Analysis of Data From the Barnett Shale Using Conventional Statistical and Virtual Intelligence Techniques

Abstract

A Barnett shale water-production data set from approximately 11,000 completions was analyzed using conventional statistical techniques. Additionally, a water/hydrocarbon ratio and first-derivative diagnostic-plot technique developed elsewhere for conventional reservoirs was extended to analyze Barnett shale water-production mechanisms. To determine hidden structure in well and production data, self-organizing maps and the k-means algorithm were used to identify clusters in data. A competitive-learning-based network was used to predict the potential for continuous water production from a new well, and a feed-forward neural network was used to predict average water production for wells drilled in Denton and Parker Counties, Texas, of the Barnett shale.Using conventional techniques, we concluded that for wells of the same completion type, location is more important than time of completion or hydraulic-fracturing strategy. Liquid loading has potential to affect vertical more than horizontal wells. Different features were observed in the spreadsheet diagnostic plots for wells in the Barnett shale, and we made a subjective interpretation of these features. We find that 15% of the horizontal and vertical wells drilled in Denton County have a load-water-recovery factor greater than unity. Also, 15 and 35% of the horizontal and vertical wells drilled, respectively, in Parker County have a load-recovery factor greater than unity.The use of both self-organizing maps and the k-means algorithm showed that the data set is divided into two main clusters. The physical properties of these clusters are unknown but interpreted to represent wells with high water throughput and those with low water throughput. Expected misclassification error for the competitive-learning-based tool was approximately 10% for a data set containing both vertical and horizontal wells. The average prediction error for the neural-network tool varied between 10 and 26%, depending on well type and location.Results from this work can be used to mitigate risk of water problems in new Barnett shale wells and predict water issues in other shale plays. Engineers are provided a tool to predict potential for water production in new wells. The method used to develop this tool can be used to solve similar challenges in new and existing shale plays. IntroductionLarge bodies of data relating to fracturing operations in gas shales exist in public databases. These databases also contain production data. In this work, we first examine production figures from the Barnett shale using conventional statistical techniques to identify trends and extract meaningful information from these trends.Second, some work has been conducted on characterizing water-production mechanisms in conventional hydrocarbon reservoirs on the basis of the analysis of water/oil-ratio (WOR) and water/gas-ratio (WGR) data over time (Chan 1995). This technique has not been applied yet to understanding the water-production mechanisms in unconventional gas reservoirs. Therefore, we

Machine learning on field data for hydraulic fracturing design optimization

“…Source Cumulative oil production 6/18 month just after the job [24] 12 months cumulative oil production [25] Average monthly oil production after the job [19] NPV [26] Comparison to modelling [28] Delta of averaged Q oil [29] Pikes in liquid production for 1, 3 and 12 months [30] Break even point (job cost equal to total revenue after the job) [32] • In [26], a procedure was presented to optimize the fracture treatment parameters such as fracture length, volume of proppant and fluids, pump rates, etc. Cost sensitivity study upon well and fracture parameters vs NPV as a maximization criteria is used.…”

Shut-in based production optimization of shale-gas systems

“…Shale-gas wells may produce at these pseudo steady-state rates for some hundred days or even for years, depending on the well and reservoir properties. Along with the slow but steady decline in the pseudo steady-state rates, comes one of the foremost operational challenges of shalegas wells, which is to prevent the state of well liquid loading (Redden, 2012;Sutton et al, 2010;Al Ahmadi et al, 2010;Awoleke and Lane, 2011;Lea and Nickens, 2004;Whitson et al, 2012). This state is reached when the pressure support in the well is insufficient to lift co-produced liquids to the surface, causing accumulation of liquids in the wellbore and thereby increased bottomhole hydrostatic backpressure.…”

“…Gas from other production pads The second reason for many wells is the characteristic production decline profile of shale-gas wells (Awoleke and Lane, 2011;Baihly et al, 2010;Jayakumar et al, 2011;Mayerhofer et al, 2005). This production profile is particularly characteristic for dry shale-gas wells, and is seen by an initial peak rate or plateau level for some time (Jenkins and Boyer, 2008), particularly if the well is chocked back, followed by an early and steep decline with subsequent low pseudo steady-state rates.…”

“…This state can be recognized by erratic and unstable rates, sharp drops in the decline curve (Lea and Nickens, 2004), or stabilization of the production rate at significantly lower levels Dousi et al, 2006), so-called meta-stable rates. Liquid loading is frequently observed at late and intermediate times for shale-gas wells (Al Ahmadi et al, 2010;Awoleke and Lane, 2011;Cipolla et al, 2010;Kravits et al, 2011;Ilk et al, 2008;Mayerhofer et al, 2005;Sutton et al, 2010), and will additionally worsen the already low late-life rates. Furthermore, operating multiple wells under liquid loading is generally undesirable, since it may give rise to sudden liquid slugs at the surface (Lea and Nickens, 2004) and thereby very unstable rates.…”

Cognitive Data-Driven Proxy Modeling for Performance Forecasting of Waterflooding Process

“…In petroleum engineering an extensive variety of neural network applications can be found [16][17][18][19], particularly in the areas of: reservoir characterization or property prediction [20][21][22][23], classification [19], proxy for recovery performance prediction [24,25], history matching [26], and design or optimization of production operations and well trajectory [27][28][29][30][31][32][33]. In particular, neural networks have been utilized in recent years as a proxy model to predict heavy oil recoveries [34][35][36][37][38][39], to perform EOR (enhanced oil recovery) screening [40][41][42]to characterize reservoir properties in unconventional plays [43], and to evaluate performance of a CO 2 sequestration process [44].…”