312 Kane and Arwade: Effect of Climbing Line and Ascent Technique on Arboricultural Climbing Loads ranges. Two of the climbing lines (Escalator [Teufel- berger Ropes, Fall River, MA, USA]; Mercury [Sam- son Rope, Ferndale, WA, USA]) were new at the start of the experiment, while the third (WorkPro [Sterling, Biddeford, ME, USA]) had been used in several pre- vious studies and was noticeably fuzzy (Table 2). The lead author performed each of the 6 combinations of climbing line and ascent technique twice—12 trials in total—in a randomized design stratified by TIP. He completed all 12 trials for each TIP on 3 separate days, resting between trials to avoid the confounding effect of muscular fatigue. The lead author completed 36 trials overall: 12 trials on each of the 3 TIPs. From each load time history, we computed the load at the following quantiles: 75th, 90th, 95th, 99th, and 100th. We normalized the value of each quantile as a ratio with the lead author’s weight on the day the time history was collected; we designated the ratio as Pi , where i is the index of quantiles. In statistical analy- ses described below, we analyzed the mean of each Pi computed from the 2 trials the lead author conducted on each combination of climbing line, ascent tech- nique, and TIP. For each Pi , there were 3 replicates (1 for each TIP) of each combination of climbing line and ascent technique. We also counted the number of peak loads that equaled or exceeded the value of each quantile of the load time history. For example, if P75 equaled 1.1 in a load time history, we counted the number of peak loads greater than or equal to 1.1 in the time history. We averaged the 2 counts of peak loads that equaled or exceeded each quantile in the load time history from the 2 trials the lead author conducted on each combination of climbing line, ascent technique, and TIP, and analyzed the mean value in the statistical analyses described below. We used a 2-way mixed model analysis of vari- ance (ANOVA) to determine whether ascent tech- nique, climbing line, and their interaction affected Pi and the number of peak loads that equaled or exceeded loads at each quantile. Random effects in the ANOVA were TIP and its interactions with climbing line and ascent technique. Tukey’s Honestly Significant Dif- ference test was used for mean separation of signifi- cant (P < 0.05) effects in the ANOVA. We used JMP statistical software (v. 15, SAS Institute, Cary, NC, USA) for the ANOVA. To determine loading frequency during ascents, we followed the methods of Kane et al. (2020) using ©2022 International Society of Arboriculture MATLAB (MathWorks 2018). Briefly, we applied the Fast Fourier Transform (FFT) method to each load time history to calculate its power spectral density (PSD). Then we used the function “findpeaks” to determine peaks in the spectra. The output of the “findpeaks” function was a set of {frequency, PSD peak} pairs that indicated the peaks of the PSD of each trial. To determine whether ascent technique and/or climb- ing line influenced loading frequency, we also used the approach of Kane et al. (2020), which is an alter- native to direct averaging of the spectra across repli- cations. The approach used here, and described briefly now, has the advantage of retaining information about the location and amplitude of the spectral peaks in individual replications, whereas in spectral averaging techniques all information regarding individual repli- cations is lost. Scatter plots of PSD peaks vs. frequency were created for each combination of climbing line and ascent technique. To identify common loading frequencies that also had significant power spectral amplitude, a kernel density function was fitted to the {frequency, PSD peak} pairs in the scatter plot using the function “ksdensity.” Peaks in the kernel density fit were identified using the “findpeaks” function; the peaks were considered the characteristic frequencies for each combination of climbing line and ascent technique. To estimate variability of the loading fre- quencies, the width of the kernel density fit at 60% of the peak amplitude was evaluated. Sixty percent of the peak amplitude is equivalent to plus or minus one standard deviation for the Gaussian probability den- sity function. The range of a loading frequency is the upper and lower frequencies on the kernel density fit intersected by a horizontal line (constant PSD peak value) at 60% of the kernel density peak. RESULTS The random effects of TIP and its interactions with climbing line and ascent technique were not signifi- cant in the ANOVAs for any Pi loads for each Pi. Values of Pi to 1.5 for P100 or the count of peak ranged from 1.1 for P75 lines or techniques for any Pi , but they did not vary between climbing (Table 3). load time histories ranged from a single peak load for P100 The count of peak loads at each Pi except P100 (on WorkPro)(Table 4). was and The count of peak loads at each quantile of the to 31 peak loads for P75 greater when ropewalking compared to footlocking; the difference was statistically significant for P75
November 2022
| Title Name |
Pages |
Delete |
Url |
| Empty |
Search Text Block
Page #page_num
#doc_title
Hi $receivername|$receiveremail,
$sendername|$senderemail wrote these comments for you:
$message
$sendername|$senderemail would like for you to view the following digital edition.
Please click on the page below to be directed to the digital edition:
$thumbnail$pagenum
$link$pagenum
Your form submission was a success. You will be contacted by Washington Gas with follow-up information regarding your request.
This process might take longer please wait
