The State of Leak Detection

Thanks to the evolution of the computer, leak detection for liquid pipelines is a technology that has developed significantly in the past couple of decades. A lot of claims are made about the effectiveness of leak detection systems (LDSs) by vendors, and some regulators have mandated use of this technology in some form or another in the operation of many pipeline systems around the world. Some operators have adopted the use of this technology as a best practice for their systems whether required by a regulatory body or not.

There are several questions that could be asked about the application of this technology:

• Has the use of this technology really improved pipeline operation and integrity?
• Does the technology provide truly reliable results?
• For what it does, can it be done better in other ways?
• For cost and benefits, does it add adequate additional value?
• Does the pipeline operating staff embrace it, tolerate it, or prefer getting rid of it?

The American Petroleum Institute (API) has published standards and recommended practices pertaining to LDS applications on liquid pipelines. API 1130 Computational Pipeline Monitoring has been adopted by the U.S. DOT-OPS as a reference document in 49 CFR Part 195 in the design, operation, and maintenance of an LDS. To date, U.S. DOT-OPS does not mandate the use of LDS technology but does require compliance with API 1130 if a system is implemented on a pipeline regulated in the United States. Some states that regulate pipelines for oil spill prevention and response do mandate use of LDS technology.

In Alaska, for instance, pipeline leak detection requirements are specified in 18 AAC 75.055(a) that require that a crude oil transmission pipeline must be equipped with a LDS capable of promptly detecting a leak. These requirements include:

• If technically feasible, the continuous capability to detect a daily discharge equal to no more than 1% of daily throughput.
• Flow verification through an accounting method, at least once every 24 hours.
• For a remote pipeline not otherwise directly accessible, weekly surveillance, unless precluded by safety or weather conditions.

There have been a number of industry and regulatory assessments of leak detection technology since the late 1990s. One conclusion that can readily be drawn from the vendor information is that there are various highly reliable means of determining a leak, and its location, in a shut-in pipeline. An operating pipeline is a much bigger challenge.

In addition, the Minerals Management Service published Worldwide Assessment of Industry Leak Detection Capabilities for Single and Multiphase Pipelines in August 2006. This report is currently available at www.mms.gov/tarprojects/409.htm. Dr. Stuart L. Scott and Dr. M. A. Barrufet of Texas A&M University directed this work. This report concludes that there are somewhat reliable means of assessing leak detection in single-phase pipelines but not for multi-phase pipelines.

LDS technology has gaps, but the study results do show how some of the holes can be plugged. Cost was not a primary consideration in the gap analysis, since the focus was the current and potential capabilities of LDS technology, at whatever cost. The performance of computational pipeline monitoring (CPM) systems will likely continue to be limited by pipeline instrumentation and pipeline conditions such as slack flow and shutdown. There are fundamental limitations in CPM technology that prevent their use for detecting very small leaks; e.g., less than 1% of daily nominal flow.

External leak detection systems may extend sensitivities but the availability of such systems and industry experience has been very limited. Perm-anent or intermittent slack line flow degrades LDS performance. Non-leak alarms are a persistent problem that have increased the cost of ownership and reduced confidence in the reliability of systems. Thus, the key question seems to be: If the best leak detection that can be achieved is 1% of daily throughput, is there a better way to determine a 1,000-bbl leak in a 100,000-bpd pipeline than LDS?