Reduction and avoidance of lubricant mist demands an integrated assessment approach† Ewald Wahlmu�ller,*a Heribert Neubauera and Wilhelm Ho�flingerb aPROFACTOR GmbH, Wehrgrabengasse 1–5, A-4400 Steyr, Austria. E-mail ewahlm@pr-steyr.ac.at bInstitute for Chemical Engineering, Fuel and Environmental Technology, Vienna University of Technology, Getreidemarkt 9/159, A-1060 Wien, Austria. E-mail: whoeflin@fbch.tuwien.ac.at Received 7th April 1999, Accepted 21st June 1999 A case study to identify major factors for lubricant mist exposure covered 15 metal machining sites.The investigation included milling, turning, drilling and grinding applications. Systematic analysis considered all relevant data concerning the machine tool, the lubricant and the suction plant. The eYciency of the implemented maintenance program at the installed filter systems was checked by concentration measurements immediately before and after service.All performed measurements of lubricant aerosol and vapor loads upstream and downstream of the installed filter systems were carried out according to VDI 2066 and BIA 3110, respectively. The selection criteria for the sites to be investigated, the systematic nature of the data acquisition and the procedure of the analysis are demonstrated in detail by performing comparisons between selected applications using emulsions and those employing straight oil for lubrication. The results of the study identify recirculation of ventilated air as the major source of workplace exposure to airborne lubricant emissions.More than 60% of the demisters investigated emit air at total lubricant loads (aerosols and vapor) above the limit of 20 mg m-3 at any time of operation; which also means immediately after service. A common reason for exceeded aerosol loads in recirculated air is e.g. the fact that the type of filter system applied is often not suitable for the separation problem. Loads of lubricant vapor are usually higher at the processes which use water emulsions as lubricant.In a quarter of the cases the limits were exceeded solely due to high vapor loads even immediately after service. The exposure can be reduced by replacing the lubricant, installation of a vapor separation plant or avoiding air recirculation. Maintenance time of the demisting system and aerosol separation eYciency of state-of-the-art demisting systems can be expanded by implementation of enhanced preliminary filter stages.This study confirms that appropriate service measures lower both aerosol emissions and lubricant vapor concentrations due to the reduction of exposed oil-wetted surfaces. The performed measurements show no significant relationship between loads of airborne lubricants and the type of machining process.Therefore, investigations at a much more detailed level have to be performed. However, the individual assessment of any workplace due to the complex situation remains essential. industry.6 The four metal working sites included in this study Introduction employed 3380 people. Approximately 45% of these employees Metal working fluids (MWFs) are necessary to secure high were exposed to lubricant emissions.The annual consumption quality surfaces and to extend tool life in metal cutting of MWF concentrates and oils of the companies investigated applications. In thermal treatment processes lubricants guaran- is 650 tons. tee defined cooling rates. Extreme process conditions cause Exposure data reported in ref. 2 provide results of several airborne emissions of MWFs into the workspace. Aerosols are studies performed during the period 1950–96 in the USA. The generated due to high peripheral speeds of workpieces and collected data represent aerosol concentrations of MWFs cutting tools. Evaporation of MWFs due to hot surfaces and measured in the ambience of diVerent types of metal cutting re-condensation as well as combustion of components causes processes.Measurement results of lubricant aerosols and vapor very fine particles with diameters down to 0.1 mm. Vapor at 13 diVerent types of metal working processes in Germany emissions occur due to the (semi-) volatility of MWF compo- during 1988 and 1994 were published in ref. 1. Three-quarters nents. Airborne emissions of cutting fluids are expected to of the samples were taken at milling, turning, drilling and have a serious health impact on exposed employees.1–5 grinding sites.Further investigations covered measurements at Condensed oilfilms in the work environment are potential 28 mist separation plants. The recorded data set contains sources of accidents due to slippery surfaces and increase the upstream and downstream loads of airborne MWFs drawn by risk of fire.Micro-organisms growing in organic compounds isokinetic sampling according to VDI 2066 and BIA 3110 as cause hygiene problems. Furthermore, environmental pol- well as particle size distributions at the investigated site. The lution as well as machine failures and elevated cleaning costs data set is related to area measurements using a sampling are relevant.In 1982 in the USA approximately 1.2 million system relying on the German standard TRGS 402. employees were potentially exposed to MWFs.3 In Austria in In contrast to earlier investigations, the aim of the work 1996 approximately 190 000 people were employed in the metal within this field study was to identify major sources of lubricant mist exposure at metal working sites by applying an integrated assessment approach. In addition to systematic measurements †Presented at AIRMON ’99, Geilo, Norway, February 10–14, 1999.J. Environ. Monit., 1999, 1, 389–394 389the investigation should include relevant aspects of the machine nation of the method is 0.25 mg m-3 for aerosols and 0.5 mg m-3 for vapor at a sampled air volume of 5.6 m3.tool, physical and chemical properties of the MWF used as well as details about the suction plant installed. An assessment Depending on the aerosol concentration of the sampled air the measurement time was 15, 30, 60, 90 or 120 min in order of the eVectiveness of currently applied maintenance programs at the filter systems by concentration measurements immedi- to avoid overloading the sampling filter.Upstream of the demisting system the measurement time never exceeded 30 min ately before and after service is the innovative aspect of this study. The goal was to check whether recirculation of venti- due to high concentration of aerosol. An important requirement of the performed measurement was to ensure that one lated air into the workroom was permissible or not.The collected data should provide a basis to optimise maintenance pair of samples, upstream and downstream of the filter system, was drawn during machining the same workpiece at the procedures and to propose eVective measures to reduce and avoid emissions. machine tool to ensure equal upstream concentration when the downstream sample was drawn.Concentration limits for airborne lubricant emissions Description of the experimental procedures The approach aimed to enable comparisons of data collected In Germany threshold limit values for airborne lubricant concentrations at the workplace were issued in 1996 for at diVerent types of metal cutting processes and data derived from diVerent companies. Therefore, systematic measurements aerosols and for the sum of aerosols and vapor (Table 1).The Austrian limit values for loads of airborne MWFs at the of aerosol and vapor loads had to be related to technical data of the configuration system investigated which includes the workplace are of the quality of recommended standards. Furthermore, the Austrian approach diVers from the German machine tool, the lubrication and the suction plant system.The capability of the integrated assessment approach applied standard in qualifying the toxic potential of the vapor phase. In contrast to the German approach the Austrian legislation within this study should be validated. The selection of the sites investigated was determined by does not limit airborne lubricant emissions of recirculated air.In Germany recirculation is permissible if the sum of aerosols the following criteria: The study shoulver investigations (i) at the four most important metal cutting processes (occur- and vapor does not exceed 20% of the threshold limit value (ZH 1/248 and VDI 2261 Bl. 3) at the workplace (Table 1). ring in regional metal industries); (ii) in diVerent companies; (iii) at applications which employ emulsions for lubrication Recirculation of ventilated air is forbidden in Austria as well as in Germany if the lubricant contains substances that are and straight oils at the same type of metal cutting process; and (iv) at sites which were expected to generate relevant expected to cause cancer.In this paper, exceedence of the limit value is deemed to emissions (visual inspection).Criteria of minor priority covered technical details such as have occurred if the Austrian limits at the workplace, which are given as 1 mg m-3 for aerosols and 20 mg m-3 for the requirements for putting up measurement equipment and production schedule at the plant. sum of aerosols and vapor, are exceeded. The goals of the study, to identify major factors for exposure to airborne lubricant emissions and to propose eVective meas- Results ures for reduction and avoidance, required detailed analyses of configuration and process parameters at the site.The type The performed study covered systematic analysis and measurements at 15 metal machining sites. Table 2 provides infor- of metal cutting process, the type of lubricant employed, the phase of operation (e.g.smooth/rough work), the temperature mation about numbers of diVerent system configurations. The investigation included milling, drilling, turning and grinding of the ventilated air, the integrity of the machine tool housing and the type of suction plant (centralised/decentralised) were processes. Drilling was either performed together with milling or turning work. One-third of the machine tools were equipped considered as relevant for upstream concentrations of the demisting equipment installed.The filter system was character- with full-integrated housing. The MWFs employed divided into two diVerent sorts of straight oil and five qualities of ised by the main eVect, part or stage utilised for demisting although most of the state-of-the-art filter systems combine water emulsion (3–6%).All decentralised suction plants were equipped with compact filter devices and recirculated ventilated diVerent mechanisms or eVects (e.g. several fiber filter stages for pre-separation of large droplets and an electrostatic force air into the workroom. Centralised plants drew oV air after separation. filter stage to collect the finest particles such as smoke and combustion residues).In addition to upstream and down- At processes where straight oil was used for lubrication, smoke appeared frequently during operation (e.g. rough work). stream parameters of the ventilated air and details about the maintenance program employed the number of working shifts These applications were either equipped with electrostatic filters or fiber filters enhanced by using a HEPA (high eYciency at the machine tool was recorded. Air samples drawn for measuring lubricant loads upstream particulate air filter, according to EN 1822) filter stage for separation of particularly fine aerosols.Sinter plate element and downstream of the demisting system installed were taken by isokinetic sampling according to VDI 2066.Loads of filters occurred only together with milling/drilling applications. All centrifugal force filters were more than 10 years old. At lubricant aerosols and vapor were measured separately by using a sampling system recommended by the German BIA one single milling/drilling site, lubricant vapor separation was performed by utilising activated charcoal. (BIA 3110).7 The sampling system relies on collecting the particle phase on a glass-fiber filter and adsorbing the vapor All four diVerent types of demisting systems installed at the sites investigated are listed in Table 3.As mentioned above phase on AmberliteA XAD-2 adsorbent resin. The analysis of the sample was performed by eluting the charged filter and they are characterised by the main eVect, part or stage utilised for separation, despite the fact that most of the devices adsorbent resin with 1,1,2-trichlorotrifluoroethane (quality valid for IR spectroscopy, Merck, Darmstadt, Germany) and combine several diVerent types of demisting stages.Table 4 gives details of applied maintenance programs and rates. quantitative FTIR spectroscopy (PE Spectrum 1000, Perkin- Elmer, Norwalk, CT, USA) in the spectral region between Maintenance rates depended mainly on the type of filter system, upstream concentrations of aerosolised lubricants, 2800 and 3000 cm-1.Spectra of standard concentrations were prepared with cutting fluid samples drawn from the machine loads of chips and dust, the number of working shifts at the machine tool, and also on specific properties of the lubricant at the site investigated.According to ref. 7 the limit of determi- 390 J. Environ. Monit., 1999, 1, 389–394Table 1 Threshold limit values of airborne lubricant emissions at the workplace and for recirculation of exhaust air after mist separation. All values in mg m-3 Austria Germany Limit at the Limit for air Limit at the Limit for air workplace recirculation workplace recirculation Aerosols 1 — — — Sum of aerosols 20 — 10 2 and vapor Table 2 Number of system configurations investigated (machine tools, lubrication and suction plant including demisting) Machine toola Lubrication Suction plantb Ventilated Housing Type Demisting systemc air Emulsion Straight Process FI PI (3–6%) oil C D EF SPE FF CFF R O Milling/ 3 4 6 1 3 4 3 4 2 — 4 3 drilling Turning/ 5 — 4 1 — 5 1 — 2 2 5 — drilling Grinding 2 1 1 2 — 3 1 — 4 2 3 — aMachine tool housing: FI, full integrated housing; PI, partial integrated housing.bSuction plant: C, central suction plant; D, decentralised suction plant; R, air recirculated to workroom; O, outgoing air. cDemisting system: EF, electrostatic filter; SPE, sinter plate element filter; FF, fiber filter; CFF, centrifugal force filter.Table 3 Service programs of industrial demisting systems Maintenance program Demisting system Rate/months Work 1–1.5 Cleaning/replacement of wiremesh or synthetic preliminary demister stages; Electrostatic filter cleaning the collector of the electrostatic stage Sinter plate element filter 1.5 Backsweeping and washing of the sinter plate elements; replacement of broken parts Fiber filter 1–6 (18) Cleaning/replacement of wiremesh or synthetic demister stages; replacement of HEPA elements Centrifugal force filter 3–6 Cleaning and/or replacement of wiremesh and/or fiber filter stages used as preliminary and/or fine filter stages Table 4 Airborne lubricant emissions of industrial demisters measured immediately before and after service (average and range, containing all measured exhaust air loads at milling, turning, drilling and grinding processes) Total lubricant load Load of lubricant aerosols (sum of aerosol and vapor load)/mg m-3 Before service After service Before service After service Minimum 0.3 0.1 Minimum 1.2 0.9 Average 1.7 1.2 Average 20 18 Maximum 6.1 3.5 Maximum 63 48 and organisational demands. Sinter plate element filters and was at the limit at service time and slightly underneath immediately afterwards.some fiber filter systems utilise the pressure drop at the system for indicating service demand. Fig. 1 gives the percentage of sites investigated where the preliminary defined limit loads for recirculation of ventilated In conformity with the results published in ref. 1, aerosol and vapor loads upstream and downstream of the demisting air were exceeded immediately before and after service. The results show that immediately before service 70% of the cases system were scattered over a wide range and were not signifi- cant for the type of metal cutting process. Vapor loads were recirculate exhaust air into the workroom at inadmissible loads of airborne lubricants. Immediately after maintenance the usually higher at sites where emulsions were used.Table 4 provides the range and average of aerosol loads and total measurements indicate that total lubricant loads are still exceeded at 60% of the sites. Regarding the aerosols solely, lubricant loads (sum of aerosols and vapor) measured in ventilated air after demisting immediately before and after 54% of the measured downstream concentrations exceeded the limit of 1 mg m-3.Before service 30% of downstream samples service. As indicated by the values of Table 4 the average aerosol load at all the sites investigated exceeded the defined exceeded the total limit value solely due to vapor loads. The applied service program reduced the exceeded concentrations limit value of 1 mg m-3 in both cases, immediately before and after maintenance.The mean value of the total loads of due to high vapor loads to a quarter of the sites checked. In the following, the systematic nature of the assessment airborne lubricants (aerosols and vapor) in the ventilated air J. Environ. Monit., 1999, 1, 389–394 391The phase shift induced by the high-pressure drop at the filter system results in zero separation eYciency for aerosol loads at the end of the service time.A comparison between applications using water emulsion for lubrication and those which employ straight oil indicates significantly that vapor loads were less when straight oil was used for lubrication. In addition to the fact that high vapor loads are usually associated with high concentrations of aerosols, high temperatures in the machine housing as well as the physical properties of the lubricant employed also determine the load of gaseous phase.This is significantly demonstrated if vapor loads at Ecocut 1 (milling/drilling, checkpoint No. F2/F3) and Hu� ller Hille 11/12 (milling/drilling, checkpoint No. F4/F5) are Fig. 1 Analysis of lubricant loads exceeding the limit values immediately before and after service. compared. For sampling volatile particles ( liquid and also solid materials) on fiber filters several points have to be considered in order to avoid measurement errors or to assess the reliability approach applied within this field study and the results achievable are demonstrated in detail by presenting some selected of the results obtained.For sampling aerosolised MWFs, there is the risk that some of the collected liquid adhering to the examples of the sites investigated. In addition to the results of the measurements performed, Table 5 provides relevant back- sampling filter will be stripped oV during sampling if the gaseous phase is not saturated with MWF vapor.Evaporative ground information about the configuration of the machine tool, the lubrication and the demisting system at the selected losses of sampled aerosols result in an underestimation of the liquid phase. However, quantification of evaporative losses is processes. For each type of process an example site employing emulsion and an application for using straight oil is listed. complex.The amount of losses depends mainly on the vapor pressure of the sampled MWF aerosols, the degree of saturation of the sampled gas, the gas velocity penetrating the Discussion and conclusion sampling filter and the sampling time. The saturation rate of the air in the suction plant depends In the following, the integrated assessment approach applied within this study will be demonstrated by performing detailed on the retention time in the connecting pipes between the machine tool and the filter system.Hence, upstream of filter analysis and comparison of the measurement results given in Table 5. Its capability to identify major sources of airborne systems in central suction plants the ventilated air can be regarded as saturated. Ventilated air upstream of filter systems lubricant emission and to formulate eVective measures will be verified. installed in decentralised suction plants often remains unsaturated.Further relevant factors determining the conditions of Changes in the quality of capturing generated emissions at the site are dominated by the decrease of ventilated air during the ventilated air are the integrity of the machine tool housing, the vapor pressure of the MWF employed and the temperature maintenance time due to the elevated pressure drop of a clogged filter system.For example, the filter at the machine of the ventilated air. Unsaturated conditions may also be present during phases of operation where workpieces are tool Ecocut 1 (milling/drilling, checkpoint No. F2/F3) was operated at approximately 50% of the nominal air flow of the changed.Ideal isokinetic sampling of volatile particles requires equal electrostatic demister installed. The intended expansion of residence time for the charged air in the demister system filtration velocities at the sampling filter, which means the same by-pass flow rate, respectively. DiVerences in the velocit- results in extremely low downstream loads of lubricant aerosols as attested by the measurement results provided (checkpoint ies of the sampling air flow penetrating the sampling filter cause diVerent amounts of evaporative losses of the collected No.F2/F3). The measured loads of lubricant vapor are extremely high at Ecocut 1, even immediately after service. lubricant. In reality it is impossible to ensure equal flow rates if isokinetic sampling is performed.The velocity in the pipe Forty-two days after service downstream loads of aerosols are still less than 1 mg m-3. The appropriate measure at this site of the suction plant and the selected diameter of the nozzle at the sampling probe determine the rate of the by-pass flow is to check the vapor pressure of the lubricant in order to avoid high loads of gaseous phase. The volume of ventilated during isokinetic sampling.Realising a constant flow rate would require continuous adjustable nozzles at the sampling air at the fiber filter, which is part of the suction plant at Gildemeister (turning/drilling, checkpoint No. D5/D6), probe. Uncertainties resulting from diVerences in the sampling time decreased by approximately 25% within less than 25 d. Even worse is the situation at Karstens (grinding, checkpoint No.have to be taken into account when measurements between upstream and downstream loads of MWF particles are com- S1/S2) where the HEPA filter stage of the fiber filter system installed was completely clogged. This example demonstrates pared. Upstream sampling time normally has to be kept short in order to avoid overloading the sampling filter.the consequences of belated service drastically: Exposure to hazardous emissions is caused due to the reduction of venti- Measurements downstream of highly eYcient filter systems may need to be performed for a prolonged period in order to lated air and by re-emission of lubricants from dirty filters (checkpoint No.S1/S2). The centrifugal force filter installed reach the detection limit of the method. However, further research is required to quantify uncertainties in the field of at the Oerlikon Boehringer (turning/drilling, checkpoint No. D1) was not serviced within 6 months. The diVerence of the sampling volatile particulate matter. The described assessment approach identifies recirculation vapor load measured upstream and downstream of the filter system indicates a measurement error due to an overloaded of ventilated air into the workroom as the major source of workplace exposure at metal working sites. More than half of sampling filter (turning/drilling, checkpoint No.D1). Sinter plate element filters cause a high pressure drop due the state-of-the-art filter equipment investigated, which was designed for aerosols separation only, was not able to ensure to the fine pores necessary for highly eYcient aerosol separation.The measured loads at Hu� ller Hille 11/12 (milling/drilling, exhaust aerosol loads below the limit for workplace recirculation. For a high upstream load of aerosols, enhanced prelimi- checkpoint No. F4/F5) indicate a certain transformation of lubricant aerosols into vapor caused by the filter system itself.nary filter stages can be an eYcient measure to elevate the 392 J. Environ. Monit., 1999, 1, 389–394Table 5 Measurement results of selected sites investigated within the described field study Demisting sy Machine tool Lubrication Technical information Measurement results Separation Upstream load/ Downstream load/ eYciencyh Nominal Measured Time mg m-3 mg m-3 (%) Exhaust Manufacturer, air flow/ Checkpoint air flow/ after Temperature/ Name Housinga Typeb Manufacturer Typec aird date m3 h-1 no.m3 h-1 service/d °C Ae Vf A+Vg A V A+V A A+V Milling/drilling— Ecocut 1 FI WE Aral Sarol EF R Ifs Industrie- 4.800 F2 2.209 15 24.5 2 55.2 57.2 0.1 44.9 45 94.2 21.4 474 EP, Aral filter-Service, F3 1.997 42 27.5 8.4 61.6 70 0.6 62.4 63 92.5 9.9 IFE 5000 D, 1996 Hu� ller Hille PI WE Estramet S SPE O Heimpel & 5000 F4 7.901 6 23.5 4.3 7 11.3 1.3 7.1 8.4 69.7 25.7 11/12 33, Oemeta Besler, F5 7.558 135 30 3.7 5.2 8.9 0.5 8.5 9 86 (-1.1) Sinterstar 7,1, >10 years Liebherr FI SO OMV CUT EF R CZECH, EP 1.700 F11 1.154 3 28 20.6 0.1 20.7 0.8 0.1 0.9 96 95.5 LC 255 XU, OMV 1000, 1998 F12 1.063 50 26 12.7 0.8 13.5 0.9 0.3 1.2 92.9 91.3 AG T urning/drilling— Oerlikon FI WE BIO KS 3 CFF R Ing.G. 1.000 D1 846 6 6 35 33.8 68.8 3.5 19.3 22.8 90 66.8 Boehringer HoVmann, Type 12, >10 years R.N.F.90 FI SO OMV CUT EF R CZECH, EP 1.400 D5 1.459 3 27 14.5 5.6 21.1 0.6 4.7 5.3 95.9 73.9 Gilde- XU 1000, 1997 D6 1.086 25 30 9.7 8.9 18.6 0.4 5.5 5.9 96.1 68.3 meister Grinding— Karstens PI WE BIO KS 3, FF R ISI-Industrie- 1.800 S1 633 9 23.5 5.7 7.3 13 1.1 7.5 8.6 80.8 33.7 Quaker Produkte, FIB S2 102 152 26 8.9 34.8 43.7 6.1 26.7 32.8 31.5 25 Chemical 1800, 1992 Gleason- FI SO Excelene 416, FF R HoVmann No info S5 471 4 31 1.5 1.2 2.7 1.8 1.3 3.1 (-20) (-14.8) Phoenix D.A. Stuart Filter, DLT 40 R, 1991 aAccording to Table 2.bWE=water emulsion; SO=straight oil. c,dAccording to Table 2. eA=aerosol. fV=vapor. gA+V=sum of aerosol and vapor. h(Upstream load-downstream load)/upstream load×100 (%). J. Environ. Monit., 1999, 1, 389–394 393aerosol separation eYciency and also the maintenance rate. References The performed comparison demonstrates that high loads of 1 HVBG, Berufsgenossenschaftliches Institut f.Arbeitssicherheit, lubricant vapor are mainly related to the physical properties BIA Report 7/96, Ku�hlschmierstoVe, St. Augustin, 1996. of the lubricant in use. Changing the MWF is a suitable and 2 US Department of Health and Human Services (NIOSH), eVective measure to reduce vapor emissions in addition to the Occupational Exposure to Metalworking Fluids—What you need to installation of a vapor separation plant or avoiding air recircu- know about, National Institute for Occupational Safety and Health (NIOSH), Publication No. 98–116, 1998. lation. This study confirms that appropriate service measures 3 US Department of Health and Human Services (NIOSH), lower both aerosol and vapor concentrations due to the Occupational Exposure to Metalworking Fluids—Criteria for a reduction of oil-wetted surfaces exposed to the air flow. In Recommended Standard, Education and Information Division of conformity with the results published in ref. 1, aerosol and the National Institute for Occupational Safety and Health vapor loads upstream of the demisting system were scattered (NIOSH), Cincinnati, Publication No. 98–102, 1998. 4 AUVA Merkblatt M 368–1097, Ku�hlschmierstoVe, Unfallver- over a wide range even for the same type of metal cutting hu�tungsdienst der Allgemeinen Unfallversicherungsanstalt, Wien, process or phase of operation, respectively. However, an 1997. assessment at each individual workplace is still essential. 5 T.Mang and H. Kraner, Mineraloel-Tech., 1978, 23 (7/8), 1. The described assessment approach is capable of analysing 6 Wirtschaftskammer O� sterreich, Statistisches Jahrbuch 1997, the complex situation at metal-working sites and of identifying Wirtschaftskammer O� sterreich, Wein,May 1998. 7 HVBG, Messung von GefahrstoVen—BIA Arbeitsmappe, Erich major sources of airborne lubricant emissions. The background Schmidt Verlag, 12th edn. IV/94, 1994. information provided about the machine tool, the lubrication and the suction plant system enables eVective measures for reduction and avoidance of exposure to be proposed. Paper 9/02768H 394 J. Environ. Monit., 1999, 1, 3