Dataset of Interactive Threat Matrix–Driven System Dynamics for Assessing CO₂ Pipeline Risk and Resilience Policy Measures
Description
Published dataset that are generated throughout the study titled, " Interactive Threat Matrix–Driven System Dynamics for Assessing CO₂ Pipeline Risk and Resilience Policy Measures". Being an established leader in the carbon management sector with one seventh of world’s active carbon management projects, Canada plans to improvise on technologically focused approaches to grow it’s CO2 capture capacity from current 4.4 Mt per year to 16.3 Mt year by 2030 and scaling up by nearly 200 times by 2050 to reach the net-zero emission target. The published study has the potential to be developed into an interpretable quantitative risk assessment framework through further extensions.
Files
Steps to reproduce
In the sphere of decarbonization, a comprehensive CO2 (Carbon dioxide) pipeline risk analysis framework is crucial for resilient long-term operations. Canadian Standards Association (CSA) updated regulations Z662:23 requires operators and regulatory bodies to develop quantitative risk assessment methodologies with probability and consequence analysis. Thus, this study is aimed at determining risk probability of CO2 pipelines across Canada, while developing a simulation tool for consecutive policy analysis. The process involves integration of threat matrix from real gas pipeline incident dataset, long-short term memory (LSTM) model and system dynamics (SD) simulation. The developed methodology is a simplified risk probability analysis tool for CO₂ pipelines, with extensible features to incorporate further consequences and economic analysis. To determine the historical incident rates, gas pipeline incident dataset has been extracted from published repository by CER. The interactive threat matrix formulation is divided into two interconnected sections depicting risk factors of level 1 and 2. Once pathway relationship and weightage strength are developed for risks among level 2 factors, this relationship is then converted to causality and the weighted strength are converted to weight co-efficient of the risk path. The developed methodology for this study incorporates only LSTM with historical incident data, for generating synthetic incident frequency, which is later used as a lookup function in built SD model. The built SD model is then simulated from 2025 until 2055 to check the baseline risk probability. In addition to structural validity, to validate the model and to check numerical model’s boundary adequacy, calibration with previously published PHMSA dataset by Xi et al., (2023) and sensitivity analysis is performed. Once the model is validated, consecutive policy measures are being simulated to check the long-term risk likelihood alongside CO2 pipeline integrity. The shared dataset contains raw judgement matrix, normalization, formulization of 38 X 38 matrix (main model) and 16 x 16 matrix (calibration Xi et al., 2023), extracting factorial weights as well as score of each primary and secondary subfactors. These dataset are then used as input in the built system dynamics (SD) model.
Institutions
- University of Regina