Integrating Risk in System-Level Asset Management : A Model-Based Systems Engineering Approach

Sammanfattning

Asset management decision in asset-intensive industries often require a balance betweenrisk, cost and performance. In complex industrial systems, these aspects are difficult tomanage separately, since local asset-related issues may affect production, maintenanceplanning, investments and long-term organizational objectives. This creates a need for astructured approach that supports traceability between operational risk information andhigher-level asset management decisions.The purpose of this thesis was to develop a traceable Model-Based Systems Engineering(MBSE) modelling approach for integrating risk into system-level asset management. Thestudy was conducted as a qualitative case study in collaboration with LKAB, whichprovided an important industrial context. The empirical material consisted of semistructured interviews, observations, internal documents and focus group feedback. Inaddition, a literature review was used to establish a theoretical foundation for assetmanagement, risk classification, risk quantification, traceability and MBSE.The findings show that risk at the case organization is mainly handled close to dailyoperations, through local experience, maintenance activities, inspections and operationalroutines. Although risk knowledge exists within the organization, the current handling ispartly fragmented, especially regarding traceability between local risks, system-level effectsand strategic objectives. The study also shows that risk can be structured by combiningrisk categories with decision levels. Operational, safety and financial risks affect assetmanagement differently depending on whether they are considered at operational, tacticalor strategic level.Based on the empirical findings and the theoretical framework, a modelling approachwas developed using CATIA Magic and the Unified Architecture Framework. The approachconnects organizational objectives, physical asset structures and measurements for risk, costand performance in a shared model structure. The measurement structure makes it possibleto relate operational risk information to cost and performance effects, such as downtime,production loss, maintenance cost and investment priorities. Scenario representation wasalso used to show how different asset management alternatives can be compared using thesame measurement logic.The main contribution of the thesis is a conceptual and traceable decision-supportstructure for linking operational risk information to higher-level asset management decisions.The proposed approach should not be understood as a validated quantitative predictionmodel, but as a way to organize and compare risk, cost and performance informationacross system levels. The results suggest that MBSE can support more transparent andcoordinated asset management decision-making in complex industrial contexts.