METHOD FOR PROCESSING SAMPLED PROPPANT DURING HYDRAULIC FRACTURING

Abstract

A method for processing proppant from a well. A plurality of proppant samples are collected during drilling of a well for imaging analysis. In addition to imaging analysis a further correction factor is determined for the samples using additional analysis on a portion of the samples. Determining and applying the correction factor to the imaging results provides a more accurate proppant log. The additional analysis can be by scanning electron microscopy, such as for differentiating the silicon proppant particles from other elemental particles resulting from the drilling.

Claims

1. A method for processing proppant from a well, the method comprising the steps of: collecting a plurality of proppant samples during drilling of a well; obtaining imaging results from the plurality of proppant samples collected; determining a correction factor from a sample subset of the plurality of proppant samples; and applying the correction factor to the imaging results to obtain corrected proppant data.

2. The method of claim 1, further comprising: determining the correction factor by elemental analysis of the sample subset.

3. The method of claim 1, further comprising: determining the correction factor with scanning electron microscopy (SEM) of the sample subset.

4. The method of claim 1, further comprising: determining a distance zone within the well for the sample subset; and applying the correction factor to all proppant samples within the distance zone.

5. The method of claim 1, wherein the correction factor is a correction curve.

6. The method of claim 1, wherein the correction factor is an element ratio of silicon to calcium and/or barium in the subset of the plurality of proppant samples.

7. The method of claim 6, wherein applying the correction factor comprises adjusting an imaging element ratio in the imaging results as a function of the element ratio of the correction factor.

8. The method of claim 1, further comprising: normalizing the correction factor as a function of a rate of drilling or penetration at a time of the sample collection.

9. The method of claim 1, further comprising: determining a number of proppant particles in the subset sample within each of a plurality of particle size ranges; and interpreting hydraulic fractures as a function of a particle size distribution within the sample subset.

10. The method of claim 1, further comprising: determining a source of proppant in the sample subset as a function of proppant particle size within the sample subset.

11. The method of claim 1, further comprising: collecting the sample subset at one of a mud and cuttings inlet, a liquids and fine solids discharge, and/or a cuttings discharge of a hydraulic fracturing mud cleaner or shale shaker.

12. A method for processing proppant from a well, the method comprising the steps of: collecting a plurality of proppant samples during drilling of a well; obtaining imaging results from the plurality of proppant samples collected; determining proppant values in the plurality of proppant samples using a SEM analysis; generating a correction factor from the SEM analysis; and validating the imaging results with the correction factor.

13. The method of claim 12, further comprising: applying the correction factor to a proppant log; and producing a corrected proppant log.

14. The method of claim 12, further comprising: determining a distance zone within the well for each of the proppant samples; and applying the correction factor to all proppant samples within the distance zone.

15. The method of claim 12, wherein the correction factor is determined by comparing the imaging results to an elemental analysis of the proppant subset.

16. The method of claim 12, wherein the correction factor is an element ratio of silicon to calcium and/or barium in the subset of the plurality of proppant samples.

17. The method of claim 16, wherein applying the correction factor comprises adjusting an imaging element ratio in the imaging results as a function of the element ratio of the correction factor.

18. The method of claim 12, further comprising: normalizing the correction factor as a function of a rate of drilling or penetration at a time of the sample collection.

19. The method of claim 12, further comprising: determining a number of proppant particles in the subset sample within each of a plurality of particle size ranges; and interpreting hydraulic fractures as a function of the particle size distribution within the sample subset.

20. The method of claim 12, further comprising: determining a source of proppant in the sample subset as a function of proppant particle size within the sample subset.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0016] FIG. 1 shows a graphical representation of interpreted hydraulic fractures compared with observed proppant to mineral ratios according to one embodiment of this invention.

[0017] FIG. 2 shows calcite and proppant particles according to one embodiment of this invention.

[0018] FIG. 3 shows SEM-EDS composition maps for identifying proppant ratios according to one embodiment of this invention.

[0019] FIG. 4 shows subsurface proppant particles according to embodiments of this invention.

[0020] FIGS. 5A and 5B shows a correction plot derived from SEM-EDS analysis before and after correction of proppant log data according to one embodiment of this invention.

[0021] FIG. 6 shows a graphical representation of correction factors identified at various depths according to one embodiment of this invention.

[0022] FIG. 7 shows a graphical representation of penetration behavior compared with proppant counts according to one embodiment of this invention.

[0023] FIG. 8 shows a graphical representation of proppant distribution from a cored interval according to one embodiment of this invention.

[0024] FIGS. 9A and 9B show graphical representations of normalization results to remove an impact of proppant outliers according to one embodiment of this invention.

[0025] FIG. 10 shows a graphical representation of normalized proppant distribution and original proppant distribution according to one embodiment of the invention.

[0026] FIG. 11 shows a perspective view of a mud/drill cuttings handling system according to one embodiment of the invention.

[0027] FIG. 12 shows a flow chart of a modified workflow with SEM analysis-based QC according to one embodiment of the invention.

[0028] FIG. 13 shows a series of charts of plotted size distributions of proppant according to one embodiment of the invention.

[0029] FIG. 14 shows a graphical representation of mesh proppant and normalized proppant distribution according to one embodiment of the invention.

[0030] FIG. 15 shows a graphical representation of proppant count vs. lateral well offset.

[0031] FIGS. 16 and 17 show an exemplary proppant log, and the cross-well interconnections between two example wells.

[0032] FIG. 18 shows another example proppant log to illustrate aspects of this invention.

DETAILED DESCRIPTION OF THE INVENTION

[0033] The present invention provides a process for accurately determining proppant particles. Proppant particles, and other minerals with similar optical properties, tend to be hard to distinguish using traditional imaging workflows. This is particularly true for particles with smaller size ranges (approximately 50 .Math.m to 120 .Math.m). FIG. 2 shows examples of proppant and calcite cement particles at these small size ranges obtained by controlled sampling. Downhole samples are more difficult to distinguish by imaging as these particles undergo abrasive action from fluid movement.

[0034] Baseline proppant results are available from an imaging workflow. The claimed invention incorporates an additional analysis step using scanning electron microscopy (SEM) analysis of samples that have already been imaged. Using SEM-EDS in accordance with the subject invention for elemental composition analysis, on selected samples, can provide enough data to allow for an accurate relative proppant indication which can then be applied on a dataset to more accurately determine final proppant counts, as shown in FIG. 1. A representative sample or multiple representative samples can be analyzed for their elemental composition and relative presence of proppant or other mineral particles can be accurately identified. Observed results can then be corrected for the original sample imaging using a correction factor identified from the SEM analysis.

[0035] SEM analysis allows for an understanding of the chemical composition of the samples. Proppant particles are rich in Si whereas other, different minerals such as sulphates or calcites do not show significant Si but are instead usually rich in Ca or Ba. Other particles, such as shales, are richer in Al, Mg, etc. With known chemical compositions of samples of interest, proppant particles can be isolated from other particles that may show proppant-like characteristics.

[0036] Most samples show natural minerals to be either calcites or sulphates (CaCO.sub.3/ BaSO.sub.4). Embodiments of the invention utilizes relative compositions of Si rich vs. Ca or Ba rich samples. Once these ratios are identified from the SEM analysis, they can be compared with observed ratios from the original imaging workflow to figure out how much correction may be necessary in the data. A relatively small amount of material is used to prepare samples to be imaged and analyzed for a bench scale SEM-EDS system. FIG. 3 shows two samples that contain varying quantities of proppant.

[0037] FIG. 4 shows another set of examples containing varying levels of proppant from negligible proppant particles to significant proppant and calcite or other mineral particles. Proppant ratios of the samples can be compared with ratios observed from original imaging analysis. Output or corrected (SEM derived) ratios can then be modeled using regression analysis.

[0038] FIG. 5 shows imaging vs. SEM ratios and a regression model that can be used for corrections. FIG. 5 shows 223 samples collected with 23 points per sample selected for SEM based QC (quality control). The corrected data vs. the original data is also shown in FIG. 5.

[0039] While the impact on individual samples in terms of corrections can be significant, an overall trend maintains the behavior in the proppant distribution data. FIG. 6 shows higher correction factors identified from samples above 15,000 feet on average, compared to samples taken at shallower depths. This corresponds to the depth at which hydraulic fractures begin showing in FIG. 1.

[0040] Another important aspect considered in the subject invention is curve normalization based on varying rates of drilling or penetration (ROP). FIG. 7 shows ROP normalization for in-fill drilling cases. More particularly, FIG. 7 shows ROP behavior across lateral compared with proppant counts. Varying ROP on final proppant logs can have a detrimental impact to the results. Specifically, at lower ROP’s, particle counts tend to be higher due to more re-circulation over a given period. At higher ROP’s, the amount of material being re-circulated is relatively lower. Therefore, it is important to maintain a consistent ROP over a logged length and if not, data normalization is necessary to recover a more accurate proppant distribution. FIG. 7 shows the impact of a drop in ROP as a lateral was built during the transition from vertical to near horizontal drilling.

[0041] Without correction, the proppant distribution from the log may only be useful or comparable from a depth of 13000 ft. MD onwards. There are also ways to normalize for changing ROP. For example, a median filtered ROP curve can be generated to remove outliers. Then, medium adjusted ROP values are computed. Since most of the ROP data is stabilized at approximately 75 ft/hour to 80 ft/hour, the normalized distribution weighs lower ROP’s lower and weighs ROP’s close to the median at approximately 1. This can then be used as a multiplier for the proppant data and can retain most of the data past 13000 ft. MD, as is, and reduce the values prior to 12000 ft. MD.

[0042] When it comes to normalization for proppant screen-out conditions, proppant distribution, particularly at very high sampling rates (such as one sample every 3 to 6 feet, for example), the impact of individual fracture morphologies and consequent localized screen-out can be significant. At lower sampling rates (such as 10's of feet per sample, for example) surface sample collection allows for moderation, however subsurface samples need to be further normalized to reduce the impact of localized screen-out behavior.

[0043] FIG. 8 shows distribution from samples collected from a core. Here, the sampling rate was every 3 feet. At the high resolution shown in FIG. 8, localized peaks are shown in the distribution as a result of fracture morphologies and complexities associated with bed boundaries and high stress contrast zones.

[0044] There are various approaches to normalization available. Often very high proppant peaks are associated with localized screen-out and therefore, these high peaks are likely not well drained. However, higher ratios of proppant particles to other minerals are a good indicator of propped fractures. Therefore, normalization according to the subject invention can occur by maximizing the ratio and minimizing maximum proppant identified at sampling locations.

[0045] FIG. 9 shows two normalization results according to the present invention to remove an impact of proppant screen-out outliers and compare those results with hydraulic fractures. In this example, results were observed to drainage behavior at various locations in cored wells. Pressure gauges were placed in the cored wells at specific depths of interest and then the pressure was monitored over a period of a few years to see how pressure drawdown varied over time. The drawdown behavior was found to be strongly correlated with proppant as well as hydraulic fractures. The normalized parameter shared here as well as other methods have shown a strong correlation with the fractures and drawdown behavior as seen in FIG. 9.

[0046] Several normalization approaches were used in the examples shown in FIG. 9. The normalization relationship can be shown by the following equations, for example.

[00001]$\text{Normalization}\text{of}\text{Proppant}=fx\left[\frac{\#P_i}{\frac{{\displaystyle .Math.{}_{i-win}^{i+win}\#P_i}}{2\times win}}\right]$

[00002]$\text{Normalization}\text{of}\text{Proppant}=fx\left[\frac{{\displaystyle .Math.{}_{i-win}^{i+wi{n}^{\#P}}\raisebox{1ex}{$i$}\!\left/ \!\raisebox{-1ex}{$\#N{C}_i$}\right.}}{\underset{i-win:i+win}{{\displaystyle \max }}{P}_i}\right]$

[0047] FIG. 10 shows a first normalized distribution curve as well as original proppant distribution before normalization. With normalization, proppant count drops at peaks. Convergence is shown between fracture data as well as pressure drop behavior. FIG. 10 shows a normalization parameter and how the parameter compares with the original proppant distribution. This normalization is preferably used when proppant sampling is at higher rates (for example, <1 sample every 6 feet).

[0048] FIG. 11 shows collection points located on a mud/drill cutting handling system 20 for a drilling rig. In order to take a proper composite sample at a wellsite that contains proppant (and other particles) and is representative of the length of the borehole drilled during the sampling period, a continuous liquid or solids slip stream collection can be used. A portion of a liquid mud return sample (that has not been filtered in any way) is redirected from the inlet 22, so as to only collect solids that settle at a bottom of a collection vessel. The collection vessel may be a bucket 30, such as shown in FIG. 11.

[0049] Another collection method includes collecting a slip stream of separated drill cuttings from a screen at a cuttings discharge location 24 and/or the liquid and fine solids discharge 26 of the mud handling system 20. This can be done by placing a bucket 30 under the falling cuttings at one location of the screen. The bucket 30 can have a fine screen 32 placed on top so as to prevent large cuttings from entering the bucket 30.

[0050] FIG. 12 shows a modified workflow with SEM analysis-based quality control in a flowchart according to the invention. The flow chart details to obtain final corrected proppant log data from original sample preparation. The left side of the flow chart includes the steps of the imaging process. The right side of the flow chart includes the steps of the further analysis according to embodiments of this invention. As shown in FIG. 12, N-number of subset samples are sent for SEM analysis in step 40. In step 42, the silicon vs. calcium/barium particles are determined for each sample by the SEM. This analysis provides the correction factor, namely correction curves in step 44, which are applied for correction of the log data from the imaging process in step 46, resulting in the corrected proppant log 48.

[0051] In embodiments of this invention, the correction curve looks to compare a relative abundance of the proppant observed in the scanned samples (from image analysis) vs. the abundance of proppant from the SEM scans, where elemental classification is possible to get accurate proppant counts vs. other minerals such as calcites and sulphates. An α parameter (α = number of proppant particles/ number of other minerals with proppant-like optical characteristics, i.e., calcites, barite, etc.) is obtained. Then one can cross-plot values from the two analysis techniques and use regression fit to get a correction factor (slope/intercept). In order to identify all potential proppant-like particles, the elemental composition of baseline samples are studied (rock with no proppant), as well as various chemical additives used in the frac fluid.

[0052] FIG. 13 shows representations of extracting and plotting size distributions of proppant. FIG. 13 shows original proppant counts as well as proppant to calcite ratios and data relative to size distribution of proppant. The size distribution can be calculated as follows:

[00003]$\Delta \text{}Proppant\#\left(Bin\right)=\frac{1-\left(Coun{t}_{100-125}-coun{t}_{125-150}-Coun{t}_{150-175}\right)}{\left(Coun{t}_{100-125}+coun{t}_{125-150}+Coun{t}_{150-175}\right)}$

Using values obtained from this calculation, various parameters can be found from proppant samples such as the size and quantity of proppant (with the ideal values being close to 100 microns of proppant). The concentration of the ideal 100 micron proppant is further shown in FIG. 14. FIG. 14 shows that having a majority of proppant particles in a higher level bin results in a higher likelihood of those particles being 100 micron mesh proppant particles.

[0053] In additional embodiments of this invention, the determined information and attributes of the samples is used to characterize proppant distribution. For example, consider the following β parameter:

[00004]${\beta }_i=SF\times \frac{C_{125-150}+C_{150-175}}{C_{100-125}+C_{125-150}+C_{150-175}}$

where C.sub.a-b indicates the number of proppant particles in a given size range [a, b], defined in microns. A higher β value indicates relatively more larger particles, whereas a smaller value indicates higher amounts of smaller particles at the sampled location. This is useful in interpreting a source of the observed proppant and/or associated hydraulic fractures at the sampled location. This is because proppant distribution is governed by vertical and lateral transport behavior which is a function of particle transport in suspended mediums. As an example, the impact of lateral transport distances can be seen in FIG. 15, where the upper Wolfcamp formation (WC) core has similar distributions of 100 mesh (smaller) and 40/70 (larger) mesh proppant (though 100 mesh is generally more compared to 40/70 mesh). On the other hand, the middle WC core shows significantly larger proportion of 100 mesh proppant compared to 40/70 mesh. The likely cause is the larger lateral offset of the sampled core location with the middle WC well (~150 ft) when compared to the upper WC well (~70 ft to 100 ft).

[0054] This template for interpretation has been verified in field tests where fracture propagation has been independently quantified using cross-well fiber (strain change) response. FIG. 16 highlights two sections along the proppant log where significant crosscutting hydraulic fractures from a well farther afield (B4H) intersects the well that is being sampled (B3H). Thus, the expectation is that significant proppant observed across stage 20 (FIG. 17) are associated with B4H fractures and not B3H fractures.

[0055] To validate the understanding from cross-well fracture arrivals, the distribution of the β parameter from proppant log is looked at to make the interpretations. Referring to FIG. 18, note that the first peak at ~ 13100′ corresponds to relatively smaller particles as per the β parameter, but later peaks such as at 13430′ correspond with larger proppant particles. Thus the first section is likely associated with fractures from B4H and the second section with more proximal B3H well. This interpretation template can be applied more generally on any proppant log and is especially useful when there is good well control in the pad where this log is being run.

[0056] The invention illustratively disclosed herein suitably may be practiced in the absence of any element, part, step, component, or ingredient which is not specifically disclosed herein.

[0057] While in the foregoing detailed description this invention has been described in relation to certain preferred embodiments thereof, and many details have been set forth for purposes of illustration, it will be apparent to those skilled in the art that the invention is susceptible to additional embodiments and that certain of the details described herein can be varied considerably without departing from the basic principles of the invention.