Browse Prior Art Database

Classification of Proppant Distribution for Stimulation Treatments by K-Means Clustering

IP.com Disclosure Number: IPCOM000240017D
Publication Date: 2014-Dec-22
Document File: 6 page(s) / 468K

Publishing Venue

The IP.com Prior Art Database

Abstract

Hydraulic fracturing stimulation treatments are accomplished by executing a planned schedule of pumping stages which are defined by a purpose, fluids, proppants, and liquid and dry additives. These stages are quantified by clean volume of each fluid, slurry pumping rate, proppant concentration, and additive concentrations. Design of the stimulation job consists of selecting these fluids, proppants, and additives, and setting fluid volumes, rates, and proppant and additive concentrations to achieve an optimal job. A statistical model can be used to measure the effectiveness of fracturing designs based on these and other input parameters.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 38% of the total text.

Page 01 of 6

Classification of Proppant Distribution for Stimulation Treatments by K-Means Clustering

Abstract

Hydraulic fracturing stimulation treatments are accomplished by executing a planned schedule of pumping stages which are defined by a purpose, fluids, proppants, and liquid and dry additives. These stages are quantified by clean volume of each fluid, slurry pumping rate, proppant concentration, and additive concentrations. Design of the stimulation job consists of selecting these fluids, proppants, and additives, and setting fluid volumes, rates, and proppant and additive concentrations to achieve an optimal job. A statistical model can be used to measure the effectiveness of fracturing designs based on these and other input parameters.

Introduction

Previous statistical model parameters associated with proppant have included proppant type(s), sieve size(s), total mass, volume fraction, and maximum and average concentration for a treatment or over all treatments on a well, and did not provide granularity within the design of the pumping schedule. Design of proppant distribution within in the pumping schedule (i.e., proppant concentration for each sub-stage of the pumping schedule) is usually performed as part of the design process based on planned distributions only. The planned did not attempt to statistically measure the impact of proppant distribution on production. This analysis provides a parameter which can be used via statistical modeling to improve well production through optimizing the actual mass of proppant and fluid volumes at multiple proppant concentrations within the pumping schedule of fracturing treatment(s).

Description

Requirements. This method uses data from the pumping schedules of previously executed fracturing (frac) treatments and a novel use of the statistical analysis technique known as K- means clustering to classify proppant distribution within a pumping schedule. This classification can then be used both in the analysis of previous treatments as well as in the design of upcoming treatments. A source of stimulation job data is needed which contains the material selections (fluids and proppants) and the actual fluid volumes, proppant masses, and slurry rates for each pumping stage of the pumping schedules for a statistically relevant number of stimulation jobs.

Methodology.

1. The actual total proppant mass pumped at each of the designed concentration ranges (e.g.
0.5 lb/gal, 1.0 lb/gal, 1.5 lb/gal, etc.) is summed from post job data files for every well in the dataset of interest.

2. The percentage (fractional proportion) of proppant mass within each concentration range of the total proppant mass for the well is calculated as shown in Figure 1.

1


Page 02 of 6

Figure 1. Table of Proppant Mass and Percentage of Total Mass at Concentration Ranges

3. Proppant mass percentages within each range are aggregated for all wells in the study area. For visualization these might be plotted on a line chart with a separate...