The pipeline industry is investigating the implementation of airborne leak detection surveys that utilise various sensor technologies in an effort to identify pipeline leaks that may be lower than the detection limits of traditional CPM (computational pipeline monitoring) methods.
These methods rely on changes in flow parameters, such as pressure, volumetric or mass flow, to identify potential pipeline leaks. To date, airborne leak detection systems have generally been used to monitor gas pipelines. The overarching goal of the current work being done by industry is to determine which of these technologies is best suited to detect small fluid leaks from pipelines carrying liquid hydrocarbon products.
Natural gas vs liquid pipeline leaks
Traditionally, liquid pipeline leaks are detected through CPM systems which compare computer models of the flow parameters to measured flow characteristics such as pressure, temperature, density and flow rate to identify irregularities that might be associated with leaks. These systems are typically limited to detecting leaks greater than approximately one per cent of the total pipeline flow, due to the sensitivity of the instrumentation used and the ability of the models to predict the flow parameters over the full range of operating conditions in the pipeline.
To complement these systems, operators have used a variety of pipeline leak detection technologies, ranging from in situ sensors such as fibre-optic sensing cables, to visual aerial patrols and aircraft mounted sensors that periodically monitor a pipeline right-of-way (ROW). The primary focus of aircraft mounted leak detection sensors has been on natural gas pipelines. Unintended releases from natural gas pipelines produce methane plumes, which are easier to detect due to the distinct absorption spectrum characteristics of methane, the short time for the gas to reach the surface of the ground, and the Joule-Thompson cooling effect produced at the leak location as the gas expands.
Leaks from liquid pipelines, however, produce very different signatures, with the leaked liquid following the path of highest permeability, until it ultimately reaches surface. The goal is to detect these pipeline leaks before the liquid reaches surface. Sensor manufacturers have been adapting and developing sensors used in other applications for this purpose.
Airborne leak detection sensors
Through a comprehensive market survey, it was determined that the majority of airborne leak detection sensors that are commercially available fall into the following categories:
Laser Absorption: This technology is based on the principle that each individual gas species absorbs electromagnetic radiation (EMR) at different wavelengths. A laser is tuned to the wavelength of the gas of interest. When the laser light passes through a gas plume of interest, the gas absorbs some of the beam’s energy, thereby weakening the reflected signal. By measuring this energy change, the sensor can determine the presence and concentration of gas along the path of the laser.
Gas Filter Correlation Radiometry: A band pass filter is used to select the excitation frequency of the gas of interest from ambient light (i.e. sunlight). This filtered EMR is passed through two gas cells. The correlation cell contains the gas of interest, whereas the reference cell is evacuated. When sunlight passes through a plume of interest and then into the cells, the reference signal from the evacuated cell will decrease from the energy absorbed in the plume, but the correlation cell signal will remain unchanged. The difference between the two cells’ signals allows the sensor to identify the presence of the gas of interest.
Thermal Imaging: This method uses a camera to measure the EMR emitted in the infrared range (i.e. thermal energy) of all objects in the field of view. The leaking products are usually at different temperatures than the surrounding air and soil due to the thermal signature of leaking fluids, evaporative processes that cool the leaked product, or frictional heating as the product passes through the small hole or crack in the pipeline. This technology is often used to detect fugitive gas emissions from oil and gas facilities. This method’s biggest advantage is that it can optically locate the leak.
Flame Ionisation: As the sensor passes through a gas plume, a portion of the gas vapour is collected and burned in a hydrogen flame. This process generates ions, which are proportional to the number of carbon atoms present in the gas sample or the hydrocarbon concentration.
Visual Inspection: High definition recording equipment, such as still and/or video cameras, are often used to record the condition of the pipeline right-of-way. The images are traditionally reviewed by an operator trained in visual inspection, but new automated methods are being developed to identify signs of leaks such as disturbance of the ground surface or leak-induced stress on vegetation overlying the pipeline.
To increase the accuracy of their surveys, and to avoid false positives, most technology vendors use a combination of the above technologies. However, there are currently no standards to verify the ability of these technologies to detect sub-surface liquid hydrocarbon leaks. Therefore, it is imperative that these sensors be independently tested to evaluate their performance on simulated subsurface liquid pipeline leaks and to determine their operational lower detection limits, which will assist pipeline operators in selecting appropriate technologies.
To evaluate these technologies effectively, C-FER Technologies and Alberta Innovates – Technology Futures (AITF), with support from Enbridge, TransCanada, Kinder Morgan (Canada) and government agencies, carried out laboratory testing and numerical modelling to gain a better understanding of possible leakage indicators produced from a liquid pipeline leak. These indicators include temperature differentials, volatile organic compound (VOC) concentrations, and induced ground displacement. The numerical models focused upon subsurface liquid hydrocarbon leak behaviour and provided information regarding the dispersion and evolution of VOCs migrating from the leakage point through the soil to the ground surface.
Once the VOCs breached the surface, the data acquired were used to develop atmospheric dispersion models to predict the VOC gas plume characteristics at several elevations within the atmosphere in the vicinity of the fluid leak location. A variety of wind conditions and altitudes were analysed to fully characterise the gas plume. In addition to VOC models, AITF also numerically modelled the fluid dispersion, and associated temperature distribution and geomechanical effects resulting from a subsurface liquid release.
The “÷ELDER’ approach
In an effort to validate these numerical models, AITF completed soil column tests and C-FER completed full-scale testing using their External Leak Detection Experimental Research (ELDER) test apparatus located in Edmonton, Alberta, Canada. The ELDER, as shown in Figure A, tests various in-situ sensors for external pipeline leak detection (i.e. fibre-optic cables, vapour sensing tubes, among others). A full-scale 24 inch pipe with various independently controlled leak ports was buried inside the ELDER and instrumented to monitor the migration of the leaked oil through the soil. Efforts were taken to ensure that the soil conditions, such as water content and compaction, were generally representative of real world conditions. Liquid hydrocarbons may be released at any specified leak port, while the leak parameters such as flow rate, pressure, and temperature are controlled.
To benchmark the VOC data obtained from the numerical models, C-FER installed isolation flux chambers on the ELDER’s soil surface as shown in Figure B. Total VOC concentrations within each flux chamber were monitored throughout the test duration (approximately four days) to determine vapour breakthrough times and expected concentrations at the soil surface. The VOC flux rates measured in the lab tests agreed well with predictions of the numerical models. Other leak characteristics such as temperature distribution and ground surface heave will be assessed in future tests.
Understanding liquid pipeline leak signatures
The lab testing and numerical modelling completed to date have served as a preliminary analysis to better understand liquid pipeline leak signatures. These test conditions have been idealised in that they do not include environmental conditions such as soil heterogeneities, snow cover, or precipitation, which might influence the performance of airborne leak detection sensors. It is paramount that these sensors be tested in the field, under less than ideal conditions.
Therefore, the ultimate goal of this project is to deploy and test these sensors in the field using multiple simulated leak stations. The leak locations and parameters would be unknown to the vendor during testing, similar to a real pipeline ROW survey. Controlled field tests may determine the performance of each sensor, based on the overall detection rate of the leaks and lower detection limits, and identify the range of application for each technology.
Continuous improvements are being made in pipeline leak detection technologies. Some vendors have already adapted and miniaturised their leak detection technology for use on unmanned aerial vehicles (UAVs). In the future, UAVs could be utilised to economically survey pipeline ROWs and complete surveys in high risk or otherwise inaccessible areas. These technologies will be developed and proven through ongoing numerical modelling, lab testing, and field trials to allow pipeline operators to detect liquid hydrocarbon leaks before they reach surface.