Hard-rock LHD cost estimation using single and multiple regressions based on principal component analysis

► Adequate estimation of equipment costs is a key factor in feasibility study of mining and tunneling projects. ► Available models are univariates (the role of other effective variables has simply been ignored) and out of date. ► A model was developed for estimating capital and operating cost of Load–Haul–Dump (LHD) machines. ► The regression based model can be considered as an accurate tool in the feasibility study of mining and tunneling projects. In feasibility studies and mine planning, accurate and effective tools and methods facilitating cost estimation play an important role. Load–Haul–Dump (LHD) machines are a key loading and haulage equipment in most of the underground metal mines and hard rock tunnels. In this paper, a cost estimation model of these vehicles has been presented in the form of single and multivariable functions. These functions have been provided on the basis of costs types (i.e. capital and operating costs) and motor types (diesel and electric). Independent variables, in the single regression analysis is bucket capacity and in Multiple Linear Regression (MLR) analysis include bucket capacity, overall width, overall machine height and horse power (HP). The MLR is conducted in three steps. First, with the help of Principal Component Analysis (PCA), correlation between independent variables is omitted. Thereafter, significant PCs are selected and used as independent variables in the MLR functions. Finally, the cost relationships are established as functions of initial LHD variables. The mean absolute error rates are 11.59% and 6.87% for the single and multiple linear regression functions, respectively.