Risk Management Module
The Risk Management module leverages industry best practices to deliver on-demand quantitative risk analysis and reporting. Models were designed and validated by our partner, C-FER Technologies, and utilize integrity data ingested and aligned using machine learning, contained within the Asset Integrity for Pipelines (AIP, formerly CIM) platform.
The Risk Management module assesses risk from two different perspectives:
- Structural Reliability models calculate the probability of failure (POF) using inline inspection (ILI) data for the following threats: external corrosion, internal corrosion, stress corrosion cracking, and mechanical damage (also known as third-party damage).
- Historical Data models calculate the POF using historical incident data from PHMSA for the following threats: equipment failure, incorrect operations, manufacturing defects, fabrication defects, and weather and outside forces.
Solutions and Features
- A centralized data platform that ingests and aligns the various data sets required for advanced risk analysis, i.e., inline inspections, external corrosion control surveys, incident reports, pipeline data from geographic information systems (GIS), and repair data.
- By leveraging the power and scalability of Microsoft’s Azure platform, large amounts of data can be effectively utilized for quantitative and probabilistic analyses much faster than by using Excel or related third-party tools.
- Industry-leading risk models were designed and validated by CFER Technologies for all major pipeline threats outlined in ASME B31.8S.
- Model types defined as best practice by PHMSA’s risk modeling working group, i.e., probabilistic and quantitative, have supported the development of CSA Z662 in defining pipeline risk and reliability.
- Models have been uniquely trained and validated on a simulation of 3.5 million mile-years of pipeline operation.
- The reliability models use Monte-Carlo simulations to perform probabilistic calculations, which inform the operator regarding the possible spectrum of POF values.
- Models calibrated to industry failure rates from the PHMSA incident database take the guesswork out of the unknown and align results with reality.
- A rolling mile risk score allows the user to pinpoint the highest risk location, what threats are most prevalent, and plan risk reduction measures accordingly.
- Risk results are displayed in pre-built dashboards and reports to provide transparency to all levels within the company.
Methodology
Structural Reliability-Based Models
Structural reliability theory requires loads and resistances to be modeled as probabilistic variables, which are then used to calculate the probability of failure (PoF).
Resistance is calculated using pipe geometry and material properties such as yield strength or fracture toughness. The load is determined by the forces applied to the pipe. This includes the internal force induced by the pressure of the product and external forces or impacts. Reliability-based models are used for the following threats:
- External Corrosion, Internal Corrosion - This model simulates individual defects or a distribution of defects. The characterization of these defects is obtained from ILI measurements of both internal and external corrosion features.
- Stress Corrosion Cracking - This model addresses two distinct forms of SCC: non-classical SCC, characterized by transgranular cracking in association with a near-neutral pH, and classic SCC, characterized by intergranular cracking in association with an alkaline (high pH) electrolyte.
- Third Party/Mechanical Damage - Mechanical damage incidents are typically caused by construction or excavation equipment working in the area of the pipeline. The failure rate is calculated by multiplying the number of hits per unit length of pipeline by the probability of failure given a hit. The number of hits is calculated using a fault tree analysis approach. In contrast, the probability of failure per hit is calculated using structural reliability models, which are based on the recognition that both the load applied to the pipe and the pipe's capacity to withstand the applied load are uncertain quantities.
Historical Based Models
Historical-based models rely on data from the PHMSA incident database to build statistical predictive models.
- Equipment Failure - Pipeline failure associated with mechanical equipment is typically the result of gasket, O-ring, seal, or packing failure, as well as the malfunction of pressure relief or regulator valves.
- Incorrect Operational Procedures - Pipeline failures resulting from incorrect operation or maintenance actions are addressed by this model. Failures due to incorrect operation are primarily attributed to errors in actions by control room personnel. In contrast, failures due to incorrect maintenance are primarily attributed to errors in actions by maintenance personnel.
- Manufacturing defects - Pipeline failure associated with manufacturing defects is primarily attributable to fatigue growth of seam weld or pipe body defects. Seam weld failures tend to occur in susceptible seam welds (i.e., seams with significant starter defects) that undergo significant stress fluctuations due to pressure variations and/or external loads. Pipe body defects (e.g., hard spots) can also serve as initiation sites for planar defects that subsequently grow to failure under fatigue. Additionally, although not strictly a manufacturing-related issue, the fatigue-induced failure of pipe body fittings and attachment welds is also a possibility.
- Weather-Related and Outside Forces – This threat addresses geotechnical hazards involving progressive or sudden ground movement, as well as exposure and loss of support (e.g., river scour). The model requires a third-party assessment of the probability of an event and the probability of failure given the event.
- Welding and fabrication-related threats - Pipeline failure associated with welding and fabrication-related defects is typically the result of failures in pipe body fittings and attachments, girth welds and couplings, or wrinkle bends. A time-independent historical-based model can be used to characterize this threat. For these defects, a common driving mechanism is secondary stresses induced by differential settlement. The geotechnical and seismic hazard models address failure due to significant progressive or sudden ground movement events. Failure due to cyclic pressure loading is addressed in the manufacturing defect fatigue model.
One Integrated Solution for Integrity and Risk
- Address your most failure-prone pipelines in the areas of highest impact.
- Calculate the probability of failure using real integrity data, allowing for a direct connection between integrity actions and risk reduction.
- Receive advanced insights compiled from all of your integrity-related data to drive risk management decisions.
- View risk at the enterprise, pipeline, and dynamic section levels.
The Risk Management module breaks down silos within teams (risk and integrity) as well as data, providing a centralized platform for ingesting, aligning and standardizing integrity data for on-demand risk analysis and reporting. By centralizing data characterizing the pipe, environmental conditions, maintenance activities, and operating conditions, the platform can leverage its proven machine learning-based alignment, along with scalable quantitative and probabilistic risk modeling, to extract deep insights. This enables operators to utilize real-time risk analysis in their day-to-day integrity decisions, ushering in a new era of pipeline safety that transitions from reactive to proactive and ultimately predictive, thereby enhancing the opportunity to reduce pipeline failures.
Risk Management Phase 2 Now Available (as of 3.45 Release), which provides:
- Probability of failure for each of the ASME B31.8S CSA Z662 Annex O pipeline threats for rupture, large leak, and small leak.
- Dashboards and reporting to compare the probability of failure to user-defined thresholds and industry-calculated failure rates.
Was this article helpful?
That’s Great!
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
Feedback sent
We appreciate your effort and will try to fix the article