Infrastructure protectionEffective, inexpensive method to detect natural gas pipeline leaks
Major leaks from oil and gas pipelines have led to home evacuations, explosions, millions of dollars in lawsuit payouts, and valuable natural resources escaping into the air, ground and water. Scientists say they have developed a new software-based method that finds leaks even when they are small, which could help prevent serious incidents — and save money for customers and industry.
Major leaks from oil and gas pipelines have led to home evacuations, explosions, millions of dollars in lawsuit payouts, and valuable natural resources escaping into the air, ground and water. In a report in ACS’ journal Industrial & Engineering Chemistry Research, however, scientists say they have developed a new software-based method that finds leaks even when they are small, which could help prevent serious incidents — and save money for customers and industry.
An ACS release reports that Gary Valtinson and Miguel Bagajewicz note that using pipelines to move oil, gas, and even water from one place to another is highly effective, for the most part. Serious, costly problems, however, arise when pipes break. Existing methods for detecting leaks are limited.
Hardware-based approaches using special instrumentation are expensive and complicated, and software-based systems don’t model pressure drops in pipelines correctly. This leads to a lot of errors, particularly for gas pipelines. Valitonson and Bagajewicz set out to fix this flaw.
The researchers developed a method that compares pressure and flow rate measurements from a pipeline with mathematical models that can accurately predict what the pressure and flow rate should be. Their technique successfully detected small leaks and reduced errors from 21 percent to 3 percent when compared to existing software. The researchers estimated that their method would have saved millions of dollars more than other leak-detection methods.
— Read more in “Leak Detection in Gas Pipelines Using Accurate Hydraulic Models,” Industrial & Engineering Chemistry Research (6 October 2014)(DOI: 10.1021/ie501322g)