Implementing Resonance-Based Acoustic Technology on Mechanical Harvesters/Processors for Real-Time Wood Stiffness Assessment: Opportunities and Considerations
Authors
Dzhamal Amishev
Glen E. Murphy
Abstract
Acoustic technology has been successfully used as a non-destructive technique for assessing the mechanical quality of various wood products and species based on stiffness. Many mechanical harvester/processor manufacturers have implemented mechanical sensors to measure tree diameter and length as well as optimal bucking algorithms on their equipment. There is a growing interest in incorporating technologies for measuring internal stem features into a harvester head. The objectives of this study, therefore, were to i) determine and investigate the factors arising from incorporating acoustic instruments on a mechanized harvester head that might influence resonance-based acoustic signal and velocity readings and quality in Douglas-fir, and ii) investigate the issues and considerations associated with suggested working strategies in regard to harvest productivity impacts and processing decisions.
After taking into account some feasibility considerations, it was determined that the hold of the machine grapple would not compromise the accuracy of resonance-based acoustic velocity readings. There were three working procedures suggested for measuring resonance-based acoustic velocity: 1) after the stem is delimbed and run through the measuring equipment, 2) once a portion of the stem is measured and the length of its unmeasured portion is forecast, and 3) after the tree is felled by the harvester but before any further processing is done.
Regardless of the working procedure, it was determined that logs produced from lower sections of the tree are stiffer than those from upper portions. If the processor head traverses the stem partially or completely, the removal of bark and branches and their effect on acoustic velocity readings should be taken into account. Forecasting routines could be developed to account for imperfect and even non-existing information about tree length with the second or third working procedure. Results yielded by the two methods used for stem height (and consequently acoustic velocity) prediction in this study (linear regression model and a k-nearest-neighbor) were considered rather promising. Testing feasibility concerns with the resonance-based acoustic technique were observed if the entire stem was intact to the very top offshoot bud.
