Rod-Pump Optimization: Big Data Helps Accelerate and Improve Operations

Abstract

This article, written by Special Publications Editor Adam Wilson, contains highlights of paper SPE 181216, “Proactive Rod-Pump Optimization: Leveraging Big Data To Accelerate and Improve Operations,” by Tyler Palmer, SPE, and Mark Turland, SPE, Denbury Resources, prepared for the 2016 SPE North American Artificial Lift Conference and Exhibition, The Woodlands, Texas, USA, 25–27 October. The paper has not been peer reviewed. This paper presents how a US onshore operator took a three-step approach to optimize more than 100 rod-pump wells. The approach involved data consolidation, automated work flows, and interactive data visualization. This approach led to increased unit run times, decreased unit cycling, improved production and equipment surveillance, and increased staff productivity. The ultimate goal was to increase profitability by decreasing lifting costs and increasing operating efficiency. Introduction The processes and tools described in this paper cover a subset of approximately 125 wells in eastern Montana and western North Dakota, but they have been designed to be applicable and scalable to any fields that use rod-pump artificial-lift systems with supervisory control and data acquisition (SCADA). Simple modifications can be made to the tools and processes for wells that do not have SCADA capabilities. While optimization efforts and best practices have been implemented for the subject rod-pump systems during the past 6 decades, many opportunities remain to create additional value. Empirical knowledge from field personnel serves as the basis for the analytical model. Categorizing and quantifying the observations made by the field personnel is critical to developing any analytical model involving oil and gas operations. On the basis of feedback from field personnel and engineers, the following areas had the most potential for improvement: data consolidation, automated work flows, and data visualization. The data-consolidation issue stems from data being located in multiple file locations, sometimes being stored in nontabular formats and initially lacking the necessary unique identifiers for mapping between databases. Automated work flows were essentially non-existent; wells were analyzed individually using deterministic, static data. Data were previously visualized in multiple locations but never integrated into a single, interactive visualization tool. The opportunities to maximize asset value led to the development and implementation of the rod-pump optimization tool (RPOT). The RPOT is a data-visualization tool that generates a single recommended optimization action (ROA) for each well being analyzed. The ROA logic calculates the optimal amount of fluid for a well to produce on the basis of its inflow, while accounting for surface and subsurface equipment constraints. General examples of ROAs include slowing wells down by making a specific sheave adjustment, speeding wells up by a specified strokes/minute (spm) amount, upsizing the downhole pump to a specific pump size, or upsizing or converting to a different artificial-lift system. If in accurate or incomplete data are brought into the database, the ROA specifies the data source that needs to be quality-checked.