High impulse acceleration levels in hand-held vibratory tools An additional factor in the hazards associated with the hand-arm vibration syndrome

STARCK J. High impulse acceleration levels in hand-held vibratory tools: An additional factor in the hazards associated with the hand-arm vibration syndrome. Scand J Work Environ Health 10 (1984) 171-178. The measurement of the root-mean-square (rms) acceleration of vibration does not take into consideration the short high peak values of the vibratory signal, which may be a hazard contributing to the development of vibration-induced disease. A method for evaluating the impulse character of vibra- tory signals is given, and impulsiveness is defined as the difference between the peak and rms signals. Measurements were taken during pedestal grinding, during chain sawing with three different generations of saws, and during chiseling with a pneumatic hammer. The measurements comprised (i) analyses done according to the International Organization for Standardization (ISO) draft, (ii) an evaluation of impulsiveness, and (iii) an analysis of the short-time history transient. The risk of vibration-induced white fingers was estimated by the IS0 method, and the results were compared with those observed. It was found that the analysis of impulsiveness provided additional data and partly explained the observed symptoms of vibration-induced white fingers. The parameters for the impulsiveness of the vibration signal agreed with the short-time history analysis.

The hazardous effects of occupational exposure to vibration are influenced by several factors related to work conditions, eg, the physical characteristics of vibration, how the tool is handled, biological susceptibility, and differences in the work involved (20). The risk assessment of hand-transmitted vibration should take all these factors into account, even though their individual importance in the etiology of vibration syndrome is not clearly known.
So far the magnitude of vibration has been expressed as either root-mean-square (rms) acceleration (m/s2) or acceleration level (dB) (4). The measurements d o not take the short and high peak values of the vibration signal into consideration. The content of the high peaks in the signal, ie, the impulse character of vibration in some occupations, is important because high-frequency components may be more hazardous in the etiology of vibration syndrome than has been thought (1,21).
The purpose of the present report is to measure and analyze the high peak level of impact vibration from different tools and to relate the results to the observed prevalence and latency data from exposed I Department of Industrial Hygiene and Toxicology, Institute of Occupational Health, Helsinki, Finland.
workers. The following three types of vibration tools were evaluated: grinding wheels (pedestal grinding), chain saws, and a pneumatic hammer.

Pedestal grinding
Castings were cleaned manually by an operator who pressed the castings against a grinding wheel. The constant circle speed of the wheel was 47 m/s. Two different types of grinding wheel were measured, one being made of corundum and the other of zirconium corundum. The corundum wheel had not caused vibration-induced white fingers in any of the grinders, whereas the zirconium corundum wheel proved to be very hazardous in that all the workers using such a wheel had contracted the syndrome within one year (18).
The roughness, hardness, and bond of the wheels were the same, but the abrasion material was either corundum or zirconium corundum mixed with silica. The eccentricity of each of the grinding wheels was the same. The weight of the castings cleaned manually varied between 0.5 and 10 kg. The measurements were taken when a casting weighing 5 kg was being cleaned. The same operator was used throughout all the measurements.

Chain saw
Three different types of chain saw were included in the measurements. They were selected in order to re-. present three generations of saws. The oldest (Homelite ZIP) was manufactured in 1958 and had no vibration isolation. The second (Partner R22) was manufactured in 1972 and represented the first type of antivibration chain saw. The third saw (Raket R 420) was the most up-to-date; it was manufactured in 1982. All of the saws were in good condition. Their technical specifications are shown in table 1.
Vibration was measured from the handle of the chain saw when slices of pine or spruce logs were being cut. The logs were horizontally supported on a bench during sawing. The operators, professional forest workers, sawed downwards and upwards in turn from four to six times (19).

Pneumatic hammer
Vibration was measured from a pneumatic hammer (Atlas Copco type BHV-22) used by stoneworkers. The diameter of the piston of the hammer was 22 mm, the length of the blow was 30 mm, the frequency of the blow was 83 strikes/s, and the hammer weighed 1.4 kg. During the measurement the operator chiseled gray granite. The air pressure operating the pneumatic hammer was 6 kg/cm2 (600 kPa).

Definition of impulsiveness
The primary quantity used to describe the magnitude of the vibration is still rms acceleration, which is normally expressed in terms of meters per second squared. If the impulses are short, the energy level remains low in comparison with the peak level. This phenomenon led to the assessment of impulsiveness on the basis of the difference between the peak and rms signals (6). The shorter the impulse, the greater the difference between the peak and the rms levels. This difference (I) is related to the crest factor of the signal and can be expressed by the formula The definition is suitable for the continuous analysis of randomly varying vibration signals. The ana- lytic functions were the cumulative distribution function F(1) for impulsiveness (equation 1) and the shorttime history of the signal. From the cumulative distribution function the impulse indices I,,, I,, and I,, were selected to compare the impulsiveness of the different vibration signals. The impulse indices gave impulsiveness I in decibels that corresponded to 10, 1, and 0.1 percentage values (17).

Vibration measurements
Vibration was measured from the casting, from the handle of the chain saw, and from the pneumatic hammer. To mount an accelerometer to the casting, a metal piece was welded, and a mechanical filter [Briiel & Kjaer (B&K) UA 05531 and an accelerometer (B&K 4375) were fastened to the metal piece by a screw. The measurements were taken from the casting only in the horizontal direction because this direction had been found to generate the highest level of vibration (18). To measure vibration from the chain saw, an accelerometer and a mechanical filter were mounted on the front handle of the saw by a hose clamp to which a steel stud had been welded (19). The direction of the accelerometer was vertical so as to approximate the direction of the operator's third metacarpal bone. Accelerometers and mechanical filters were fixed to the chisel head of the pneumatic hammer by the method used for the casting and to the barrel of the hammer by the method used for the handle of the chain saw. The main axis of the accelerometers was in the direction of the blow of the hammer.
The signal was amplified (B&K 2635) and recorded on a tape recorder (B&K 7003) for further laboratory analysis. The frequency range of the sampling was 2 Hz-10 kHz; in this range the frequency response was linear within 1 dB (figure 1, part a).

Laboratory analysis
The recordings were replayed on a real-time analyzer (B&K 2131), and the total acceleration level and weighted acceleration level according to the IS0 draft (4) were calculated from the average one-third octave spectrum of 20 to 50 spectra with the use of an averaging time and sampling interval of 0.5 s (figure 1, part b). The length of the sample analyzed for the pedestal grinding and chiseling was 10 s, and that for the chain sawing, from 10 to 25 s. It was not possible to take longer continuous recordings because the work phase, ie, the cleaning, in the case of the castings lasted this period. For the operation of the chain saws, one sample comprised the cutting of several slices of a log. The results for pedestal grinding were calculated from one sample because all the castings were identical. Five samples were collected for sawing, two for chiseling. The vibration levels (L) were based on the formula L = 20 log a/a, dB, (eq 2) where a = rms acceleration (m/s2) and a, = reference acceleration of 10W6 m/s2. The assessment of risk according to the IS0 draft (4) should be based on a daily exposure time of 4 h. For different daily work periods the level of exposure to vibration was normalized to 4 h with the formula a,, = (T/4)0.5 a,,, (eq 3) where a,, = the frequency-weighted 4-h energy equivalent acceleration (m/s2) and a,, = the frequency-weighted equivalent acceleration (m/s2) for a time of T hours. The probability distribution function for I (equation l) was analyzed with a fast Fourier transform function analyzer (Hewlett Packard 5420A) that replayed the signal through the rms peak detector (Wartsila L674) (figure 1, part c). The cumulative probability distribution function F(1) of I was calculated by a microcomputer (Luxor ABC SO), which also showed the selected impulse indices I,,, I,, and I,,,. The impulse indices were calculated from the same number of experiments as the total and weighted acceleration values.
A detailed time-history analysis was carried out with a transient recorder (B&K 7502) and a microcomputer (Luxor ABC 80) for signals of 100-ms duration ( figure 1, part d). These short-time histories, 100 ms in duration, were displayed by the joining of recurrent 10-ms samples, which were analyzed at the bandwidth of 12.8 kHz.

Results
High impulsiveness occurred for vibration at the low probability levels. The cumulative distribution function was shown only at probabilities from 0 to 15 %, as these probabilities were regarded to be the most suitable for this study ( figure 2).
The impulsiveness increases from left to right in figure 2. The highest value was measured from the chisel head, and the lowest from the third-generation chain saw. In pedestal grinding the impulsiveness of the vibration caused by the zirconium corundum wheel proved to be significantly higher than that of the corundum wheel. The impulse character of the vibration of the chain saws decreased from generation to generation. The weighted accelerations were about the same in the second-and third-generation saws.   Table 2 shows the vibration parameters as measured for different tools and calculated according to the I S 0 draft (4) and according to the amount of impulsiveness (figure 2). The pneumatic hammer and, especially, the chisel head generated the highest vibration with respect to both the equivalent acceleration and im~ulsiveness. The corundum and zirconium corundum grinding wheels gave off the same weighted acceleration values, but the impulsiveness of the vibration generated by the zirconium corundum wheel was higher than that generated by the corundum wheel. Of the different chain saws the first-generation saw gave off the highest weighted and unweighted acceleration, as well as the greatest degree of impulsiveness. No differences were observed in the weighted acceleration of the second-and third-generation chain saws, whereas the unweighted acceleration value and the impulsiveness of the vibration were higher for the second-generation chain saw than for the third-generation saw.
The weighted acceleration a,, and impulsiveness I, analyzed from the measurements of vibration for each chain saw are presented in figure 3. The vibration of the chain saws was separated according to the generation of the saw. The range of the weighted acceleration in the different experiments was 8.0-14.4 m/s2 for the first generation, 1.8-2.7 m/s2 for the second generation, and 2.2-2.8 m/s2 for the third generation. The corresponding ranges for impulsiveness I, were 15.3-18.6 dB, 13.6-16.0 dB, and 12.1-12.9 dB. The weighted acceleration was about the same for the second-and third-generation chain saws, but the impulsiveness was the lowest for the third-generation saw.
Transient analyses were done in order to find an explanation for the results shown in figure 2 and table 2. The chisel head generated short peaks that occurred regularly at intervals of 12 ms and corresponded to a blow frequency of 83 Hz (figure 4F).
The peak acceleration in the barrel of the chiseling hammer was reduced to about one-eighth of the acceleration of the chisel head (figures 4F & 4G). In Table 2  pedestal grinding the short peaks occurred randomly ure 4C). In the second-generation chain saw the peaks at intervals of several milliseconds (figure 4B). For were strong!y reduced (figure 4D), and in the thirdthe chain saws the transient analysis showed the at-generation saw the peaks disappeared ( figure 4E).
tenuation of vibration that occurred over the three The risk assessment of vibration-induced white fingenerations. The first generation chain saw had regu-gers was done according to the I S 0 method (4). Both larly occurring peaks at 8-ms intervals that corre-the 10 and 50 percentiles were evaluated. Table 3   Comparisons were made between the expected and observed exposure times for different tools. Agreement was found for chain saw vibration. In the case of pedestal grinding, however, the I S 0 draft (4) underestimated the hazardous effects of vibration. The prevalence of vibration-induced white fingers among the grinders would presume a weighted acceleration of about 55 m/s2, but in this study it was only about 15 m/s2.

The hazard of vibration-induced white fingers in relation to vibration level
Standardized methods for the analysis of vibration are restricted to measurements of frequency-weighted vibration acceleration and daily exposure times. Safe exposure limits are not defined, but the draft is based on a latency period for vibration-induced white fingers at probability levels of 10 to 50 %. This estimation is given as a function of the weighted acceleration value, which should then be normalized to correspond to a daily work period of 4 h (table 2).
The I S 0 draft (4) provides a good method for estimating the risk of vibration-induced white fingers in lumbering. In a follow-up study of vibration-induced disease among forest workers, the prevalence of vibration-induced white fingers has decreased from 40 to 7 Vo between 1972 and 1980 (13). At the same time the weighted acceleration decreased from about 10 to 2 m/sZ because of the use of antivibrating chain saws (3,14,15,21,22). At the beginning of the 1970s the latency period for vibration-induced white fingers was, on the average, five years; it is now longer (13). These findings agree well with the IS0 proposal for the risk assessment of vibrating tools for chain saws used before and after 1972.
The prevalence of 7 % in 1980 may be an overestimation of the hazardousness of vibration generated by the third-generation chain saws, since the prevalence was also still probably influenced by secondgeneration saws.
However, I found that the I S 0 draft (4) was not suitable for the risk assessment of human exposure to vibration in different occupations. In a foundry where pedestal grinding is used to clean small castings, new and more effective grinding machines with a different type of grinding wheel caused vibration-induced white fingers in all the workers after a period of one year (18). The weighted acceleration values did not differ significantly between the old and new grinding wheels (17). Moreover, the vibration in pedestal grinding proved to be almost the same as that of the uninsulated chain saws of the 1960s.
The new grinding machines allow the castings to be pressed towards the wheel with more force than the old machines. The transmission of vibration from the casting to the hand is increased (ll), as is the static load on the muscles during the grinding operations. These factors should also be considered when the risk of vibration-induced disease is being assessed.
The present results confirmed that conventional methods of measuring vibration do not reveal all its hazardous characteristics. I observed high impact components of vibration in the tools regarded to be the most hazardous. The highest impacts were observed for the pneumatic hammer, particularly the chisel head, followed by pedestal grinding wheels, and chain saws.
The additional data concerning the high values of peak acceleration measured for the tools may partly explain the increased hazard associated with certain tools. For example high impulse indices were ob-tained for pedestal grinding when the zirconium error in I,,. The highest impulses may not be decorundum wheel was used. As reported earlier (18), scribed by I,,. These limitations would support the this wheel also caused a drastic outbreak of vibra-selection of I,, if only one index is t o be shown tion-related symptoms aniong grinders. Thus the ( figure 3). impact component of some tools can be a factor The short-time recordings were 100 ms in duration. crucial to the severity of vibration-induced white At the lowest sufficient bandwidth of 12.8 kHz, the fingers.
digital signal analyzer (Hewlett Packard 5420A) can handle samples 10 ms in duration. For samples longer

Physiological aspects of impulsive vibration
In previous reports it was suggested that the impulsiveness of vibration may create shock waves in tissues, and these waves may be transmitted at a higher velocity and to larger body areas than nonimpulsive vibration (11, 18). Thus high peak accelerations may contribute to the hazards associated with vibration by two different mechanisms. High impacts may powerfully distort the tissues and shear the endothelium of capillary beds, an event that leads to leakage and capillary dilation (9). This situation has been reported in connection with laboratory animals exposed to vibration (7, 16), and it may be responsible for the erythema and swelling of the fingers observed after exposure to vibration (22). Furthermore, since the tactile end organs of the fingers, ie, quickly and slowly adapting skin receptors and Pacinian corpuscles, respond to sinusoidal vibration in a phaselocked manner (5, 8), these receptors respond to one impulse with one spike. Since impulsive vibration spreads to larger body areas, more receptors are also recruited. The sensory responses to vibration are enhanced through temporal and spatial summation, and, as a result, impulsive vibration may facilitate somatosensory reflexes and thereby contribute to the injuries caused by vibration (12).
than 10 ms a transient recorder (B&K 7502) had to be used. The short-time history was analyzed with recurrent 10-ms samples which were joined together (figure 4). The peaks were then displayed without leakage caused by the limited number of points on display. This method was applied to display several recurrent impulses in order to obtain information about the randomness of the signal with respect to both peak amplitudes and the repetition rate. The parameters selected for the impulsiveness of the vibration signal agreed well with the short-time transient analysis. The highest impulsiveness was analyzed in the chisel head, and the recorded values corresponded with the results for the short-time history analyses (figure 4).
In previous studies vibration was measured also from the wrist (2,18). It was found that the transmission of vibration at high levels of acceleration is not necessarily linear. This nonlinear transmission from the vibrating tool to the wrist occurs when impulsive vibration is being generated, the reason being the short duration of the high peak values in the vibration signal. Our method of measuring vibration in the wrist is not suitable for high impulse vibration because of the lower limiting frequency response caused by the mounting of a n accelerometer to the wrist.

Conclusion
General purpose accelerometers of delta shear con-Insufficient measurements have been carried out to struction were used for the measurements to allow date to allow conclusions or practical proposals to be the measurement of peak of up made on how the impulse character of vibration sig-250,000 m/s2. The accelerometers were mounted on nals should be taken into account in the assessment a low-pass mechanical filter, a procedure which is reof risk. The analysis of high impulse acceleration commended to avoid the nonlinear effects caused by peaks obtained by the method presented in this report subjection to extreme accelerations, eg, those gencould provide additional data necessary to improve erated by percussive tools. Otherwise the linearity of the assessment of the risk involved in exposure to the measurements is in doubt due to the directhazardous hand-transmitted vibration. current shift in transducers (10). The cumulative distribution function was selected to describe impulsiveness according to equation 1. Only that part of the function which revealed the distribution 'of the values for the highest impulsiveness has been shown. In addition the indices for impulsiveness were selected to describe the highest such values. The results in figure 2 and table 2 show that any of the selected indices are appropriate for describing impulsiveness, as the difference between the chosen index for different tools remains almost constant. However, limitations in the dynamic range, and possible disturbances in the signal, can cause