Abstract

Coiled tubing (CT) erosion can occur during CT fracturing operations. Resulting wall loss and fatigue can limit CT life and prevent safe wellsite operations. An artificial neural network (ANN) has been successfully developed for accurately predicting wall loss resulting from erosion.

This paper presents a case history in which ANN technology was used to successfully manage tubing strings during CT fracturing operations. During these operations, wall loss affects CT pressure ratings, tensile strength, and fatigue, all of which are critical performance parameters used for determining CT life and identifying a safe operating envelope. An ANN can predict erosional wall loss and quantify critical performance parameters for specific applications.

Introduction

Hydraulic fracture-stimulation treatments performed through CT provide a cost-effective method of stimulating wells in which producing intervals have multiple stringers. This technique is being successfully applied in many areas and is being used on a daily basis in the shallow gas fields of southern Alberta. Fracturing operations in these shallow gas fields typically require a 2 3 /8 -in. (0.203-in.) QT-900 tubing string that is 2,625 to 2,789 ft (800 to 850 m) in length. The zones targeted for stimulation are located at depths of 984 to 2,362 ft (300 to 720 m). Typical treatments involve three to eight stages in which a total of 110,231 to 220,462 lb (50 to 100 tonnes) of proppant is pumped. In addition, most treatments involve the use of a gas-assist with either carbon dioxide or nitrogen added to the stimulation fluid.

Operating within the safe working envelope of the CT is critical to the success of these operations. Wall loss from erosion affects operational CT parameters, and previously identified patterns of erosion 1 required a better understanding of this phenomenon. Three of the 35 strings used during CT fracturing operations performed in the year 2000 were evaluated for wall loss. Magnetic wall-loss detection was performed with Hall-effects sensors. The strings were measured at 82-ft (25-m) intervals and at 30° intervals around the circumference of the string. The data from one string clearly indicate maximum erosion loss on the outer radius of the pipe and minimal wear on the inner radius of the pipe ( Table 1). Wall thickness for all three strings is plotted in Figs. 1, 2, and 3. The wall loss from the tests demonstrates a repetitive pattern similar to those previously identified. 1 Pipe thickness affects critical CT operating parameters, including fatigue, internal pressure capacity, and tensile/compressive loading. A method for accurately predicting erosion could help operators use traditional CT pipe-management simulators to effectively manage CT fracturing strings.

Using an ANN to Develop a Wall-Loss Model

Several technical papers have been written about the development of ANNs. Neural nets have been given many different names, including black boxes, empirical models, universal approximators, and parallel models. The basic function of an ANN is to map data from one multidimensional space to another. Fig. 4 shows an example of a simple ANN with three inputs, two hidden layers with two neurons in each layer, and one output. The example has 12 unknown weights. The weights must be determined so that the inputs can be properly mapped to the corresponding outputs.

ANNs are constructed of neurons organized in layers. An ANN can consist of a single layer or multiple layers, and each layer consists of one or more neurons. Artificial neurons receive, consider, and calculate input from other sources. This information is then displayed by an algorithmic process or transfer function. The number of layers, neurons in each layer, and connecting transfer functions depend on the problem that must be solved.

This content is only available via PDF.
You can access this article if you purchase or spend a download.