Determinação e análise de taxas de deterioração de pontes rodoviárias do Brasil

• Ano de defesa: 2019
• Autor(a) principal: Caroline Buratto de Lima e Oliveira
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; ENG - DEPARTAMENTO DE ENGENHARIA ESTRUTURAS; Programa de Pós-Graduação em Engenharia de Estruturas; UFMG
• Palavras-chave em Português: Pontes; Gestão; Modelos de Deterioração; Cadeias de Markov; Engenharia de estruturas; Markov, Processos de
• Link de acesso: http://hdl.handle.net/1843/41718

Collapses, cracks, and structural damages are usually associated with an inadequate maintenance strategies and they have been occurred more often, enforcing the engineering to understand these structures performance to provide a more suitable management. In order to guarantee bridges in service is required regular inspections and a routine repair work - usually restricted by limited resources. This scenario encourages researches in the area to provide more analytical and efficient management methodologies, especially important in large stocks administration. In this context, predict the future condition of the bridge stock is the most crucial knowledge to develop reliable management programs. This could assist to estimate cost and activities of maintenance, repair and reinforcement (MR&R actions) required in short, medium and long term, also, it is basilar in risk assessment. This research collected and registered inspection reports performed by Brazilian road owners sum up to 1.707 bridges over seven years in 16 road segments. These reports were dispersed and were not registered on a common base or on specific bridge management software (BMS). The data was suited to enable stochastic simulations based in Markov Chains method with particular approaches in order to estimate Brazilian bridges deterioration rates. To this end, a specific data filtering criteria was created and were established specific routine for data entry and information recording using the “R” – a statistical computing language. Finally, this thesis proposed a method to compare the results of diverse rates results and to determine the deterioration model accuracy. The results demonstrated good predictions for short-term, although the limited data and it was confirmed its highly dependence to the number of inspection history data cycles. In addition, the results were compared with possible degradation agents not related in data. The collection of Brazil bridge inspections contributed to promote the inventory of these structures and enables to identify these most basilar characteristics. Further, the thesis presents the current management practices and its diversity in Brazil, the efficacy of their utilization and suggests future challenges to provide efficient management of national bridges.