Maintenance Cost and Residual Value Prediction of Heavy Construction Equipment

  • Author / Creator
    Zong, Yi
  • Equipment cost represents a large expenditure for construction companies. Making economic decisions such as when to replace or rebuild is a complicated and difficult task that often relies on the experience of an equipment manager. This research analyzed the historical maintenance cost of heavy construction equipment to simulate and predict the maintenance cost of different types of machines. Based on historical maintenance data of 15 fleets, including 250 heavy machinery units from a construction company, regression models were developed for each type of machine to compare the relationship between the maintenance cost and machine age. A second-order polynomial expression of the Cumulative Cost Model developed by Mitchell (1998) was used to identify optimum economic decisions such as replacement and retirement. As equipment owning companies often design maintenance policies according to specific operating hour intervals, different datasets based on varying service meter reading (SMR) intervals were created to provide equipment managers with different equations, thereby providing a guideline to help the company revise maintenance policies for each type of machine. Statistical analyses were conducted for each dataset and it was found that the best regression model performance was obtained at 500 and 1000 SMR intervals. Residual plots indicate that the models can be improved by including other variables despite the high R2 values. Besides, the residual value of heavy construction equipment is of great significance for equipment owning companies. Many factors such as the manufacturer, model, machine age, operating hours, and even macroeconomic indicators might have direct or indirect impacts on the price of machines in an auction market. In current practice, machine-owning companies use rule-of-thumb opinions or single-linear functions to make predictions, providing a very rough estimate to decision makers. This study considers the current state of knowledge on residual value estimation for used heavy construction equipment and introduces two effective data mining methods, k nearest neighbor (KNN) and random forest (RF) with comparisons to a single regression tree. The proposed methods are exemplified based on a dataset of articulated trucks. Equipment specifications are considered as predictive features, and a feature selection algorithm is implemented to provide a rank of different factors. Distinct models are built after multiple runs and cross validation. Compared to the single regression tree method which has been studied by other researchers, the KNN and RF methods demonstrated better performances in terms of accuracy as well as running time.

  • Subjects / Keywords
  • Graduation date
    Spring 2017
  • Type of Item
  • Degree
    Master of Science
  • DOI
  • License
    This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.