A computational intelligent fuzzy model approach for excavator cycle time simulation

The tracked hydraulic excavator is one of the most versatile and widely utilised piece of earthmoving equipment. In many instances, the 'excavator' represents the first choice of earthmoving plant for both construction managers and estimators, since when properly employed (i.e. with a competent operator and in an appropriate working environment), it offers high production rates at economical cost. Nonetheless, predicting machine production performance is difficult; given the typical multiple operational parameters (e.g. machine weight, machine configuration, ground conditions, operator ability) that can apply. Consequently, determination of accurate cost estimates and predicted contract durations are subject to considerable inaccuracy, especially where a significant amount of site work is needed. To address this inadequacy, this paper presents a computational intelligent 'fuzzy' model with the ability to forecast excavator cycle time. In this context, a cycle is defined as one complete revolution, from 'place empty bucket in dig material' through 'fill bucket', 'move charged bucket to target', 'empty charged bucket' and 'return bucket to dig material'. The developed model is based upon 70 separate cycle time observations obtained from four plant manufacturers. These data provide a representative spread of machine cycle times since they include a range on a continuum from optimum to adverse operational parameters. Tests on the derived model identified that its accuracy was acceptable; but the accuracy could be improved using larger samples and a more comprehensive and exhaustive range of variables to predict machine cycle time.