April 2014, Vol. 241 No. 4
Features
Computing Risk: Leak Detection Strategies
Kalamazoo River, Mayflower, Grand Marsh, Bonga Field, Little Buffalo – all places that suffered pipeline spills in the last several years. The recurrence of significant incidents has many countries spending more and more time investigating the pipeline industry’s safety practices. This resurgence comes at a time when oil and natural gas initiatives around the world are putting pressure on new pipeline development.
This scenario creates a significant risk/reward situation for the midstream industry. On one hand, continued safety failures from the midstream industry risks lost development permits and increased government regulation, alongside billions in fines, fees and lost revenue. On the other hand, companies that vigorously take on safety and put protocols and systems into place that meet industry best practices will be rewarded with contracts and development rights to capitalize on a booming market.
At the core of this discussion is effective leak detection. While many different factors affect leak detection, including training, control room protocol and controller overload, the fundamental tool in any pipeline operator’s toolbox is its computational pipeline monitoring (CPM) software. But just as pipeline operations vary widely across the globe, CPM systems vary in complexity and function. Selecting the right solution depends on that same risk/reward relationship and pinpointing where the two intersect for your company.
Defining CPMs
American Petroleum Institute (API) Publication 1130 defines CPM systems as those that are internally based, use field sensor outputs that monitor internal pipeline parameters, such as pressure, temperature, viscosity, density, flow rate, product sonic velocity and product interface locations. The systems use real-time information from the field to estimate volume, mass or the hydraulic behavior of the product being transported. Calculated results are then compared to field references to identify an unexpected anomaly that might signal product loss, whether an accidental rupture or an unlawful tap.
Different CPM systems are designed to monitor different parameters and use varying logic complexities to detect product loss. The common CPM systems in use today are:
• Line balance: measures the imbalance between the receipt and delivery meters. This is your basic “meters in, meters out” comparison and is the least sensitive to anomaly capture as volume in the pipe is not part of the balance.
• Volume balance: similar to the line balance but also includes a limited pipe volume adjustment. An overall pipeline density is calculated based on pipeline boundary pressure and temperature.
• Modified volume balance: similar to volume balance except a different volume adjustment is used for each product in the pipeline. This takes into consideration where the density changes are in the pipeline.
• Pressure/flow rate modeling: essentially what a controller does by nature – looking for unexplained large drops in pressure or flow – but an automated application monitors for these anomalies to ensure large changes are not missed.
• Acoustic/negative pressure wave: detects a leak by quickly sensing a pressure wave and using those detections to triangulate the location of the event/leaks.
• Statistical leak detection: pressure and flow inputs that define the perimeter of the pipeline are statistically evaluated in real time for the presence of patterns associated with a leak.
• Real-time transient model: all the fluid dynamic characteristics are modeled, including line pack, slack, shut-in and transients, under all pipeline flow conditions. This is a detailed configuration with fast calculations and the ability to model hydrocarbons in any phase.
Factors To Consider When Selecting CPM
The most important objective in selecting a leak detection system is that the chosen solution must assist the controller in identifying a leak event quickly enough that mitigates the safety and environmental risk to a level that is acceptable for the stakeholders in a particular pipeline. At the same time, it must meet the operator’s overall business objective and risk threshold. This includes the potential value of product lost, the cost of clean-up and regulatory fines, detriment to surrounding environments, reputation damage and potential impact on future projects.
The multitude of choices for leak detection systems is necessitated by the vast array of pipelines in use. Therefore, a comprehensive analysis is necessary to identify which CPM technologies and methods are best suited for the pipeline assets in question. A simple A-to-B pipeline route will have less complex operations than a pipeline with many active route connections and elevation changes, multiple receipt and delivery points, and reversible flow. As the complexity increases, the CPM to be applied needs to be flexible enough to handle all operational scenarios in the pipeline’s operation.
High-Risk Factors To Consider
Determining at what point business objectives and acceptable risk intersect relies on a thorough review of pipeline characteristics. Assessing any challenges to effective leak detection, as well as the risk level associated with a leak of a given size are inherently tied to pipeline type, terrain and location.
When analyzing which leak detection systems to employ, the question of whether the pipelines being monitored are located in or close to high consequence areas (HCAs) must be answered. HCAs are defined as areas where a pipeline leak will have a significant effect on people, property, the environment or all three. Whether through regulation, local authority or community requirements and expectations, HCAs typically demand higher levels of leak detection capability and sensitivity to mitigate the higher risk of significant consequences from a leak.
Other factors present particular challenges that need to be considered when selecting the CPM that meets the standards set by an operator’s threshold for risk. These include the types of products flowing through the pipeline, pipeline construction and geography, control room factors or controller related-factors.
Some of these challenges require a wide variety of internal pipeline parameters to be measured and analyzed by the CPM and controller. For example, batched systems, transient flow conditions, elevation-induced hydraulic variations and slack line static conditions all create widely varying hydraulic models the CPM must account for when analyzing the difference between normal conditions and a potential leak. On the other hand, outages in the communications network or high employee turnover also require operators to select CPM systems that are able to overcome these challenges.
If the pipeline network has locations close to or in HCAs – or if it is a large and complex network with diverse operating parameters and products – more than one leak detection system might be warranted. A multi-tiered approach might be necessary to satisfy specific requirements. Of course, it is always important to remember that other factors, such as controller overload and fatigue, will rely on the implementation of control room management, human machine interface and training best practices.
Industry Tools And Standards
Operators are not on their own in deciding what detection capabilities are acceptable. Although an American standard, the API 1149 is a best practice evaluation around the world. This is a theoretical analysis of a given leak detection system’s ability to find a leak of a given size, based on the specifications of a given pipeline. It weighs the time to detection against the size of the leak.
While not always necessary to employ techniques that achieve the lowest theoretical capability, API 1149 provides a measure to weigh against when analyzing the cost of leak detection systems against the risk of undetected leaks. Of equal, if not greater, importance is that the API 1149 calculations can assist pipeline operators in determining the benefit of specific pipeline infrastructure enhancements on their leak detection capability. For example, what increase in leak detection sensitivity can be achieved by adding, replacing, or upgrading instrumentation on all, or sections of a pipeline?
The Association of Oil Pipelines (AOPL) has created a Leak Detection Rupture Monitoring project as part of its “Pipeline Leadership Initiative” to develop additional strategies for continued improvements in leak detection.
A key area of improvement the initiative has identified is execution of the “3 R’s”: Recognition, Response and Reporting. The association has developed performance standards for the industry to follow in this area, with the target goal of 30 minutes for complete “3 R” execution:
• Recognize: Recognize a 50% flow change within five minutes.
• Response: A response to the rupture must come within five minutes of the recognition. Responses can be tailored to particular situations.
• Report: A report must come within 20 minutes.
While these may seem like basic standards, meeting each is important to the industry’s goal to be great in the execution of the “3 R’s” of rupture detection.
Now What?
Having assessed all the elements of the leak detection capability to risk threshold relationship, selecting the right CPM or mix of CPMs requires consideration of two primary categories:
• Volume balance CPMs can offer effective, cost-efficient solutions for operators with low risk/low consequence systems, such as pipelines generally operating in steady state conditions and non-HCAs. They can also be appropriately sensitive for pipelines with shorter segments or as an interim solution while a higher fidelity model is being deployed or upgraded. Because volume balance CPMs monitor basic flow statistics for anomalies or imbalances, they are a strong solution for simple pipeline systems and basic detection capabilities over sections of pipe. However, they lack the sophistication to effectively monitor complex pipeline systems or specify the geographic location or estimated size of a given leak.
• Real-time transient CPMs provide the best platform to support highly effective leak detection for the risk-averse pipeline operator, complex systems or those systems in close proximity to an HCA. Because it incorporates all the fluid dynamic characteristics, it is adaptable to steady state, transient, slack line and shut-in conditions, and can account for drag-reducing agents. The advanced modeling also allows for accurate estimation of both the location of a detected leak and its size. Its accuracy allows for detection of leaks that are well below 1% of flow while demonstrating significantly lower false alarm rates compared to other CPMs.
While environmental and public safety will always be chief concerns of the pipeline industry, it is this last point of continued permitting and development that strikes at the industry bottom line. Damaging, costly and public leak detection failures like Kalamazoo River and Bonga Field have long-lasting effects on the industry in the form of increased regulatory scrutiny and public aversion to pipeline development. Therefore, selecting the right leak detection method is not just an exercise in risk management for individual operators; it is an exercise in reputation management for an entire industry.
Comments