FCTM Esope

S1.3- Bringing HPC-Driven Engineering Analysis to the Cloud

Oct 6, 2021 | 12:00 PM - 12:30 PM

Scénario

Description

In the petrochemical and refining industries various forms of corrosion pose constant threats to mechanical integrity, safety, and profitability. To better understand and mitigate the risks associated with such aging energy infrastructure, strong technical analysis capabilities, combined with optimized life-cycle decision making, are critical. Today, this is accomplished via complex simulations and data analysis. To this end, High-Performance-Computing (HPC) software is leveraged and integrated into simple and intuitive cloud-based web tools, to lower the barrier for new users, increase the ease of access for experienced users, and ultimately lead to smarter life-cycle decisions in the field. Each life-cycle decision has associated costs and benefits that must be fully considered to determine the best overall decision strategy for asset management. Figuring out these trade-off decisions for complex industrial applications (and across an entire facility) requires High Performance Computing (HPC) simulations. In this paper, we discuss HPC accelerated approaches for improved life-cycle decision optimization of plant-wide fixed-equipment and piping subject to corrosion. The HPC accelerated approaches discussed herein utilize Bayesian Decision Networks. Additional HPC accelerated topics discussed include the analysis of large-scale pigging data for corrosion detection and local thinning area segmentation. The challenges here are multiscale in nature as the data from a pigging scan results in kilometers of data, while local thinning areas may only be a few inches in diameter. Once these local thinning areas are found, they can be sent to a cloud-hosted FEA solver to perform local thinning burst pressure analysis, e.g. B31G. Here, HPC parallelization expedites computation times and by utilizing the cloud, it reduces the compute infrastructure cost as well. Closing the loop, these results are then used as inputs to the life-cycle decision optimization framework for managing plant-wide fixed equipment and piping subject to corrosion.

Presented by