NOTE: The sponsor of this content may contact you with more information on this topic. Click here to opt out from sharing your email address with this sponsor. (This link will not unsubscribe you from any other BIC email list).
Semi-quantitative RBI has delivered value for decades, but its limits are clear with most of the value realized by users. Incremental technology improvements to tweak risk calculations may slightly change inspection priorities but may not result in improved reliability or significant risk reduction. The next generation of RBI is quantitative.
Pinnacle’s Quantitative Reliability Optimization (QRO) is a fully quantitative model that combines the best of First Principles modeling with the best of data science to deliver precision forecasting to pinpoint failures in minutes. By integrating decades of inspection history, real-time operating data, and advanced statistical modeling, QRO delivers precise, actionable insights that semi-quantitative methods cannot.
The result:
- Stronger reliability: Accurate, predictive insights that prevent failures before they happen
- Optimized spend: Every inspection, maintenance and operating decision tied directly to risk reduction.
- Actionable precision: Uncertainty is quantified, producing realistic failure predictions. Plans built at the CML level, eliminating guesswork and waste with clear ROI and risk-reduction impact.
- Future-ready operations: Reliability, maintenance, and operations make faster, better-informed decisions with integration of IOW excursions, process deviations, and advanced NDT data for live, dynamic risk management.
- Data integration: CML/TML, NDT, and process data are unified for continuous model updates.
The data and technology is here to deliver the precision, scalability, and reliability impact that semi-quantitative RBI was never built to provide. The future of mechanical integrity is here and it’s quantitative.
Below is a summary from an extensive analysis published in Inspectioneering Journal by Lynne Kaley that takes an in-depth look at RBI’s limitations and the need for change:
RBI was built from opportunity & collaboration
In 1993, the industry was undergoing significant change due to several industry incidents. Recognizing the need for more effective inspection strategies, an API-led Joint Industry Project was initiated to develop a risk-based approach for prioritizing the inspection of fixed equipment. While the group believed that a fully quantitative methodology was beyond reach because of computing limitations (personal computing was in its infancy in the early 90s), there was a belief that it was possible to provide simplified methods for estimating failure rates and consequences of pressure boundary failures in terms of safety, monetary loss, and environmental impact.
Since its release, the recommended practices outlined in API RP 580 and 581 have changed the way inspections are planned and managed, transitioning to the use of overall risk as opposed to condition-based approaches.
Time for the evolution of RBI
Owner-operators who have implemented RBI know that the benefits are realized in the first 5-10 years of implementation. Taking the next significant step in reducing fixed equipment failures is unlikely to be achieved by making minor tweaks and adjustments to the current semi-quantitative methodology.
A successful transition to a new era using a next-generation risk-based approach is required to continue to improve fixed equipment reliability, reduce in-service related failures, and identify and manage risks.
API 581 current limitations
While a quantitative methodology was envisioned from the start, access to data and technology to adequately leverage that data into a fully quantitative model did not exist at the time.
As we progress to the next generation of models, I anticipate a higher level of performance in loss of containment event reduction, maintenance and inspection spend optimization and more confident and aligned decision-making.
There are several aspects of RBI the next generation of fully quantitative models needs to address. Outlined below:
Deterministic vs. statistical modeling
- API 581 semi-quantitative limitations: Single-value inputs oversimplify probability of failure (POF), sometimes causing over- or under-conservatism.
- Fully quantitative reliability: More statistical with uncertainty quantified to more accurately calculate probability of failure automatically. The model should use ranges of specific variables, providing a calculated uncertainty based on a blend of historical data, such as inspection, process, or other variables
CML/TML data management
- API 581 semi-quantitative limitations: Deterministic models underutilize historical corrosion data making it difficult to harness the value of the CML data sets and engineer out the value provided by the old data.
- Fully quantitative reliability: Statistical models can screen poor data, calculate uncertainty related to SME rates and historical thickness data, and generate probabilistic corrosion rate ranges, turning decades of CML data into actionable insights.
Piping modeling
- API 581 semi-quantitative limitations: Managing piping is cumbersome with current piping model complexity over-simplifying the risk calculation. It limits risk accuracy and ROI.
- Fully quantitative reliability: Individual CML-level analysis provides realistic damage rates and failure rate predictions and identifies unexpected damage rates and types.
Inspection recommendations specificity
- API 581 semi-quantitative limitations: Inspection plan recommendations leave room for interpretation, waste and risk as the plan is conducted on user-by-user basis sometimes without complete understanding of the implications of selecting one location over another.
- Fully quantitative reliability: Inspection planning is down to the CML level and quantifies reduction in uncertainty, evaluates cost-benefit, and ensures actions directly mitigate risk.
Process data and integrity operating windows (IOWs)
- API 581 semi-quantitative limitations: IOW models are based on exceedances and an alter is initiated to take an action, like performing an inspection, but the quantitative impact on risk is not accounted for in the risk model.
- Fully quantitative reliability: Linking operational excursions to degradation and risk models allows real-time updates to Probability of Failure and risk, enabling near-instantaneous operational decision-making and action while determining cost-benefit tradeoff of unit operational changes nearly real-time.
Non-thickness data management
- API 581 semi-quantitative limitations: Inspection data management limited to spot UT data. Advanced NDT data are not managed by most platforms and not harnessed in the semi-quantitative model.
- Fully quantitative reliability: High-resolution data sets are integrated into probabilistic models, improving analysis, quantification of uncertainty and the impact on Probability of Failure
Damage Mechanisms
- API 581 semi-quantitative limitations: Guidance is limited for certain damage mechanism and mitigation strategies. The HTHA (API 941) module has not been updated. Additions should include damage assessment guidance as well as mitigation activities that are not limited to nondestructive examination techniques, such as credit for engineering analysis and monitoring methods.
- Fully quantitative reliability: Models cover all possible damage mechanisms (including non-refining industries) with guidance with guidance on risk-reduction activities that include nondestructive and other methods, expanding applicability beyond refining.
The case for sunsetting API 581
Today, it is clear to me that the ongoing efforts to improve the current technology incrementally will not help us take these important next steps. This is the reason I’m recommending sunsetting API 581.
“Sunsetting” is not about abandoning semi-quantitative RBI, but rather, it’s about recognizing the current limitations, realizing that additional improvements in asset reliability are necessary and possible, and challenging ourselves to take the next step-change in industry practices.
Incremental improvements to API 581 will not achieve next-level gains. The industry should focus its time and energy on the advancement of quantitative approaches to initiate the next big step in asset management.
Read the complete analysis in Inspectioneering Journal.
For more information, visit pinnaclereliability.com.
