Summer 2010

Using Long-Term Conductivity Tests to Differentiate Proppants

Are We Talking Apples to Apples?

With the growth of proppant choices, it has become increasingly difficult to determine if one proppant is superior to another or whether they are in fact interchangeable. This analysis becomes particularly important at times in which proppant suppliers are struggling to meet demand. In an effort to address this situation, service and production companies have found themselves relying more on product performance data generated at an independent testing facility. The “long-term conductivity test” has often been viewed as the ultimate measure of a product’s performance properties. This test combines the elements of temperature, stress and time in an attempt to give a more realistic assessment of a proppant’s conductivity/permeability. As this trend has become more widespread, proppant suppliers have increased the use of the third-party testing facilities to generate the data that is necessary to define their product’s capabilities.

With an increasing amount of proppant data carrying the label “Long-term Conductivity Data Performed Following Industry Accepted Procedures,” has it become easier to differentiate between products or to determine if they are interchangeable? Unfortunately, the answer is “only if you are comparing apples to apples.” The value of such a comparison is directly related to your understanding of the factors that can influence (if not control) the product data you are examining. There are three major factors that can significantly influence the results of a long-term conductivity test.

These factors include:

  • Size of the Database
  • Sample Gradation/Sieve Distribution
  • Test Duration

In the following paragraphs, each of the factors will be described and their potential impact estimated.

Size of the Database

The data plotted in Figure 1 shows the results of nine runs from the same sample of product. In this case, the product was a precured resin coated sand. The sample was retested in an effort to establish the repeatability of the long-term conductivity test procedure. All samples were sent in using a code designation to ensure the outside lab (performing the test procedure) would not have a preconceived expectation of the test results. The test results show a ±25% spread (compared to the average of the runs) in conductivity values for the identical samples. However, if a proppant supplier based their reported data on a single test run, it could mean overestimating this key performance parameter by a significant amount (potentially as much as 50%).

Figure 1

Figure 2 shows the spread (and average) for a similar compilation of data for a curable resin coated sand. In this case, the data spread (from three runs) equates to ±12% from the curve representing the average.

Figure 2

Sample Gradation

Sample gradation or sieve distribution plays a critical role in the conductivity results from long-term tests. In general, samples containing a higher percentage of coarse material (within the specified screen designation) will yield a higher conductivity value. Figure 3 illustrates the affect that particle size distribution has on fracture conductivity (for a precured resin coated sand). At 6000 psi closure stress, there is a > 50% increase in conductivity (equates to an additional 850 md-ft), as the 20/30 cut in the test sample is increased 10% (from 65 to 75%).

Figure 4 illustrates (that in a 20/40 mesh curable coated sand) a similar trend exists. At 8000 psi closure stress, the observed conductivity doubles (from 900 to 1800 md-ft) as the 20/30 cut increases from 30% to a maximum of 75%.

Having established the sensitivity (of results from a long-term conductivity test) to sieve distribution, it is important to know not only what was tested but also to be sure that it is representative of product delivered to the wellsite. This is a primary reason that Hexion’s Oilfield Technology Group (OTG) initiated its practice of generating a detailed sieve analysis on every truckload shipment of product from its plants and transloads. The critical nature of sieve analysis is also the primary reason OTG developed PropQA™, an internal web-based database used to verify proppant sieve distribution and generate a certificate of analysis for each load.

Test Duration

A final factor (that can significantly affect conductivity test results) is deviating from so-called “baseline” procedures. Years of consortium testing have yielded a long-term test procedure that is based on:

  1. Proppant concentration
  2. Test temperature
  3. Test fluid (usually 2% KCl)
  4. Range of stress conditions (usually starting at 2000 psi and increasing in 2000 psi increments to maximum conditions)
  5. Duration of time spent at each stress level (each level is tested for 50 hours consecutively)

To get a true comparison of two products, the long-term conductivity data must be generated using an identical test procedure. To accomplish this result all five of the aforementioned parameters must be accounted for in the test procedure. This is particularly true with items 4 and 5 above.

Figure 5 illustrates the impact of deviating from accepted test protocol in an example in which the conductivity at 8000 psi is the desired data point. In this example, identical test samples were tested using the same values for items 1-3, but in one case, the test was started at 6000 psi instead of the 2000 psi called for in a “baseline” procedure. In this comparison, the desired 8000 psi data point varied from 750 to 1350 md-ft (an increase of 80% for the shortened test).

The results of a long-term conductivity test are very much dependent on the test temperature, and in particular, the stress history and length of time the proppant samples are subjected to test conditions.

In summary, to properly utilize long-term conductivity data to compare proppant technologies (either to select a proppant or to determine if proppants are interchangeable), you need to be sure that you understand the potential effects of differences in the way the data is collected and presented. It is also critical that the data being examined is representative of the product being delivered to the wellsite. Careful attention must be given to the data being compared.

This attention would include:

  • Determining if your data represents averaged values (from multiple test runs) or data from a single test
  • Determining if the test results were generated from a sample with a sieve distribution that represents what will be delivered to the wellsite
  • Determining if the data to be compared was generated following an identical set of test conditions and procedures

For more information on proppant testing and Hexion's quality focus, contact your Hexion representative or e-mail otg.customerservice@hexion.com.

  © 2010 Momentive   |   Privacy Policy   |   Email    |   Unsubscribe