Using an Expert System to Diagnose Formation Damage Mechanisms and Design Stimulation Treatments for Gas Storage Wells

Abstract

In gas storage wells, many different types of formation damage can occur that dramatically curtail injection and withdrawal rates. Some of these damage mechanisms are similar to producing wells (mud/cement damage during drilling, completion problems, etc.); however, some types of damage are more specific to gas injection and storage (bacterial growth, contaminants, etc.). All these different damage mechanisms require different stimulation treatment methods and fluids to increase injectivity and deliverability. However, diagnosing the correct type of formation damage is not a simple task. The process requires gathering specific types of data and interpreting the results. Correct diagnosis of the actual damage mechanism(s) and design of the appropriate treatment requires expertise and experience. Over the years, numerous studies have been performed on formation damage, and a vast amount of experience has been obtained on the design of stimulation methods and fluids to remove damage. However, this knowledge and experience is not always available or accessible to engineers dealing with stimulation design, especially with the unique types of problems associated with gas storage reservoirs. This paper describes a comprehensive computer model designed for gas storage wells to help engineers diagnose formation damage and select the best stimulation treatment. The model combines domain knowledge bases with the best available expertise using fuzzy logic and expert system technologies. After diagnosing the most likely formation damage mechanism(s), based upon input data, the program will select the best treatment method and recommend treatment fluids and additives for the stimulation. Introduction The gas storage industry uses more than 400 reservoirs and 15,000 wells to store and withdraw gas, making it a significant contributor to gas supply in the United States. It is typically observed that many gas storage wells show a loss of deliverability each year caused by a variety of damage mechanisms. In fact, the American Gas Association estimates a deliverability loss of approximately 3 Bscf/D each year. Tens of millions of dollars are spent each year to recover or replace lost deliverability, including both the drilling of new wells and the stimulation/remediation of existing wells. Recently, the Gas Research Institute (GRI), working with gas storage operators, conducted a research project to address the damage mechanisms occurring in storage wells.1 In this project, multiple wells were evaluated in more than 10 gas storage fields. Numerous downhole tests were performed in those wells, including well tests to quantify skin damage. Downhole samples of fluids, wellbore solids, and core were also obtained to characterize the damage mechanisms. As a result of this study, damage was identified in many of the study wells. The types of damage were primarily:bacteriainorganic precipitates, including iron compounds, salts, calcium carbonate, and barium sulfatehydrocarbons, organic residues and production chemicalsparticulate plugging Several other damage mechanisms were identified, including completion/stimulation fluid effects, relative permeability effects, sanding, and mechanical obstructions. GRI initiated this project to assist gas storage operators in diagnosing these potential damage mechanisms and designing optimal treatments to remove the damage. The goal of the project (and the subject of this paper) was to develop an expert system that could capture the knowledge and experience from previous studies, and the industry in general, so that these processes could be automated. Technical Approach On the basis of our experience with developing a comprehensive expert system,3–6 we outlined the following technical approach for the project.

Disaggregation and discretization methods for formation damage estimation in oil and gas fields: an overview

“…However, it does not specify the phenomena of formation damage caused by thermodynamic effects such as pressure, temperature, and stress changes.On the other hand, expert systems have been developed to simulate complex decision making fed by multivariate systems and expert opinions [43]. In this sense, Xiong et al [44] developed a computational model, designed for gas storage wells, to disaggregate formation damage and select the best stimulation treatment. The model combines expert systems, neural networks, and algorithms together with knowledge, experience, results of numerous well tests, technical literature, and historical cases to identify the damage mechanisms present in a well.…”

“…As a result of the model, the possible types of damage that are presented in the well are obtained based on input data, categorized from higher to lesser importance, select the possible treatment, fluids, and additive for intervention that could be used by any engineer with ease. Even so, it has the disadvantage of not considering the damage caused by geo-mechanical phenomena, being designed only for gas storage wells, and relying on excessive information and input data [44]. On the other hand, Garrouch et al [45] develop an expert system for the identification and quantification of formation damage by clay swelling, fines migration and deposition of organic and inorganic scales divided into two modules, one for clay swelling, deposition of inorganic and organic scales and another for quantification of the sludge formed by asphaltenes based on empirical, technical knowledge and results found in the literature.…”

A review of fuzzy logic applications in petroleum exploration, production and distribution operations

“…Fuzzy logic applications in E&P operations have been reported in different areas such as engineering design and control (Nikravesh et al 1997;Sengul and Bekkousha 2002;Mohaghegh et al 2005;Widarsono et al 2005;Cao et al 2006;Taheri 2008;Weiss et al 2001;Zarei et.al. 2008;Murillo et al 2009) and also in operations of production facilities, systems and optimal well operations (Rivera and Farabee 1994;Dumans 1995;Xiong et al 2001;Alimonti and Falcone 2004;Garrouch and Lababidi 2005)…”

An expert decision support system for sandstone acidizing design

“…The damage degree is then evaluated based on the results from the pressure test. However, this method is not sufficient to tell the damage radius, degree, and impact of each cause (Xiong and Holditch 1995;Xiong et al 2001). Since these factors determine acidizing parameters such as acid fluid concentration, formula, and applicable range, the developed system adopts a quantitative method to simulate the permeability and damage radius via modeling, which greatly accelerates the precise diagnosis process and provides reliable support to the design of acidizing.…”