News and events

Back to News & Events
Posted: 27 Jan, 2014

Atmos LDS betters the leak detection systems cited in the DOT Leak Detection Study

Executive Summary

The vast majority of the operators’ events cited in the Leak Detection Study – DTPH56-11-D-000001 [1] that reported leaks in the period between 1 January 2010 and 7 July 2012 had no CPM leak detection systems (LDS) or only very basic LDS. Gas operators who use no CPM for leak detection accounted for 43% of the events, instead gas operators rely on rate of change monitoring in SCADA.

The majority of the few liquid pipelines with more sophisticated CPM systems used Real Time Transient Models (RTTM), many of which failed to detect the leaks. The report blames the failures on a recurring theme of false alarms and the inability of the cited CPM systems to maintain sensitivity during transient operations. Some of these companies are now installing Atmos Pipe to improve their chances of detecting and locating a leak.

The other methods, including RTTM failed because they issue false leak alarms or automatically reduce sensitivity or suspend leak detection during transient operations. Frequent false leak alarms destroy operator confidence. This happened when the CPM analysts did not believe the CPM that alarmed the largest leak addressed in the report.

In stark contrast, the patented statistical algorithms allow Atmos Pipe alone to maintain a low false leak alarm rate and not reduce the sensitivity during transient operations. High reliability and high sensitivity, particularly during transients gave Atmos Pipe a 100% success rate in detecting real leaks, including the four recent leaks. On January 23, 2014 Atmos Pipe detected a real leak on a copper concentrate pipeline, and the pipeline was shut down immediately, minimizing the damage from the leak.

Atmos Pipe has never failed to detect a real leak larger than the quoted minimum detectable leak size. In the period from 7 July 2012 to 27 January 2014, four pipelines fitted with Atmos Pipe did suffer leaks, all of which were detected and alarmed by Atmos Pipe, allowing the operators to shut-down operations immediately. In addition, three pipeline operating companies that had major leaks since this DOT report are now installing the Atmos Pipe system.

Key Study Findings

The study finds that it is important that dependable leak detection systems are used to promptly identify when a leak has occurred so that appropriate response actions are initiated quickly. The swiftness of these actions can help reduce the damage. The study infers that many hazardous liquid operators are deploying some form of LDS but the incident reports reviewed from the PHMSA incident database suggest this may not be the case and that many of the CPMs deployed were not suitable for the task.

The study finds that leak alarms are rarely black-and-white or on/off situations. The length of time to take actions on a leak alarm depends on the operator’s confidence in the CPM and that frequent false leak alarms destroy this confidence. Pipeline operators with Atmos CPMs have succeeded in minimizing the consequence of a leak because the high reliability of the system gave the operators the confidence to shut-down the pipelines when Atmos Pipe issued a leak alarm.

The report suggests that the reliability will be improved by a combination of technologies – utilizing multiple, redundant and independent LDS. Atmos chose this path several years ago with the development of other CPM methods, including negative pressure/rarefaction wave leak detection and RTTM leak detection to compliment Atmos Pipe. We can combine these to accelerate the detection time and increase the sensitivity provided by the highly reliable Atmos Pipe system. The report states that LDS are engineered systems. This means that precisely the same technology, applied to two different pipelines, can have very different results. This supports the Atmos Philosophy of having a team of professional and experienced engineers who understand pipeline operations and have the knowledge to recommend the best solution for each pipeline.

The report finds that many technologies were adopted from other process industries that involve fluid movement, including storage, chemical process industries, water distribution, and the nuclear industry. Atmos Pipe was conceived and developed by Atmos’ Managing Director, Dr. Jun Zhang, in Shell Research specifically to provide leak detection during both transient and steady-state pipeline operations in the Oil and Gas industry, to minimize false alarms and maintain leak size sensitivity during transients. These are all of the features recommended by this DOT study.

The report shows that only a very small percentage of all these pipelines (14%) had any CPM installed. The 1,337 incident pipelines are divided to 766, 295 and 276 for hazardous liquids, gas transmission and gas distribution respectively. For 795 onshore hazardous liquid pipeline incident reports CPM was reported as in place for 192 (25%), not in place for 286 (38%) and not reported for 281 (37%). None of the gas pipelines are reported to have had CPM.

Thus, in reality, only 14% of all the events occurred on pipelines with CPM leak detection (Figure 1). Many of the CPM systems on these lines were poor performers: rate of change monitoring, volume balance or RTTM systems.


Figure 1: Only 14% of all lines had any CPM at all (Gas lines had no CPM)


The number of leak events detected by CPM systems was 28 (4%). The reason for this low performance is that many of the CPM systems were not functional at the time of the leak. As an example, see the chart below from the study [1] for the incidents of liquid gathering lines.


The study infers that many hazardous liquid pipeline operators are deploying some form of LDS but the incident reports reviewed from the PHMSA incident database suggest this may not be the case. The study also infers that many natural and other gas operators rely on SCADA as an LDS.

Within these CPM systems deployed, the study breaks down the technology as follows:


These results show that the CPM systems that were installed on the pipelines studied were mainly Real Time Transient Models and the simple SCADA based systems such as volume balance, pressure/flow monitoring. These systems failed to detect pipeline leaks larger than the suggested minimum threshold of 1% of flow rate

Indeed it is obvious that the inability of the Real Time Transient Models, volume balance and Rate of Change technologies to detect leaks during transients without false leak alarms often inhibits them from detecting even large leaks and ruptures.


The vast majority of the pipelines that experienced a leak in the study had either no CPM leak detection or inferior leak detection technologies. The simple SCADA based systems are installed because it is cheap and easy to do so. The more complex RTTM based LDS are older systems that have been thoroughly discredited by this report. These systems lacked the key features recommended by the study, namely the ability to maintain good sensitivity during transient operations and a low false alarm rate that gives the operator the confidence to shut down the pipeline when a leak alarm is issued. The report states that as the first trigger for mitigation, a leak detection system that prioritizes rapid detection and high sensitivity is particularly valuable. Atmos Pipe provides the desired feature as pipeline operators recently had the confidence to shut down four different pipelines, minimizing the consequences of the leaks. Atmos Pipe has met the stringent German requirement; and can be combined with any one of several other Atmos CPM solutions to satisfy the study recommendation for two or more leak detection systems on a pipeline.

Following the leak events cited in the study, many pipeline operators are installing Atmos CPM systems, assuring high reliability, high sensitivity and fast leak detection.


1. U.S. Department of Transportation, Pipeline and Hazardous Materials Safety Administration, Final Report No. 12-173, “Leak Detection Study – DTPH56-11-D-000001”, Dr. David Shaw, Dr.Martin Phillips, Ron Baker, EduardoMunoz, Hamood Rehman, Carol Gibson, Christine Mayernik, December 10, 2012

Web Analytics