Exposure to power-frequency electromagnetic fields in Denmark

SKOTIE to OBJECTIVES- The purpose of the study was to assess exposure to power-frequency electromagnetic fields in various groups with normal or high exposure in occupational and residential environments. METHODS - Exposure to power-frequency (50 Hz) electric and magnetic fields was measured for 301 volunteers (396 measurements) in periods of 24 h in both occupational and residential environ ments. The study included electrical utility workers (generation, transmission, distribution , substation), office and industrial workers, and people living near high-power transmission lines. Electric and mag netic fields were measured with personal dosemeters, and the mean values were calculated for work and nonwork periods. RESULTS - The work-period magnetic field exposure, as the geometric mean of the distribution of the work-period means, for a group of selected industrial workers with high exposure was 6 [geomet ric standard deviation (GSD) 4.6] ~T. The exposure level was 0.10 (GSD 2.4) ~T for "normal" in dustrial workers and 0.09 (GSD 1.8) ,.LT for office workers. For electrical utility workers the corre sponding values were 0.72 (GSD 2.5) ~T for substation workers, 0.52 (GSD 4.2) ~T for generation workers, 0.36 (GSD 3.5) IlT for transmission workers, and 0.15 (GSD 2.9) IlT for distribution work ers. The magnetic field exposure in normal residences was 0.04 (GSD 2.1) ~T, and in residences near high-power lines it was 0.29 (GSD 2.8)I1T. Corresponding results on exposure to electric fields are given in the study. COSCLUSIONS- All of the measurements of exposure to electric and magnetic fields were below the values normally used as guidelines.

Until rece ntly, the assessment of exposure to extremel y low-frequenc y or power-frequency electromagneti c fields has primarily been based on spot or point-in-time measurement with direct-reading handheld instrumentation. In the last few years dosemetel's for the continuous recording of personal exposure have become commercially available and have made it possible to measure and describe long-term exposure (I, 2).
The purpose of this study was to assess exposure to power-frequency electromagnetic fields in both occupational and residential environments through a measurement program with personal dosemeters. An attempt was made to measure and compare the exposure of people in jobs with "high" and "normal" power-frequency field s and the exposure of people living in normal residences and those living in residence s near high-power lines.

Subjects
The study included elect ric utilit y work ers, people living near high-po wer lines, office workers, industrial workers, and others, for a total of 30 I subjects and 396 measurements (table I). There were 47 subjects with three measurements and one subj ect with two measurements. The multiple measurements, which were made for information on variations between work periods, were taken with a time interval of more than one week.
All of the subjects were volunteers and almost everyone asked was willing to parti cipate.
Electric utiliti es. Workers were selected from the following five primary work environments: generation facilities, transmi ssion lines (voltage above or equal to 50 kV), distribution lines (voltage less than 50 kV), substations, and other electrica lly and nonelectrically related job s. For inform ation on the number of workers engaged in these types of work environments a questionnaire was sent to all Dani sh utilit y co mpanies. Workers from six of the companies were then selec ted to participate in the study. The subjects were not chosen accordin g to a formal random scheme but were , instead, selected in col-laboration with the utilities to cove r a broad range of different job types and tasks. The following types of jobs were represented: electricia ns, linesmen, facilit y operators, boilermen, smiths, engineer s, mechani cs, labor ers, gardeners, survey ors, and others. The number of subje cts in the stud y accounted for 2.5% of the workers in the five types of work environments in Danish utilit ies.
Residences near power lines. Seven different types of high-power line sections with voltages of 50/60, 132/150, and 400 kV and two 132 kV cable s were selected. There were different com bination s of single and double lines and close ly spaced parallel lines with different voltages. The residences were randomly selected within lao m from 400 kV lines, 50 m from 132/150 kV lines, and 25 m from 50/60 kV lines, and 5 m from 132 kV cables. A few examples (five) from residences nearby substation facilities were added to this group. Two of them were 60 kV outdoor substations selected by the utiliti es and the last three were 10-kV indoor substations known by the investigators.
Offices. The group working in an office environment included clerk s, secretaries, manager s, technic al personn el, and others (outsi de the utilit y companies) . The subjects were selected from eight public and private workplaces in 14 different sections. The subjects were chosen from workplaces known to the investigators through colleag ues.
Industry. Two groups of industrial workers were formed, one with high exposure and one with normal exposure. The first was a specially se lected group with j obs expected to have high exposure. It was made up of welders in a shipyard, electric furnace workers and elec tricia ns in a steel mill plant, electric furnace workers in a chemical plant, electric railroad engineers, and a labor atory technician working with a spectrophotometer. The workplaces in this group were selected in colJaboration with the Danish Working Environment Service. The normal exposure group comprised workers with exposure in an automobile repair shop, a plate shop, a machine shop, a laboratory, and a plant producing telecommunication equipment. Th is gro up of industrial workplaces was based on contacts established before the study for other reason s than power-frequ ency field s and was found con venient to include in the study to represent very common industrial activities.

Instrumentati on
Eight personal dosemeters of the Positron type were used. The dosemeters measure 50-H z electric and magnetic field s and high-frequency (5-20 MHz) electromagnetic fields. The dynami c range was Scan d J Work Environ Health 1994. vol 20, no 2 0.0 1-200 JlT for the magneti c fields and 0.6-10 000 V . m" for the electr ic fields, and the values were classified into 16 logarithm ic equally spaced bins (4-bit resolution). This type of dosem eter has a narrow band frequen cy response (3 dB points at 42 Hz and 60 Hz) and measures magnetic fields in three perpendicular directions.
One magnetic field meter of the type Combino va MFM I 0 was used. Thi s handheld meter measures magnetic fields in three direction s in the range of 0.0 I JlT-1 0 mT. It is a broad-band instrument (5-1000 Hz).
One dosemeter of the type Emde x II was used. It is a broad-band magnetic field dosemeter with a dynamic range of 0.0 1-300 JlT and a resolution of 0.1% (10 bits). With this dosemeter simultaneous measurements can be made in the broad-band range of 40-800 Hz with power-frequency harmonics and in the range of 40-100 Hz without power freq uency harmoni cs.
Before the measurement program was started, the characteristics of the dose meters were tested by checki ng calibration, linearity, freque ncy response, directivity, the timing and influence of tempe rature variations, and electromag netic interference. The calibration check was made according to standard procedures (3), and it was carried out regularly (after every five measurements).

Measure ment procedu res
All measurements of expos ure were carried out with the Positron dosemet ers. The Comb inova magnetic field meter was used for additional point-in-time measurements for industrial workers in cases in which high levels of magnetic fields capable of over- loading the dosemeters were found . The Emdex II dosemet er was used in additional measurements to quantify the influ ence of power-frequency harmonics.
The study consi sted in 24-h measurements with dosemet ers sampling field strength every 5 s. The dosemeter was normall y worn in a leather case attached to a belt at the hip. In a fe w cases the dosemeter was worn in a trouser or shirt pocket. The display of the dosemeter was switched off during the measurement.
The subjec ts wore the dosemeter at work, at home, dur ing travel, and so forth , and they recorded the times for these acti vities in a personal logbook . The subjects were instructed to put the dosemeter near their bed at night, but not close to an electric watch , a clockradio, or the like. Immed iately after the measurement was finished, the data were transmitted to a portable computer, after which the recordings were shown and explained to the subj ect.
During the dosemeter measurement in residences near high-power lines the utility companies recorded the current load of the lines, and afterwards the distance from the right-of-way center to the nearest part of the residence was measured.

Data analysis
After the measurements the dosem eter data and the logbook record of the subjec ts were checked for any indic ations of violation of the measurement protocol. For this purpose the recordin gs of electric field strength were useful because normall y the electric field shows considerable fluctu ations if measured at the surface of the human body , owing to body movements and field perturbations. Contrarily, a stationary measurement away from the body will show long periods with con stant electric field recordings. When such recordings were found, it was taken to sugges t that the dosemeter had not been worn and these period s were excluded from furth er analysis.
According to the logbook the dosemeter measurements were divided into period s according to the following four types of activities: (i) work (ie, time at the workplace), (ii) periods neither at the workplace nor at home (commuting, shopping, visiting, and other nonwork, nonre sidential activities) , (iii) periods at home with the dosemeter worn (daytime), and (iv) periods at home without the dosemeter bein g worn (nighttime).
Normally a dosemeter measurement had more periods of the same type, espe ciall y types I , 2, and 3, depending on the starting time of the measurement. All periods with the same type of activit y were combined, and the distribution of the measurement values in the 16 bins of the dosemeter was calc ulated for the four types of activities. In this way every dosemeter measurement containing a large number of single values was reduced to dat a on four distributions corresponding to the four types of activities for the subject. From these distributions on personal exposure, seve ral statistical parameters co uld be calculated, but in this study it was decided to use the arithmetic mean as the param eter describing the exposure of the subj ects.
All multipl e measurements made for some of the subjects were included in the analysis in the same way as the single measurement for the rest of the subje cts.

Results
The total approv ed time of measurement was 8894 h, an average of 22.5 h per dosemet er measurement. Figure I shows the work-period magnetic field means for the generator facility workers and the nonwork magnetic field means for all of the subjects with residences distant to high-power lines. The distribut ions of the mean s were highly skewed. Therefore the exposure of the group s has been charac terized in table s 2-4 with the follo wing statistica l parameters: maximum ; minimum; 5th, 25th, 50th, 75th and 95th percentiles ; mean; standar d deviation ; geometric mean; and geometric standard deviation. The distributi ons of the work -period means are given in table 2 for the main group s and in table 3 for the subgroups of utilit y workers. The corresponding mean s for the nonwork periods are given in table 4. On the basis of the data in table s 2-4, figur e 2 compares the distributions of the work-period means for the groups and also shows the aggregated distribut ion of the means for the non work period s for the subj ects living far from power lines.  a Twen ty -five in dividuals were unemployed or we re omitt ed becaus e of their workhours were less than 4 h the day of the measuremen t. b Eight dos emeter measurements were corrected for dosemeter overrange values on the basis of measur ements wi th the cornbinova magne t ic field meter. a One perso n belon ging to th e ut ili ty group, who happened to liver near a high-power li ne, is not included in the analy si s.

Magnetic fie ld
The exposure of the group with residence near high-power lines and that of the group with residences far from high-power lines is compared in figure  3. The distance of the near-line residences to the neares t high-power line varied betwee n 0 and 68 m with a mean of 23 m. The average current load during the exposure measurements was 60-610 A with a mean of 290 A. In 26 of the 396 dosemeter measurements there were short periods of magnetic fields greater than 200 /-IT (ie, above the dynamic range of the dosemeter). These peaks occurred especia lly in the industria l group with high exposu re ( 14 measurements) . Where possible, additional measureme nts of the true peak level (and duration) were made with the Combinova field meter, and the magnetic field means were correc ted accordingly. Peak levels at 1-2 mT were found for the electric furnace workers .
Fifteen measurements in periods of 4 to 24 h were made with the Emdex II meter. The mean magnetic field showed , on the average , a ]2% higher (maximum 25%) value when the harmonics to the power frequency were incl uded than without the harmo nics included in the calculation of the mean.

Electric fie ld
Data on electric fields measured at the same time as the magnetic fiel ds are given in table s 5 and 6. The distrib utions of the electric work-period means for the main groups and the residential means are given in table 5, and the distribu tions of the electric field work-period means for the subgroups of utility workers are presented in table 6. Figure 5 compares the distributions of the work-period means for the groups and also shows the aggregated distribution of the means for the nonwork periods for subjects living far from power lines. The nightti me without the dosemeter was exc lude d in the calculation of the means.  Table 6. Distribution of the work-period electric field means 0/' m-1 ) for the subgroups of utility workers. (N = number of rnea. surements, Min = minimum, Max = maximum, SD = standard deviation, GM = geometric mean , GSD = geometric standard deviation)  roughly the same as for people living in residences near power lines (approx imately 5 llT · hours), but the relationship was reversed between work and nonwork exposure. Th e average "dose" in the high-exposure industrial group was roughl y 35 times higher than the average "dose" of the office and normalexposure industrial gro ups. Figure 6 compares the results of this study with the results of the American Emdex project (2) and shows that the exposure of the generat or facility and substation workers was nearly the same but that the exposure of the tran smission and distribution workers in this study was half the exposure found in the Emdex proj ect. Th is differenc e might have been caused by differences in the transmission and distribution systems or by differences in work conditions. The exposure in Danish residences is about half the exposure in Amer ican residences. A part of this difference might have been caused by harmon ics to the power frequency, because the dosemeter used in this study was a narrow-band type (42-60 Hz) and the dosemeter in the American study was a broad-band type (Emdex I, 35-300 Hz). The measurements made in this study with the Emd ex II instruments showed average difference s of 12% between measurements with versus without harmonics; therefore the main cause of the differ ences in magnetic field Onl y six dosemeter measurements, of which four were for linemen , showed short period s with readings greater than 10 kV . m' (ie, above the dynamic range of the dosemeter).

Magnetic fi eld
Large variations in the work period means for the magnetic fields were evident in the utility and highexposure industrial groups, while comparably smaller variations were found for the other gro ups. The smallest variations were found for the office workers.
The exposure of workers in offices and the normal-exposure industrial groups was similar and very much lower than in the high-exposure industr ial group. In the electric utilities the subgroups with the highest work-period exposure were the substation workers and the generator facility worke rs, and the group with the lowest was the distribution workers. The large variations in exposure among the transmissio n line workers and the ge nerator facility workers were related to variations in tasks, which sometimes involved work close to live power systems or work on grounded systems distant to live power systems.
There were no differences in the nonwork exposure of the groups for all of the subjects with residence s far from power lines. The re was a slight tendency towards lower nightti me exposure than daytime expo sure. During commuting by car, train, and the like, the exposure was the same in all of the groups.
In a comparison of the exposure in residences ncar power lines versus exposure in those far from power lines (fig ure 3), it was found that individuals with a mean exposure of 0.1 u'I' appeared in both groups, but the geometric mean for the near-line group was seve n times higher than the geo metric mean for the res idences far from power lines. Average exposure above 0.2 u'T, which was typical for near-l ine residences, was infrequent in residences far from power lines. Exposure below 0.05 u'T, which was typical for residences far from power lines , was infrequent in residences near power lines.
The 24-h " magnetic field dose" plot in figu re 4 shows that the average "dose" for the generator facility, transmission line, and substation workers was  Figure 6. Comparison of results on magnetic field exposure in this study using Positron dosemeters (DK) and the American Emdex project using broad-band width dosemeters (Emdex) using data on the geometric mean for instantaneous measurement values. The data on the utility groups and the office group are valid for the work periods, and the residential data are the correspond ing aggregated measurements for these groups.
expos ure in Dani sh and Americ an residences was probably due to differences in wiring and the current load of the power distr ibution sys tems. The analyses in tables 2-4 were repeated with the excl usio n of the second and third measurem ent s mad e for 47 subjects with multiple measurements (mainly in the generation , distribution, substation, office and residence gro ups) (table 1). In all of the groups except the gen eration group, this procedure resulted in insignificant changes. In the generation group the geo metric mean of the work-period mean s chan ged from 0.5 2 to 0.36 IlT when the mul tiple measurement s were exc luded. Thi s change indicates that the mult iple mea surement s in this group were somew hat biased by high exposure . Gen erall y it should be noted that the subjects in the study were selec ted from volunteers and some of the groups were sma ll; therefore bias cannot be ruled out. Especiall y the two industrial groups should only be considered as preliminary examples of indu stri al exposure.

Electric fields
Measurements of electric fields with personal dosemeters involve large uncer taintie s because the bod y, the posture, the po sition of the do semeter, and the like ca n subst antially influence the electric fie ld (2,4). For this reason the results should be inter preted with great caution, especially if compared with those of oth er studies .
The influence of the body, dosemeter position, and the like is normally described by the enhancement factor, wh ich is defined as the ratio of fie ld measured by the dosemeter on the bod y and the unpe rturbed electric field. Laboratory experiments (5) have shown that 1.0 would be a reasonable aver age enhancem ent fac tor for a dosemeter worn attached to a belt at the hip.

138
As expected, the work -period expos ure was ge nera lly the highest for the tran smission line workers, but there were large vari ations in thi s group and in the substation gro up. Th e variatio ns were related to differ ent tasks involving work on live or gro unded systems. Typical work-period exposure for the office and industrial workers and aver age residential exposure was 10-20 V . m-I . There wa s no significant difference in the me an electric fie ld expo sure between residences near power lines and those far from power lines.

Compliance with standards
All of the result s of the measurement s of exposure to elec tric and magn etic field s in this study were clear ly below norm ally used guideli nes on lim its of expos ure to power-frequency fie lds. The IRPAI INIRC (International Non-ionizin g Radiation Committe e of the International Radi ation Protection Associa tion) guidel ines (6) recommend that continuous occ upational expos ure durin g the workday should be limited to 10 kV . m-I for electr ic fields and 500 IlT for magneti c fie lds . The highest mean expo sure for work-period electric fields was approximately 1.5 kV . rrr" for a substation wor ker, and the highest wor k-period mean ex pos ure for magnetic fields was approximately 70 IlT for a welder and a electric furnace worker.