ADVERTISEMENT

Prescriptive Analytics Aids Completion Optimization in Unconventionals

This paper discusses a prescriptive analytics framework to optimize completions in the Permian Basin. The methodology involves data processing, ingestion into databases, and data cleansing; application of automated machine learning (AutoML) to generate an accurate machine-learning model; and numerical optimization of decision parameters to minimize an economic objective. Constructing a Pareto front enabled decision makers to select a strategy that minimized cost without sacrificing too much of the initial 12-month oil production.

Overview

A typical way to assess the performance of an oil well is to compute the total cost per barrel of oil produced at ultimate recovery. However, in cash-constrained operating environments, including many unconventional plays, other measures, such as $/bbl of oil produced in the first year, may also be used. Because of the intensive nature of running detailed physics-based models to optimize decision parameters related to wellbore placement and well completion, a pure simulation-based optimization strategy typically is not feasible. The complete paper describes a data-driven machine-learning model that was selected to predict the oil production profile of a new well in an unconventional play because of its efficiency and its ability to directly learn from abundant historical data.

Data analytics are typically applied in three phases:

  • Descriptive analytics, which makes sense of data
  • Predictive analytics, which helps build an accurate machine-learning model to forecast a performance measure or an output on the basis of input parameters
  • Prescriptive analytics, which helps to develop recommendations to improve performance

Input parameters to the machine-learning model presented in the complete paper included system parameters (e.g., well location and trajectory, existence and type of artificial lift) and decision parameters (e.g., number of stages, amount of stimulation material). The authors used an optimization algorithm to explore systematically the space of decision parameters. Training data included drilling, completion, and production data from previous wells in the same formation. Initial input features were extracted and prepared on the basis of domain expertise. The AutoML process searched automatically among various machine-learning algorithms (e.g., neural networks, random forest) to find the best algorithm and the best associated hyper-parameters.

A multiobjective optimization process was implemented to optimize the 12‑month oil production (Qoil) and the completion cost divided by Qoil simultaneously. The methodology was applied to a real case in the Permian Basin, in collaboration with Diamondback Energy. The results demonstrated that the two objectives were conflicting. The construction of the Pareto front, which showed a set of optimal solutions, provided a visual tradeoff for decision makers and enabled them to select a strategy that minimized cost without sacrificing much of the initial 12-month oil production. The predictive model was then used to optimize the completion design of new wells to either maximize the cumulative oil produced, or minimize the completion cost per barrel of oil produced, in the first year.

This article, written by JPT Technology Editor Judy Feder, contains highlights of paper SPE 195311, “Prescriptive Analytics for Completion Optimization in Unconventional Resources,” by Mehrdad Gharib Shirangi, SPE, Yagna Oruganti, and Thomas Wilson, Baker Hughes, et al., prepared for the 2019 SPE Western Regional Meeting, San Jose, California, 23–26 April. The paper has not been peer reviewed.
...
This article is reserved for SPE members and JPT subscribers.
If you would like to continue reading,
please Sign In, JOIN SPE or Subscribe to JPT

Prescriptive Analytics Aids Completion Optimization in Unconventionals

01 April 2020

Volume: 72 | Issue: 4

No editorial available

ADVERTISEMENT


STAY CONNECTED

Don't miss out on the latest technology delivered to your email weekly.  Sign up for the JPT newsletter.  If you are not logged in, you will receive a confirmation email that you will need to click on to confirm you want to receive the newsletter.

 

ADVERTISEMENT

ADVERTISEMENT

ADVERTISEMENT

ADVERTISEMENT