Construction equipment reliability analysis and failure prediction

Abstract

Construction equipment plays an important role in civil engineering works, particularly in infrastructure projects such as railways and bridge construction. Unexpected failures of equipment can cause serious consequences such as increased cost, project period extension, or even safety issues in some cases. Even though different maintenance and reliability prediction methods have been applied by contractors on site, a significant proportion of equipment repairs still follow unexpected failures. To bridge the gap between failure and preventive maintenance, it is important to discover scientific and precise methods for analyzing and predicting the failures before they happen. Traditionally, there are a number of standard distribution functions which can be used for reliability analysis. However, a number of books and papers have stressed that the usual non‐repairable reliability methodologies, such as the Weibull distribution, are not appropriate for repairable system reliability analyses and have suggested the use of Non‐homogeneous Poisson Process (NHPP) models. Most construction equipment and their components are considered to be in the category of repairable system. Apart from the traditional distributions introduced above, researchers have applied more sophisticated data mining methods to equipment reliability analysis. Time series modeling is one of the more advanced techniques which this research is focused on. Time series analysis can be used to describe and model the historical data, and forecast the future values of the series based on the past values. Construction equipment failure follows the time series pattern, and thus it can be adopted. The aim of this research is to study the possible methods which can be used to analyze reliability and predict the failures of construction equipment in order to bridge the gap between preventive maintenance and repairs and help to make managerial decisions on equipment allocation and maintenance. The objectives are: a) To increase the understanding of the nature of failure patterns of the selected construction equipment; b) To estimate the reliability characteristics of construction equipment in precise quantitative terms by using power law models and time series models; c) Compare the advantages and disadvantages of the traditional power law models with those of time series models in construction equipment reliability analysis; d) To give recommendations on construction equipment management and maintenance based on the research findings. The major works for this research comprise of literature review, data collection, data preparation, quantitative analysis, time series prediction and case studies. A comprehensive literature review on the fields of reliability and construction equipment has been conducted. Quantitative analysis is used in this study including data collection, modelling and validation with the aid of computer software packages. Time series is the main method adopted for reliability analysis and failure prediction while traditional power law models are used as baseline for comparison. Case studies are employed to study the reliability of construction equipment with real maintenance data collected from construction site. The major findings of the research include: the investigation and analysis of the importance of reliability and failure prediction for construction equipment from the aspects of cost, time and safety; testing the traditional power law model and time series model on failure prediction for construction equipment based on real industry data; studying the construction equipment reliability and failure from both the systems and subsystems levels; taking related factors into consideration and evaluating the importance of these factors (e.g. impact from Time to Repair) in the modelling process of failure prediction; comparing the advantages and disadvantages of power law model and time series model. Based on the results and findings of the data modelling and analysis in this research, advice is given for managerial decisions on construction equipment maintenance to promote the practice of repair before failures.