Forecasting Building Maintenance Using the Weibull Process
Abstract
Companies budget and spend a large amount of money each year on maintaining or repairing their buildings. Currently, most companies use historical data to determine an average amount spent and add any known work to forecast their future budgets. This empirical model may have great variance from the actual amount of money spent each year. Budgets are based on the amount of maintenance to be done, so this study examines the prediction of maintenance actions and not the cost of the maintenance. Further studies can link the prediction of maintenance to the cost of that maintenance. A statistical model(the Weibull Process)has been proven to predict the failures of repairable systems such as electronics and automobiles. It was assumed that buildings could be classified as repairable systems since they are repaired rather than thrown away the first time a component breaks. A linear regression model is also examined as a possible method of predicting maintenance. The Weibull Process and this linear regression model were used to test their applicability to predicting building maintenance. The tests found that neither the linear regression or the Weibull Process model could accurately be used to predict the occurrence of maintenance on a set of buildings. The data set used is assumed to be the major reason for these results. Further study of the Weibull Process should be done using variations of the data set. Theses.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 1988
- Accession Number
- ADA196312
Entities
People
- Ann K. Yeoman
Organizations
- Air Force Institute of Technology