Timber harvesting on wetlands in the South is normally conducted in the dry season. The unavailability of machines with appropriate traction devices hinders wood extraction of these wetlands during the wet season. Because of this extend of woodlands in the Costal Plain region, maintaining a continuous flow of wood to the mills requires unlimited scheduling of mechanized logging operation on wetlands. Since the evolution of skidders in the 1960s, tire manufacturers have developed special logger tires to contend with the harsh environment experienced during wood extraction. The main objective of this paper is to report the results of training a neural network using field data to predict the skidder pull-load relationship. These predictions would then be compared with the results of a regression-based model. A 4-5-3-1 neural network simulated the pull-load relationship successfully and was appropriate for predicting results with conditions not used in network formulation (tire inflation pressure of 138 kPa and tire size of 30.5LX32). The network converged with a total sum of squares error of 0.332. The network was capable of predicting vehicle performance when presented with unseen data (tire size and inflation pressure). The line pull predicted by the network increases linearly with increases in load weights. Linear equations relating the model line pull to the loads for the three tire size tested were in close agreement with the experimental results. No effects were noticeable due to changes in tire size or tire inflation pressure on the line pull. Future studies should explore the use of cross validation and/or stop training methods for improving the neural network generalisation of field data.


    Access

    Access via TIB

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Application of artificial neural networks to skidder traction performance


    Additional title:

    Die Anwendung künstlicher neuronaler Netzwerke auf die Traktionsleistung von Holztransportern


    Contributors:
    Tohmaz, A.S. (author) / Hassan, A.E. (author)

    Published in:

    Journal of Terramechanics ; 32 , 3 ; 105-114


    Publication date :

    1995


    Size :

    10 Seiten, 6 Bilder, 1 Tabelle, 13 Quellen




    Type of media :

    Article (Journal)


    Type of material :

    Print


    Language :

    English





    Application of artificial neural networks to skidder traction performance

    Tohmaz,A.S. / Hassan,A.E. / North Carolina State Univ.,Dep.of Forestry and Biological and Agricultural Engng.,US | Automotive engineering | 1995


    Caterpillar's new track skidder

    British Library Online Contents | 1999


    Component Nomenclature—Skidder-Grapple

    SAE Technical Papers | 2007