A multilevel decision-making approach for the resilience assessment of industrial plants in seismic prone areas

Mussini, Nicola; Salvatore, Walter (Thesis advisor); Hoffmeister, Benno (Thesis advisor); Spacone, Enrico (Thesis advisor); Vayas, Ioannis (Thesis advisor)

Aachen (2018, 2019)
Dissertation / PhD Thesis

Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2018


In the present work the problem of the seismic resilience of industrial plants, characterized by a sensitive interconnection among single facilities, has been investigated. A brief introduction about the typical problem the industrial facilities are affected after a seismic event was provided. By this means it was possible to clarify the dimension of the problem and the extension, spatial and temporal, of the economic exposure industrial plants are affected. Then, the focus of the current method of analysis of industrial plants has been proposed, showing the differences among the vulnerability, risk and resilience. It was observed that: vulnerability methods are not suitable to provide information of the economic exposure; risk analysis can provide an insight about the economic exposure immediately after the seismic event, but does not provide any information about its temporal evolution; resilience analysis can fill the lack of information provided by the risk analysis, giving an overview of the temporal evolution of the economic exposure. Not withstanding the importance of the resilience analysis, a lack in the current scientific literature was observed, evidencing that only risk analysis is usually adopted, providing several methodologies, qualitative and quantitative as well. Then, the aspect of the resilience analysis was deepened, specializing the general concepts for industrial plants. In particular the problem was faced from two sides. At first the behaviour of stand-alone facilities and the methods for the resilience assessment was analysed. Specific methods for the estimation of the consequences associated with the resilience analysis, i.e. theresidual functionality q(t), the recovery time trec and the recovery path, were defined. Therefore, the interest moved to inter-connected facilities, for which the concepts of the reliability of systems are enriched with further aspects to make them applicable for the resilience assessment. With such investigations, a detailed method, based on the PBEE framework, for the resilience assessment of industrial plants and facilities prioritization was proposed, filling the gap of current literature. Notwithstanding the evident complexity to perform detailed analyses on a huge set of facilities, variously connected among themselves, could be to ocumber some, a multilevel procedure is proposed. The aim of such a procedure is to provide a global, but rough, overview of the results at the first level, performing simplified analyses (for the definition of the structural performance sand the estimation of consequences ), and specializing the results at the next levels by means of detailed analyses targeted on few critical structures. The multilevel procedure is intended to be a tool that simplify the detection of critical elements within the whole Plant, performing on them detailed assessment. The procedure is based on three levels:• the first level is based on the execution of simplified analyses at the plant level; for the estimation of consequences a specific method, based on a tabular format is developed;• the second level focuses on refining the structural performance and the estimation of consequences of critical facilities;• the third level provides a specific definition of the assessment of the structural performance of the critical details, providing the effectiveness on the plant resilience of possible upgrading solutions, that aim at reducing the vulnerability, operating on the structure, or increasing the resilience, enhancing the method of interventions after the seismic damage. Each level provides also a prioritization of facilities, based on the concept of Resilience Indicators (RI). The Resilience Indicators, from a practical perspective, scales the consequences of each facility to achieve a optimal resilience condition decided by the owner of the plant and defined through the break even point. The Resilience Indicators allow to establish which facilities require in deep assessment from one step to the next. Two case studies, representative of two actual industrial realities, a SteelMill and a Chemical Plant, were carefully selected to apply and validate the proposed multilevel procedure. The Steel Mill is characterized by a huge in plan extension, and is characterized by drift sensitive facilities, easily to schematize as single degree of freedom structures. The Chemical Plant has a compact configuration that develops over the height; it is characterized by a main braced building that support several vessels, mainly sensitive to peakfloor acceleration. The results of the detailed resilience assessment, applied on both the case studies, are used as a reference to validate the multilevel procedure. The multi level procedure provided results that well fit the detailed resilience assessment, regarding both resilience curve and the facilities prioritization.Furthermore, the convergence of results was achieved with a reduced numberof steps.Clearly, the application of the multilevel procedure strongly reduced theamount of detailed analyses, optimizing the in-field surveys and simplifyingconsiderably the achievement of satisfactory results in terms of Plant’sresilience.