A Kinematic and Dynamic Analysis of Shoveling Snow by José Andrés DeFaria An Engineering Project Submitted to the Graduate Faculty of Rensselaer Polytechnic Institute in Partial Fulfillment of the Requirements for the degree of Master of Engineering Major Subject: Mechanical Engineering Approved: _________________________________________ Ernesto Gutierrez-Miravete, Thesis Adviser Rensselaer Polytechnic Institute Hartford, Connecticut May 2016 (Second Progress Report: March 2016) i © Copyright 2016 by José Andres DeFaria All Rights Reserved ii CONTENTS A Kinematic and Dynamic Analysis of Shoveling Snow ................................................... i LIST OF TABLES ............................................................................................................ iv LIST OF FIGURES ........................................................................................................... v ACKNOWLEDGMENT .................................................................................................. vi ABSTRACT .................................................................................................................... vii 1. Introduction.................................................................................................................. 1 2. Background .................................................................................................................. 2 3. Methodology ................................................................................................................ 6 3.1 Initial Phase ........................................................................................................ 7 3.2 Second Step ........................................................................................................ 8 3.3 Third Step ........................................................................................................... 9 3.4 Fourth Step ....................................................................................................... 10 3.5 Fifth, Sixth, and Seventh Step .......................................................................... 10 4. Results........................................................................................................................ 11 5. Discussion .................................................................................................................. 13 6. Conclusions................................................................................................................ 14 7. References.................................................................................................................. 15 8. Appendices ................................................................................................................ 16 8.1 Appendix A - Determination of Body Segment Weights ................................ 16 iii LIST OF TABLES Table 1: Errors associated with estimation in step three ................................................... 9 Table 2: Average body heights and lengths [8] ............................................................... 16 Table 3: Average body segment weights [8] ................................................................... 16 Table 4: Average heights and weights for across the adult human population ............... 16 Table 5: Body segment lengths and heights across the human adult population ............ 17 Table 6: Body segment weights across the human adult population ............................... 17 iv LIST OF FIGURES Figure 1: A typical straight-shaft shovel and an ergonomic bent-shaft shovel [3] ............ 1 Figure 2: Regular and modified shovel design for dirt used in study [4] .......................... 2 Figure 3: Flexion angle and loads which induce moments in the lower back [5] ............. 3 Figure 4: Dimensions of the snow shovels used in [6] ...................................................... 4 Figure 5: Mean L5/S1 extension moment-time curves (sagittal plane) [6] ....................... 5 Figure 6: Mean upper body flexion-time curves (sagittal plane) [6] ................................. 5 Figure 7: Heaviest and lightest weights for subjects of given heights [7] ......................... 6 Figure 8: Initial positions of upper arm (green) and lower arm (blue) for the Lewinson Average .............................................................................................................................. 7 Figure 9: Calculation of moment arm ................................................................................ 8 Figure 10: Position of back and arms at the end of the second step for the straight-shaft (left) and the bent-shaft (right) shovels ............................................................................. 9 v ACKNOWLEDGMENT Type the text of your acknowledgment here. vi ABSTRACT Type the text of your abstract here. Current summary of my progress: (Second Progress Report) It is now my second progress report and I am pleased to say that things are perfectly on track! Based on my preliminary schedule from the project proposal, I am supposed to have completed my model and generated preliminary results. I have succeeded in meeting this goal and am actually further along with the model than I originally expected to be. As discussed with you via email, I decided to switch from Abaqus to Maple for the software that will run my analysis. Abaqus has a wonderful graphical user interface; however it is not adequately suited for this type of analysis. To complete a kinematic analysis in Abaqus, a multitude of points must be drawn and then connected with connector elements. This is what was done in the Beaman report about throwing a football. Although Abaqus permits for parameterization of certain geometries, point cannot be parametrically located. This means that a new model would need to be constructed for each and every height that I wished to examine. On the other hand, Maple is a computer algebra system. The height and weight of a subject can be defined as variables, allowing the results to be recalculated for a subject of a different height/weight in a matter of seconds. Since everything must be coded into Maple, it requires a deeper understanding of the geometric motions completed in the analysis. The model has been simplified into line segments (such as the upper arm, which has a fixed length and can only rotate at its ends) which means most of this geometry is limited to three-dimensional trigonometry. Separate Maple worksheets have been created for the straight shaft and the bent shaft shovels. Some preliminary data are presented in the Result section, although there is no discussion about it yet. Although I am confident that the model does not have any significant sources of error, I am aware of a minor source which, on average, may misplace some of the positions by about a quarter of an inch. This is not expected vii significantly affect the end results, but it will be looked into over the course of the next month. Per the MANE6970 Key Deadlines document, the following sections of this report were expected to have content by today's date: title page, list of tables, symbols, contents, etc, introduction, methodology, results, discussion, references, and appendices. As discussed above, most of these sections do have content. The only true exception is the results and discussion section, which is currently populated by tables and data only. The data will be further evaluated before the discussion section is written. Future Work before the next progress report The next progress report (4/25) is expected to be a complete preliminary final report. In order to meet this goal, the Maple models will be further examined in an attempt to reduce some of the known sources of error. Additionally more content will need to be written to support this document. This is not expected to be an issue since the primary analytical model is complete, this will permit additional time to focus on the report writing. viii 1. Introduction Snow shoveling is a routine winter task for many Americans, especially those living in southern New England. Boston and Hartford both receive in excess of 30 inches of snowfall per winter [1]. While corporations and municipalities rely on plowing or snowblowers for snow removal, private residences are typically shoveled by the owners or occupants [5]. Since snow shoveling is not the primary occupation of many of the participants, they do not receive training as a full-time laborer might. As a result, many of these residents may incorrectly use the snow shovels they have, such as lifting with their back, rather than with their knees. This type of misuse is likely the reason why approximately 11,500 individuals are treated in US emergency rooms each year due to injuries sustained while shoveling snow [2]. Many of these injuries occur in the lower back. Ergonomic snow shovels exist which are purported to reduce stress in the back due to their bent-shaft design, but many households continue to use straight-shaft shovels. Figure 1: A typical straight-shaft shovel and an ergonomic bent-shaft shovel [3] 1 2. Background There have been some previous studies into the impact of ergonomic shovel use. Some of these studies relied on qualitative surveys to determine the effectiveness of each shovel design. Other studies used mechanical means in an attempt to measure lifting forces. A select few studies used data collection equipment to take quantitative information into the bending of the back during shoveling tasks. In one of the earlier studies found, a different style of ergonomic shovel was evaluated [4]. This shovel was designed for dirt rather than snow. Instead of using a bent shaft design, similar to that depicted in Figure 1, it used a handle mounted to a secondary shaft, as shown in Figure 2. Only qualitative data was collected via surveying study participants; however, the results indicated that the modified shovel did reduce perceived pain in the participants. The authors also noticed that the modified shovel appeared to reduce the tendency for the participants to stoop. It is these extended periods of bending that are considered the primary cause of muscle fatigue and back pain or injury. It should be noted all of the study participants were male, and all were industrial workers who completed shoveling or digging tasks in their daily work. Figure 2: Regular and modified shovel design for dirt used in study [4] In another study, published ten years later, shovels similar to those depicted in Figure 1 were used exclusively for snow [5]. Quantitative data was collected for this study, including the trunk flexion angle, lateral bending angle, and rotation angle. The results 2 showed usage of the bent shaft shovel significantly reduced the trunk flexion angle. The average trunk flexion angle was 41.4 degrees with the bent shaft shovel compared to 49.2 degrees for the straight shaft. This difference in bending angle is significant because it decreases the moment placed on the lower back by the upper body as shown in Figure 3. It should be noted that all study participants were male. Figure 3: Flexion angle and loads which induce moments in the lower back [5] Despite the fact that ergonomic shovels are marketed as such, as recent as 2014, no scientific evidence supported claims that the shovel would reduce mechanical loading on the lower back [6]. Another study, [6], continued the work from [5], with an attempt to determine the reaction moment in the lower back, referred to as the L5/S1 extension moment. L5 and S1 refer to a particular location within the vertebrae, between the lumbar and the sacrum, essentially the base of the spine. 3 Figure 4: Dimensions of the snow shovels used in [6] The dimensions of the snow shovels used in [6] are presented in Figure 4. It should be noted that out of all the previous studies examined, this was the only study which included women in the sample size, although the results were not broken down by gender. The study found the L5/S1 peak extension moment was 0.627 N·m/kg·m for the bent-shaft shovel compared with 0.703 N·m/kg·m for the straight-shaft shovel. Additionally, the peak upper body flexion for the bent-shaft shovel was 74.3°, reduced from a value of 84.8° with the straight-shaft shovel. These results indicate that the ergonomic shovel does reduce mechanical loading on the lower back. Graphs of the L5/S1 extension moments and the upper body flexion are presented in Figure 5 and Figure 6, respectively. In an attempt to account for the different body proportions, the extension moment was divided by the total mass and height of each participant. 4 Figure 5: Mean L5/S1 extension moment-time curves (sagittal plane) [6] Figure 6: Mean upper body flexion-time curves (sagittal plane) [6] Although this study did include women in the study, its presentation of the final results in the form of averages speaks only to the average individual. A female of the 5th percentile height is fifteen inches shorter than a male of 95th percentile height [7]. A 95th percentile height male who is also in the 95th percentile in weight can weigh more than three times more than a 5th percentile female in height and weight. These discrepancies in height should greatly affect the flexion angle and the discrepancies in weight should greatly affect the extension moment. 5 3. Methodology This evaluation will expand upon previously completed work in [6]. Will ergonomic shovels reduce mechanical loading of the lower back across the human adult population? Figure 7 shows the weight range of North American and European adults within various height ranges. While the subjects selected in Figure 7 are at the extremes for their heights, the 5th percentile female is 4 feet, 11 inches tall and weighs on average 113 lbs (150 cm, 51.26 kg) and the 95th percentile male is 6 feet, 2 inches tall and weighs on average 246 lbs (188 cm, 111.58 kg) [7]. Figure 7: Heaviest and lightest weights for subjects of given heights [7] To understand the ergonomic value of the bent-shaft shovel, it must be proven to reduce the mechanical loading of the lower back in a great proportion of the adult population. A small female will not need to bend over much to reach the snow. Additionally, her upper body will weigh significantly less than that of a large male. The extension moment is theorized to increase with both flexion angle and upper body weight; therefore, ergonomic advantages gained by use of the bent-shaft shovel may diminish as height and weight decrease. 6 To evaluate the effectiveness of the ergonomic value across the adult human population, a kinematic analysis will be completed. This analysis will use Maple (a computer algebra program) to compute the body joint positions throughout the shoveling motions. Kinematic analysis will first be accomplished for an adult matching the average height and weight of the subjects in [6] (1.77 m, 73.5 kg) to reproduce the results shown in Figure 5 and Figure 6. When this kinematic model is complete, the variables defining the height and weight can be modified to represent a 5th percentile female and a 95th percentile male and the results will be compared. 3.1 Initial Phase Two separate Maple worksheets were created, one each for the standard shovel and the bent shovel. A global coordinate system is defined with the ground below the base of the trunk as (0,0,0). The x-direction is the front-to-back direction, the y-direction is the vertical direction, and the z-direction is the left-to-right direction. The height and weight of the subject can be input in inches and pounds. Using these values, and the ratios discussed in Appendix A, the body segment weights and lengths can be computed. The subject is assumed to stand with their upper arm rotated 10 degrees from vertical, and their lower arms parallel to the ground. This can be seen in Figure 8. Knowing these angles permits the calculation of the base of the trunk, the top of the trunk, both shoulders, both elbows, both hands, the shovel shaft end, and the center of the shovel blade. The x-, y-, and z-coordinates of these points are calculated. For the bent shovel, additional points are necessary to define the two vertices which comprise the bent shaft. Figure 8: Initial positions of upper arm (green) and lower arm (blue) for the Lewinson Average 7 The total moment on the base by summing the contributions of each item. For simplicity, the center of mass for the trunk and head was considered to be the supersternale height. The center of mass for the upper and lower arms was calculated to be at the center of each arm. The center of mass for the shovel was located at the shaft termination and the snow load was applied to the blade position. To determine the contributions each of these items have towards the total moment, the moment arm was determined using the distance from the base in the x- and z-directions as shown in Figure 9. Only the x-direction contribution is considered for the upper and lower arms as the zdirections are counteracted by the left and right symmetry. Figure 9: Calculation of moment arm 3.2 Second Step In the second step, the trunk rotates at a constant angular velocity until the peak trunk flexion angle (84.8° for the straight shovel, 74.3° for the bent shovel) is reached. The 8 upper arm rotates another 20° with respect to the vertical and the lower arm rotates until it is perpendicular to the ground. The positions of these arms is shown in Figure 10. Figure 10: Position of back and arms at the end of the second step for the straight-shaft (left) and the bent-shaft (right) shovels 3.3 Third Step During the third step, the left arm lowers while the right arm raises to reduce the blade height to zero. This is accomplished by estimating what angle the shovel needs to be rotated to in order to reduce the blade height at the end of the second step to zero. Since the lower arm remains perpendicular to the ground, this raising and lowering is accomplished by increasing and decreasing the angle at the shoulder. As a result, the right arm moves closer to the body and the left arm moves further away, causing a rotation of the shovel in the x-z plane as well as the y-z plane. The estimation used in determining these angles is a small source of error in the overall calculation. Theoretically, the blade height (y-direction) should be zero at the end of this step. The actual blade heights for the three body types are shown in Table 1. Table 1: Errors associated with estimation in step three Straight shovel Bent shovel 5th percentile female NEED -0.04672 in Lewinson average NEED -0.39272 in 9 95th percentile male NEED -0.54481 in 3.4 Fourth Step During the fourth step, the weight of the snow linearly ramps over the time period. Based on [6], the weight of the snow is assumed to be 6 pounds. 3.5 Fifth, Sixth, and Seventh Step During the remaining steps, the motions are completed in reverse. This is accomplished by setting the position values equal to the appropriate position value in the corresponding step. 10 4. Results The graphs and tables below are PRELIMINARY. That is why they are not yet captioned or discussed in great detail. Moment for the Lewinson Average individual 70000 60000 Momnet (in*lbs) 50000 40000 Straight Shovel 30000 Bent Shovel 20000 10000 0 0 1 2 3 4 5 Time (seconds) Similar style curves to what is seen in Figure 5. Peak Moment Straight shovel Bent shovel 32.5e3 in*lb 26.7e3 in*lb 5th percentile female (average) 37.1e3 in*lb 31.0e3 in*lb Lewinson average 57.3e3 in*lb 50.2e3 in*lb 95th percentile male (average) 86.7e3 in*lb 78.0e3 in*lb 95th percentile male (heavy) 101.4e3 in*lb 91.9e3 in*lb 5th percentile female (light) 11 6 Peak Moment, normalized for height and weight using the Lewinson methodology Straight shovel Bent shovel 6.269 5.134 5th percentile female (average) 5.571 4.648 Lewinson average 5.079 4.450 95th percentile male (average) 4.762 4.286 95th percentile male (heavy) 4.615 4.182 5th percentile female (light) Graph showing Moment over time for all five subjects with both shovels 120000 100000 A1 Momnet (in*lbs) 80000 A2 A3 A4 60000 A5 B1 B2 40000 B3 B4 B5 20000 0 0 1 2 3 4 5 6 Time (seconds) A= straight shovel, B=bent shovel, 1-5 = the five subjects listed in the order of the table above. 12 5. Discussion This is a discussion of the results. 13 6. Conclusions conclusions will go here. 14 7. References [1] National Oceanic and Atmospheric Administration. (2014, February). National Overview - February 2014, Winter Snowfall Departure from Average. Retrieved February 23, 2016, from https://www.ncdc.noaa.gov/sotc/national/2014/2/supplemental/page-4/ [2] Watson, D. S., Shields, B. J., & Smith, G. A. (2011). Snow shovel-related injuries and medical emergencies treated in US EDs, 1990 to 2006. American Journal of Emergency Medicine, 29(1), 11-17. [3] Huang, C., & Paquet, V. (2002). Kinematic evaluation of two snow-shovel designs. International Journal of Industrial Ergonomics, 29(6), 319-330. [4] Degani, A., Asfour, S. S., Waly, S. M., & Koshy, J. H. (1993). A comparative study of two shovel designs. Applied Ergonomics, 24(5), 306-312. [5] McGorry, R. W., Dempsey, P. G., & Leamon, T. B. (2003). The effect of technique and shaft configuration in snow shoveling on physiologic, kinematic, kinetic and productivity variables. Applied Ergonomics, 34(3), 225-231. [6] Lewinson, R. T., Rouhi, G., & Robertson, D. G. E. (2014). Influence of snow shovel shaft configuration on lumbosacral biomecahnics during a load-lifting task. Applied Ergonomics, 45(2), 234-238 [7] Dowell, B., & Gscheidle, G. (2003). The Evolution of Anthropometrics and User Control: The Science and Research Behind the Mirra 2 Chair. Retrieved February 21, 2016, from http://hermanmiller.com/research/solution-essays/theevolution-of-anthropometrics-and-user-control.html [8] Clauser, C. E., McConville, J. T., & Young, J. W. (1969). Weight, Volume, and Center of Mass of Segments of the Human Body (Tech. No. AMRL-TR-69-70). Wright-Patterson AFB, Ohio: USAF Aerospace Medical Research Laboratory. 15 8. Appendices 8.1 Appendix A - Determination of Body Segment Weights Weights, volumes, and center of masses for the human body and body segments are presented in [8]. The body mass segments in this study will be used to determine the mass of each segment as a percentage of the total body weight. Using this percentage, the body mass segments of any body type can be extrapolated proportionally. Similar extrapolation can be accomplished for body segment lengths. Table 2 presents body lengths and heights while Table 3 presents body segment masses. Table 2: Average body heights and lengths [8] Total height Head + trunk Suprasternale height Chest breadth Trunk Arm Length (cm) 172.72 81.92 141.05 33.23 57.89 77.45 Percentage of Height 100% 47.43% 81.66% 19.24% 33.52% 44.84% Table 3: Average body segment weights [8] Total body Head + trunk Total arm Trunk Upper arm Forearm and hand Mass (kg) 65.606 38.061 3.216 33.312 1.730 1.483 Percentage of Weight 100% 58.01% 4.901% 50.78% 2.64% 2.26% The weights and heights presented in Table 4 will be used in this study. Table 4: Average heights and weights for across the adult human population Body Type Light 5th percentile female [7] Average 5th percentile female [7] Lewinson Average [6] Average 95th percentile male [7] 16 Height (in) 59.00 59.00 69.69 74.00 Weight (lbs) 88 113 162 246 74.00 297 Heavy 95th percentile male [7] To calculate the respective heights and lengths across the adult population, the height from Table 4 will be multiplied by the segment height of length from Table 2 and then divided by the total height from Table 2. These results are shown in Table 5. Table 5: Body segment lengths and heights across the human adult population Total height Head + trunk Suprasternale Chest breadth Trunk Arm Light 5th Average Lewinson percentile 5th Average female percentile female 59.00 in 59.00 in 69.69 in 27.98 in 27.98 in 33.05 in 48.18 in 48.18 in 56.91 in 11.35 in 11.35 in 13.41 in 19.77 in 19.77 in 23.36 in 26.46 in 26.46 in 31.25 in Average 95th percentile male 74.00 in 35.10 in 60.43 in 14.24 in 24.80 in 33.18 in Heavy 95th percentile male 74.00 in 35.10 in 60.43 in 14.24 in 24.80 in 33.18 in To calculate the respective body segment weights across the adult population, the height from Table 5 will be multiplied by the segment height of length from Table 3 and then divided by the total height from Table 3. These results are shown in Table 6. Table 6: Body segment weights across the human adult population Total body Head + trunk Total arm Trunk Upper arm Forearm + hand Light 5th Average Lewinson percentile 5th Average female percentile female 88.00 lbs 113.00 lbs 162.00 lbs 51.05 lbs 65.56 lbs 93.98 lbs 4.31 lbs 5.54 lbs 7.94 lbs 44.68 lbs 57.38 lbs 82.26 lbs 2.32 lbs 2.98 lbs 4.27 lbs 1.99 lbs 2.55 lbs 3.66 lbs 17 Average 95th percentile male 246.00 lbs 142.72 lbs 12.06 lbs 124.91 lbs 6.49 lbs 5.56 lbs Heavy 95th percentile male 297.00 lbs 172.30 lbs 14.56 lbs 150.80 lbs 7.83 lbs 6.71 lbs