Review of productivity models for harvesters, feller-bunchers, and harwarders published in 2013-2023

Sammanfattning

Accurate predictions of work productivity are crucial for fully mechanized harvesting. Predictive models are thus employed to, for instance, estimate the duration of a harvesting operation. If these predictions are inaccurate, planning failures of forest companies may occur. The employed models are developed based on time studies or follow-up studies, and due to the diverse work conditions and the choices of the modeler, models’ characteristics, such as independent variables and predictive capabilities, will differ considerably. Despite the models’ significance, there are few publications that have scrutinized models for fully mechanized harvesting in terms of the models’ independent and dependent variables, mathematical structures, coefficients of determination, the quantity of operators used for collecting data, the equipment used for time measurement, incorporation of delays, and potential regional or harvesting method-based differences. Therefore, a literature review was conducted to scrutinize productivity models for harvesters, feller-bunchers, and harwarders, aiming to better understand and explain the observed variability in models. One-hundred-and-fifteen publications containing models were identified, and the publications originated from six continents, 22 countries, and covered a wide range of forest types and work conditions. The review identified several differences in terms of equipment type used, with hand-held computers being the most common. However, more than a third of the publications did not specify equipment type. The number of operators studied also varied considerably, where most publications focused on a single operator, but the number of operators ranged from 1-120. In more than a third of publications, the number of operators could not be ascertained. Regarding their experience, it ranged from only a couple of months to several decades of expertise. In the publications, 422 models were identified, and several commonalities as well as differences between models and publications were identified. For example, the majority of publications used piece size (tree volume or diameter at breast height) as a predictor, but there was a large variation in which additional predictors were used. Regional, as well as harvesting method-based differences, were also identified. It was, for instance, more common to model work cycle time and include movement/distances travelled-based and work cycle-based variables in North American publications, as well as in publications focusing on full-tree harvesting, than in other publications. Notable regional differences were also identified regarding models’ mathematical structure. Publications from Asia, Eastern Europe, and North America predominantly developed purely linear models, a trend not apparent in other regions. Predictive capabilities of models also varied considerably, where some explained almost nothing, while others explained almost all variation. In terms of delays, most publications excluded them, while, when they were included, several approaches could be observed. This review identified several regional and harvesting method-based differences in modelling practices for harvesters, feller-bunchers, and harwarders. Not only did this review identify differences, it is also a compilation of productivity models which facilitates easier access and identification of available productivity models published in 2013-2023.