This paper documents a data-mining study of well, hydraulic fracture treatment, and production parameters for horizontal wells in the north Texas Barnett Shale play. In this study, the authors have analyzed well and production data from more than 13,400 producing Barnett wells. A subsample of over 3,300 horizontal wells was characterized with respect to detailed well architecture data such as drift direction and angle, lateral length, perforations, etc. The study uses Geographical Information System pattern-recognition techniques in conjunction with more traditional statistical techniques to interpret hidden trends in otherwise scattered data sets. This work provides a case study in the practical use of data-mining techniques to address questions of best practices in Shale Gas reservoirs. It is made possible because the availability and quality of public domain well and production data has increased significantly in the past few years. Simple cross plotting of production data against well and treatment variables normally leads to broad scattering of results. This study takes advantage of the largest, richest well and production data set available from the gas shales and identifies key lessons-learned. Relevant trends, such as the impact of toe up versus flat versus toe down, horizontal well length, and drift angle variability on gas production rate are presented. This work is significant in that it shows that the application of practical data-mining methods to a large Shale Gas data set can result in learning key lessons that may not be apparent when working with small data sets. This work is also significant in the use of merged reservoir quality proxies, well architecture data, completion data, and stimulation data, against which production results are placed in geographical perspective for improved interpretation.

This paper documents follow-on work to an original data mining study of horizontal wells in the North Texas Barnett Shale play. In this study, the authors have analyzed well and production data beginning with over 15,000 producing Barnett wells. Study wells were grouped for similar map-based reservoir properties, normalized for the effects of well architecture, and normalized for production. The study used statistical and data mining techniques plus Geographical Information System pattern-recognition techniques to aid in interpretation of production result trends. The principal focus of the follow-on study is in the areas of completion and hydraulic fracturing. It is intended to demonstrate lessons-learned from careful analysis of large volumes of data, some of which can only be examined by proxy. Data sets from public and proprietary data sources were collected and merged into a common database. Data elements were subjected to statistical quality control methods in order to minimize the potential for inaccurate data to be included in the analysis. Short-term production proxies were studied and decided upon in order to adequately compare well productivity values over many years of horizontal well drilling across the Fort Worth Basin. Previous work showed that the application of practical data mining methods to a large Shale Gas data set resulted in learning key lessons that were not apparent from small data sets. The previous work also showed that certain popular beliefs did not hold up to examination of the available data. Similar to past work, this work is significant in the use of merged reservoir quality proxies, well architecture data, completion data, and stimulation data, against which production results are placed in geographical perspective for improved interpretation. This work is also significant in the application of advanced data mining methods beyond routine statistical work.

It has been observed that pumping a mini-frac prior to a TSO Frac-pack can impact the effectiveness of the frac-pack. The calculated fluid loss parameters determined in the diagnostic test are often not valid for the main fracture design due to the residual effect of the mini-frac and/or step-rate fluids. A technique will be presented in this paper which allows the calculated fluid loss parameters from the diagnostic test to be used reliably without excessive waiting time for the reservoir to recover to its original leak off characteristics. Fifty plus treatments were evaluated to develop a technique which makes this possible. The use of this technique resulted in a significant change in the success of the TSO designed treatments - success being a TSO type pressure increase while pumping. The success rate to achieve designed TSO, by incorporating the changes described in the paper, was increased over 20 percent with a reduction in time between diagnostic tests and the main frac. In the wells associated with this paper, a borate-crosslinked fluid was used for a mini-frac treatment followed by a step-rate test prior to the main proppant laden frac-pack. The fluid was designed with minimal polymer loading for the well conditions. The resulting mini-frac tests had low fluid efficiencies. It was originally thought that using this fluid, followed by injection of a linear step-rate fluid, would minimize the changes observed in fluid efficiency between the diagnostic test and the main fracture treatment. However, the effect of the diagnostic test on fluid leak off still resulted in less than desired TSO predictability. A technique of adding a pH control additive into the final portion of the step-rate test fluid was found to successfully allow the use of the observed diagnostic test results, honoring the efficiency from the mini-frac test. The quantity and placement of the pH control agent in the step-rate protocol were dependent upon well conditions. The waiting time between the diagnostic test and the main treatment was reduced since a positive, controlled change was applied. The optimum pH reduction for the desired effect was determined in the laboratory and designed into each treatment depending upon well conditions.

the distribution of production results for wells completed with fracturing sleeves and packers, plug and perforated, or complex completions to determine whether differences in productivity existed and needed to be factored into completion recommendations.Trends examined in the project in addition to completion type included treatment parameters such as fracturing fluid types and quantities, proppant types and quantities, number of completion stages and stage lengths, perforation cluster spacing and length, and calculated perforation friction drop. All parameters analyzed were examined for statistical importance.This work is significant in that it shows that the application of practical data-mining methods to an intermediate-size Shale Oil (light, tight oil) well data set can result in learning key lessons that may not be apparent when working with small data sets. This work is significant in the use of merged reservoir quality proxies, well architecture data, completion data, and stimulation data, against which production results are placed in geographical perspective of the Bakken Formation for improved interpretation. The work is also significant in that it may be used to allow selection of completion systems on the basis of completion time and cost balanced against concerns over differences in well production impact of one system over another, e.g., frac sleeves versus plug and perf type and complex completion systems.

No abstract