Evaluating Production Performance from Horizontal Well Refracs in Saskatchewan

Transcription

1 Evaluating Production Performance from Horizontal Well Refracs in Saskatchewan Michael J. Nelson 1 Information from this publication may be used if credit is given. It is recommended that reference to this publication be made in the following form: Nelson, M.J. (2017): Evaluating production performance from horizontal well refracs in Saskatchewan; in Summary of Investigations 2017, Volume 1, Saskatchewan Geological Survey, Saskatchewan Ministry of the Economy, Miscellaneous Report , Paper A-5, 10p. Abstract Refracking, a form of well completion, is the process of performing a hydraulic fracture stimulation treatment on a well that has previously been hydraulically fractured. This can be done with the objective of fracturing a new interval within the well, fracturing a previously attempted interval, or a combination of both. The major advantage of this process is that recompleting a well can be done at a fraction of the cost of drilling a new well, and in some cases the production rate following the refrac has met or even exceeded the initial well production rate. Refracking stimulations have been attempted in unconventional reservoirs both in Canada and the United States. The focus of this paper is evaluating the production performance of refractured horizontal wells in Saskatchewan, and how successful or unsuccessful the refracking has been. This study was done using publicly available data from geologic Systems. Well production data, well production rates and well cumulative recoveries before and after the refrac completion were compared and evaluated to determine the level of success of the recompletion. In addition, the data was broken down by formation in which the re-stimulation was attempted, to determine which formations showed more potential for successful refrac completions. The majority of the analyzed refrac completions were in Bakken Formation pools in southeastern Saskatchewan and Viking Formation pools in the west-central area of the province. Keywords: refrac, completion, unconventional reservoir, Bakken Formation, production rate, Viking Formation, recompletions, estimated ultimate recovery 1. Introduction Refrac completions have become an increasingly popular method to boost production rates and estimated ultimate recovery (EUR) in unconventional horizontal wells. The most attractive feature of a refrac completion is that it requires considerably less capital investment than drilling a new well, approximately 20 to 35% of the cost to drill and complete a new well, according to some estimates (Endress, 2015). So, if well production rates after refrac can approach initial production rates and EUR significantly increases, the economics of refrac completions become desirable. It is estimated in a report by IHS Inc. that up to 11% of horizontal wells in the U.S. could eventually be refractured (IHS Inc., 2014). This paper will examine the outcomes of refracs in Saskatchewan, nearly all of which have been completed in the Bakken Formation, with a few in the Viking Formation. Initial research was carried out in the fall of 2016, follow-up research was carried out in early 2017, and final results were presented at the annual Williston Basin Petroleum Conference in 2017 (Nelson, 2017). 2. Data Acquisition and Sorting Using data from geologic Systems Well Completions and Frac Database and geoscout, a 57-well dataset was obtained. To choose wells to be examined, a search was performed on the frac database to identify all wells in 1 Formerly Saskatchewan Ministry of the Economy, Saskatchewan Geological Survey, 201 Dewdney Avenue East, Regina, SK S4N 3Z8 Although the Saskatchewan Ministry of the Economy has exercised all reasonable care in the compilation, interpretation and production of this product, it is not possible to ensure total accuracy, and all persons who rely on the information contained herein do so at their own risk. The Saskatchewan Ministry of the Economy and the Government of Saskatchewan do not accept liability for any errors, omissions or inaccuracies that may be included in, or derived from, this product. Saskatchewan Geological Survey 1 Summary of Investigations 2017, Volume 1

2 Saskatchewan where a second workover was recorded. The list of wells was then narrowed by removing wells where the second workover was not a refrac, or where there was no production history prior to the second fracture completion. After the list of wells with refrac completions was established, the production history of each well was obtained through geoscout. The detailed information on refracture completions from the frac database was compared to the production data from geoscout, to ensure that the information obtained from the two sources was consistent. For example, the production rate history from geoscout had to correlate with the dates recorded in the frac database for each completion attempted on the well; there should have been an increase in production rate immediately after the completion of the workover; and there should have been a shut-in period in the production history that matched the time period when the workover was performed, according to completion records. This process resulted in a dataset of 57 wells. These consist of 45 refrac completions in the Bakken Formation, 11 in the Viking Formation, and 1 in the Shaunavon Formation (Figure 1). Figure 1 Base map (upper left) showing the location of the study area (red polygon) in southern Saskatchewan. The map of the study area (lower centre) shows the wells used in this study (green crosses within magenta polygons). Refractured wells used in the study include 45 wells in the Bakken Formation in southeastern Saskatchewan; 11 wells in the Viking Formation in west-central Saskatchewan; and 1 well in the Shaunavon Formation in southwest Saskatchewan. Saskatchewan Geological Survey 2 Summary of Investigations 2017, Volume 1

3 As there is only one well with a refrac completion in the Shaunavon Formation, it would be difficult to conclude how successful refrac completions are for this formation. The remainder of this paper will therefore be focussed on wells in the Viking and Bakken formations. 3. Bakken Formation Analysis The analysis presented in this paper will focus on oil production rates. By comparing the oil production rate immediately after the refrac has been completed to the oil production rate immediately before the refrac, we can observe how much of an increase in production rate the refrac was able to generate. By also comparing the postrefrac oil production rate to the initial well completion oil production rate, it can be determined the extent to which the refrac was able to restore the well to the initial oil production rate. Any resulting incremental increase in EUR was also determined, by comparing the EUR before and after refrac completion. The change in EUR was difficult to quantify for the wells in the Bakken pools, however, as large areas of the Bakken Formation are under a waterflood enhanced oil recovery (EOR) program. The refrac and waterflood would both have some impact on the increase in EUR, and it would be difficult to calculate the impact of one recovery scheme over another. Where possible, a traditional Arps decline curve (IHS Inc., 2014) with an abandonment rate of 0.5 cubic metres per day (m 3 /day) was fit to the data to estimate EUR for the Bakken wells. Where a reasonable decline curve fit was not possible, the actual cumulative production was used. On initial breakdown of the 45 wells in the Bakken Formation, it is worth noting that a large number of the refracs took place in 2009, on wells that were originally completed in Many of those wells changed ownership in an acquisition deal, which suggests that the company acquiring the wells felt that the original completion strategy could be improved upon. After the spike in refrac completions in 2009, there was only about one refrac per year, with the exception of 2013, which saw an increase to six refracs (Figure 2). Figure 2 Histogram showing the timing of the frac completions (initial and refrac) between 2006 and 2016 for the 45 wells studied in the Bakken Formation in southeast Saskatchewan. As can be seen in Figure 3, the majority of Bakken Formation wells were refractured after the oil production rate dropped below 4 m 3 /day, which suggests that the economic limit of Bakken Formation wells is somewhere between 0 and 4 m 3 /day. In 44 out of 45 cases, the well s oil production rate improved after refrac; in the majority of cases Saskatchewan Geological Survey 3 Summary of Investigations 2017, Volume 1

4 production improved to at least 10 m 3 /day. There is one case, however, where the oil production rate after refrac was actually lower than before the refrac. This shows that refracs, like any operational process, carry the risk that they will not be successful. Figure 3 Graph comparing oil production rates (m 3 /day = cubic metres per day) prior to and after refrac for wells drilled into the Bakken Formation in southeast Saskatchewan up to the end of The majority of Bakken Formation well refracs were performed after oil production rate dropped below 4 m 3 /day. The blue line represents the equal rate line, where the oil production rate was the same immediately before and after the refrac was performed, in other words, the line where the refrac completion resulted in no change to the rate of oil production. Wells plotting above the blue line had a higher rate of production after refrac than before the refrac. Only one well (symbol plotting below the blue line) had a decreased rate of oil production after the refrac. In Figure 4 it can be seen that, for a large number of the Bakken Formation wells in the study, the initial oil production rate following the initial frac completion averaged between 15 and 30 m 3 /day. After refrac, the oil production rate for the majority of these wells was restored to between 50 and 100% of the the initial oil production rate. Some of the wells did achieve over 150% of the initial oil production rate after the refrac, however, the initial oil production rate on the majority of them was below average to begin with. The wells that had an initial oil production rate over 30 m 3 /day saw their rates after refrac become nearly 50% of the initial oil production rate. This is not surprising given that, in Figure 3, it appears that very seldom does a refrac result in a production rate above 27 m 3 /day. In the 45 Bakken Formation wells studied, the average initial oil production rate was 21.4 m 3 /day (135 barrels of oil per day (bopd)), and the average initial EUR was 6700 m 3. At the time of refrac completion, the average oil production rate had dropped to 2.4 m 3 /day (15.1 bopd). After refrac operations, the average initial oil production rate was 16.8 m 3 /day (105.7 bopd) and the average EUR was about m 3. On average, refrac completions restored the oil production rate to 79% of the initial oil production rate, with a median value of 80%, and appear to have added m 3 to the average EUR. Refrac completion costs were only reported in the frac database for the two most recently completed wells, in 2015 and 2016; the average cost was C$587,000 per well. Saskatchewan Geological Survey 4 Summary of Investigations 2017, Volume 1

5 Figure 4 Graph showing initial oil production rates (m 3 /day = cubic metres per day) relative to production rates after refrac as a percentage of the initial production rate, for the 45 wells in the Bakken Formation in southeast Saskatchewan that were examined in this study. Several of the refrac completions in the Bakken Formation restored the well s oil production rate to 50% or more of the initial production rate. 4. Viking Formation Analysis The wells used for the analysis of refrac performance in the Viking Formation were initially completed between 2010 and 2012, and the majority were refractured between 2012 and 2015 (Figure 5). The one exception was well 191/ W3/00; 10A128, which was fractured in July 2010 and refractured in December of the same year. Figure 5 Histogram showing the timing of frac completions (initial and refrac) between 2010 and 2015 in the 11 wells studied in the Viking Formation in west-central Saskatchewan. Saskatchewan Geological Survey 5 Summary of Investigations 2017, Volume 1

6 Although refrac operations did improve the oil production rate in all 11 wells in the Viking Formation, Figure 6 shows that the improvement was only marginal in two wells (those plotting near the blue equal rate line in Figure 6), and the post-refrac rate failed to exceed 2 m 3 /day in two other wells. In addition, no refrac completion improved the oil production rate by more than 4 m 3 /day. Initial oil production rates in the Viking Formation are considerably lower than those in the Bakken Formation, however, so high oil production rates in the Viking after a refrac operation would not be expected. Incremental increases in EUR can also be largely attributed to the refrac completion, as there is no EOR recovery scheme contributing to the EUR increase in the Viking wells studied. Figure 6 Graph comparing the oil production rates (m 3 /day = cubic metres per day) prior to and after refrac for 11 wells in the Viking Formation in west-central Saskatchewan. All wells exhibited a higher production rate after refrac operations, as they all plot above the blue equal rate line; however, the increase was only marginal in a few wells. In the 11 wells in the Viking Formation that had refrac operations performed and where oil production data was available, the average initial oil production rate was 7.4 m 3 /day (47 bopd) and the average initial EUR was 2000 m 3. At the time of refrac completion, the average oil production rate had dropped to 0.8 m 3 /day (5.0 bopd). After refrac operations, the average initial oil production rate was 2.6 m 3 /day (16.3 bopd) and the average EUR was 3100 m 3. On average, refrac completions restored the oil production rate to 35% of the initial oil production rate, with a median value of 26%, and appear to have added 1100 m 3 to the average EUR. Figure 7 illustrates that the difference between the average and median production value for Viking Formation wells after refrac can be attributed to two wells in particular (wells 191/ W3/00; 10A128 and 191/ W3/00;10H135, the two blue diamonds plotting on and above the 100% line on Figure 7), which both showed a higher percentage increase in their oil production rate relative to the other Viking Formation wells, caused largely by lower initial oil production rates from the initial completion. Seven of the 11 wells studied failed to restore the oil production rate to even 40% of the initial rate. Refrac completion costs were reported in the frac database for 2 of 11 wells; the average value was C$379,000. Saskatchewan Geological Survey 6 Summary of Investigations 2017, Volume 1

7 Figure 7 Graph showing initial oil production rates (m 3 /day = cubic metres per day) relative to production rates after refrac as a percentage of the initial production rate, for wells in the Viking Formation in west-central Saskatchewan. Refrac completions failed to restore production to 40% of initial oil production rate in 7 of the 11 wells studied. 5. Viking Formation Refracs in 2016 During the follow-up research carried out in early 2017, it was discovered that seven additional wells in the Viking Formation, covering a very small area within a Viking pool between Townships 29 and 30 and Ranges 26W3 and 27W3, had refrac workovers performed during 2016 that targeted the same interval as the initial completions. Because these wells were all within the same small area (in contrast to the original 11 Viking Formation wells studied, which were much more widely scattered), and because they had all had the exact same frac and refrac operations performed on them, they were considered an ideal test group for evaluating the success of refrac completions in this part of the Viking Formation. Although there wasn t sufficient production data to accurately calculate EUR after the refrac, it was possible to compare pre- and post-refrac oil production rates. The results of this comparison showed that, although the average oil production rate on these wells after refrac was over 100% of the initial oil production rate, the initial oil production rates were lower than the average initial oil production rate in the 11 Viking Formation wells targeted in the research in the fall of Further investigation showed that all seven wells were initially completed with the same two-stage completion method, where 84 tonnes of frac fluid was pumped into the rock during each stage, and refractured with a 14-stage completion, where only 12 tonnes of frac fluid was pumped into the rock during each stage. In all seven wells, both completions (initial and refrac) were performed over the same measured depth interval. In Figure 8, these seven additional wells are added to the graph shown in Figure 6, to compare their pre- and postrefrac production rates to those of the 11 Viking Formation wells discussed earlier in this paper. It is clear that the post-refrac production rates for the seven wells refracked in 2016 are better overall than those of the initial 11 wells examined. Saskatchewan Geological Survey 7 Summary of Investigations 2017, Volume 1

8 Figure 8 Graph comparing oil production rates (m 3 /day = cubic metres per day) prior to and after refrac for the 11 wells refracked prior to 2016 (blue diamonds) and the 7 additional wells refracked in 2016 (red squares) within the Viking Formation in west-central Saskatchewan. The majority of the additional wells refracked in 2016 show better oil production rates after refrac than do any of the wells refracked prior to In Figure 9, the Viking Formation wells refracked in 2016 are added to the wells shown in Figure 7, which were refracked prior to Although the later refracs appear to have outperformed the earlier refracs in terms of percentage of initial oil production rate, the initial completion rates for the wells refracked in 2016 are poor in comparison to that of the other wells in the Viking pools. These higher post-refrac oil production rates for the wells refracked in 2016 can be used to suggest that the refrac potential of the Viking Formation is better than the wells refracked prior to 2016 suggest. It is also possible, however, that the improved production after refrac on these seven wells was the result of an improved completion design, implemented on underperforming wells whose initial completion design was far below optimal. To assist in determining whether the improved production rate in these seven wells indicates a better refrac potential for the Viking pools overall, or whether it is simply the result of a better completion design for these seven wells, the initial oil production rates were examined for 82 horizontal offset wells located within the same small area of westcentral Saskatchewan as the seven wells being discussed that were only fractured at the onset of oil production. Figure 10 illustrates that the majority of these offset wells had an initial oil production rate that exceeded 6 m 3 /day, with an average of 8.9 m 3 /day. This is in strong contrast to the initial oil production rates recorded in the well production history for the seven wells refracked in 2016, which were all below 5 m 3 /day and averaged 4.0 m 3 /day. It appears that the lower initial oil production rates and subsequent better-than-average post-refrac rates of these seven wells refracked in 2016 was more likely due to a sub-optimal design for the initial completion that was followed by an improved design in the recompletion, rather than an improved outlook on refrac potential in the Viking Formation in general. That being said, this is a good example of how a sub-optimal initial completion presents a good opportunity for a better refrac design to improve a well s performance. It is also worth noting that the average cost of the seven refrac completions in 2016 was C$306,000. The decrease in cost from the previous refrac operations in the Viking Formation is consistent with the concurrent trend of falling prices within the oilfield services industry, particularly the pressure pumping sector, thereby improving the economics of refrac completions in the Viking Formation. Saskatchewan Geological Survey 8 Summary of Investigations 2017, Volume 1

9 Figure 9 Graph showing initial oil production rates (m 3 /day = cubic metres per day) relative to production rates after refrac as a percentage of the initial production rate (blue diamonds indicate wells with a refrac prior to 2016; red squares indicate wells with a refrac in 2016), for wells in the Viking Formation in west-central Saskatchewan. The figure clearly shows that refracking substantially improved the low initial production rates for the wells refracked in Figure 10 Histogram showing initial production rates for 82 offset wells in the Viking Formation and the seven Viking Formation wells that were refractured in 2016; all wells in this figure are located within the same small pool. It can be seen that all seven of the Viking Formation wells that were refractured in 2016 had initial oil production after the initial frac completion that was well below the average for the area, as indicated by the production rates for the offset wells. Saskatchewan Geological Survey 9 Summary of Investigations 2017, Volume 1

10 6. Conclusions The results of this study indicate that refrac completions appear to be more effective in the Bakken Formation than in the Viking Formation. It is not surprising, therefore, that more wells in the Bakken Formation have been refracked. It also appears that wells with the highest quality reservoir, or those with sub-optimal completion designs, present the best opportunities for refrac success; however, that does not exclude other wells from having improved post-refrac production. The seven Viking Formation wells that had refracs performed in 2016 show how identification of an initial poorly designed completion can lead to production of oil that would otherwise have been left behind, if a betterdesigned completion is performed. Producers have done a better job in recent years in designing and planning frac operations, meaning that fewer suboptimal frac jobs are completed. Increasing the number of stages, and smaller frac spacing, has improved recovery factors, but there are still potential candidate wells for refrac in the Bakken Formation. In the Viking Formation, refrac completion results to date have not been sufficient to justify extensive refrac completions, but improved technology and lower completion costs could allow investment in refracs in this formation to become viable in the future. 7. Acknowledgments The author would like to thank the staff at the Saskatchewan Geological Survey s Subsurface Geological Laboratory, especially Arden Marsh and Dan Kohlruss for their valuable input and discussions related to this paper. 8. References Endress, A. (2015, November): As industry seeks cheaper ways to boost production, horizontal refractures gain prominence; retrieved October 1, 2016, from Drilling Contractor: IHS Inc. (2014): Traditional decline analysis theory; retrieved 2016, from Fekete.com: is_method_theory/traditional_decline_theory.htm Nelson, M. (2017): Evaluating production performance from horizontal well refracs in Saskatchewan; talk presented at the Williston Basin Petroleum Conference, May 2-4, 2017, Regina, Saskatchewan. Saskatchewan Geological Survey 10 Summary of Investigations 2017, Volume 1