A Generalized Prognosis and Health Monitoring Platform for Network-Based Systems Using Hybrid Causal Logic Modeling Approach

Over the past few years, the authors have developed a system-level Prognosis and Health Monitoring (PHM) modeling framework for gas pipeline integrity management. The original framework (Chalgham et al. 2020), based on Hybrid Causal Logic (HCL) with embedded Dynamic Bayesian Networks (DBNs), provided a comprehensive methodology and software platform for real-time risk assessment, failure prediction, and optimal mitigation planning. While effective for gas pipelines, the framework's domain-specific structure limited its applicability to other complex infrastructure systems. In this paper, we introduce a generalized PHM framework designed to model and monitor a broader class of systems that involve pipelines and similar transmission networks such as electric power grids, water distribution systems, and transportation infrastructure. This new implementation retains the HCL and DBN-based architecture but introduces configurable abstractions to enhance flexibility and reusability. Systems are modeled as networks composed of segments and station points, each may experience failure through diverse mechanisms. Segment failures can be modelled using templates based on fault trees or event sequence diagrams, with failure mechanisms (e.g., corrosion, wear, overloading) modeled as modular components linked to customizable Bayesian belief networks. These modules support evidence injection and model updating based on periodic observations, facilitating more accurate and adaptive risk quantification. The framework is implemented in the NSIM software platform (Network System Integrity Management Platform) that supports interactive configuration and real-time health monitoring. Its flexibility makes it suitable for a wide range of applications beyond oil and gas pipelines, including smart grid reliability, water infrastructure resilience, and logistics network risk management. This work represents a step toward a unified, data-informed decisionsupport system for integrity management across complex engineered systems.

Keywords: Network Systems, Prognosis and Health Monitoring, Hybrid Causal Logic, Dynamic Bayesian Networks, Risk Assessment, Transmission Infrastructure, Consequence Modeling, Geospatial Integration.