INTERNATIONAL JOURNAL OF OCCUPATIONAL SAFETY AND ERGONOMICS 1997, VOL. 3, NO. 3^1, 109-118

Torque Production Using Handwheels of Different Size During a Simulated Valve

Operation Task

Lawrence J.H. Schulze David Goldstein Anupam Patel

Edward Stanton James Woods

University of Houston, USA

Opening and closing valves in industrial facilities often requires operators to use bars and wrenches as levers (cheaters) in order to overcome initial actuation forces. In order to determine more appropriate operational specifications, the maximum torque production capability was measured when 12 male participants used 4 different valve handwheels at3 different heights and 2 different angles (in relationship to the coronal plane). The results indicate that the participants produced significantly greater torque when the largest of the4 wheels (40.6 cm diameter) was used than when the medium (22.9 cm), small (20.3 cm), and handled (17.8 cm) handwheels were used. Although the main effect of heights was found to be statistically significant, post-hoc analyses between the heights found them to be, essentially, equal. In addition, the vertical and horizontal wheel orientations were not found to be different. The results are applicable to all industries where handwheels are used and applicable to valve manufacturers for designing operational torque specifications below the values found in this study.

valve operation handwheels torque production capability

1. IN TRO D U CTIO N

In industry many workers are required to perform a variety of handwheel turning tasks such as opening and closing valves in chemical process and waste water treatm ent facilities (Figure 1 a and b). In many cases, the torque required to begin movement of the wheel ( breaking”) is greater than can be applied by an individual. In these cases, other methods are employed to assist the operator in opening the valve (e.g., “cheater bars”). These methods introduce safety hazards, and can cause physical injury, equipment damage, or both. If these valve systems are designed properly, workers will be able to perform wheel turning tasks safely and efficiently.

Correspondence and requests for reprints should be sent to Lawrence J.H. Schulze, Department of Industrial Engineering, University of Houston, 4800 Calhoun Street, Houston, TX 77204-4812 USA E-mail: < LJHS@UH.EDU > .

System designers and equipment specifiers should consider the following in order to optimize the design of handwheel turning subsystems: (a) torque required to turn the handwheel in its expected operating environment (location, orientation, flow resis­tance); (b) wheel size (diameter); (c) wheel location relative to user; and (d) other environmental factors (e.g., footing, interference, etc.). By evaluating trade-offs between these items and system requirements, systems designers can integrate safe and efficient handwheel turning tasks into the overall system.

110 L.J.H. SCHULZE, D. GOLDSTEIN, A. PATEL, E. STANTON, AND J. WOODS

Figure 1. (a) Typical handwheel configuration for valve operations, and (b) extrem e position.

2. BACKGROUND

Although Van Cott and Kinkade (1972) and Rogers (1983) provide information regarding handwheel design, these authors do not provide information regarding torque production capabilities of operators in terms of operational or design specifications.

Several studies have evaluated handwheels with respect to the isometric strength production capabilities of participants. Woldstad, Rockwell, Johnson, McMulkin, and M cM ahan (1992) evaluated the isometric strength capability of 125 male and 125 female participants in a vertical handwheel turning task. The task was a simulated rail road handbrake turning task. These authors found that females could produce approximately 60% of the torque produced by their male counterparts and found that grip strength was highly correlated with torque production capability.

Johnston and W oldstad (1993) looked at estimates of compressive force on the

lower back (L5/L4 joint) and found that gender, ladder rung, and torque level were significant factors in a simulated handbrake turning task. These authors found that males produced approximately twice as much torque as did the female participants with estimated compressive forces on the L5/L4 nearly 3 times the levels estimated for the female participants.

M cM ulkin, W oldstad, M cM ahan, and Jones (1993), later reported by M cM ulkin and W oldstad (1995), compared the design of the new handwheels with a standard railroad handbrake wheel in a simulated handbrake turning task. These authors found that the 12 male and 12 female participants produced greater isometric torques with a zig-zag handwheel than with the other handwheels (wheel with spheres, cylindrical wheel with knurling on the surface, and the standard rail road handbrake wheel, respectively).

A m ore related study was conducted by Sanders (1981) to examine peak and sustained isometric forces applied to a truck steering wheel during a simulated tire blow-out scenario. M ean peak torque was found to be 147 Nm, which decayed 12% in the first 5 s, 10% after 10 s, and 21% after 15 s. Unfortunately, this study also examined isometric forces.

A Russian standard (State Committee of the USSR for Standards, 1984) for the operation of control handwheels and steering wheels provides maximum exertion forces in kilograms for the number of handwheel operations per shift. The specifica­tions included in the standard are presented in Table 1.

TORQUE PRODUCTION USING HANDWHEELS 111

TABLE 1. Control Handwheel and Steering Wheel Exertion Forces from State Committee of the USSR for Standards (1984)

Activity DescriptionFrequency of Use per Shift

5-16 < 5

Primarily by fingers — -----

By hand with forearm 3 kg 6 kgEntire arm (muscles of hand, forearm,

and shoulder) 4 kg 15 kgTwo arms 6 kg 2 0 kg

In order to study the kinematic operation of handwheels in the workplace, a simulation of a valve control operation was conducted at the Anthropometries and Biomechanics Lab at the National Aeronautics and Space Administration (NASA) Johnson Space Center. The results of this study provide information regarding maxi­mum torque production capabilities of participants using different sized handwheels.

3. METHOD

3.1. Participants

The test population was composed of 12 right-handed males employed at the NASA Johnson Space Center. The average age of the participants was 30, the mean stature

112 L.J.H. SCHULZE, D. GOLDSTEIN, A. PATEL, E. STANTON, AND J. WOODS

was 179.73 cm, and mean weight was 86.8 kg. A summary of the anthropometric measurements of the participants is presented in Table 2.

TABLE 2. Anthropometric Data of Participants

Height WeightFunctional

ReachShoulder

HeightForearm

Circumference

Participant Age (cm) (kg) (cm) (cm) (cm)

1 27 196.4 99.8 91.4 163.1 30.0

2 23 178.2 87.1 82.6 148.2 29.0

3 31 184.4 89.1 88.4 150.2 29.0

4 35 185.3 74.4 89.4 154.6 29.2

5 36 171.5 73.9 83.4 138.5 26.7

6 29 176.2 76.2 79.0 144.4 28.5

7 25 176.2 62.6 83.3 144.9 28.3

8 38 177.4 85.3 85.0 144.6 30.5

9 27 182.0 75.3 84.4 151.4 27.9

1 0 32 181.9 93.0 81.3 151.6 31.1

1 1 31 168.8 78.0 84.1 139.0 28.6

1 2 30 178.5 69.9 82.0 147.9 27.9

Mean 30.3 179.7 80.4 84.5 148.2 28.9

3.2. ApparatusFour handwheels used to operate valves were used in the study (Figure 2). The handwheels were assigned Large (40.6 cm), Medium (22.9 cm), Small (20.3 cm), and Small-Handled (crank-type handle on wheel, 17.8 cm) designations (beginning with the largest and moving counterclockwise in Figure 2). The handwheels used in the study were actual valve handwheels and were representative of the types and sizes found in both chemical process and waste water treatment facilities.

A biomechanical testing machine designed by LIDO™ (Loredan Biomedical, 1992) was used to define optimum design parameters for a variety of wheel turning tasks. The apparatus consisted of a dynamometer, an analog-to-digital converter (A/D), and a computer.

The dynamometer was an extremely flexible instrument. The height, the pitch angle, the distance from the participant, and other variables could all be adjusted with ease. The heights were adjusted using a crank on the back mounting beam. The pitch angle was changed by simply rotating the dynamometer at the two side attachm ent points. For this experiment, only the height and pitch angles were changed. The dynamometer used in the experiment is presented in Figure 3.

The dynamometer was a self-contained unit that gave a voltage output. The A/D board converted the analog voltage to a digital torque output. The digital torque was then displayed on the computer screen as a bar that increased as the torque increased, and decreased as the torque decreased. The maximum torque value was also displayed after each cycle. This peak torque value was the measurement used in this study.

TORQUE PRODUCTION USING HANDWHEELS 113

Figure 2. Handwheels used in this study.

Figure 3. Dynamometer equipment used in this study.

The m otion of the dynamometer was controlled by a LIDO™ software package. The software allowed for many types of biomechanical tests with varying conditions. The isokinetic testing part of the software was used for this experiment. The software also calibrated the dynamometer before each test by spinning the wheel through two

114 L.J.H. SCHULZE, D. GOLDSTEIN, A. PATEL, E. STANTON, AND J. WOODS

complete revolutions. In the calibration, the forces and moments caused by the wheel and the dynamometer were negated from the test data.

3.3. ProcedureParticipants were screened for any pre-existing physical conditions or cumulative traum a injuries that might affect them during the study. Qualified test participants were then briefed on testing procedures and allowed time to familiarize themselves with the handwheel and operation of the dynamometer. Anthropometric measure­ments were collected between tests to allow additional rest time.

Each participant performed five maximum effort clockwise rotations with a given handwheel at a specified combination of handwheel height and orientation. After each task the wheel, the height, the orientation, or both were changed and the participants performed five maximum effort clockwise rotations with the new equipment configuration. Two minutes of rest were allowed between test conditions.

Test participants were provided with gloves and were allowed to grasp the wheels as they desired as they would in the industrial setting. No foot restraints were provided, nor foot positions prescribed. Also, no clothing or shoe types were prescribed.

The participants were free to perform the tasks in the posture that was m ost comfortable to them. Letting the participants choose their own body posture and position relative to the wheel allowed for a more worksite-representative study because workers will tend to use their entire body to complete the task, and usually will not have the benefit of foot restraints or be constrained to just using their arms.

T h e hand w h eel cou ld ro ta te u p to a m ax im um ra te o f 390°/s. T o rq u e (dependen t variab le ) w as m easu red by the d y n am o m eter and disp layed on a co m p u te r screen. T h e m ax im um to rq u e p roduced by each partic ip an t during five ro ta tio n s w as recorded from th e co m p u te r screen.

3.4. Experimental Design and Data Analysis

Four different valve handwheels with diameters of 17.8 cm, 20.3 cm, 22.9 cm, and 40.6 cm were interchangeably attached to the rotational dynamometer. The handwheels were positioned at three different heights (81 cm, 102 cm, and 122 cm) and two different angles (0° and 90°) with respect to the coronal plane. Therefore, a total of 24 different task combinations were completed by each participant. All test paradigms were randomized using Latin Square methodology (Myers, 1972).

All data, once collected, were entered into a spread sheet for sortation and statistical analysis. An Analysis of Variance (ANOVA) was conducted on the main effects of and interaction between handwheel size, height, and angle (handwheel orientation) of this within subjects design. In such experimental designs, all main effects and interactions are tested using the interactions between the m ain effects and participants as the error term. (Montgomery, 1976; Keppel, 1973; Winer, 1971).

TORQUE PRODUCTION USING HANDWHEELS 115

4. RESULTS

The ANOYA summary table for the analysis of the main effects and their inter­actions is presented as Table 3. As can be seen, the main effect of Wheel and Height were statistically significant (p < .001 and p < .0067, respectively). No other main effects nor any interactions were statistically significant (p > .05).

TABLE 3. Analysis of Variance Summary Table

Source df SS MS F Value P

Between ParicipantsParticipants 1 1 10966.872 996.988

Within ParicipantsHandwheel Size 3 35670.260 11890.067 95.54 .0 0 0 1

S*P 33 4107.031 124.456Height 2 1059.257 529.628 6.34 .0067H*P 2 2 1838.243 83.557Angle 1 12.971 12.971 0 . 1 0 .7576A*P 1 1 1424.992 129.545S*H 6 479.271 79.878 1.95 .0851S‘ H*P 6 6 2698.563 40.887S*A 3 10.775 3.592 0.06 .9807S*A*P 33 1992.512 60.379H*A 2 305.821 152.910 2 . 0 0 .1598H*A*P 2 2 1685.641 76.620S*H*A 6 199.518 33.253 0.60 .4273S*H*A‘ P 6 6 3640.270 55.156

Total 288

Note. A— angle, H— height, P— participant, S— handwheel size.

The m ain effect of Wheel is illustrated in Figure 4. As can be seen, participants were able to produce significantly greater torques (p < .05) when the large wheel was used (40.6 cm, 61.8584 Nm) than when either the medium (22.9 cm, 34.8929 Nm), small (20.3 cm, 25.5338 Nm), or small-handled (17.8 cm, 24.4613 Nm) were used. Further, participants were able to generate more torque with the medium sized wheel 22.9 cm, 34.8929 Nm) than with the two smallest wheels (20.3 cm, 25.5338 Nm and 17.8 cm, 24.4613 Nm, respectively). In addition, participants were able to generate, essentially, equivalent torques (25.5338 Nm and 24.4613 Nm, respectively) with the two smallest wheels (20.3 cm and 17.8 cm, respectively). The horizontal bar indicates no statistical difference (p > .05) between levels. The means and standard deviations associated with the different handwheel sizes are presented in Table 4.

The main effect of Height is illustrated as Figure 4. Although the main effect of Height was statistically significant (p < .001), post-hoc analysis revealed that there was no statistical difference (p > .05) between or among the three wheel heights. The means and standard deviations associated with the different handwheel heights are presented in Table 5.

TORQ

UE

(Nm

) TO

RQUE

(N

m)

116 L.J.H. SCHULZE, D. GOLDSTEIN, A. PATEL, E. STANTON, AND J. WOODS

70

65 ■■

60 -

55 -

50 ••

45 ••

40 -

35 -

30 ■■

25 ■■

20

15

10

5

0

50

45

40

35

30

25

20

15

10

5

0

40.6 22.9 20.3 17.8HANDWHEEL SIZE (diameter, cm)

Figure 4. Main effect of handwheel Size (S).

TABLE 4. Mean and Standard Deviation Torque Values for the Four Handwheel Sizes.

HandwheelSize (cm) M (Nm) SD (Nm)

40.6 61.8584 22.099522.9 34.8929 11.318220.3 25.5338 8.923917.8 24.4613 9.5882

81 102 122 HANDWHEEL HEIGHT (cm)

Figure 5. Main effect of handwheel Height (H).

TORQUE PRODUCTION USING HANDWHEELS 117

TABLE 5. Mean and Standard Deviation Torque Values for the Three Handwheel Heights

HandwheelSize (cm) M (Nm) SD (Nm)

81 39.6856 22.26771 0 2 37.0310 20.48211 2 2 33.3446 18.5297

5. DISCUSSION

This study provided the opportunity to compare different types of handwheels with participants’ maximum torque generation ability. The largest handwheel diameter was associated with the greatest torque generation, followed by the smaller diameter handles. The two smallest diameter handles, one with and one without a crank-type handle, were associated with similar torque production capabilities (small wheel with handle was associated with slightly lower torque production). This probably occurred because the participants could not effectively use both hands on the wheel with the handle. The handwheel with the handle was only large enough for one hand, unless a hand-over-hand grip was used. The test participants generally used both hands to turn the wheels except the wheel with the handle.

Although three different handwheel heights were evaluated, no statistical difference was found between or among the three different levels. This may be due to allowing participants to assume individually determined body postures so as to complete the required task, as would be the case in the operational environment.

Participants described feeling fatigued at the end of the test session. Generally, participants had a 2-min rest between each task. The participants complained that they were tired towards the end and felt that they were going to be sore because of the exertions. This may well represent field situations in which multiple handwheels m ust be operated, one after the other, during start-up, shut-down or process adjustment operations within process facilities.

6. CONCLUSIONS

Based on the conditions simulated in this study, larger rather than smaller valve handwheel are the best choice for an application where maximum torque is the prim ary consideration, such as “ breaking” valves in the process industry. Therefore, operators would be able to use their whole body weight to rotate the larger handwheel. In addition, following the laws of physics, torque produced by a force is increased with a greater moment arm. Therefore, when designing any valve that will require a significant amount of torque to operate, data from this study indicates that larger wheels should be used.

Although participants were able to exert twice the amount of torque using the largest wheel as compared to the two smallest wheels and 77% more than the

midsized wheel, the participants were still required to bend over while operating the handwheels at the lowest height, possibly stressing the L5/S1 portion of the lower back. As a result, a compromise between body position and torque exerted should be established in the operational environment.

In cases where a large wheel is impractical, the results of this study can be used to predict the maximum torque that an unaided operator may be expected to exert to operate a valve, given a combination of wheel size, height, and angle.

REFERENCES

Johnson, C.A., & W oldstad, J.C. (1993). Optimization-based biomechanical evaluation of isometric exertions on a brake wheel. Proceedings o f the Human and Ergonomics Society 37th Annual Meeting (pp. 693-696). Santa Monica, CA: The Hum an Factors and Ergonomics Society.

Keppel, G. (1973). Design and analysis: A researcher’s book. Englewood Cliffs, N.J.: Prentice-Hall. Loredan Biomedical. (1992). Simulation. Evaluation and training. Davis, CA: Author. M cM ulkin, M .L., & Woldstad, J.C. (1995). Effects of wheel design on the torques applied to

large hand wheels. International Journal o f Industrial Ergonomics, 75(3), 205-213. M cM ulkin, M .L., W oldstad, J.C., M cM ahan, P.B., & Jones, T.M . (1993). Wheel turning

strength for four wheel designs. Proceedings o f the Human Factors (pp. 730-734). Santa M onica, CA: The Hum an Factors Society.

M ontgomery, D.C. (1976). Design and analysis o f experiments (2nd ed.). New York: Wiley. Myers, J.L. (1972). Fundamentals o f experimental design (2nd ed.). Boston, MA: Allyn & Bacon. Rogers, S.H. (Ed.). (1983). Ergonomic design fo r people at work. Volume I. New York: Van

N ostrand Reinhold.Sanders, M.S. (1981). Peak and sustained isometric forces applied to a truck steering wheel.

Human Factors, 23(6), 655-660.Schulze, L.J.H., Stanton, E., Patel, A., & Cheli, M. (1997). Torque production with hand

cranks in a simulated gear-operated valve opening task. International Journal o f Occupa­tional Safety and Ergonomics. 3(1-2), 31-40.

State Committee of the USSR for Standards. (1984). General ergonomic requirements (State Standard of the USSR, GOST 21752-76). Moscow, USSR: Author.

Van Cott, H.P., & Kinkade, R.G. (Eds.). (1972). Human engineering guide to equipment design (2nd ed.). Washington, DC: U.S. Government Printing Office.

Winer, B.J. (1971). Statistical principles in experimental design (2nd ed.). New York: McGraw-Hill. W oldstad, J.C., Rockwell, C.J., Johnson, C.A., McMulkin, M .L., & M cM ahan, P.B. (1992).

Isometric strength capability for a vertical wheel turning task. Proceedings o f the Human Factors Society 36th annual meeting (pp. 664-668). Santa Monica, CA: The H um an Factors Society.

118 L.J.H. SCHULZE, D. GOLDSTEIN, A. PATEL, E. STANTON, AND J. WOODS