- Email: [email protected]
Handle displacement and operator responses to pneumatic nutrunner torque buildup
ARTICLE IN PRESS
Applied Ergonomics 37 (2006) 367–376 www.elsevier.com/locate/apergo
Handle displacement and operator responses to pneumatic nutrunner torque buildup Jia-Hua Lin, Raymond W. McGorry, Patrick G. Dempsey, Chien-Chi Chang Liberty Mutual Research Institute for Safety, Hopkinton, MA, USA Received 23 March 2004; accepted 10 May 2005
Abstract The objective of this study is to investigate the workstation and tool effects on the responses of the powered hand tool operator reacting against the impulsive reaction forces that may be associated with upper extremity musculoskeletal disorders. The study demonstrated a means of direct measurement of force at the interface between the tool and the operator. Fifteen experienced male operators performed three independent work configurations: pistol grip and right angle tools used on the horizontal surface and pistol grip tools used on the vertical surface, in the laboratory. A full factorial experiment consisting of 36 conditions was designed to examine the effects of working height, distance, tool, and fastener joint hardness on handle displacement and grip forces. The results indicate that operator responses were affected by different factors depending on the work configuration. When pistol grip tools were used on the vertical surface, the mean handle displacement decreased from 9.91 to 7.31 as the working height increased from 30 cm below shoulder to 30 cm above shoulder. When right angle tools were used, the greatest handle displacement (51.1 mm) and grip force (84.7% MVC) during torque reactions were measured at 30 cm below elbow and 40% forward reach away from the operator. This study provides quantitative information that can be used for workstation design and tool selection to reduce the torque reaction experienced by powered nutrunner operators. r 2005 Elsevier Ltd. All rights reserved. Keywords: Assembly tools; Occupational safety; Occupational biomechanics
1. Introduction An early survey (Myers and Trent, 1988) showed that among hand-tool related injuries in the United States in 1983 (129,399 cases), 22% were caused by powered hand tools. This trend is comparatively constant throughout recent years (Bureau of Labor Statistics, 2004). Struck by or struck against objects, and over-exertion were the two main exposures leading to powered hand tool related accidents. Unlike manual tools, in which the operator can adjust the power output (force or torque), powered hand tools use power sources (air, electricity) that sometimes require demands that exceed the capacity of human operators. These tools can potentially produce undesired forces that are a risk factor for upper extremity muscuCorresponding author: 71 Frankland Rd., Hopkinton, MA 01748, USA. Tel.: +1 508 497 0228. E-mail address: [email protected] (J.-H. Lin).
0003-6870/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.apergo.2005.05.013
loskeletal disorders (UEMSDs). A better understanding of the operator–tool–task interface may help reduce the associated morbidity. Fastener-driving tools are commonly used in manufacturing and assembly jobs. At the beginning of a fastening process, the operator holds and supports the tool. From the instant the fastener joint starts to be compressed, the tool starts to build up torque. When the tool torque overcomes the operator, the hand and arm exert a force to stabilize the tool in a direction opposite to the tool movement. Once the torque exceeds operator capacity, the tool becomes unstable and displacement occurs. Radwin et al. (1989) found that for powered hand tool operation, forearm muscle electromyography (EMG) during the torque buildup phase was greater than other operation phases. The operators associated greater handle displacement resulting from the impulsive torque reaction with increased discomfort and perceived exertion (Kihlberg et al., 1993; Lindqvist, 1993). Kilhberg et al. (1995)
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
368
concluded that for a right angle tool to be accepted by 90% of the operators, the tool should cause handle displacement of less than 30 mm. In addition to handle displacement, grip force subjectively rated by the operators during tool use was correlated to the overall tool preference (Armstrong et al., 1989). However, these studies did not identify the factors that affect the tool handle displacement and grip force. For fastener-driving tasks, work location or posture has been associated with overall operator discomfort (Armstrong et al., 1989; Ulin et al., 1992). These previous studies used psychophysical measures or subjective assessments to identify the operator discomfort that resulted from the general screw-driving experience. Nevertheless, it is uncertain whether discomfort was associated with posture alone, or with other aspects of the screw-driving task that varied with posture, such as the displacement caused by the impulsive tool reaction torque (Freivalds and Eklund, 1993; Kihlberg et al., 1993; Lindqvist, 1993) or required grip force. Workstation design was found to affect operator hand–arm stability (Oh and Radwin, 1997). It is plausible that participants preferred different tool types in different work locations and orientations because of different mechanical advantages for reacting against impulsive forces. Lin et al. (2003b) developed a single–degree-of-freedom model for common powered hand tool shapes. The human operator was represented as a single mass, spring, and damper system. It was demonstrated that the model parameters were greatly affected by working posture. These mechanical characteristics possessed by human operators directly affect the exposure resulting from the impulsive tool torque. In addition to posture or workstation effect on powered tool use, ergonomics research has provided some guidance regarding powered hand tool reaction torque using criteria such as subjective ratings (Freivalds and Eklund, 1993; Kihlberg et al., 1993), ground reaction force (Kihlberg et al., 1993, 1995), handle displacement (Kihlberg et al., 1993, 1995; Lin et al., 2001, 2003a; Oh et al., 1997), and EMG or grip force derived from EMG (Armstrong et al., 1999; Forsman et al., 2002; Freivalds and Eklund, 1993; Radwin et al., 1989). However, those studies only considered limited work locations, tool shapes, or tasks. Therefore, this study aims to integrate the most common factors
involved in powered hand tool usage: the effects of tool characteristics, task (fastener joint hardness), and posture (working height, working distance) and examine their effects on the handle displacement resulting from torque reactions for three work configurations: pistol grip tools used on the vertical and horizontal surfaces, and right angle tools used on the horizontal surface. Furthermore, ISO 6544 (International Organization for Standardization, 1981) states that there is no known device for measuring force at the interface between the tool and the operator. This study attempts to provide a means of such direct measurement using an established grip force sensor (McGorry, 2001). Thus the results derived in the study can provide further insights for workstation design and tool selection that may reduce risks for UEMSDs. 2. Method 2.1. Participants Fifteen right-handed male participants were recruited to participate in the experiment after giving their informed consent. They were all free of musculoskeletal disorders. All of the participants were experienced in assembly, manufacturing, maintenance, or other industries where powered tool use was part of their jobs. Their professional experience was validated using a custom proficiency questionnaire. In this questionnaire they were asked about their current or past jobs, tool use frequency and experience, understanding of fastener-joining procedures, and to name brands of the tools they had used. Their average experience with fastener-driving tools (e.g., screwdrivers, nutrunners) was 14.73 years. Anthropometric information of these participants is summarized in Table 1. 2.2. Apparatus The task joint and reaction torque measurement were established according to ISO 6544 guidelines (International Organization for Standardization, 1981). Four commercially available pneumatic nutrunners with automatic shutoff mechanisms were used in the experiment. They were two pistol grip tools (PA and PB) and two right angle tools (RA and RB) (Fig. 1). Their specifications are listed
Table 1 Anthropometry measurements of the experiment participants
Mean SD Min Max a
Age (years)
Weight (kg)
Height (cm)
Elbow heighta (cm)
Shoulder heightb (cm)
Forward Reachc (cm)
Exp (years)
41.2 7.4 28 56
84.18 12.38 61.7 104.4
176.87 4.98 168 185
112.73 3.30 106.5 118
146.23 3.97 140.5 154
83.76 5.08 71.4 92
14.73 7.60 5 30
Elbow height was measured from the floor to the lateral epicondyle with elbow bending at 901 Shoulder height was measured from the floor to the center of the gleno-humeral joint. c Forward reach was measured from ankles to horizontally forward pinching finger tips according to Gordon et al. (1989). b
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
in Table 2. The two pistol grip tools had similar torque setting with varying motor speed that affected the torque buildup times. The two right angle tools had similar motor speed but different handle length and torque setting. They were driven by 630 kPa (6.3 bar) of air pressure. For the two right angle tools, a tool balancer was used to partially support the weight. Joint hardness is characterized by the time or the spindle rotation required from the joint initial contact to reaching the preset tool torque. The softer the joint, the longer time or the greater spindle rotation the tool takes to reach its preset torque. Two fastener joint simulators, hard joint and soft joint, were used to emulate actual assembly tasks in the current study. Joint hardness was manipulated using different numbers of Belleville spring washers. Fig. 2 shows an example of the effect of joint hardness on the tool torque profiles. The joint simulators were placed on a height-adjustable platform that allowed various task orientations and working heights (Fig. 3). Each tool was equipped with a simulated handle. The handle has a stainless-steel core instrumented with strain gauges that allow measurement of the grip forces and reactive moments acting on the handle (McGorry, 2001).
369
Molds were made from the nutrunner handles and replicas were cast in polyurethane, cut, and mounted to the handle core. For the pistol grip tools, the handle was positioned posterior to the original handle on the extension of the spindle (Fig. 3A). For the right angle tools, the handle was affixed parallel to its original, on the plane of the tool torque (Fig. 3C). The trigger was pressed by the subject to initiate the tool operation and data collection simultaneously. A torque transducer was positioned between the joint simulator and the tool bit to monitor the tool torque output. A cable potentiometer with a 10 cm measurement range was used to record the handle position. The potentiometer, equipped with a recoil spring, was mounted on the joint simulator, and the cable was attached to a bar extending from the body of the tool. The sampling rate for all sensors was set at 1000 Hz. 2.3. Experiment design and procedures Three common work configurations were treated as independent experiments and tested in three random sessions: pistol grip tools on a horizontal surface Torque
7 6
Torque (Nm)
5 4 3 2 1 0
Fig. 1. Four pneumatic nutrunners adapted with mounting hardware used in the current experiment.
100
200
300 Time
400
500
Fig. 2. Torque buildup profiles for tool PA. The solid line represents a hard joint and the dotted line a soft joint. The double-arrow lines indicate the torque buildup phase.
Table 2 Specifications of the tools used in the current study and their torque buildup characteristics on the two joints. Tool
PA PB RA RB a
Handle lengtha (mm)
65 65 260 350
Weight (kg)
1.4 1.7 1.6 2.4
Torqueb (Nm)
6 5.6 16 23.8
Measured from the spindle axle to the handgrip center. Preset torque levels for tool shutoff. c Spindle angle rotating during torque buildup. b
Free Speed (rpm)
555 570 400 400
Buildup Time (ms)
Buildup anglec (deg)
Hard joint
Soft joint
Hard joint
Soft joint
60 44 307 685
400 254 1671 3647
28 29 145 216
547 473 1478 2126
ARTICLE IN PRESS 370
J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
(Fig. 3A), pistol grip tools on a vertical surface (Fig. 3B), and right angle tools on a horizontal surface (Fig. 3C). Only one session was tested on a given day. For each work configuration, the full factorial experiment consisted of independent factors of working height (V), working distance (H), tool (Tool), and joint hardness (Joint) (Table 3). Working heights and distances were chosen to cover a reasonable working range and to avoid awkward postures. These working posture assignments were determined based on individual operator anthropometric measurements. The dependent measures were grip force and handle displacement for each trial. In order to normalize the grip force measurement, maximum voluntary contraction (MVC) was recorded as reference at each assigned posture using the respective handle orientation in each experiment session. Each subject exerted his maximum grip on the instrumented handle at the specific posture for 5 s according to the Caldwell regimen (Caldwell et al., 1974). The period between MVC measurements was at least 2 min to avoid fatigue. The grip force at the middle 2 s was averaged to obtain the respective MVC values. Participants practiced with all the tools prior to the experiment so they could become accustomed to the tasks. The test conditions were presented in a random order to reduce the potential of confounding associated with learning or fatigue effects. Two repetitions were performed at each condition. At their self-selected pace, the participants were instructed to operate the tools using only their right hands, and to initiate a trial by pressing the trigger. Each trial consisted of a short run-down phase before the fastener started to compress. Each subject performed a total of 72 trials in a session. Three dependent variables were evaluated for each trial. Handle displacement due to torque reaction was defined as the net change of the handle position between the start and end of the torque buildup. For pistol grip tools, angular units were reported because of the nature of the forearm rotation motion. This value was converted from the linear displacement measured using the positional potentiometer. For right angle tools, linear displacement at the center of the handgrip was reported because the motion was similar to a push–pull task. Normalized grip force prior to the torque reaction (Grippre) was the grip force averaged for the 50 ms prior to the start time and divided by the corresponding MVC at the given posture. Normalized grip force during torque buildup (Gripbu), was averaged for the whole torque buildup period, and divided by the corresponding MVC for the given posture. These three measures were then averaged for the two repetitions at each condition. 2.4. Data analysis Fig. 3. Experiment setup: (A) tool PB on the horizontal surface; (B) tool PA on the vertical surface; (C) tool RB on the horizontal surface. The subject was holding the instrumented handle.
All signals were first smoothed using a 5 ms centered moving average window. Operator responses during torque buildup were the focus of this experiment, therefore it was necessary to define the start and end time frames for each
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
371
Table 3 Factors studied in the present experiment Tool configuration
Working height (V)a (cm)
Working distance (H)a (% of forward reach distance)
Jointb
Tool
Pistol grip/horizontal surface
Elbow30 Elbow10 Elbow+10 Shoulder30 Shoulder Shoulder+30 Elbow30 Elbow Elbow+30
25 50 75 25 50 75 40 60 80
Hard, Soft
PA, PB
Hard, Soft
PA, PB
Hard, Soft
RA, RB
Pistol grip/vertical surface Right angle/horizontal surface a
Defined at the handgrip center: the working height above the floor, and the working distance horizontally from the ankles. According to ISO 6544 (1981).
b
7
Tool torque (Nm)
6
Start
were treated as a random factor. In addition, two-way ANOVA was used to compare operator grip force prior to and during tool torque buildup.
End
Rundown
Buildup
Shutoff
5 4
3. Results
3 2
3.1. Pistol grip tools on the horizontal surface
1
120
28
100
26
80
24
60
22
40
20
20
18
0 300
600
Hand Position (degree)
Grip Force (N)
0
16 900
Time (ms)
Fig. 4. A typical trial for pistol grip tool PA on the soft joint. The upper thick trace is the tool torque; the lower thick trace is the handle position, and the thin trace is the grip force. Dash lines are the start and end of the torque buildup.
trial post hoc. A trial started with the rundown phase where there was little resistance. The start time of the impulsive torque buildup was identified when the tool torque, increasing due to the compression of the joint, reached beyond the threshold of 2.5% of the peak torque. The end time was when the torque reached its peak with shutoff phase followed (Fig. 4). Analyses of variance (ANOVA) for the main factors and their two-way interactions on dependent variables were performed using SPSS for Windows v. 11. Participants
The effects of main experimental factors and their twoway interactions are summarized in Table 4. For pistol grip tools used on a horizontal surface, the mean handle displacement due to torque reaction decreased from 8.81 when the tools were held at 25% of the reach distance, to 7.31 when the tool was held at 75% of the reach distance, a 17% decrease. Tool PA, which resulted in greater torque impulses due to greater torque setting and longer buildup times, caused more handle displacement (9.8177.21) than tool PB (6.1175.81) did. Joint was not a significant factor for handle displacement. However, the interaction between Tool and Joint was significant, as demonstrated in Fig. 5. For tool PA, the hard joint caused greater mean handle displacement (10.61) than the soft joint (9.01), while for tool PB, the effect was reversed (4.71 and 7.51, respectively). Normalized grip force during torque buildup (Gripbu) was significantly greater than during rundown phase (Grippre) (t (5 3 6) ¼ 18.43, po.001), although they were highly correlated (Pearson r ¼ 0.98, po.001). Both grip force measurements were significantly affected by working height, distance, and their interaction. The least Grippre and Gripbu occurred at 10 cm above elbow and 50% forward reach away (59.6 and 63.3% MVC, respectively). The greatest Grippre was found at 10 cm above elbow and 75% forward reach away (68.9% MVC); while the greatest Gripbu was measured at 30 cm below elbow and 25% forward reach away (73.2% MVC). On average, participants gripped tool PB harder (Grippre ¼ 67.7% MVC, Gripbu ¼ 70.9% MVC) than tool PA (Grippre ¼ 62.7% MVC, Gripbu ¼ 68.0% MVC). During the joint rundown phase, the operators held the handle with similar level of exertion (65.2725.8%) regardless of
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
372
Table 4 Analysis of variance results for the effects of all main factors and their two-way interactions on handle displacement (Disp), normalized grip force before (Grippre) and during torque buildup (Gripbu) for three work configurations. Asterisks () denote factors of statistical significance (po0.05) Factor
Pistol grip/horizontal
Working height (V) Working distance (H) Tool Joint VH V Tool V Joint H Tool H Joint Tool Joint
Right angle/horizontal
Disp
Grippre
Gripbu
Disp
Grippre
Gripbu
Disp
Grippre
Gripbu
Tool PA
Tool PB
10 8 6 4 2
Normalized grip force (%MVC)
Handle displacement (degrees)
12
Pistol grip/vertical
90%
Working height:
Shoulder
Shoulder+30
60%
30%
0% 25%
0
Shoulder-30
50%
75%
Working distance (%forward reach)
Hard
Joint
Soft
Fig. 5. The effects of tool and joint on handle angular displacement (mean and one standard error) due to impulsive torque buildup when pistol grip tools were used on a horizontal surface.
the joint hardness. During torque buildup, however, the operators increased their Gripbu more on soft joint (71.1728.0%) than on hard joint (67.8727.8%). 3.2. Pistol grip tools on the vertical surface When pistol grip tools were used on the vertical surface, the mean handle displacement decreased from 9.91 (7 8.01) to 7.31 (7 5.51) as the working height increased from 30 cm below shoulder to 30 cm above shoulder. The displacement resulting from tool PA (9.8177.21) was greater than by tool PB (6.1175.81). In addition to Tool and Joint factors, the Tool Joint interaction was also significant. When tool PA was used, the operators allowed more displacement on the hard joint (12.3175.41) than on the soft joint (8.7179.01), which was contrary to the finding for tool PB where greater mean displacement was recorded on the soft joint (7.21) than on the hard joint (6.71). The MVC recordings of one subject in this experiment session were failed due to hardware difficulty and discarded. Therefore, the analyses for normalized grip
Fig. 6. The effects of work posture on normalized grip force (mean and one standard error) during impulsive tool torque buildup when pistol grip tools were used on a vertical surface.
force were performed on the remaining 14 participants. Grippre and Gripbu were highly correlated (r ¼ 0.99) and both were affected by working height, working distance, and their interactions (their effects on Gripbu were demonstrated in Fig. 6). The greatest level of force (Grippre ¼ 69.0% MVC and Gripbu ¼ 73.2% MVC) was exerted by the operators at 30 cm below shoulder, and 75% of reach distance. The least grip force was exerted when operators held the tools at 30 cm above shoulder and at 50% of reach distance (Grippre ¼ 52.1% MVC and Gripbu ¼ 55.8% MVC). Participants exerted significantly higher mean Grippre for tool PB (65.1%) than for PA (61.9%), but Tool had no significant effect on Gripbu. Both Grippre and Gripbu were not affected by Joint for this work configuration. 3.3. Right angle tools on the horizontal surface All main factors were significant for hand displacement when right angle tools were used on the horizontal surface. The effects of working distance and height on handle
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376 Table 5 Mean and standard deviation (S.D.) of hand displacement (Disp) and normalized grip force (Gripbu) during torque buildup for various working postures when right angle tool RB was used horizontally on the hard joint V (cm)
Elbow30
Elbow
Elbow+30
H (% forward Disp (mm) reach) Mean SD
Gripbu (% MVC) Mean
SD
40 60 80 40 60 80 40 60 80
84.7 64.4 58.1 63.4 58.4 57.6 62.1 54.6 52.9
57.8 21.2 31.3 26.8 19.1 17.5 22.1 20.5 19.4
51.1 39.5 36.1 38.9 34.3 29.5 42.6 35.8 36.6
22.2 19.0 12.2 14.8 18.5 12.8 13.2 18.4 22.2
displacement and grip force for tool RB used on the hard joint is summarized in Table 5. For this work configuration, the operators allowed more hand displacement on the hard joint (32.5716.2 mm) than on the soft joint (14.4712.6 mm). Tool RB, which delivered greater torque output, caused more displacement (26.9719.9 mm) than tool RA did (20.1712.9 mm). Among the three working heights, participants allowed the greatest mean displacement at 30 cm below elbow for both joints (37.1 mm for the hard joint and 15.5 mm for the soft joint). However, the least displacement for the hard joint was measured at elbow height (29.8 mm); while for the soft joint it was measured at 30 cm above elbow (13.3 mm). Among the three working distances, participants allowed the greatest mean displacement at 40% forward reach away for both joints (38.6 mm for the hard joint and 16.6 mm for the soft joint). However, the least displacement for the hard joint was measured at 80% reach (28.0 mm); while for the soft joint it was measured at 60% reach (12.0 mm). Gripbuwas significantly greater than Grippre by 14.9% (t(539) ¼ 18.875, p o0.001) and these two measures were highly correlated (r ¼ 0.94, p o0.001). The operators were found to exert the greatest Gripbu (73.7740.7%) when tools were held at 30 cm below elbow and at 40% of reach distance. The least Gripbuwas exerted (54.3721.8%) when tools were held at 30 cm above elbow and at 80% of reach distance. Participants exerted significantly higher mean Grippre and Gripbu for tool RA (56.5 and 63.1% MVC, respectively) than for RB (50.4 and 59.8% MVC, respectively). Participants exerted significantly less mean Grippre for the soft joint (49.9% MVC) than for the hard joint (56.5% MVC); but Joint had no significant effect for grip force during torque buildup. 4. Discussion For right angle tools used on a horizontal surface, the resultant peak hand displacement was less when the tools were held farther away from the operator (Table 5). This
373
trend corresponds to a previous finding on right angle tool kinematics (Oh and Radwin, 1997). Furthermore, the observed displacement was greater when the tool was held 30 cm below the elbow, and was less when the tool was held at elbow height. It is hypothesized that working height and distance may affect the biomechanics of the upper extremity, and changing the capability to react against torque reactions. As the forward reach distance decreased, the upper arm started to bend, potentially allowing for more displacement of the upper extremity linkage system. Table 5 also demonstrates that the handle displacement was the greatest when the right angle tool was held at 30 cm below elbow. The right angle tool delivered its reaction torque to the operator as a horizontal force pulling the arm away from the body. The moment arm, the perpendicular distance from the shoulder to the tool handle, was the greatest as the vertical location of the tool handle decreased to 30 cm below elbow. The impulsive torque was amplified for the operator due to the longer moment arm compared to when the tool was held higher. In addition, Lin et al. (2003b) found that operator upper extremity joint stiffness increased significantly with increasing working height in the range from 80 to 140 cm above the floor for right angle tool operations. Therefore, analogous to Hooke’s law, for the same torque impulse input, less displacement would be expected for greater stiffness when the tool was held higher. When pistol grip tools were used on a vertical surface, resultant angular handle displacement was affected by work posture and it decreased from 9.91 to 7.31 as the working height increased from 30 cm below shoulder to 30 cm above. The difference may seem without practical significance. However, considering the accumulative effect for the operator, the difference may be meaningful. Lin et al. (2001) constructed a mechanical model to represent the human operator during pistol grip power tool use on a vertical surface. They found that the hand–arm stiffness, which is an important mechanical property that resists motion, increased as the working height increased from 93 to 190 cm.These working heights approximately correspond to the levels used in the current study. Therefore, the observed trend may be attributed to the operator hand–arm stiffness associated with various postures. Ulin et al. (1993) measured psychophysical responses for screw driving tasks when different shapes of tool were used at various work locations. It was found that for right angle tools, as the working height decreased from 140 to 89 cm, the perceived exertion increased. This corresponded to our findings that the handle displacement and grip force increased as the working height decreased from 30 cm above shoulder to 30 cm below (Table 5). However, for pistol grip tools used on a vertical surface, Ulin et al. (1992) observed an increasing perceived exertion as the working height increased from 114 to 191 cm. Our findings show that the handle displacement and grip force (as shown in Fig. 6) were less when the tool was held higher. In their study the participants used the tool only at a fixed distance, and the tool torque setting (3.2 Nm) was much less than in
ARTICLE IN PRESS 374
J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
the present study. Also, the screws were driven on a perforated sheet metal where a very hard joint was formed, hence very short buildup times and small impulses were expected. Considering these differences, it was possible that the exertion perceived by their participants might reflect more the posture than the torque reactions. It was demonstrated through mechanical analysis that tools of the same shape but of varied geometries, such as handle length, or varied torque settings impose different static physical demands on the tool operator (Lin et al., 2003c; Radwin et al., 1995). Current results further show that tools of the same shape resulted in different dynamic responses. Right angle tool RB had greater torque setting and longer handle than tool RA. It is intuitive that greater torque caused greater handle displacement. Additionally, the longer the handle is, the greater the translational distance the hand has to travel for a given angular displacement resulting from the torque reaction. Although tool RB had a longer handle which provided better leverage for the operator, it can be derived from Table 2 that the equivalent peak reaction force transmitted to the hand at the handgrip was greater for tool RB (68 N) than tool RA (61.5 N). All of these factors resulted in greater handle displacement for tool RB. Pistol grip tools PA and PB delivered similar torque outputs. Their motor running speed governed their behaviors on the two joints and resulted in different torque buildup times (Lin et al., 2003c), and hence different magnitudes of torque impulse. These results suggest that tool selection should not only consider tool performance and task requirements, but also the avoidance of unnecessary physical exposure such as handle displacement or grip force. Normalized grip force prior to torque buildup, Grippre, is the bracing force that participants develop before torque buildup (Armstrong et al., 1999). This grip force was mostly affected by posture (working height, distance, and their interaction) for all three work configurations in the experiment (Table 4). It is not conclusive from the current study if the operators adjusted or determined the grip force for each posture, or if the posture limited the grip force. The high correlation for grip force before and during toque buildup suggests that operators adjusted their grip force to react against the impulsive torque buildup. Handle displacement and grip force during impulsive torque reaction were anticipated to be greater when tools were used on the soft joint because the torque impulse (area under the curve of Fig. 2) was greater than the hard joint. Our results show that the mean handle displacement was 7.21 for the soft joint and 6.71 for the hard joint when tool PB was used on the vertical surface. For this tool configuration, Lin et al. (2001) tested two different levels of joint hardness for the pistol grip tool and showed that hard joints resulted in less displacement than soft joints. Magnitude of torque impulses, which is determined by the peak torque level and joint hardness, was correlated with subjective ratings of perceived exertion (Freivalds and Eklund, 1993). Interestingly, for tool PA, the effect of joint
hardness was reversed where more displacement was allowed on the hard joint (12.3175.41) than on the soft joint (8.7179.01). Oh and Radwin (1997) observed a decrease in handle displacement when torque buildup time increased beyond a threshold. Armstrong et al. (1999) found a similar trend for in-line screwdrivers. Considering that the torque buildup times for PB were much less than for PA (Table 2), it is understandable that the effect of joint hardness interacted with tool. It was speculated that when buildup times were longer, the participants were able to respond by exerting enough force timely to counteract with the torque impulses. For certain right angle tool trials, the tool could be maintained in equilibrium by the operator and the tool displacement was minimized. A further look into the moment generated by the operators could improve understanding of the dynamics of the tool-joint-operator system. Current findings provide quantitative measures on handle displacement and grip force at various work locations and can used as criteria when designing workstation or selecting tools. For example, it is shown that when pistol grip tools are to be used on a vertical surface, higher location may be preferred because of less handle displacement and less grip force. When selecting a pistol grip tool to be used on a horizontal surface, tool PB might be favored because it caused less handle displacement. In addition to relative comparisons between factors, if standards or predefined thresholds are available, the current findings can be used to identify acceptable work conditions for a group of tool operators. However, as with any job or workstation modification, all of the ergonomic implications of the redesign must be considered. In the example above, when using the pistol grip tool on a vertical surface, greater work height would be expected to reduce exposure to tool handle displacement. But the advantage of reduced exposure to handle displacement must be carefully weighed against the potential increase in exposure due to increased shoulder elevation. Kihlberg et al. (1995) tested four right angle nutrunners and concluded that for a tool to be accepted by 90% of the operators, it should produce handle displacement of less than 30 mm.The current results demonstrated that for a given tool running on a given joint, using different working postures resulted in different handle displacements. For example, when tool RB was used on a hard joint held at elbow30 cm high and 40% of reach distance, the mean handle displacement was 51.1 mm (Table 5). This would exceed the proposed 30-mm limit. Using the same tool on the same hard joint, if the task was raised to elbow height, and the operator stood at 80% of reach distance, the displacement decreased to 29.5 mm, which would be within the proposed limit. Therefore, right angle tool workstation configuration can be arranged to minimize handle displacement. Currently there is no similar handle displacement exposure limit or recommendation for pistol grip tools, and there is no grip force information regarding powered tool operations. Future research is warranted
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
because such establishments can provide better guidelines for workstation design and tool selection. Grip force during torque buildup was affected by joint hardness only for pistol grip tools used on horizontal surface but not for the other two configurations. Originally, it was anticipated for all work configurations because the greater the torque reaction was, the greater grip strength was needed to overcome the reaction. Participants in the current study were professional tool users. We therefore speculate that because they were very familiar with the task settings and the tools, the grip force they exerted did not vary significantly for the joint. Lin et al. (2003a) tested powered tool operations using six university students and reported that for right angle tools used at a soft joint on the horizontal surface, the handle displacement was 59.11 and 79.02 mm for two work locations. Many factors were different, but the displacement measured for the soft joint in the current study was much less (14.4712.6 mm). User experience might contribute to the difference. Malchaire et al. (1996) identified working experience as a predictor for wrist musculoskeletal disorder prevalence. Ha¨kka¨nen et al. (2001) found that new workers had a higher rate of sick leave due to UEMSDs than their experienced colleagues did in a trailer assembly factory. Similarly, Park et al. (1996) found that the incidence rate ratios in UEMSDs are higher among employees who were newly hired or given new assignments among various assembly tasks. Further, experience is considered a factor influencing hand tool performance (Rockwell and Marras, 1986) because of different techniques adapted by experienced and novice users. Due to these differences, it is speculated that for the similar powered hand tool usage, novice users would have different handle displacement and grip force possibly leading to higher exposure. For example, their responses could be affected by task joints because the novice users do not anticipate different magnitudes of torque impulses and hence do not adjust their techniques accordingly. More handle displacement is expected to be imposed on the novice tool operators and result in more unwanted exposures. Thus, there is a need to examine tool usage among novice users.
5. Conclusion This study investigates how nutrunner operators respond to the impulsive reaction force for various working postures, tasks, and tools. This is the first study to instrument a grip sensor for direct measurement of the force between the tool and the operator. The results including the handle displacement and grip force associated with various torque reactions provide quantitative insight to understand tool use. The information makes it possible to reduce physical exposures that are associated with UEMSDs through tool selection, and task and workstation design.
375
References Armstrong, T.J., Punnett, L., Ketner, P., 1989. Subjective worker assessments of hand tools used in automobile assembly. Am. Ind. Hyg. Assoc. J 50, 639–645. Armstrong, T.J., Bir, C., Foulke, J., Martin, B., Finsen, L., Sjøgaard, G., 1999. Muscle responses to simulated torque reactions of hand-held power tools. Ergonomics 42, 146–159. Bureau of Labor Statistics, Department of Labor, 2004. Case and demographic characteristics for work-related injuries and illnesses involving days away from work. http://www.bls.gov/iif/oshcdnew.htm, (accessed 25.03.04). Caldwell, L.S., Chaffin, D.B., Dukes-Dobos, F.N., Kroemer, K.H.E., Laubach, L.L., Snook, S.H., et al., 1974. A proposed standard procedure for static muscle strength testing. Am. Ind. Hyg. Assoc. J 35, 201–206. Forsman, M., Cyre´n, H., Mo¨ller, T., Kadefors, R., Mathiassen, S.E., 2002. Activity in five muscles during joint securing using pneumatic nutrunners. Int. J. Ind. Ergon. 29, 21–32. Freivalds, A., Eklund, J., 1993. Reaction torques and operator stress while using powered nutrunners. Appl. Ergon 24, 158–164. Gordon, C. C., Churchill, T., Clauser, C. E., Bradtmiller, B., McConville, J. T., Tebbetts, I., et al. 1989. 1988 Anthropometric Survey of US Army Personnel: Methods and Summary Statistics (No. NATICK/ TR-89/044). Anthropology Research Project, Inc. Yellow Springs, OH Ha¨kka¨nen, M., Viikari-Juntura, E., Martikainen, R., 2001. Job experience, work load, and risk of musculoskeletal disorders. Occup. Environ. Med. 58, 129–135. International Organization for Standardization. 1981. ISO 6544 Handheld pneumatic assembly tools for installing threaded fastener— reaction torque and torque impulse measurements. International Organization for Standardization, Geneva. Kihlberg, S., Kjellberg, A., Lindbeck, L., 1993. Pneumatic tool torque reaction: reaction forces, displacement, muscle activity and discomfort in the hand–arm system. Appl. Ergon. 24, 165–173. Kihlberg, S., Kjellberg, A., Lindbeck, L., 1995. Discomfort from pneumatic tool torque reaction: acceptability limits. Int. J. Ind. Ergon. 15, 417–426. Lin, J.-H., Radwin, R.G., Richard, T.G., 2001. Dynamic biomechanical model of the hand and arm in pistol grip power handtool usage. Ergonomics 44, 295–312. Lin, J.-H., Radwin, R.G., Richard, T.G., 2003a. Handle dynamics predictions for selected power hand tool applications. Hum. Factors 45, 645–656. Lin, J.-H., Radwin, R.G., Richard, T.G., 2003b. A single degree-offreedom dynamic model predicts the range of human responses to impulsive forces produced by power hand tools. J. Biomechan 36, 1845–1852. Lin, J.-H., Radwin, R.G., Richard, T.G., Fronczak, F.J., 2003c. Forces associated with pneumatic power screwdriver operation: statics and dynamics. Ergonomics 46, 1161–1177. Lindqvist, B., 1993. Torque reaction in angled nutrunners. Appl. Ergon. 24, 174–180. Malchaire, J.B., Cock, N.A., Robert, A.R., 1996. Prevalence of musculoskeletal disorders at the wrist as a function of angles, forces, repetitiveness and movement velocities. Scand. J. Work, Environ. Health 22, 176–181. McGorry, R., 2001. A system for the measurement of grip forces and applied moments during hand tool use. Appl. Ergon. 32, 271–279. Myers, J.R., Trent, R.B., 1988. Hand tool injuries at work: a surveillance perspective. J. Safety Res 19, 165–176. Oh, S., Radwin, R.G., 1997. The effects of power hand tool dynamics and workstation design on handle kinematics and muscle activity. Int. J. Ind Ergon. 20, 59–74. Oh, S., Radwin, R.G., Fronczak, F.J., 1997. A dynamic mechanical model for hand force in right angle nutrunner operation. Hum. Factors 39, 497–506.
ARTICLE IN PRESS 376
J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
Park, R.M., Krebs, J.M., Mirer, F.E., 1996. Occupational disease surveillance using disability insurance at an automotive stamping and assembly complex. J. Occup. Environ. Med. 38, 1111–1123. Radwin, R.G., VanBergeijk, E.M., Armstrong, T.J., 1989. Muscle response to pneumatic hand tool torque reaction forces. Ergonomics 32, 655–673. Radwin, R. G., Oh, S., Fronczak, F. J., 1995. A mechanical model of hand force in power hand tool operation. In: Proceedings of the Human Factors and Ergonomics Society 39th Annual Meeting. Human Factors and Ergonomics Society, Santa Monica, CA: (pp. 548–552)
Rockwell, T.H., Marras, W.S., 1986. An evaluation of tool design and method of use of railroad leverage tools on back stress and tool performance. Hum. Factors 28, 303–315. Ulin, S.S., Snook, S.H., Armstrong, T.J., Herrin, G.D., 1992. Preferred tool shapes for various horizontal and vertical work locations. Appl. Occup. Environ. Hyg. 7, 327–337. Ulin, S.S., Armstrong, T.J., Snook, S.H., Monroe-Keyserling, W., 1993. Perceived exertion and discomfort associated with driving screws at various work locations and at different work frequencies. Ergonomics 36, 833–846.
Applied Ergonomics 37 (2006) 367–376 www.elsevier.com/locate/apergo
Handle displacement and operator responses to pneumatic nutrunner torque buildup Jia-Hua Lin, Raymond W. McGorry, Patrick G. Dempsey, Chien-Chi Chang Liberty Mutual Research Institute for Safety, Hopkinton, MA, USA Received 23 March 2004; accepted 10 May 2005
Abstract The objective of this study is to investigate the workstation and tool effects on the responses of the powered hand tool operator reacting against the impulsive reaction forces that may be associated with upper extremity musculoskeletal disorders. The study demonstrated a means of direct measurement of force at the interface between the tool and the operator. Fifteen experienced male operators performed three independent work configurations: pistol grip and right angle tools used on the horizontal surface and pistol grip tools used on the vertical surface, in the laboratory. A full factorial experiment consisting of 36 conditions was designed to examine the effects of working height, distance, tool, and fastener joint hardness on handle displacement and grip forces. The results indicate that operator responses were affected by different factors depending on the work configuration. When pistol grip tools were used on the vertical surface, the mean handle displacement decreased from 9.91 to 7.31 as the working height increased from 30 cm below shoulder to 30 cm above shoulder. When right angle tools were used, the greatest handle displacement (51.1 mm) and grip force (84.7% MVC) during torque reactions were measured at 30 cm below elbow and 40% forward reach away from the operator. This study provides quantitative information that can be used for workstation design and tool selection to reduce the torque reaction experienced by powered nutrunner operators. r 2005 Elsevier Ltd. All rights reserved. Keywords: Assembly tools; Occupational safety; Occupational biomechanics
1. Introduction An early survey (Myers and Trent, 1988) showed that among hand-tool related injuries in the United States in 1983 (129,399 cases), 22% were caused by powered hand tools. This trend is comparatively constant throughout recent years (Bureau of Labor Statistics, 2004). Struck by or struck against objects, and over-exertion were the two main exposures leading to powered hand tool related accidents. Unlike manual tools, in which the operator can adjust the power output (force or torque), powered hand tools use power sources (air, electricity) that sometimes require demands that exceed the capacity of human operators. These tools can potentially produce undesired forces that are a risk factor for upper extremity muscuCorresponding author: 71 Frankland Rd., Hopkinton, MA 01748, USA. Tel.: +1 508 497 0228. E-mail address: [email protected] (J.-H. Lin).
0003-6870/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.apergo.2005.05.013
loskeletal disorders (UEMSDs). A better understanding of the operator–tool–task interface may help reduce the associated morbidity. Fastener-driving tools are commonly used in manufacturing and assembly jobs. At the beginning of a fastening process, the operator holds and supports the tool. From the instant the fastener joint starts to be compressed, the tool starts to build up torque. When the tool torque overcomes the operator, the hand and arm exert a force to stabilize the tool in a direction opposite to the tool movement. Once the torque exceeds operator capacity, the tool becomes unstable and displacement occurs. Radwin et al. (1989) found that for powered hand tool operation, forearm muscle electromyography (EMG) during the torque buildup phase was greater than other operation phases. The operators associated greater handle displacement resulting from the impulsive torque reaction with increased discomfort and perceived exertion (Kihlberg et al., 1993; Lindqvist, 1993). Kilhberg et al. (1995)
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
368
concluded that for a right angle tool to be accepted by 90% of the operators, the tool should cause handle displacement of less than 30 mm. In addition to handle displacement, grip force subjectively rated by the operators during tool use was correlated to the overall tool preference (Armstrong et al., 1989). However, these studies did not identify the factors that affect the tool handle displacement and grip force. For fastener-driving tasks, work location or posture has been associated with overall operator discomfort (Armstrong et al., 1989; Ulin et al., 1992). These previous studies used psychophysical measures or subjective assessments to identify the operator discomfort that resulted from the general screw-driving experience. Nevertheless, it is uncertain whether discomfort was associated with posture alone, or with other aspects of the screw-driving task that varied with posture, such as the displacement caused by the impulsive tool reaction torque (Freivalds and Eklund, 1993; Kihlberg et al., 1993; Lindqvist, 1993) or required grip force. Workstation design was found to affect operator hand–arm stability (Oh and Radwin, 1997). It is plausible that participants preferred different tool types in different work locations and orientations because of different mechanical advantages for reacting against impulsive forces. Lin et al. (2003b) developed a single–degree-of-freedom model for common powered hand tool shapes. The human operator was represented as a single mass, spring, and damper system. It was demonstrated that the model parameters were greatly affected by working posture. These mechanical characteristics possessed by human operators directly affect the exposure resulting from the impulsive tool torque. In addition to posture or workstation effect on powered tool use, ergonomics research has provided some guidance regarding powered hand tool reaction torque using criteria such as subjective ratings (Freivalds and Eklund, 1993; Kihlberg et al., 1993), ground reaction force (Kihlberg et al., 1993, 1995), handle displacement (Kihlberg et al., 1993, 1995; Lin et al., 2001, 2003a; Oh et al., 1997), and EMG or grip force derived from EMG (Armstrong et al., 1999; Forsman et al., 2002; Freivalds and Eklund, 1993; Radwin et al., 1989). However, those studies only considered limited work locations, tool shapes, or tasks. Therefore, this study aims to integrate the most common factors
involved in powered hand tool usage: the effects of tool characteristics, task (fastener joint hardness), and posture (working height, working distance) and examine their effects on the handle displacement resulting from torque reactions for three work configurations: pistol grip tools used on the vertical and horizontal surfaces, and right angle tools used on the horizontal surface. Furthermore, ISO 6544 (International Organization for Standardization, 1981) states that there is no known device for measuring force at the interface between the tool and the operator. This study attempts to provide a means of such direct measurement using an established grip force sensor (McGorry, 2001). Thus the results derived in the study can provide further insights for workstation design and tool selection that may reduce risks for UEMSDs. 2. Method 2.1. Participants Fifteen right-handed male participants were recruited to participate in the experiment after giving their informed consent. They were all free of musculoskeletal disorders. All of the participants were experienced in assembly, manufacturing, maintenance, or other industries where powered tool use was part of their jobs. Their professional experience was validated using a custom proficiency questionnaire. In this questionnaire they were asked about their current or past jobs, tool use frequency and experience, understanding of fastener-joining procedures, and to name brands of the tools they had used. Their average experience with fastener-driving tools (e.g., screwdrivers, nutrunners) was 14.73 years. Anthropometric information of these participants is summarized in Table 1. 2.2. Apparatus The task joint and reaction torque measurement were established according to ISO 6544 guidelines (International Organization for Standardization, 1981). Four commercially available pneumatic nutrunners with automatic shutoff mechanisms were used in the experiment. They were two pistol grip tools (PA and PB) and two right angle tools (RA and RB) (Fig. 1). Their specifications are listed
Table 1 Anthropometry measurements of the experiment participants
Mean SD Min Max a
Age (years)
Weight (kg)
Height (cm)
Elbow heighta (cm)
Shoulder heightb (cm)
Forward Reachc (cm)
Exp (years)
41.2 7.4 28 56
84.18 12.38 61.7 104.4
176.87 4.98 168 185
112.73 3.30 106.5 118
146.23 3.97 140.5 154
83.76 5.08 71.4 92
14.73 7.60 5 30
Elbow height was measured from the floor to the lateral epicondyle with elbow bending at 901 Shoulder height was measured from the floor to the center of the gleno-humeral joint. c Forward reach was measured from ankles to horizontally forward pinching finger tips according to Gordon et al. (1989). b
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
in Table 2. The two pistol grip tools had similar torque setting with varying motor speed that affected the torque buildup times. The two right angle tools had similar motor speed but different handle length and torque setting. They were driven by 630 kPa (6.3 bar) of air pressure. For the two right angle tools, a tool balancer was used to partially support the weight. Joint hardness is characterized by the time or the spindle rotation required from the joint initial contact to reaching the preset tool torque. The softer the joint, the longer time or the greater spindle rotation the tool takes to reach its preset torque. Two fastener joint simulators, hard joint and soft joint, were used to emulate actual assembly tasks in the current study. Joint hardness was manipulated using different numbers of Belleville spring washers. Fig. 2 shows an example of the effect of joint hardness on the tool torque profiles. The joint simulators were placed on a height-adjustable platform that allowed various task orientations and working heights (Fig. 3). Each tool was equipped with a simulated handle. The handle has a stainless-steel core instrumented with strain gauges that allow measurement of the grip forces and reactive moments acting on the handle (McGorry, 2001).
369
Molds were made from the nutrunner handles and replicas were cast in polyurethane, cut, and mounted to the handle core. For the pistol grip tools, the handle was positioned posterior to the original handle on the extension of the spindle (Fig. 3A). For the right angle tools, the handle was affixed parallel to its original, on the plane of the tool torque (Fig. 3C). The trigger was pressed by the subject to initiate the tool operation and data collection simultaneously. A torque transducer was positioned between the joint simulator and the tool bit to monitor the tool torque output. A cable potentiometer with a 10 cm measurement range was used to record the handle position. The potentiometer, equipped with a recoil spring, was mounted on the joint simulator, and the cable was attached to a bar extending from the body of the tool. The sampling rate for all sensors was set at 1000 Hz. 2.3. Experiment design and procedures Three common work configurations were treated as independent experiments and tested in three random sessions: pistol grip tools on a horizontal surface Torque
7 6
Torque (Nm)
5 4 3 2 1 0
Fig. 1. Four pneumatic nutrunners adapted with mounting hardware used in the current experiment.
100
200
300 Time
400
500
Fig. 2. Torque buildup profiles for tool PA. The solid line represents a hard joint and the dotted line a soft joint. The double-arrow lines indicate the torque buildup phase.
Table 2 Specifications of the tools used in the current study and their torque buildup characteristics on the two joints. Tool
PA PB RA RB a
Handle lengtha (mm)
65 65 260 350
Weight (kg)
1.4 1.7 1.6 2.4
Torqueb (Nm)
6 5.6 16 23.8
Measured from the spindle axle to the handgrip center. Preset torque levels for tool shutoff. c Spindle angle rotating during torque buildup. b
Free Speed (rpm)
555 570 400 400
Buildup Time (ms)
Buildup anglec (deg)
Hard joint
Soft joint
Hard joint
Soft joint
60 44 307 685
400 254 1671 3647
28 29 145 216
547 473 1478 2126
ARTICLE IN PRESS 370
J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
(Fig. 3A), pistol grip tools on a vertical surface (Fig. 3B), and right angle tools on a horizontal surface (Fig. 3C). Only one session was tested on a given day. For each work configuration, the full factorial experiment consisted of independent factors of working height (V), working distance (H), tool (Tool), and joint hardness (Joint) (Table 3). Working heights and distances were chosen to cover a reasonable working range and to avoid awkward postures. These working posture assignments were determined based on individual operator anthropometric measurements. The dependent measures were grip force and handle displacement for each trial. In order to normalize the grip force measurement, maximum voluntary contraction (MVC) was recorded as reference at each assigned posture using the respective handle orientation in each experiment session. Each subject exerted his maximum grip on the instrumented handle at the specific posture for 5 s according to the Caldwell regimen (Caldwell et al., 1974). The period between MVC measurements was at least 2 min to avoid fatigue. The grip force at the middle 2 s was averaged to obtain the respective MVC values. Participants practiced with all the tools prior to the experiment so they could become accustomed to the tasks. The test conditions were presented in a random order to reduce the potential of confounding associated with learning or fatigue effects. Two repetitions were performed at each condition. At their self-selected pace, the participants were instructed to operate the tools using only their right hands, and to initiate a trial by pressing the trigger. Each trial consisted of a short run-down phase before the fastener started to compress. Each subject performed a total of 72 trials in a session. Three dependent variables were evaluated for each trial. Handle displacement due to torque reaction was defined as the net change of the handle position between the start and end of the torque buildup. For pistol grip tools, angular units were reported because of the nature of the forearm rotation motion. This value was converted from the linear displacement measured using the positional potentiometer. For right angle tools, linear displacement at the center of the handgrip was reported because the motion was similar to a push–pull task. Normalized grip force prior to the torque reaction (Grippre) was the grip force averaged for the 50 ms prior to the start time and divided by the corresponding MVC at the given posture. Normalized grip force during torque buildup (Gripbu), was averaged for the whole torque buildup period, and divided by the corresponding MVC for the given posture. These three measures were then averaged for the two repetitions at each condition. 2.4. Data analysis Fig. 3. Experiment setup: (A) tool PB on the horizontal surface; (B) tool PA on the vertical surface; (C) tool RB on the horizontal surface. The subject was holding the instrumented handle.
All signals were first smoothed using a 5 ms centered moving average window. Operator responses during torque buildup were the focus of this experiment, therefore it was necessary to define the start and end time frames for each
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
371
Table 3 Factors studied in the present experiment Tool configuration
Working height (V)a (cm)
Working distance (H)a (% of forward reach distance)
Jointb
Tool
Pistol grip/horizontal surface
Elbow30 Elbow10 Elbow+10 Shoulder30 Shoulder Shoulder+30 Elbow30 Elbow Elbow+30
25 50 75 25 50 75 40 60 80
Hard, Soft
PA, PB
Hard, Soft
PA, PB
Hard, Soft
RA, RB
Pistol grip/vertical surface Right angle/horizontal surface a
Defined at the handgrip center: the working height above the floor, and the working distance horizontally from the ankles. According to ISO 6544 (1981).
b
7
Tool torque (Nm)
6
Start
were treated as a random factor. In addition, two-way ANOVA was used to compare operator grip force prior to and during tool torque buildup.
End
Rundown
Buildup
Shutoff
5 4
3. Results
3 2
3.1. Pistol grip tools on the horizontal surface
1
120
28
100
26
80
24
60
22
40
20
20
18
0 300
600
Hand Position (degree)
Grip Force (N)
0
16 900
Time (ms)
Fig. 4. A typical trial for pistol grip tool PA on the soft joint. The upper thick trace is the tool torque; the lower thick trace is the handle position, and the thin trace is the grip force. Dash lines are the start and end of the torque buildup.
trial post hoc. A trial started with the rundown phase where there was little resistance. The start time of the impulsive torque buildup was identified when the tool torque, increasing due to the compression of the joint, reached beyond the threshold of 2.5% of the peak torque. The end time was when the torque reached its peak with shutoff phase followed (Fig. 4). Analyses of variance (ANOVA) for the main factors and their two-way interactions on dependent variables were performed using SPSS for Windows v. 11. Participants
The effects of main experimental factors and their twoway interactions are summarized in Table 4. For pistol grip tools used on a horizontal surface, the mean handle displacement due to torque reaction decreased from 8.81 when the tools were held at 25% of the reach distance, to 7.31 when the tool was held at 75% of the reach distance, a 17% decrease. Tool PA, which resulted in greater torque impulses due to greater torque setting and longer buildup times, caused more handle displacement (9.8177.21) than tool PB (6.1175.81) did. Joint was not a significant factor for handle displacement. However, the interaction between Tool and Joint was significant, as demonstrated in Fig. 5. For tool PA, the hard joint caused greater mean handle displacement (10.61) than the soft joint (9.01), while for tool PB, the effect was reversed (4.71 and 7.51, respectively). Normalized grip force during torque buildup (Gripbu) was significantly greater than during rundown phase (Grippre) (t (5 3 6) ¼ 18.43, po.001), although they were highly correlated (Pearson r ¼ 0.98, po.001). Both grip force measurements were significantly affected by working height, distance, and their interaction. The least Grippre and Gripbu occurred at 10 cm above elbow and 50% forward reach away (59.6 and 63.3% MVC, respectively). The greatest Grippre was found at 10 cm above elbow and 75% forward reach away (68.9% MVC); while the greatest Gripbu was measured at 30 cm below elbow and 25% forward reach away (73.2% MVC). On average, participants gripped tool PB harder (Grippre ¼ 67.7% MVC, Gripbu ¼ 70.9% MVC) than tool PA (Grippre ¼ 62.7% MVC, Gripbu ¼ 68.0% MVC). During the joint rundown phase, the operators held the handle with similar level of exertion (65.2725.8%) regardless of
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
372
Table 4 Analysis of variance results for the effects of all main factors and their two-way interactions on handle displacement (Disp), normalized grip force before (Grippre) and during torque buildup (Gripbu) for three work configurations. Asterisks () denote factors of statistical significance (po0.05) Factor
Pistol grip/horizontal
Working height (V) Working distance (H) Tool Joint VH V Tool V Joint H Tool H Joint Tool Joint
Right angle/horizontal
Disp
Grippre
Gripbu
Disp
Grippre
Gripbu
Disp
Grippre
Gripbu
Tool PA
Tool PB
10 8 6 4 2
Normalized grip force (%MVC)
Handle displacement (degrees)
12
Pistol grip/vertical
90%
Working height:
Shoulder
Shoulder+30
60%
30%
0% 25%
0
Shoulder-30
50%
75%
Working distance (%forward reach)
Hard
Joint
Soft
Fig. 5. The effects of tool and joint on handle angular displacement (mean and one standard error) due to impulsive torque buildup when pistol grip tools were used on a horizontal surface.
the joint hardness. During torque buildup, however, the operators increased their Gripbu more on soft joint (71.1728.0%) than on hard joint (67.8727.8%). 3.2. Pistol grip tools on the vertical surface When pistol grip tools were used on the vertical surface, the mean handle displacement decreased from 9.91 (7 8.01) to 7.31 (7 5.51) as the working height increased from 30 cm below shoulder to 30 cm above shoulder. The displacement resulting from tool PA (9.8177.21) was greater than by tool PB (6.1175.81). In addition to Tool and Joint factors, the Tool Joint interaction was also significant. When tool PA was used, the operators allowed more displacement on the hard joint (12.3175.41) than on the soft joint (8.7179.01), which was contrary to the finding for tool PB where greater mean displacement was recorded on the soft joint (7.21) than on the hard joint (6.71). The MVC recordings of one subject in this experiment session were failed due to hardware difficulty and discarded. Therefore, the analyses for normalized grip
Fig. 6. The effects of work posture on normalized grip force (mean and one standard error) during impulsive tool torque buildup when pistol grip tools were used on a vertical surface.
force were performed on the remaining 14 participants. Grippre and Gripbu were highly correlated (r ¼ 0.99) and both were affected by working height, working distance, and their interactions (their effects on Gripbu were demonstrated in Fig. 6). The greatest level of force (Grippre ¼ 69.0% MVC and Gripbu ¼ 73.2% MVC) was exerted by the operators at 30 cm below shoulder, and 75% of reach distance. The least grip force was exerted when operators held the tools at 30 cm above shoulder and at 50% of reach distance (Grippre ¼ 52.1% MVC and Gripbu ¼ 55.8% MVC). Participants exerted significantly higher mean Grippre for tool PB (65.1%) than for PA (61.9%), but Tool had no significant effect on Gripbu. Both Grippre and Gripbu were not affected by Joint for this work configuration. 3.3. Right angle tools on the horizontal surface All main factors were significant for hand displacement when right angle tools were used on the horizontal surface. The effects of working distance and height on handle
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376 Table 5 Mean and standard deviation (S.D.) of hand displacement (Disp) and normalized grip force (Gripbu) during torque buildup for various working postures when right angle tool RB was used horizontally on the hard joint V (cm)
Elbow30
Elbow
Elbow+30
H (% forward Disp (mm) reach) Mean SD
Gripbu (% MVC) Mean
SD
40 60 80 40 60 80 40 60 80
84.7 64.4 58.1 63.4 58.4 57.6 62.1 54.6 52.9
57.8 21.2 31.3 26.8 19.1 17.5 22.1 20.5 19.4
51.1 39.5 36.1 38.9 34.3 29.5 42.6 35.8 36.6
22.2 19.0 12.2 14.8 18.5 12.8 13.2 18.4 22.2
displacement and grip force for tool RB used on the hard joint is summarized in Table 5. For this work configuration, the operators allowed more hand displacement on the hard joint (32.5716.2 mm) than on the soft joint (14.4712.6 mm). Tool RB, which delivered greater torque output, caused more displacement (26.9719.9 mm) than tool RA did (20.1712.9 mm). Among the three working heights, participants allowed the greatest mean displacement at 30 cm below elbow for both joints (37.1 mm for the hard joint and 15.5 mm for the soft joint). However, the least displacement for the hard joint was measured at elbow height (29.8 mm); while for the soft joint it was measured at 30 cm above elbow (13.3 mm). Among the three working distances, participants allowed the greatest mean displacement at 40% forward reach away for both joints (38.6 mm for the hard joint and 16.6 mm for the soft joint). However, the least displacement for the hard joint was measured at 80% reach (28.0 mm); while for the soft joint it was measured at 60% reach (12.0 mm). Gripbuwas significantly greater than Grippre by 14.9% (t(539) ¼ 18.875, p o0.001) and these two measures were highly correlated (r ¼ 0.94, p o0.001). The operators were found to exert the greatest Gripbu (73.7740.7%) when tools were held at 30 cm below elbow and at 40% of reach distance. The least Gripbuwas exerted (54.3721.8%) when tools were held at 30 cm above elbow and at 80% of reach distance. Participants exerted significantly higher mean Grippre and Gripbu for tool RA (56.5 and 63.1% MVC, respectively) than for RB (50.4 and 59.8% MVC, respectively). Participants exerted significantly less mean Grippre for the soft joint (49.9% MVC) than for the hard joint (56.5% MVC); but Joint had no significant effect for grip force during torque buildup. 4. Discussion For right angle tools used on a horizontal surface, the resultant peak hand displacement was less when the tools were held farther away from the operator (Table 5). This
373
trend corresponds to a previous finding on right angle tool kinematics (Oh and Radwin, 1997). Furthermore, the observed displacement was greater when the tool was held 30 cm below the elbow, and was less when the tool was held at elbow height. It is hypothesized that working height and distance may affect the biomechanics of the upper extremity, and changing the capability to react against torque reactions. As the forward reach distance decreased, the upper arm started to bend, potentially allowing for more displacement of the upper extremity linkage system. Table 5 also demonstrates that the handle displacement was the greatest when the right angle tool was held at 30 cm below elbow. The right angle tool delivered its reaction torque to the operator as a horizontal force pulling the arm away from the body. The moment arm, the perpendicular distance from the shoulder to the tool handle, was the greatest as the vertical location of the tool handle decreased to 30 cm below elbow. The impulsive torque was amplified for the operator due to the longer moment arm compared to when the tool was held higher. In addition, Lin et al. (2003b) found that operator upper extremity joint stiffness increased significantly with increasing working height in the range from 80 to 140 cm above the floor for right angle tool operations. Therefore, analogous to Hooke’s law, for the same torque impulse input, less displacement would be expected for greater stiffness when the tool was held higher. When pistol grip tools were used on a vertical surface, resultant angular handle displacement was affected by work posture and it decreased from 9.91 to 7.31 as the working height increased from 30 cm below shoulder to 30 cm above. The difference may seem without practical significance. However, considering the accumulative effect for the operator, the difference may be meaningful. Lin et al. (2001) constructed a mechanical model to represent the human operator during pistol grip power tool use on a vertical surface. They found that the hand–arm stiffness, which is an important mechanical property that resists motion, increased as the working height increased from 93 to 190 cm.These working heights approximately correspond to the levels used in the current study. Therefore, the observed trend may be attributed to the operator hand–arm stiffness associated with various postures. Ulin et al. (1993) measured psychophysical responses for screw driving tasks when different shapes of tool were used at various work locations. It was found that for right angle tools, as the working height decreased from 140 to 89 cm, the perceived exertion increased. This corresponded to our findings that the handle displacement and grip force increased as the working height decreased from 30 cm above shoulder to 30 cm below (Table 5). However, for pistol grip tools used on a vertical surface, Ulin et al. (1992) observed an increasing perceived exertion as the working height increased from 114 to 191 cm. Our findings show that the handle displacement and grip force (as shown in Fig. 6) were less when the tool was held higher. In their study the participants used the tool only at a fixed distance, and the tool torque setting (3.2 Nm) was much less than in
ARTICLE IN PRESS 374
J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
the present study. Also, the screws were driven on a perforated sheet metal where a very hard joint was formed, hence very short buildup times and small impulses were expected. Considering these differences, it was possible that the exertion perceived by their participants might reflect more the posture than the torque reactions. It was demonstrated through mechanical analysis that tools of the same shape but of varied geometries, such as handle length, or varied torque settings impose different static physical demands on the tool operator (Lin et al., 2003c; Radwin et al., 1995). Current results further show that tools of the same shape resulted in different dynamic responses. Right angle tool RB had greater torque setting and longer handle than tool RA. It is intuitive that greater torque caused greater handle displacement. Additionally, the longer the handle is, the greater the translational distance the hand has to travel for a given angular displacement resulting from the torque reaction. Although tool RB had a longer handle which provided better leverage for the operator, it can be derived from Table 2 that the equivalent peak reaction force transmitted to the hand at the handgrip was greater for tool RB (68 N) than tool RA (61.5 N). All of these factors resulted in greater handle displacement for tool RB. Pistol grip tools PA and PB delivered similar torque outputs. Their motor running speed governed their behaviors on the two joints and resulted in different torque buildup times (Lin et al., 2003c), and hence different magnitudes of torque impulse. These results suggest that tool selection should not only consider tool performance and task requirements, but also the avoidance of unnecessary physical exposure such as handle displacement or grip force. Normalized grip force prior to torque buildup, Grippre, is the bracing force that participants develop before torque buildup (Armstrong et al., 1999). This grip force was mostly affected by posture (working height, distance, and their interaction) for all three work configurations in the experiment (Table 4). It is not conclusive from the current study if the operators adjusted or determined the grip force for each posture, or if the posture limited the grip force. The high correlation for grip force before and during toque buildup suggests that operators adjusted their grip force to react against the impulsive torque buildup. Handle displacement and grip force during impulsive torque reaction were anticipated to be greater when tools were used on the soft joint because the torque impulse (area under the curve of Fig. 2) was greater than the hard joint. Our results show that the mean handle displacement was 7.21 for the soft joint and 6.71 for the hard joint when tool PB was used on the vertical surface. For this tool configuration, Lin et al. (2001) tested two different levels of joint hardness for the pistol grip tool and showed that hard joints resulted in less displacement than soft joints. Magnitude of torque impulses, which is determined by the peak torque level and joint hardness, was correlated with subjective ratings of perceived exertion (Freivalds and Eklund, 1993). Interestingly, for tool PA, the effect of joint
hardness was reversed where more displacement was allowed on the hard joint (12.3175.41) than on the soft joint (8.7179.01). Oh and Radwin (1997) observed a decrease in handle displacement when torque buildup time increased beyond a threshold. Armstrong et al. (1999) found a similar trend for in-line screwdrivers. Considering that the torque buildup times for PB were much less than for PA (Table 2), it is understandable that the effect of joint hardness interacted with tool. It was speculated that when buildup times were longer, the participants were able to respond by exerting enough force timely to counteract with the torque impulses. For certain right angle tool trials, the tool could be maintained in equilibrium by the operator and the tool displacement was minimized. A further look into the moment generated by the operators could improve understanding of the dynamics of the tool-joint-operator system. Current findings provide quantitative measures on handle displacement and grip force at various work locations and can used as criteria when designing workstation or selecting tools. For example, it is shown that when pistol grip tools are to be used on a vertical surface, higher location may be preferred because of less handle displacement and less grip force. When selecting a pistol grip tool to be used on a horizontal surface, tool PB might be favored because it caused less handle displacement. In addition to relative comparisons between factors, if standards or predefined thresholds are available, the current findings can be used to identify acceptable work conditions for a group of tool operators. However, as with any job or workstation modification, all of the ergonomic implications of the redesign must be considered. In the example above, when using the pistol grip tool on a vertical surface, greater work height would be expected to reduce exposure to tool handle displacement. But the advantage of reduced exposure to handle displacement must be carefully weighed against the potential increase in exposure due to increased shoulder elevation. Kihlberg et al. (1995) tested four right angle nutrunners and concluded that for a tool to be accepted by 90% of the operators, it should produce handle displacement of less than 30 mm.The current results demonstrated that for a given tool running on a given joint, using different working postures resulted in different handle displacements. For example, when tool RB was used on a hard joint held at elbow30 cm high and 40% of reach distance, the mean handle displacement was 51.1 mm (Table 5). This would exceed the proposed 30-mm limit. Using the same tool on the same hard joint, if the task was raised to elbow height, and the operator stood at 80% of reach distance, the displacement decreased to 29.5 mm, which would be within the proposed limit. Therefore, right angle tool workstation configuration can be arranged to minimize handle displacement. Currently there is no similar handle displacement exposure limit or recommendation for pistol grip tools, and there is no grip force information regarding powered tool operations. Future research is warranted
ARTICLE IN PRESS J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
because such establishments can provide better guidelines for workstation design and tool selection. Grip force during torque buildup was affected by joint hardness only for pistol grip tools used on horizontal surface but not for the other two configurations. Originally, it was anticipated for all work configurations because the greater the torque reaction was, the greater grip strength was needed to overcome the reaction. Participants in the current study were professional tool users. We therefore speculate that because they were very familiar with the task settings and the tools, the grip force they exerted did not vary significantly for the joint. Lin et al. (2003a) tested powered tool operations using six university students and reported that for right angle tools used at a soft joint on the horizontal surface, the handle displacement was 59.11 and 79.02 mm for two work locations. Many factors were different, but the displacement measured for the soft joint in the current study was much less (14.4712.6 mm). User experience might contribute to the difference. Malchaire et al. (1996) identified working experience as a predictor for wrist musculoskeletal disorder prevalence. Ha¨kka¨nen et al. (2001) found that new workers had a higher rate of sick leave due to UEMSDs than their experienced colleagues did in a trailer assembly factory. Similarly, Park et al. (1996) found that the incidence rate ratios in UEMSDs are higher among employees who were newly hired or given new assignments among various assembly tasks. Further, experience is considered a factor influencing hand tool performance (Rockwell and Marras, 1986) because of different techniques adapted by experienced and novice users. Due to these differences, it is speculated that for the similar powered hand tool usage, novice users would have different handle displacement and grip force possibly leading to higher exposure. For example, their responses could be affected by task joints because the novice users do not anticipate different magnitudes of torque impulses and hence do not adjust their techniques accordingly. More handle displacement is expected to be imposed on the novice tool operators and result in more unwanted exposures. Thus, there is a need to examine tool usage among novice users.
5. Conclusion This study investigates how nutrunner operators respond to the impulsive reaction force for various working postures, tasks, and tools. This is the first study to instrument a grip sensor for direct measurement of the force between the tool and the operator. The results including the handle displacement and grip force associated with various torque reactions provide quantitative insight to understand tool use. The information makes it possible to reduce physical exposures that are associated with UEMSDs through tool selection, and task and workstation design.
375
References Armstrong, T.J., Punnett, L., Ketner, P., 1989. Subjective worker assessments of hand tools used in automobile assembly. Am. Ind. Hyg. Assoc. J 50, 639–645. Armstrong, T.J., Bir, C., Foulke, J., Martin, B., Finsen, L., Sjøgaard, G., 1999. Muscle responses to simulated torque reactions of hand-held power tools. Ergonomics 42, 146–159. Bureau of Labor Statistics, Department of Labor, 2004. Case and demographic characteristics for work-related injuries and illnesses involving days away from work. http://www.bls.gov/iif/oshcdnew.htm, (accessed 25.03.04). Caldwell, L.S., Chaffin, D.B., Dukes-Dobos, F.N., Kroemer, K.H.E., Laubach, L.L., Snook, S.H., et al., 1974. A proposed standard procedure for static muscle strength testing. Am. Ind. Hyg. Assoc. J 35, 201–206. Forsman, M., Cyre´n, H., Mo¨ller, T., Kadefors, R., Mathiassen, S.E., 2002. Activity in five muscles during joint securing using pneumatic nutrunners. Int. J. Ind. Ergon. 29, 21–32. Freivalds, A., Eklund, J., 1993. Reaction torques and operator stress while using powered nutrunners. Appl. Ergon 24, 158–164. Gordon, C. C., Churchill, T., Clauser, C. E., Bradtmiller, B., McConville, J. T., Tebbetts, I., et al. 1989. 1988 Anthropometric Survey of US Army Personnel: Methods and Summary Statistics (No. NATICK/ TR-89/044). Anthropology Research Project, Inc. Yellow Springs, OH Ha¨kka¨nen, M., Viikari-Juntura, E., Martikainen, R., 2001. Job experience, work load, and risk of musculoskeletal disorders. Occup. Environ. Med. 58, 129–135. International Organization for Standardization. 1981. ISO 6544 Handheld pneumatic assembly tools for installing threaded fastener— reaction torque and torque impulse measurements. International Organization for Standardization, Geneva. Kihlberg, S., Kjellberg, A., Lindbeck, L., 1993. Pneumatic tool torque reaction: reaction forces, displacement, muscle activity and discomfort in the hand–arm system. Appl. Ergon. 24, 165–173. Kihlberg, S., Kjellberg, A., Lindbeck, L., 1995. Discomfort from pneumatic tool torque reaction: acceptability limits. Int. J. Ind. Ergon. 15, 417–426. Lin, J.-H., Radwin, R.G., Richard, T.G., 2001. Dynamic biomechanical model of the hand and arm in pistol grip power handtool usage. Ergonomics 44, 295–312. Lin, J.-H., Radwin, R.G., Richard, T.G., 2003a. Handle dynamics predictions for selected power hand tool applications. Hum. Factors 45, 645–656. Lin, J.-H., Radwin, R.G., Richard, T.G., 2003b. A single degree-offreedom dynamic model predicts the range of human responses to impulsive forces produced by power hand tools. J. Biomechan 36, 1845–1852. Lin, J.-H., Radwin, R.G., Richard, T.G., Fronczak, F.J., 2003c. Forces associated with pneumatic power screwdriver operation: statics and dynamics. Ergonomics 46, 1161–1177. Lindqvist, B., 1993. Torque reaction in angled nutrunners. Appl. Ergon. 24, 174–180. Malchaire, J.B., Cock, N.A., Robert, A.R., 1996. Prevalence of musculoskeletal disorders at the wrist as a function of angles, forces, repetitiveness and movement velocities. Scand. J. Work, Environ. Health 22, 176–181. McGorry, R., 2001. A system for the measurement of grip forces and applied moments during hand tool use. Appl. Ergon. 32, 271–279. Myers, J.R., Trent, R.B., 1988. Hand tool injuries at work: a surveillance perspective. J. Safety Res 19, 165–176. Oh, S., Radwin, R.G., 1997. The effects of power hand tool dynamics and workstation design on handle kinematics and muscle activity. Int. J. Ind Ergon. 20, 59–74. Oh, S., Radwin, R.G., Fronczak, F.J., 1997. A dynamic mechanical model for hand force in right angle nutrunner operation. Hum. Factors 39, 497–506.
ARTICLE IN PRESS 376
J.-H. Lin et al. / Applied Ergonomics 37 (2006) 367–376
Park, R.M., Krebs, J.M., Mirer, F.E., 1996. Occupational disease surveillance using disability insurance at an automotive stamping and assembly complex. J. Occup. Environ. Med. 38, 1111–1123. Radwin, R.G., VanBergeijk, E.M., Armstrong, T.J., 1989. Muscle response to pneumatic hand tool torque reaction forces. Ergonomics 32, 655–673. Radwin, R. G., Oh, S., Fronczak, F. J., 1995. A mechanical model of hand force in power hand tool operation. In: Proceedings of the Human Factors and Ergonomics Society 39th Annual Meeting. Human Factors and Ergonomics Society, Santa Monica, CA: (pp. 548–552)
Rockwell, T.H., Marras, W.S., 1986. An evaluation of tool design and method of use of railroad leverage tools on back stress and tool performance. Hum. Factors 28, 303–315. Ulin, S.S., Snook, S.H., Armstrong, T.J., Herrin, G.D., 1992. Preferred tool shapes for various horizontal and vertical work locations. Appl. Occup. Environ. Hyg. 7, 327–337. Ulin, S.S., Armstrong, T.J., Snook, S.H., Monroe-Keyserling, W., 1993. Perceived exertion and discomfort associated with driving screws at various work locations and at different work frequencies. Ergonomics 36, 833–846.