Predictive modelling of building degradation as influenced by contributing factors

Enviar para

Título:
Predictive modelling of building degradation as influenced by contributing factors
Autor: DE SILVA, Dehinga Tharanga
Assuntos: Building inspection ; Condition Monitoring ; Deterioration modelling ; Maintenance prioritisation ; Markov process
Descrição: Buildings are complex infrastructure systems with diverse types of components ranging from structural, architectural, and services components to fittings and fixtures. Due to this complexity, managing the building infrastructure while maintaining the level of service provided can be challenging. Asset management of buildings, therefore, requires a systematic process with optimized investment strategies. The proposed research aimed to address the challenge through the development of a number of new methodologies. Asset management of buildings typically requires an inspection strategy for building components, predictive modelling of the deterioration of the components considering influencing factors, prioritization of maintenance activities considering an available budget, and forecasting the cost of intervention. Condition rating methods are an essential component of the management process, and a well-developed condition rating method enhances the accuracy of the subsequent stages of the asset management program. The research presented here has conducted a comprehensive review of the literature to establish the gaps in knowledge and practice of current building asset management methods. Major gaps identified include (a) the inability of the standard condition rating methods based on visual inspections to arrive at a realistic capture of the condition of a building element and (b) the inability of the current methods of deterioration prediction to consider the random nature of deterioration and also lack of understanding of the impact of influencing factors on the degradation of elements, (c) need for a systematic maintenance prioritization method and (d) forecasting costs accurately based on the predictive models of degradation of building elements. A comprehensive research program was conducted to address the gaps identified in knowledge and practice. A new condition rating method that considers the impact of defects, their criticality and the extent for structural and architectural building elements has been developed. Moreover, condition rating method that considers the external and functional condition of HVAC systems are also introduced. A survey with 120 building asset managers was used to establish the impact of defects on the service provided by the building components. A new iPad app was developed to enable field data collection using the proposed approach. 32 buildings in the City of Kingston were inspected using the new app to demonstrate the application of the new condition rating method and the iPad application. A similar methodology was developed for HVAC systems, and inspections were conducted in collaboration with the inspection company: and the Airmaster team on RMIT buildings to demonstrate the application of the new method. Collected condition inspection data are clustered and used as input to Markov process-based deterioration prediction models to predict the future condition of selected building assets. Three Markov transition matrix calibration techniques have been used in this research: Markov Chain Monte Carlo simulation, Nonlinear optimization, and regression-based optimization. Derived models are subjected to qualitative and quantitative analysis to demonstrate the superiority of the proposed new condition rating method over the existing condition rating method. To demonstrate the methodology for analysis of the impact of influencing factors on the degradation of building components, a study was conducted to understand the effect of run time and environment on the deterioration of HVAC systems. Outcomes clearly indicated the value of considering the influencing factors in predicting future conditions of building components. The newly proposed condition assessment method leads to the establishment of defects, which enables accurate capture of maintenance activities. A novel maintenance prioritization method is proposed and demonstrated through a case study. The last element supporting decision-making in asset management is cost forecasting. Three case studies are presented where the derived models are utilized in predicting future costs. Scenario-based cost forecasting method has been incorporated into a cloud-hosted software platform developed by RMIT for asset management of buildings. The Central Asset Management System (CAMS) has been updated to integrate the outcomes of the research presented here, enabling immediate improvement of the accuracy of building asset management using CAMS. Source: TROVE
Data de criação/publicação: 2022
Idioma: Inglês

Predictive modelling of building degradation as influenced by contributing factors

DE SILVA, Dehinga Tharanga

2022

Buscando em bases de dados remotas. Favor aguardar.