Reliability-based design has been in extensive use since the early 1960's, and has been applied by operators for well design during the last ten years. Reliability-based casing design quantifies drilling and production loads and tubular performance to define reliability. Different reliabilities are useful based on the consequence of failure: connector leak- 99.9684% reliability (10−3.5 probability of failure), pipe yield- 99.9990% reliability (10−.5 probability of failure), pipe rupture- 99.99997% reliability (10−6.5 probability of failure). The intent of setting rupture reliability requirements very high is to make the likelihood of rupture statistically insignificant. For reference, current offshore structure (jackups) reliability is 99.9684%. Reliability is the quantification of the overlap between stochastic load and stochastic strength distributions for a particular failure mode and load combination. The determination of these stochastic distributions requires load and strength data. Load data comes from actual measurements from wells, while strength data is derived from mechanical and dimensional variations within the material.


A well will experience a multitude of loads during its life. Loads can be intentional, unintentional, of known or unknown magnitude, be of short duration or extend over the life of the well. Each of these loads will affect the reliability of a well quite differently. One easily understands the probability of an intentional load occurring (say pressure test) is 100%. However, the probability of an unintentional load occurring (say tubing leak) is far less than 100%. There is also a large variability in the magnitude of loads. The variability in pressure test loads, from actual to design, is relatively small in comparison to kick loads. The time factor has to be incorporated in the reliability determination since some casing will only be exposed to a particular load for a few days, while other casing strings will be exposed to a particular load for many decades. In addition, the time sequence of when a load will occur will impact the reliability. Some loads will only occur during drilling operations, while others can only occur at the end of the well life.

Tubing and casing performance is dependent on dimensional control and mechanical properties; parameters known to vary with manufacturer and batch, even along the length of a tube. Performance-based acquisition stochastically quantifies tubular performance based on relevant attributes, providing a basis for managing total costs while ensuring performance for critical applications. Recent physics of failure developments, combined with accurate data management, provide the basis for stochastic design with the following benefits:

  1. addressing sources of failure (increase performance),

  2. optimize schedule, performance, and cost objectives,

  3. save suppliers and users time and labor, and

  4. improve user-supplier relationships. In this specific application, the needs were acute and included (1) tubulars required by conventional design methods were unavailable, (2) the operator was interested in fast tracking 25 MM/day gas production, (3) the advanced performance tubulars minimized costs; tubulars typically run 20% of well budget.

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