the absorption of light by the fluid and the composition of the fluid, which is known in the optics as the Beer-Lambert element. With the application to oil or gas mixtures of the reservoir and OBM filtrate, the Beer-Lambert law states that the optical density of a mixing OD can be 0 expressed through the optical densities of its components OOq and ODf where OD = cODf + (1 - C) 0d0

Here, c is the molar concentration of the OBM filtrate in the mixture. For simplicity, we assume below that the molar concentration c is equal to the volumetric concentration of the laine filtrate and, therefore, contamination n (t) can. be 5 expressed through the measured optical density OD (t) of the produced fluid, as: rj (t) = [OOq - 0D (t)] / '[0D0-0Df] (2) Both individual optical densities, OOq and ODf ,

The are generally unknown, however, the optical density of the widely used fluid-based OBM is in the order of 0-0.1, meaning that the OBM filtrates are almost completely transparent to light.

Since the OD0 optical density of the formation oil cannot be reliably determined until a clean sample of oil is recovered, contamination cannot be estimated by measuring optical density using only Eq. (2). This uncertainty, po'réin, can be reduced by using the asymptotic solution for the variation of the contamination with the time of Eq- (1). Replacing the Rq. (1) in Eq. (2), it appears that the expression for the measured optical density of the mixture produced in long times is: oD (t) = a -bt ". A = OD ,,, b = B (ODÁ - OD,). 1 + OQ .... ..... (3) Eq. (3) can be used for installation in measured optical density data by determining suitable a and b parameters. of the oil reservoir becomes known, and the evolution of contamination over time, which is determined by Eq. (1) with parameter B = b / (OD0 - ODf), is fully calibrated.

For a given N contamination target, the cleaning production time can be estimated as: t, = {B / n, r .... ...... .. (4) The analysis of this formula, in the However, it reveals a deficiency. In fact, since t / Ct = 3/2 or 12/5, the cleaning time estimate tQ must be very sensitive to parameter B = bl (OD0-ODf), which is expressed through the adjustment parameters a and b, especially for low contamination targets T) q. If r) 0 ~ 0, small errors in setting adjustment parameters can result in large errors in cleaning time estimates. © Another limitation of the described procedure is related to the fact that it does not allow quick estimates of cleaning time, due to the asymptotic nature of Eq. (1) and Eq. (2). A relatively long sequence of the cleaning history has to be accumulated before any prediction can be made. The practical requirements, unfortunately, are completely different, in the sense that estimates of pumping time are needed as soon as possible - for example, preferably after the eruption of virgin Dara a

V probe. The solutions for the first phase of cleaning, however,

are not available and the pumping time estimates, which would be based on data from early cleaning monitoring, still represent c) main challenge during the sampling operation. Obviously, this limitation can be removed by creating simulation capabilities for realistic patterns of flow and contamination transfer.

Despite the great success of the delivery fluid delivery technology, it is necessary to recognize that the cleaning monitoring system currently implemented is basically qualitative, not quantitative. It usually provides excessively large estimates for cleaning time. These estimates can also be inaccurate and unreliable, especially for sampling tasks in adverse conditions, or trivial, if cleaning is essentially finished. The current technique of estimating the cleaning time does not use a complete set of optical data obtained during the cleaning monitoring, except in its final sequences, leading to longer maneuver times for sampling operations.

A new concept of cleaning monitoring and interpretation of optimal data, introduced here, is exempt from many of the above limitations. Its relevance and schedule coincide with the needs of the sector in the optimization of sampling operations, especially for drilling environments with severe limitations.

An example of DO measurements used for cleaning monitoring is shown in Figure 2. These data represent the DO measured across five selected channels of the multi-channel OFA. They are usually viewed on a real-time monitoring screen 5, as vertical plots on multiple columns of selected ODs versus the pumping time elapsed in seconds. Since the contrast of OD between OBM filtrate and forming fluid is generally unknown in advance, different OFA channels with pre-selected wavelengths of light can be used for the most sensitive monitoring of clean production7. The data plotted in Figure 2 allows us to recognize three distinct phases of cleaning production. The initial phase, which lasts about 350 s in this example, generally corresponds to the mud and filtrate bornbinding. The second phase begins, when the DO (in channels 4 and 5 in Figure 2) drops abruptly and then

W stabilizes, indicating that the formation fluid has erupted into the probe. This phase continues for about 1600 s of the time elapsed and ends, when the DO reaches a plateau. During the third cleaning phase, the flow pattern and ODS in the monitoring channels vary very slowly, although the OD profiles become less noisy. This final phase of optical monitoring is currently used for the quantitative prediction of cleaning progress, by adjusting equation (3) to the smooth part of the DO profile, indicated by ellipse A in Figure 2, and then estimating the cleaning time. pump versus target contamination, using Eq. (4).

Despite the record of * solid performance, the 5 ongoing cleaning monitoring techniques do not use the full extent of the information, which can potentially be extracted from DO measurements. For example, the initial secrets of OD profiles are not involved in the current cleaning monitoring workflow.

These initial segments of the OD data, in eritanto, contain information about an important monitoring event, related to the eruption of the formation fluid to the probe, or to the first appearance of hydrocarbons, + which is the main signature of the depth of the unknown. invasion of the OBM filtrate and flow pattern. It is also interesting to note that the DO variation scale immediately after the eruption is generally some orders of magnitude greater than during the final stage of clearness.

This part of the DO profiles is surrounded in Figure 2 by the ellipse B. The duration of the second cleaning phase can also be considered as a signature of the contamination flow and transport patterns, which should be the target for monitoring during cleaning. + In order to use the initial parts of the OD profiles for q quantitative monitoring of cleaning, the appropriate solutions for c) progressive problem of contamination transport during the production of cleaning had to be obtained. This problem, however, represents great challenges to solve with the existing reservoir simulation tools, if the analysis of total sensitivity with respect to some unknown parameters has to be carried out, with the objective of providing the basis for the inversion of the profile of OD. The scheme of the contamination transport problem is shown in Figure 3. The pumping is done through the probe leaning against the dry mud, sealing the well wall.

Initially, the well is surrounded by a cylindrical invasion zone, saturated with QBM filtrate, and only mud filtrate is produced at the beginning of the cleaning operation, leading to non-uniform contraction of the invasion zone with a volume of borne. As the forging fluid reaches the probe's production area, the mixture of sludge filtrate and virgin forming fluids is produced. The cantamination of the formation fluid by Iama filtrate normally decreases rapidly, shortly after the eruption, and then slowly, resulting in large volumes of cleaning production to achieve low levels of contamination.

The numerical siniulation of contamination transport has to deal with the geometrical and hydrodynamic singularities of the 3D problem, which has multiple characteristic scales, mixed limit conditions, a strongly non-uniform speed field, and a moving contamination front. The characteristic scales are represented by the radius of the probe's production area, rp, which is also referred to as a probe radius, the initial depth of the invasion, dt, and the radius of the simulation domain, r ,. These scales can be different from each other in an order of. greatness, as when j; p «dt << r ,. The flow velocity is unique at the boundary of the probe's production area, ond, and the velocity changes direction at 90 °.

The contamination front is smooth at first, but is interrupted after the formation fluid has erupted into the probe, moving with the time very close to the limit of the probe's producing area. The exact calculation of contamination during the final cleaning phase requires high resolution numerical modeling close to the probe's producing area, adding another characteristic scale, which can be very

Small V when compared to the probe radius. The viscosity contrast between the OBM filtrate and the formation fluid also contributes to the challenges of numerical modeling.

Although success in modeling cleaning production has been reported recently, it is still a challenging task that requires significant time, special skills, and simulation tools.

Below are the results of the cleaning production simulation obtained with the Comsol Multíphysics FEM package. The details of the podelo are shown in the

Figures 4A and 4B. The simulation field was embedded in a spherical volume around the well, with a circular production area representing a probe and a cylindrical invasion zone.

5 The contamination and flow transport simulation was based on a single-phase displacement flow model of the piston type, regulated by Darcy's law and the mass conservation equation. The total number of finite elements used during the simulations varied between 100,000 and 250,000. The CPU time for a single simulation of the cleaning production history was about 1-3 hours on a Dell Precision 670 computer, shorter for a shallow invasion and longer for a deep invasion. The calculations were continued, until the contamination of the produced fluid reached about 2-3%. The example of the distribution of simulated contamination around a probe is shown in Figures 5 and 6, where the light areas represent the formation fluid, and the dark areas represent the OBM filtrate.

The evolution of contamination C (° 5) with the dimensionless volume produced Q = vl Vbt is shown in Figure 7 for the dimensionless dimension invasion i5 = d, j rw equal to 0.125, 0.25, 0.5, 1 and 2 All curves undergo a joint deformation during the initial cleaning production phase, after the formation fluid has erupted into the probe. Ern logarithmic coordinates (log10 Q, log10 C), that

20 "evolution of the initial contamination corresponds to the law of empirical power C - C2", with the exponent a = 5/12. Over time, the cleaning dynamics changes to the power law with the exponent cl = 2/3. As for the deep invasion, 5 early this transition occurs in terms of the dimensionless pumping volume Q, = Vt / Vbt, due to the rapid increase in the volume of Vbt irruption with the depth of invasion. The function Q, (15) shown in Figure 8 was obtained, * by finding the intersection of two asymptotes capturing the 1 () early and late contamination behavior.

Likewise, the initial contamination trend remains longer for a shallow invasion. For example, it extends below the contamination level of 10% to 6 = 0.25, which corresponds to the invasion depth of about 1 "for the hole diameter of 8.5". The switching of the initial power law C - q-S / 12 to the spherical flow asymptote C - q-2/3 reflects the variation in the transport pattern of the contamination during the production of cleanliness. + Initially, the transport of the predominant contamination occurs circumferentially, as shown in figure 6, but then the contamination comes mainly from the top and bottom along the well. Iama's filtrate has to travel long distances to reach the probe. This change in contamination transport results in a faster reduction of contamination with the volume produced, compared to the initial cleaning phase.

The simulation results can be represented schematically as the niapa of cleaning regimes in relation to the pumping volume and the depth of irritation.

An example of this map is shown in Figure 9, for q case

And 5 equal viscosities of filtrate and formation fluids. It illustrates two sequential phases of cleaning production, the initial (X) and the final (Y) phase, with the transition zone (Z) between them. The analysis indicates that the transition zone is relatively narrow, especially for deep invasion. For this reason, it can be replaced by the transition point ç) ,. In the absence of viscosity contrast, Cj transaction point Q, depends on the dimensionless invasion depth õ = di / rw, as shown in Figure 8.

O It is interesting to note that the modification of the contamination transport regime during cleaning production, from the power law. From C - q_S / 12 to the power law C ~ q-2/3 is a fundamental feature of the solution, that has been ignored until now. It reflects the physics of contamination transport during cleaning production, and creates a robust event, which provides valuable information on the flow pattern for predicting cleaning dynamics. Since the change in the cleaning regime occurs relatively early in the shallow invasion, the detection of this © event, as well as the inrush of the formation fluid to the probe, must be the main target during monitoring.

W of the production of lirripeza. On the other hand, as indicated by the diagram in Figure 8, the circumferential cleaning regime cannot be distinguished for the depth of invasion deeper than 5 well radii. In this case, the inrush of the formation fluid into the sampling tool becomes a main monitoring event.

The information given in Figure 8 and Figures 4A and 4B allow to represent contamination C (%) approximately as: "= '®" | O :::'; 3. ':::: i'};). 0 ... F (5) The outburst volume, Vbt, is plotted in Figure 10 versus the dimensionless invasion depth õ. It is standardized on the amount O "3 1 / kFi / 4 '·, to take into account the porosity of formation 0 and the anisotropic ratio of permeability kj kv to a vertical well in the formation with different horizontal and vertical permeabilities. of the curve in logarithmic coordinates is slightly less than 3, since Vbt - di3 for © shallow depth of invasion, if the probe is replaced by a point collector.

The monitoring technique, which can be used for the initial forecast of cleaning production, is described below, assuming a relatively shallow invasion of the mud filtrate, õ <5. It is also assumed that the optical channel.,

which is used for monitoring cleaning production, has already been selected. Its OD data allows to distinguish the formation fluid (petroleum) from the filtrate from OBM. This means that the range of variation of the QD 5 offers sufficient resolution to detect changes in the composition of your mixture. The example of this optical profile is shown in Figures 1IA and 11B: The point identified in Figure IIB corresponds to the visually detected variation of the inclination of the DO curve, probably representing the change from the circumferential cleaning regime to vertical cleaning.

The monitoring technique involves the following six steps, which are also illustrated in Figure 12:

1. Determination of the volume of production of the Vbt outbreak.

2. The detection of the production volume Vt

D corresponding to the transition from the cleaning regime.

3. The calculation of the normalized production volume in the transition from the cleaning regime Q, = Vt / Vbt

4. The estimate of the dimensionless invasion depth õ = di / rw, which is the log10 Cl ,, correspondence, using the diagram in Figure 8.

5. The estimate of the normalized production volume Qp for the given contamination target Cq, using the diagram in Figure 9.

6. The calculation of the production volume Vp = \ / BT Qp corresponding to the contamination target.

Figure 12 illustrates a method of monitoring cleaning and forecasting in real time, aiming at estimating the volume of boníbeio versus the final contamination, according to the aspects of the current disclosure. A similar workflow can be used within the scope of this Order, to estimate contamination at the end of cleaning production for a given pumping volume.

It is interesting to note that this cleaning control and forecasting technique does not imply any numerical differentiation of extrapolation or noise data.

It is based on the detection of two major events while pumping out. These two events are the inrush of the formation fluid into the sampling tool, and the transition from the cleaning regimen, DredoIninantamente from circumferential to vertical. They characterize the transport patterns of contamination and flow in the reservoir.

The currently available solutions allow us to reconstruct the initial depth of the invasion, in the absence of a visible contrast. They have to be extended, including the viscosity ratio in the inversion process.

The results of the treatment of the optical monitoring data shown in Figures 11A and 1IB are shown below in Table 1. The production volume of the Vt transition used to predict the cleaning dynamics

25 · corresponds to the point identified on the DO curve shown in Figure 118. The last column in Table 1 contains the measured contamination value in the PVT laboratory.

Vw, cc VJiters Pt di / rv loglQ (QJ V ,, liters Ce, t, ¶ 'C ,,,,% 720 15 20.8 lO 1.32 118 6.0 7.2 5 Table 1: Clean Production Forecast Results These results validate the new technique for monitoring and forecasting cleaning production using real field data, regarding the fact that the viscosity contrast was neglected during the modeling of the progressive problem, and also that the porosity and the anisotropy ratio of the permeability did not known results, the cleaning dynamics forecast result seems promising.

Despite reasonable predictions of final contamination in other investigated cases, attempts to match the initial C - qS / 12 cleaning trend during the initial cleaning phase have so far failed, indicating that the actual patterns of transport flow and contamination, most likely , were different from theoretical solutions.

The concept of monitoring all cleaning, however, has merits, since most of the information on the pattern of flow and transport of contamination is only available during the initial cleaning phase, that is, from the disruption of the cleaning process. forming fluid for the sampling tool until the end of the circumferential cleaning @. During this first cleaning phase, the scale of variation in optical density is large, the variation in the flow pattern in the presence of viscosity contrast 5 can be significant, and the solution of the progressive problem has the greatest sensitivity to the depth of invasion of the mud filtrate.

In contrast, during the final phase of cleaning production, the scale of variation in optical density becomes very small and the patterns of flow and transport of contamination change very slowly. All of this makes it extremely difficult to capture the trend of cleaning dynamics and to perform the inversion of the cleaning monitoring data. It also takes much longer to acquire enough data for historical correspondence.

At the same time, the volume of the reservoir involved in the tarrbérn cleaning control is expanding over time, increasing uncertainty about c) contamination flow and transport patterns, due to the possible presence of 20 large-scale heterogeneities represented, for example, by blades, fractures, impermeable barriers, and reservoir boundaries. Adequate modeling of the flow and transport of contamination in these circumstances, and the level of uncertainty, become more difficult, if not impossible.

25 The advanced hysterical correspondence technique, developed in well tests for the analysis of pressure accumulation, provides a powerful means for the interpretation of well test data. The special type curves developed, which are based on pIotages derived from logarithmic pressure, allow the capture and recognition of characteristics different from the flow patterns created during the formation tests. It may be advantageous to apply this technique for the interpretation of the optical density profiles obtained during the cleaning production.

Using Eq. (2), the relationship between the measured optical density of fluid produced OD (t) and contamination

The unknown rj (t) can be expressed as: 0 (t) = 0, [1-n (t)] (6) 0 (i) = oD (t) - op ,. q ', = ODq -ODf Here,' t'o represents the total contrast of the OD, and 0 (t) is the current contrast of the OD between the mud filtrate and the mixture produced. Parameter 0 is unknown, since it depends on the unknown OD of the formation fluid, OOq. The current contrast of OD, 0 (t), can be measured directly or estimated during cleaning production, as long as the optimum density of the OBM, ODf filtrate becomes. if known.

Using the logarithm of the two parts of equation (6) and then differentiating c) result, the relationship between the contamination rj (t) and the measured DO contrast, 0 (t), which can be represented in the factored form:: "((:; = -ln: 1}) ···" ° '·' · v ········ (7) Here, the apostrophe (') means the derivative of 5 tempo. For the variable production rate during cleaning, c) time t must be replaced by the volume produced V.

If the solution to the flow and transport problem

W of contamination during cleaning production is known, the function rj (t) and its derivative rj '(t) can be calculated. This means that the function on the right side of equation (7) can also be calculated. Thus, Eq. (J) can potentially provide the basis for the interpretation and inversion of optical data obtained during cleaning monitoring. This could be achieved by adjusting the right side of Eq. (7), which represents the theoretical solution for the measured function 0 '(t) /' P (t) on the left side of this equation. The list of adjustment parameters should include all unknowns, such as the depth of the OBM filtration invasion, the permeability anisotropy ratio, the viscosity contrast, and others, if their effects on the solutions can be quantified.

However, experience indicates that the strategy outlined, based on historical correspondence, has little chance of success in real life, for example, for monitoring the sampling of reservoirs because of the natural variability of their flow properties. The explanation for this situation lies in the fundamental differences in the physics of two phenomena, the propagation of the pressure wave and the transport of contamination. by flow in and 5 porous media, which is reflected in its control equations. The pressure accumulation in the formation is controlled by the diffusion type parabolic equation with a tendency to mediate and level the reservoir response in the presence of hornogeneity of its flow properties. In contrast, the transport of contamination in the large scale is governed by the hyperbolic equation of the wave type having a long memory of the irregularities of the flow pattern, which are induced by a heterogeneity of natural reservoir. This possibility results in deviations

Significant W of the optical density measured during the cleaning production, from the theoretical solutions obtained for homogeneous formations or formations with known regular variation of the flow properties. The differentiation-based approach to otic density profiles, tested with various sets of field data, does not show resistance and often leads to significant errors of interpretation. Factors that contribute to cleaning monitoring challenges include standard flow uncertainty, strong sensitivity to contamination transport

W for homogeneity at multiple scales, contrast of unknown DO, noise of the DO profiles, and time constraints.

In view of the above and of the figures, people skilled in the art will recognize that the present disclosure aims to introduce a method of evaluating the formation, which includes the lowering of a sampling tool in a well pit penetrating the formation underground, establishing a fluid communication with the formation, and estimating a depth of invasion. A cleaning model is then selected, based on the invasion estimate, and the cleaning model is used to determine a parameter related to the fluid sample. The estimate of the depth of invasion may include the detection of an outbreak volume. The parameter related to the fluid sample can be at least one of a contamination level and a pumping volume, to reach a contamination target. The selection of the cleaning model based on the estimated invasion may include the choice between a substantially circumferential cleaning model and a substantially vertical cleaning model. Selecting a model

Cleaning W based on the estimated invasion may include detecting a transition from a substantially circumferential cleaning regime to a substantially vertical cleaning regime. The use of the cleaning model to determine the parameter related to the fluid sample may include modification of the estimated depth of invasion.

The present disclosure also introduces a method of evaluating the formation, which includes the lowering of a sampling tool in a well bore penetrating an underground forging, establishment of a fluid communication between the formation and a sampling tool, the detection the inrush of the formation fluid to the sampling tool, and the detection of the transition from the cleaning regime, from a predominantly circumferential cleaning regime to a predominantly vertical cleaning regime. Such a method may also include the characterization of flow patterns and contamination in formation, based on the detection of the outbreak and the detection of the transition. The method may also include estimating an initial depth of invasion, before detecting the outbreak and transition. The estimate of the depth of the initial invasion can be performed in the absence of viscosity contrast. the method can also include the reconstruction of the initial depth of invasion, after detecting the outbreak and transition, based on the detected outbreak and transition. The reconstruction of the initial depth of the invasion can be performed in the absence of viscosity contrast.

The present disclosure also presents a method of monitoring the irrigation and forecasting in real tempQ, aiming at the estimation of the pumping volume versus final contamination, including the determination of a volume of irrigation production, the detection of a first volume of production corresponding to the transition from a first cleaning regime to a second cleaning regime, and determination of a

The first normalized production volume corresponding to the transition to the cleaning regime, where the determination of the first normalized production volume is based on the production volume of the irruption and the first production volume.

A first estimate of the depth of invasion corresponding to the first normalized production volume is then determined. A second normalized production volume corresponding to a predetermined contamination target is then determined, in which the determination of the second standardized production volume is based on the> stamped depth of invasion. A second production volume (corresponding to the contamination target) is then determined, in which the determination of the second production volume is based on the second normalized production volume and the production volume of the outbreak. The first cleaning regime can be a predominantly circumferential regime, and the second cleanness regime can be a predominantly vertical regime. The determination of the first invasion depth estimate corresponding to the first normalized production volume

The may include determining a logarithmic relationship between the first normalized production volume and the relationship between the depth of invasion and the radius of the well. '

The above presents characteristics of several modalities, so that people skilled in the art can better understand the aspects of the present invention. Those skilled in the art should understand, that they can easily use the present disclosure as a basis for designing or modifying other processes and structures for the realization of the same purposes and / or realization of the same advantages of the modalities introduced here. Those skilled in the art must also realize that such equivalent constructions do not depart from the spirit and scope of this disclosure, and that they can make various changes, substitutions and changes to this document, without departing from the spirit and scope of this disclosure.

The nomenclature used here includes: a, b = adjustment parameters B = coefficient in the approximate cleaning solution c = filtrate concentration C = filtration contamination, ° s Cest = estimated filtrate contamination,% CLak = measured filtrate contamination in Lab PVT, '% dí = depth of invasion Kh = horizontal permeability kv = vertical permeability OD = optical density

ODr = optical density dQ filtered from OBM

OD0 = optical density of the virgin fluid of the formation rp = radius of the probe

5 r, = radius of the simulation domain rw = radius of the well 0 t = time

V = pumping volume

Vbt = inrush volume

Vp = production volume

Vt = pumping volume in the transition to the cleaning regime oc = exponent in the cleaning solution api: oximada

8 = dimensionless depth of invasion çtl = porosity çp = contrast of optical density

0, = contrast of DO between virgin formation fluid and OBM filtrate rj = contamination rI0 = contamination target

C2 = pumping dimension

Qp = dimensionless production volume

Q, = dimensionless pumping volume in the transition to the cleaning regime

Claims (15)

I - CLAIMS - 1. EVALUATION METHOD "OF TRAINING, characterized by the fatcj of understanding: lowering of a sampling tool 5 inside a well bore penetrating an underground formation; establishment of fluid communication with the supply; estimation of an invasion depth; selection of a cleaning model, based on the estimated invasion; and W use the cleaning model to determine a parameter related to the fluid sample. 2. Method according to claim 1, characterized by the fact that the depth of invasion strain includes the detection of an outburst volurne. 3. Method, according to claim 1, characterized in that the parameter related to the fluid sample is at least one of a contamination level and a volumetric volume to reach a contamination target. 4. Method, according to claim 1, and characterized by the fact that the selection of the cleaning model based on the stretched invasion includes the choice between a substantially circumferential cleaning model and a substantially vertical cleaning model. 5. Method, according to claim 1, characterized by the fact that the choice of the cleaning model based on the estimated invasion includes the detection of a transition from a substantially circumferential cleaning regime to a substantially vertical cleaning regime. 6. Method, according to claim 1, characterized by the fact that the use of the Olympia model to determine the parameter related to the fluid sample includes the modification of the estimated depth of invasion. 7. TRAINING EVALUATION METHOD, characterized by the fact that it comprises: lowering a sampling tool into a borehole penetrating an underground formation; establishing fluid communication between training and a sampling tool; detection of the formation fluid inrush to the sampling tool; and detecting the trisposition of the cleaning regime, from a predominantly circumferential cleaning regime to a predominantly vertical cleaning regime. 8. Method, according to claim 7, characterized by the fact that it also includes the characterization of the flow and transport patterns of contamination in the formation, based on the detection of the outbreak and the detection of the transition. + 9. Method, according to claim 8, characterized by the fact that it still includes the estimate of 5 an initial depth of invasion, before the detection of the outbreak and the transition. 10. Method, according to claim 9, characterized in that the estimate of an initial depth of invasion is made in the absence of viscosity contrast. 11. Method, according to claim 9, characterized by the fact that it still · includes the reconstruction of the initial depth of invasion, after the detection of the outbreak and transition, based on the detected outbreak and transition. 12. Method according to claim 11, characterized in that the reconstruction depth of the initial invasion is carried out in the absence of viscosity contrast. 13. CLEANING MONITORING METHOD R RPÂL WEATHER FORECAST, aiming at estimating the pumping volume versus final contamination, characterized by the fact of understanding: determining an irruption production volume; V detection of a first volume of production corresponding to the transition from a first cleaning regime to a second cleaning regime; determination of a first normalized production volume 5 corresponding to the transition to the cleaning regime, where the determination of the first normalized production volume is based on the production volume of the outbreak and the first production volume; determination of a first estimate of the depth of penetration corresponding to the first volume 0 of normalized production; determining a second normalized production volume corresponding to a predetermined contamination target, where the determination of the second normalized production volume is based on the first depth of stamped invasion; and determination of a second production volume corresponding to the contamination target, where the determination of the second production volume is based on the second standardized production volume and the production volume of the Inrush. 14. Method, according to claim 13, characterized by the fact that the first cleaning regime is a predominantly circumferential regime, and the second cleaning regime is a predominantly vertical regime. 15. Method according to claim 13, characterized by the fact that the determination of the first estimate of the depth of invasion corresponding to the first volume of normalized includes the determination of a logarithmic relationship between the first volume of standardized production and the relationship between the depth of invasion and the radius of the well. And the. q 126 ,,) \\ "'° 16 22- \ 1r"!' "TSq4í ,, <T,: sZz" '"R") "jy6 //;" bv, | j! "" \ "" A / x \ &:, '., g4t l | E & -' 9 Mp '/ e X i: 23) / Í "Á jz :: Imi" Fig. 1 Signature of flow and transport patterns of contamination, depth of invasion ,, X ,, + short term of monitoring O 0 B | · "0 · 0 0: ° INPUT F" | LE: "BF MDT OFA 193LTP.CSV - RAW DATE 5p Ch 1 4-5j —--- Ch 2 Ch 3 - · ------- "· Ch 4! 3. :) C, r'TTT = T ':' r, 'Ch 5: 2: lJ '(()) lL ,, e ,,,)' ···· 0 "" "" '0 ···· 00 ...)), j..i · rl jj · 0 2l Q ): ii "I · à 0 2 1j Q 0 0 j'q l, 0 0.5 I- 0 0 · \. · V ~ j 'Lt 0; Òj e *, b 0 d 0L "' 4p ''" "% '. ·" L' ®CF ° 2QbD "·· .. .3000 · · "° 4000 5000 0 TIME: + Very slow variations in DO and flow pattern + long monitoring time Fig. 2 "hm ^:% = Slkc ~ mà * ~ íâ N ¥: L.0pq, y - <a" "" "! 7" "'N \, IK · -; "" '"" "j' ^. M.,: .-" j! K "..!.. '' MY,) i, ,, A '" "" Ú ". &" 't' '"' '" Á' "" b 'i' · "'' S ·; 1 '..' -uú 'e: .i · 1 to 4:, ·" "' ¥" "" ' ":.: ',' +: * + .L? {, SST '"' 4 "" "" "" '' · 'X:'} r; F? Sg. -...... ! cl: ', d ¶ "" "X' 'S"' "'N, W" ?? ""% "" "'" "" - - ""' "'"' '' á "" k: (3 "í; i" ',. T:' "- 'l 4": 7. < ;, m: àa .., È 'a ">: n": "L'.", í '":" - Z: "" - Y:,:?' ">" "'' '" ·· .,, · Wq = b & "X» "':: j' - ,, i: '% - A r +, yu",: - -' "* µ" ^ bN. + "» 4, ";» " N "Yr" - '""' -. > "j ~ 4U" ^ "l ¢. 'vt .. ... 4' '" "%: W> J" "" ""' "" -: <2; "'' L" "f <ÁY '" "^" üR. ^ M "n Fig. 5 t» eeWF? NÃyLi * · mNi? M * R &', í-tg »d & ey-mm = Rmm / K [ ": k #« è.i * kHgm: i) NHí "'j" 9çgf!% z ± iIÈ? i ¢ í9!:', |: iK '::' "T!)) | ij} 'è" 2 =: j: íl ') ::' '-5 -.?.5 -í -.Lf l · ák u QS' nµ ": y '!') Ll w 'X LS Z. ZS .7" - . ' Z k'iir '% im))) YES!)) "· 3 -Q, S -: 44 -; - ¥ Ü 9 $ Á Âís R 2. $ .3 Âiít. -C: IN) TÊnmw% Ê · l'í'Õ; .w®'¶9zs "? á 'H, | FÇ'%} í3-" út'r & h.gUMm) r¶ ['* C' [rkKàk .. m im'm "E ' rArkNMM'R Hey "mV3i))) p) '" "g) i |').) '))) t!:': :), ';' c3) n." µ ^ | Êt. ' "l '" k': m. 'r "'" '"·" ··· -% · à &: y>: "'>, &. ' <? -US g 'Xf mãS - 9 ·' 'Lí' $ 'A 2 .W - "' Àm, gõ: ío> ·? q-s "Z '" ·' k $ íj "= S" O :: '[LS t "U 3" làG v> .i &. Fig. 6 10 '—Ô = 0.125 —0 = 0.25 —0 = 0.5 r 7 -'0 = 1 »' r" <"--- - 8 = 2 *% Y% cP B r rnc1ínaçú * I 1Q 'Slope $ 2/3 "YX> <> S = -5 / 12« ·' + 6 = 0.125 U 0 = 0.25 6 = 0.5 c5 =] 6 = 2 io¶- 10 '10' 10 '10' VQ = m IR BIO / VQUME ~ O Fig- 7 APPROXIMATION OF CLEANING SCHEME TRANSITION 4 | 'Ê k: q: Cl Simulation - Interpolation | 3-5 I "O L O ExtrapD1açàg" | j ° y)% :: t 'È L = e2a OS Ci = m £ erencia1 t: J "' Sòg j {} i mj INVASION FUNDITY / WELL RADIUS Fig- 8 Invasion ~ & proEunda n T _ [, nv, a, s, à, o, ma, í, $ m f, a3, â jQ2 ~ \. ' "1·· .!... \ "; 'l "::' Approximation of the Transition 7 ¥! \": <'"." i:': j J ') of the Mint of Cleaning 4' "rb, Si '.' '-'. ' · ': /% DP I%% 4 4 "____m'" ..: j% LíIrE) and "the vertícal H"%. RK: '. .. "'. '. <"~ s" = -2/3 ""> "Kxrr x r r" "'' fl I% k" "% ¶. "" ''% '""'. '":':" - '\ K n "" "---,> x <¶, 4, ¶,,% á 1 Qll I" "" "' K. g—,, i | ) .Í).).) 7:?,%,. . ,, \ x "% x r>?> -.% U I TÁ.npez a · r b X ~ X. b .. Ci == lferencia.1.% b G ,,? ^ W b s, = -5 / 1A7 & '"" 4% ·% x ·' 0 "" _> r yDde1o of "'" -, \ r ^ X. [.jr "Z. í spherical flow" "" "· ..., y> G IjOQ 10 ° 10 '10' lO 'PUMP VOLUME' TRRUPTION VOLUME Fig 9 10 't FII j I ¢? U" l , Ip "~ üc 10 'LH ,,% 0 ~ f-> m' i '°°! L10"' l 10 "2,: è I 1 r 10" '10 ° iü' INVASION depth, 5 = d ./r [w Fig. lO 2 | P \ i [ij '·' F "qi | i | liq ÍJ1m ,, z 'i 1-4 i Stop of 1 => eza Q.2 ff í _ I. 0 O 1000 2000 3ODQ 4000 5000 GC) ® 7QW TEMPQ DECQKRIDO, S Fig IIA} 5 KÇ 1 ' U Í.L gives the free regime! ! t I (j 20 40 60 80 ioa 120 volume of exchange, liters Fig IIB ) | t - "'|)) Í |)) 4' 8 m â rp fl )) |))) |))) i'j) '))' l)) | È N In jG, .3 5i) \)) IÍ) m ¢