Modelling functional service life using fuzzy logic. Application to heritage buildings

  1. Prieto Ibáñez, Andrés José
Dirigida por:
  1. Francisco Javier Alejandre Sánchez Director/a
  2. Juan Manuel Macías Bernal Director/a

Universidad de defensa: Universidad de Sevilla

Fecha de defensa: 27 de octubre de 2017

Tribunal:
  1. Carles Serrat Presidente/a
  2. Antonio José López Tarrida Secretario/a
  3. José António Raimundo Mendes da Silva Vocal
  4. Vicente Flores Alés Vocal
  5. Rocío Ortiz Calderón Vocal

Tipo: Tesis

Teseo: 501153 DIALNET lock_openIdus editor

Resumen

This doctoral thesis makes a contribution to the field of preventive conservation of heritage buildings through the analysis of their functional degradation. In this sense, we have continued to advance and develop the Fuzzy Building Service Life - (FBSL) expert system designed at the Department of Building Construction (ETSIE) of the University of Seville. This fuzzy logic-based method is able to predict building functionality, managing vulnerabilities and risks that affect their performance. This research addressed new theoretical and practical applications that contribute to the applicability and general implementation of the model. The practical case study focused on churches built between the 13th and 17th centuries and located in the province of Seville in southwest Spain. The FBSL methodology was standardized based on an analysis of the ISO 31000:2009 international standard on risk management and the European standard EN 31010:2011, demonstrating that the model complies with the specifications of these standards. Additionally, the National Institute of Cultural Heritage ("Instituto de Patrimonio Cultural de España" - IPCE - in Spanish) in its National Preventive Conservation Plan ("Plan Nacional de Conservación Preventiva" - PNCP - in Spanish) recommends the use of ISO 31000 as a tool for the management, assessment, monitoring and control of risks in heritage conservation. Subsequently, a linear multiple regression analysis was performed to classify the input variables of the system based on the influence they had on the final result of the model, namely the estimation of building functionality, also considering a reduction in the number of entries and highlighting those presenting greater weight in the output parameter. For this purpose, a total of 100 historic buildings located in the south of Spain were evaluated. The FBSL model was verified by correlating it with another system for predicting the functional service life of building materials (which establishes an index for assessing the overall degradation of a building element, called "degradation severity" (Sw)), developed by the School of Engineering (IST) of the University of Lisbon (Portugal). In this sense, the conditions of physical and functional degradation of façade construction systems in Lisbon (Portugal) were analyzed together, obtaining strong determination coefficients for the entire sample analyzed, and allowing the scale of intervention priority to be associated with a specific period for performing maintenance activities. Finally, the main types of deterioration affecting these historical buildings were identified, analysing their most frequent causes and evaluating the influence of these pathologies on the level of functionality of the buildings analyzed. To carry out this analysis, a total of 390 historical time series records were gathered in a sample of 20 parish churches located in the province of Seville. This application provides very useful information for defining future preventive maintenance plans.The main contribution of this doctoral thesis is an evolved expert system for predicting the functional useful life of buildings - FBSL2.0 -, with six (6) main improvements having been made to the initial system: 1) detailed mathematical development of the fuzzy methodology; 2) standardization with the ISO 31000:2009 international risk management standard; 3) new theoretical and practical applications; 4) linear multiple regression analysis to statistically identify the weight of each variable in the system output, enabling the simplification of the model; 5) validation with another widely tested and published functional useful life prediction methodology; and finally, 6) the study of historical time series data to evaluate the influence of past pathological situations on the future functionality status of buildings.