Document RJE3pj5720DbdXg7Noa4XjwEz

FILE NAME Brakes BRK DATE 1983 June DOC BRK254 DOCUMENT DESCRIPTION Conference Presentation - Simulation of Auto Brake Wear Dynamics and Estimation of Emissions SAE Resource The EngineeringEnginering For Advancing Mobility 400 COMMONWEALTH DRIVE WARRENDALE PA 15096 SAE Technical Paper Series 831036 Simulation of Automobile Brake Wear Dynamics and Estimation of Emissions Soyoung Cha and Philip Carter Northrop Services Inc. Research Triangle Park NC Ronald L. Bradow U.S. Environmental Protection Agency Research Triangle Park NC Passenger Car Meeting Dearborn Michigan June 6-9 1983 2 The appearance of the code at the bottom of the first page of this paper indicates SAE's consent that copies of the paper may be made for personal or internal use or for the personal or internal use of specific clients This consent is given on the condition however that the copier pay the stated per article copy fee through the Copyright Clearance Center Inc. Operations Center 21 Congress St. Salem MA 01970 for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law This consent does not extend to other kinds of copying such as copying for general distribution for advertising or promotional purposes for creating new collec- tive works or for resale Papers published prior to 1978 may also be copied at a per paper fee of 2.50 under the above stated conditions SAE routinely stocks printed papers for a period of three years following date of publication Direct your orders to SAE Order Department To obtain quantity reprint rates permission to reprint a technical paper or permission to use copyrighted SAE publications in other works contact the SAE Publica- tions Division ISSN 0148-7191 Copyngnt 1983 Society of Automotive Engineers Inc. 831036 Simulation of Automobile Brake Wear Dynamics and Estimation of Emissions Soyoung Cha and Philip Carter Northrop Services Inc. Research Triangle Park NC Ronald L. Bradow U.S Environmental Protection Agency Research Triangle Park NC ABSTRACT Asbestos emissions from automobile brakes were measured under conditions simulating downtown city driving Reference data for city driving was obtained by measuring vehicle speed time of brake application cool- down time between applications brake hydraulic pressure and pad temperature Data from 1800 braking applications were then analyzed to provide a statistical distribution of representative braking cycles We constructed a controlled brake emission test rig that simulated road braking operation of a and collected airborne front wheel disc brake wear debris Repre- sentative braking cycles were programmed on this system to experimentally estimate brake emissions under a variety of braking con- ditions Realistic braking operations produced particle and asbestos emission rates in close agreement with those measured by Williams and Muhlbaier Asbestos concentration Was not correlated with mechanical work done in braking Application of the asbestos emission rates to air quality models confirmed Williams and Muhlbaier's finding that about % of asbestos fibers in city core districts originated from disc brake wear added to maintain a proper range of friction coefficient or to control noise properties Friction materials in automobiles have been considered atmospheric emission sources and have been investigated by several re- searchers to define particulate and asbestos emission rates There has been increasing evidence that the exposure of human beings to asbestos fibers can cause serious diseases , including mesothelioma asbestosis and cancer * Jacko et al 2 estimated the annual asbestos emissions from cluded that 3.2 of the automobiles and contotal automobile asbestos emissions are airborne contributing only 2.3 yr to the atmosphere Williams and Muhlbaier 3,4 found that airborne asbestos emissions from automobiles can account for only about 4.9 yr or 0.23 of the asbestos emitted from all sources airborne Using the lead tracer model for automobiles they pre- dicted that automobile tribute a small amount asbestos emissions . con- to the ambient asbestos level in New York City of automobiles estimated The small contribution in these reports was primarily due to the low asbestos content of brake wear debris The asbestos content ranged from 0.005 to 15 in other reports 4 This report presents a systematic approach to simulating brake applications and defining particu- late and asbestos emissions TECHNICAL APPROACH TYPICAL ASBESTOS FRICTION PRODUCTS used in automobile brake linings and disc pads consist of three classes of materials reinforcing agents friction and wear modifiers and organic binders Reinforcing agents make up 40 to 50 by weight of the friction products and consist almost exclusively of chrysotile asbestos Organic binders are primarily phenolic resins selected for high binding strength and comprise from 30 to 50 of the materials Friction modifiers which include polymeric metallic or ceramic materials are and In the past most asbestos emissions brake abrasion were estimated rates by analyzing trapped debris in brake systems or by choosing arbitrary braking cycles and collecting corresponding samples Accurate esti- mation however requires rigorous selection of representative braking cycles RBC's based on statistics appropriate sample collection Numbers in parentheses at end of paper designate references 0606-1036 Copyright 1983 Society of Automotive Engineers Inc. 2 and analysis This section will describe technical approach used to find the RBC's this study the in BRAKE WEAR MECHANISM - Brakes convert the kinetic energy of automobiles to heat energy by friction work at the stator interface Most of the heat generated there is absorbed by the rotor and dissipated to the atmosphere through conduction and convection The heat generated raises the temperature at the interface and alters the composition of microstructures in the brake pad and rotor surfaces forming friction heat affected layers Therefore the wear rate of friction material depends not only on mechanical properties but also on the chemical properties of friction heat affected layers Although there are various wear mechanisms abrasion by rough surface contact adhesion of materials at contacting surfaces and thermal decompo- sition are believed to control overall wear of friction materials At low brake tempera- tures friction materials undergo ploughing and mechanical damage before thermal decomposition begins At such temperatures abrasion and adhesion wear dominate and the wear rate is fairly stable under constant mechanical conditions Organic binders in brake pads start to degrade at approximately 250 At high temperatures thermal decomposition of organic binders in the outermost layer occurs accompanied by deformation and phase change in the alloy at the rotor surface For this reason thermal wear occurring above 250 increases the wear rate drastically The most interesting material in this study is the portion of chrysotile asbestos that survives thermal de- gradation wear processes and is emitted along with other airborne particulates Chrysotile asbestos is a magnesium silicate 25102 and exists in fibrous form The thermal decomposition of chrysotile follows a stage dehydroxylation at approximately 600 to 780 and anhydride breakdown to nonfibrous forsterite 2MgO.SiO2 and silica 5102 During light braking average temperatures are generally less than 250 However in localized spots temperatures can be high enough to cause thermal degradation of chrysotile asbestos Previous tests showed considerable alteration of asbestos at temperature braking suggest- ing that actual contact temperature is above the thermal degradation temperature Conse- quently thermal heat and mechanical forces at the friction interface produce complicated phenomena that cannot be simply yet accurately described The wear phenomena of friction materials are too complicated to be described micro- scopically because we would require knowledge of local physical properties chemical compo- sition roughness binding structure and interaction of friction forces But macro- scopically we can ditions with three represent factors the wear consurface pressure sliding speed and temperature Temperature is the primary factor governing surface con- dition and friction work is determined by surface pressure and sliding force OBJECTIVE AND METHODOLOGY . Our primary interest was not in localized phenomena but in the general contribution of automobile brakes to atmospheric emissions Considering this objective our testing procedure was organized to achieve confidence in brake emissions since the results estimating of previous investigations varied greatly 4 To imple- ment the objective we chose inner driving to estimate emissions By choosing this specific region we can investigate the braking and emissions more thoroughly and thereby make more accurate estimates Frequent brake applications are also expected while driving in a city a region where congestion is common We can measure emissions of composite driving schedules on a real car but a sampling procedure and experimental setup would be desirable if it had the following advantages easy to run braking cycles to establish a data bank good repeatability to estimate emissions accurately and generation of airborne particulates in reasonable doses compared to those in the ambient air The experimental setup also requires convenience in installation sample collection and sample load control To achieve these advantages we used a dynamometer equipped with a size brake system to simulate brake applications by using typical RBC's whose combination represents a composite driving pattern in an inner city The samples were analyzed by electron microscopy to determine asbestos concentration In addition to various local physical properties in a brake system other conditions affect brake wear including climate cumulative time in use and alignment These conditions are however of minor importance major factors affecting global brake emissions are brake pressure automobile perature and duration and speed lining temfrequency of brake applications Selection of RBC's was therefore based only on these factors Estimation of particulate and asbestos emissions for automobile brakes was divided into four steps First real braking cycles were collected by recording temperature brake pressure and speed while driving a real car in a downtown area Second the real braking cycles collected were analyzed to find RBC's and their corresponding distribution by speed pressure temperature and cycle duration Third these RBC's were run on a brake dynamometer equipped with a sampling to define particulate emission rates application Asbestos concentration system per brake and emission rates per brake application for these cycles were determined by electron microscope analysis Fourth emission rates per mile were estimated using emission rates of RBC's and their corresponding frequency distributions STATISTICAL ANALYSIS FOR REPRESENTATIVE BRAKING CYCLES To find RBC's during inner driving we recorded and processed real braking cycles This section presents the procedures followed facilities used and the results obtained RECORDING AND READING OF INNER DRIVING - A 1972 Chevrolet Impala V engine was selected as of the project the test Vehicle vehicle for this phase speed was measured with a de tachometer Nucleus Corporation run by a fifth wheel mounted on the rear bumper of the car The brake pressure in the hydraulic lines was monitored with a differential pres- sure transducer and a digital transducer indi- cator both from Valedyne A type thermocouple Chromel Alumel was installed in a brake pad according to the instructions described in Reference 5. A cold junction compensator compensated for the ambient tem- perature This configuration cannot measure the contact temperature at the interface but gives a reference value close to the real tem- perature The same configuration was used for dynamometer simulation to measure the reference temperature All electrical signals from these instruments for speed pressure and temperature were recorded on a channel magnetic tape recorder Hewlett The fourth channel was used as a voice channel to provide other necessary information such as driving location climate traffic conditions and voltage reference during instrument cali- bration Prior to testing the following procedures were used to provide speed pressure and tem- perature calibration data The fifth wheel was driven at discrete constant speeds 0 10 20 30 and 40 the and 50 mph on a chassis dynamometer reference signals from the tachometer for these speeds were recorded A pressure gauge was connected to a hydraulic line of the brake system The brake pedal was actuated to 100 200 and 300 psig as measured by the pressure gauge while reference signals from the pressure transducer demodulator were recorded For temperature calibration the ambient temperature was recorded prior to driving the car then additional reference signals were generated by applying a potential difference corresponding to a certain temperature instead of the thermocouple In all cases during calibration the voice channel was used to describe which reference signal was being generated The driving route covered downtown Raleigh NC and the outskirt road connecting the downtown area and U.S. Highway 64. This area is mostly a commercial and only partially a residential zone Though some parts of the terrain were hilly most of the distance was flat The recorded speed brake pressure and temperature data were retrieved by the magnetic tape recorder giving corresponding dc outputs which were converted to binary code by an digital converter Computer Products The sampling interval converter was 0.01 s of the digital The binary code was then processed by a minicomputer Texas Instruments All of the brake applications were read cycle The reference values stored before recording the real driving cycles were used to the dc to find actual values corresponding outputs Due to background noises in each channel recorded negligible brake applications in some cases or dragging brake applications at almost zero speed we discarded the braking cycles with a maximum speed and brake pressure of less than 2.0 mph and 50 psig respectively For each cycle data were reduced to mation was 100 points the processed inforstored on magnetic tapes for statistical analysis The braking cycles were collected during seven days by driving the car once in the morning and again in the afternoon A total of 1806 braking cycles were recorded while driving about 358 miles This brake applications per mile at yielded 5 05 an average driving speed of 17 mph verifying the congestion of downtown driving Cha and Carter 6 have detailed the and recorded braking cycles driving routes NORMALIZATION AND FREQUENCY DISTRIBUTION - During the recording process initial final maximum minimum and average values of a braking cycle were found for temperature pressure speed duration of braking cycle and interval of brake application from the previous cycle To find representative cycles terns we must define typical braking patFor real braking cycles the braking patterns are quite different making ficult to extract reasonable RBC's it difFor this reason normalized NBC's defined by the were used to find RBC's frequency distribution braking cycles following quantities and the corresponding - St 1 " = n Ju 2 a11a a11a A a11a 3 where at = duration of braking cycle final time minus initial time Jv = speed reduction initial speed minus final speed t = time v = speed brake pressure subscript n : normalized value superscript - : average value of brak- ing cycle subscript i : initial value Maximum temperatures of all recorded braking cycles were lower than 150 There- fore we assumed that no thermal wear occurred and the temperature did not affect the braking cycles or their wear rates The sequence of braking cycles is also believed to be unimpor- tant As seen in Eqs 1 through 3 the important parameters governing NBC's are duration of braking cycle At initial speed v speed reduction AV and average pressure p When we analyzed the real braking cycles we found that the braking cycle's duration is the most important factor determining the braking cycle patterns Therefore the representative normal~-zed braking cycles RMBC's were classified using this parameter when the denominator in Eq 2 is close to zero background noises become too large after normalization To fied mathematical errors eliminate caused by the ampli- this con- dition only those cycles that had speed reduction greater than 5 mph were used for speed normalization For Eqs 1 and 3 the lack of noise in time sampling or high average pressure made the normalization stable In this study three RNBC's were found by averaging the NBC's of those typical subsets defined by durations of braking cycles 0 - 15 15 - 30 and 30 - 45 s The RNBC's are plotted in Figures 1 corresponding normalized and other data are shown through 3 and the values statistics in Reference 6 For braking cycles of short duration 0 - 15 s the speed reduction has the strongest correlation coefficient -0.86 with the product of average pressure and cycle duration which is approximately proportional to the momentum change of the car by brake friction forces For braking cycles of long duration 15 - 30 and 30 - 45 s the above statement is no longer true and the speed reduction has the strongest correlation coefficients -0.97 and -0.99 respectively with initial speed v These statistics imply that most short brake applications were snubs brake applications not leading to full stop For long braking cycles the brake was applied continuously after the car was almost stopped This is further evidenced by Figures 1 through 3. In Figure 1 the speed continuously decreased to reach the final speed But for long cycles Figures 2 and 3 the speed reached the final value within 8 to 9 The normalized brake pressures were also relatively high during the first 8 s then were reduced for long cycles The later part of long braking cycles with the normalized speed almost constant is not believed to contribute an appreciable amount of wear debris due to small percent occurrences and lower speed in most cases Considering this fact and the almost constant slope during the initial deceleration the RNBC's can be approximated by truncating the later part with a constant speed and replacing the first part corresponding to the initial 0 to s with constant deceleration and pressure used by some previous investigators Initial speed and average pressure do not differ as much with respect to cycle duration The frequency distribution of real brak- ing cycles was found according to the subsets classified by initial speed speed reduction and cycle duration The frequency distribution was used to generate RBC's and is summarized in Table 1 with other statistics The subsets used in Table 1 could have been classified further by average pressure But the average pressures had only slight variation for the braking cycles in a subset This can be explained by the fact that when the driving road is flat braking cycles have three degrees of freedom Therefore there was no need of further subdivision Table 1 shows that a majority of the cycles were less than 7.5 s in duration amounting to approximately 64 of the cases and 23 had brake applications from 7.5 to 15 s The cycles with brake applications from 15 to 30 s or from 30 to 45 s comprised only a small portion totaling 12 and % respectively Also most of the braking cycles had initial speed from 15 to 30 mph About half of the brake applications were snubs Overall the brake applications were quite frequent resulting in an average interval of 26 s GENERATION OF REPRESENTATIVE BRAKING CYCLES - Based on the frequency distribution and statistics shown in Table 1 real cycles representing an individual subset were generated by calculating real time speed and pressure from Eqs 1 through 3 Median values of intervals that specify subsets in Table 1 were employed for cycle duration initial speed and speed reduction Using these values we could calculate the RBC values from the normalized values shown in Figures 1 through 3 When the brake and sampling system were installed general tests were performed using the RBC for all collected cycles Details of the tests are explained later This representative cycle was generated similarly from the RNBC for all braking cycles The calculated values of RBC's were converted to binary code by the computer and punched on paper tapes by ASCII for later use in simulation experiments TEST FACILITIES AND RUN OF REPRESENTATIVE BRAKING CYCLES This section presents a description of the facilities for running RBC's the measure- ments of their emission rates and the esti- mation of emissions for inner driving 101 PRESSURE PRESSURE PRESSURE 4 PRESSURE PRESSURE PRESSURE PRESSURE PRESSURE NORMALIZED 0 NORMALIZED NORMALIZED NORMALIZED NORMALIZED NORMALIZED NORMALIZED p NORMALIZED FI4 Standard Deviation Bar 0 0 a NORMALIZTEIMDE Normalized pressure and standard deviation _ r 10 10 SPEED 4 SPEED SPEED SPEED SPEED NORMALIZED NORMALIZED 05 NORMALIZED NORMALIZED NORMALIZED NORMALIZED 4 L711 T T Stangard Deviation Bar 0 +r 0 10 NORMALIZED TIME b Normalized speed and standard deviation Figure 1 Representative normalized braking cycles cycle duration 0 - 15 s TEST FACILITIES AND GENERAL OPERATION CONDITIONS - An engine dynamometer was equipped with a scale brake system since Preston and Forthofer 7 have shown that only scale brake inertia dynamometers are sufficiently correlated to test vehicles Small sample friction machines cannot duplicate the vehicle brake duty cycle The experimental setup to test the RBC's consists of a driving mechanism brake assembly electronic control system and sample collection system The electronic control system receives the instructions for speed and brake pressure compares them with measured values then actuates the engine or dynamometer and brake master cylinder to achieve the desired braking conditions Figure 4 is a schematic diagram of the setup used in this study The RBC's punched on paper tapes were read by a tape reader and the information was converted into de voltage by the analog function generators The dc output for speed was compared with the speed by the speed controller which was measured by a tachometer attached to the dynamometer shaft For acceleration and deceleration the engine throttle was driven by a servomotor During deceleration the engine throttle was controlled first and the excessive deceleration which could not be handled by the brake and engine was compensated for by the eddy current dyna- mometer For the brake pressure control a master cylinder was actuated by the piston of a diaphragm air cylinder The pressure in air cylinder was controlled by comparing the the 454 @ PRESURENORMALIZED - NORMALIZED inn H Standard Deviation Bar 0. 0 7 Tr Y T 05 NORMALIZED TIME a Normalized pressure and standard deviation : 1.0 10 4 SPEED SPEED SPEED SPEED NORMALIZED NORMALIZED NORMALIZED NORMALIZED NORMALIZED NORMALIZED 4 NORMALIZED H Standard Deviation Bar 0 0 + r NORMALIZED TIME b Normalized speed and standard deviation 7 10 Figure 2 Representative normalized braking cycles cycle duration 15 - 30 dc voltage for pressure command from the analog function generator and measured voltage from the pressure tranducer in the hydraulic line The speed and pressure responses were checked using two real braking cycles collected and showed an accuracy within 2 mph and 5 psig respectively errors that were due to overshooting or response delay The temperature of the brake pad was monitored constantly with a strip chart recorder The thermocouple in the brake pad was installed in the same manner as that used for real cycle recording 5 If the temperature rose beyond 120 the brake application was stopped until the system cooled down The brake system was shrouded with a sampling box The effective volume of the sampling box box volume less space occupied by brake system was about 23 1. An absolute filter Cambridge Filter Corporation DOP Efficiency 99.99 was used to produce dust- free air which was supplied to the sampling box All airborne particulates generated by the brake system were collected on two mm diameter filters Samudra et al 8 recommended polycarbonate membrane filter of pore size less than or equal to 0.4 m as a better choice in electron microscope analysis for counting or measuring asbestos fibers Therefore a 0.2 pm pore size and mm diameter polycarbonate membrane filter Nuclepore Corporation and a mm diameter fluorocarbon glass fiber filter Pallflex Inc. were used for sampling par- ticulates The air flow ratio between these depth and membrane filters was approximately 30 This ratio was determined by first measuring the flow through the depth filter with the valve for the membrane filter closed then adjusting the valve for the membrane filter to get a desired total flow The flow rates were measured with a laminar flow element 4 15- PRES URE PRESSURE PRES URE PRESSURE PRESSURE NORMALIZED NORMALIZED NORMALIZED NORMALIZED NORMALIZEDNORMALIZED ql NORMALIZED NORMALIZED 05 INg l T TA J T mee emmme we| ~~ } Standard Deviation Bar t T T T T ~ T NORMALIZTE IMDE a Normalized pressure and standard deviation . 10 10- SPEED SPEED SPEED SPEED NORMALIZED 4 NORMALIZED NORMALIZED NORMALIZED NORMALIZED NORMALIZED 08- NORMALIZED lL ~ = a om _ 4 Standard Deviation Bar rf T T T T T 05 10 NORMALIZED TIME bl Normalized speed and standard deviation Figure 3 RepreRepresenstative entative Representative normalized braking cycles cycle duration 30 ~ 45 s Merian Instruments together with a manometer The total air flow through the sampling system was about 87 7 residence time in the min giving air a sampling system of about 20 s The air was sampled continuously for about 20 min after the brake application was finished to collect the remaining dust in the system The deposition of brake wear debris arises mainly from diffusion by Brownian motion interception by particle surface con- tact and inertial impaction Therefore the ratio of airborne particulates to deposited particulates debris deposited on the sampling box and entrapped in the brake system depends on the residence time of air in the sampling box Even under the real circumstances de- fining airborne or deposited particulates is difficult because airborne debris departing from the brake deposit on the system ground gradually starts as a function of to time The difficulty arises in how much air flow is allowable to have an appropriate residence time Jacko et al 2 used shrouded brake systems on real cars The flow rate of sampling dir is believed to be less than the one used in this study Williams and Nuhlbaier 9,10 employed an enclosure tube in which the brake system was installed Air flow through this system was approximately 12,000 min Table 1. Subsets of Real Braking Cycles and Their Frequency Distribution Initial Speed mph Speed Reduction mph Braking Cycle Duration % Average Brake Pressure psig Cycle Interval s 15.2 0-15 0-7.5 7.5-15 15-30 30-45 111 116 102 +o 10.2 15-30 45.3 15-30 0-7.5 7.5-15 15-30 30-45 0-7.5 7.5-15 15-30 30-45 124 137 117 112 163 154 128 oo: 23.3 31.3 0-15 30-45 15-30 30-45 0-7.5 7.5-15 15-30 30-45 0-7.5 7.5-15 15-30 30-45 0-7.5 7.5-15 15-30 30-45 124 135 ore cee 163 165 148 122 226 188 149 138 32.7 43.7 52.1 % is % occurrences corresponding to subsets Subset or RBC No. 1233 1233 1233 4 56700 56700 56700 8 2012 2012 2012 2012 2525 2525 2525 2525 7890 7890 7890 20 2222 2222 23 2222 Therefore the sampling conditions of the current study including the ratio of airborne debris to total debris generated are in the range of the results from these two investiga- tions The top and bottom plates of the sampling box were removable Before a test they were removed and both the brake system and the sampling box were cleaned by blowing with compressed air Then the bottom plate was covered with tared aluminum foil and assembled to the box together with the top plate After a test the top plate was replaced with the one equip- ped with a tight rubber glove and vacuum and compressed air lines The vacuum line was connected to a filter With a combination of vacuum cleaning and compressed blow the deposit debris was collected on the filter After cleaning the aluminum foil and filters were measured to find the weight of debris collected BRAKE PAD CONDITIONING AND MEASUREMENT OF PARTICULATE EMISSIONS BY RUNNING REPRESENTATIVE BRAKING CYCLES - After the experimental setup for running various braking cycles on the engine dynamometer was installed the vibration and other indicators of mechanical stability were checked Then the RBC's were run to determine particle size distributions and emission rates Because new brake pads and a new rotor were used special tests were run to determine the effect of new parts on emissions and to generate an use brake system before running various representative cycles The detailed test procedure shown in Figure 5 will be described below wmnn zeajag Of Sue xog apy veay waisdsSusjdwes ayeg dnqos AlupapinggsajluowmeUAg 4 Dynamoetr Japsozay peyoutradx 2eunwe) suawtz po ayesweadip ainjosery Membrane Filter Membrane any pasds bm cy twaish BUI) att] 1093).Buiang yeuBsig ON397 JNBIPA 30 jeowysa 20 JY 77 soem Depth (O1e want g GojtuyUID TT 7 QD sapeay ode) 10 ( START ) Run representative braking cycle for all braking cycles collected to test new brake system NO Stability of brake application YES Generate use brake system Run representative braking cycle for all braking cycles Stability of brake application YES Find size distribution | Run various representative braking cycles and define emission rates Figure 5 Test procedure used to determine properties of brake wear debris To determine the general wear behavior of a new brake system and to estimate the approximate average wear rate and particle size distribution we used the RBC generated for all collected braking cycles We needed to measure the approximate wear rate because the appropriate amount of sample collected on a membrane filter for electron microscope analysis can be estimated from the number of braking cycles applied and the sample flow ratio of two filters By using the RBC of all braking cycles we performed the initial wear test for a new brake system until the wear rate per brake application was stabilized Each run consisted of about 75 brake applications After about 200 brake applications the initial wear rate appeared stable During this test emission rates of airborne particulates ranged from roughly 0.3 to 0.5 brake application Therefore initial wear can be considered un- important Preston 11 indicated that the repeat- ability of performance data obtained from inertia dynamometer tests can be materially influenced by the preparation of the friction elements of brakes It is important then in brake wear tests to prepare the friction materials in a uniformly used condition Therefore before collection of samples for various RBC's the brake system was conditioned to generate use brake pads and discs according to the following procedures 12 1 Inspect whether the disc surface has 1020 1020 1020 ...m surface finish 2 Run 200 stop burnish 40 - 0 mph at 3 sdeceleration operating at 120 3 Run 10 stop fade 60 4.5 sdeceleration intervals - 0 mph at with 35 4 Run 12 stop recovery 30 - 0 mph at 3 sdeceleration starting 2 min after the last fade stop and at min intervals thereafter The above procedure was programmed on a paper tape and run with the top and bottom cover plate of the sampling box removed Because the braking cycles were applied to the brake system on the dynamometer without a wind effect the ambient air was blown into the box by a fan at a constant speed of approximately 30 mph After system conditioning the wear rate of the brake system was rechecked by running the RBC of all braking cycles The test consisting of 30 - 40 braking cycles was repeated until the wear rate stabilized Then the particulate size distributions of the brake wear debris were measured with two Anderson impactors The first impactor was directly connected to the line of the sampling box by removing filters as shown in Figure 4. The sampling probe of the second impactor was located right after the brake pad This sampling probe was made of in stainless steel tubing with a sharp edge Figure 6 is a schematic of the test arrangement The required air flow rate of each impactor was 28.3 min giving a total flow rate of 56.6 m~-n This value differs from the original sampling condition of 87.7 min The sampling probe of the second impactor col- lected airborne particles mixed with debris born directly from the pad Therefore the airborne particles passing through the first sampling probe cannot exactly represent those of the system originally designed as depicted in Figure 4. We believe that samples col- lected by the first sampling probe in Figure 6 give a reasonable approximation of the airborne particle samples of the original system Based on the above discussion we assumed that the samples through the first and second probes represented airborne particles and airborne particles added with initial wear debris deposited dust plus airborne dust We will denote the size distribution functions of par- ticulates from the first and second probes as W and W respectively where d stands for aerodynamic diameter The aerodynamic diameter of a particle is defined as the diameter of a sphere of density 1 cmwith the same falling speed in air as the particle The particle size distribution W. d of the initial wear debris can be estimated by M M WdW ( + = 4 M where M = weight of particulates col- lected by first probe M = weight of particulates collected by second probe Similarly the particle size distribution Wild of deposited dust can be calculated by measuring the corresponding weight My d W = M+ M W d -M W -M W -72 d M 5. Samples for determining particle size distribution were collected by running about 20 braking cycles Table 2 shows the values of particle size distributions W and W2 measured with Anderson impactors Table 2 also includes the estimated values of size distributions of W. d and W calculated from Eqs 4 and ) Deposited dust weighed 34.73 mg M. while the sampled particulates weighed 7.69 mg M and 15.26 mg % The maximum in particle size distribution of air- borne dust occurred at of 2.1 - 3.3 pm while the the aerodynamic diameter initial wear debris had double peaks at 3.3 - 4.7 pm and 1.1 - 2.1 ...m The deposited dust had double peaks in the same range as the initial wear dust The average aerodynamic diameters of airborne deposited and initial wear dust did not show appreciable differences having values of 3 03 3.06 and 3.13 ...m respectively The size distribution of airborne particulates was close to that measured by Williams and quite Muhlbaier 9 who reported the maximum distri bution in aerodynamic diameter in the same range 1.1 - 4.7 ...m Table 2. Measured and Estimated Particle Size Distributions of Various Particulates Impactor Stage Aerodynamic Diameter ...m Measured Values Airborne Mixed Wear Particulates Particulates Wd W2 Estimated Values Initial Wear Particulates Deposited Particulates W d d 1 7.0 - 11 2 4.7 -7 3 3.3 - 4.7 4 2.1 - 3.3 5 1.1 2.1 6 0.65 " 1.1 7 0.43 0.43 - 0 65 Backup Filter TOTAL 0.0 - 0.43 + Average Aerodynamic Diameter +++ 0.105 0.081 0.212 0.226 0.219 0.048 0.015 0.094 1.0 3.68 0.102 0.083 0.209 0.149 0.296 0.082 0.025 0.054 1.0 -+- 0.099 0.080 0 206 0.071 0.374 0.117 0.035 0.014 1.0 3.13 0.090 0.088 0.203 0.003 0 442 0.147 0.147 0.043 0.022 10 3.06 12 Filtered ( ~) Ambient Air Air Sampling Second Sampling Probe First Sampling Probe 1st Impactor || 2nd Impactor -- Filter Valve + ig 4 Brake Disc Brake Pad AlumAilnuumminum , FoFoiill ? Sampling Box Ly To Vacuum Pump Figure 6 Schematic diagram of Anderson impactor arrangement for particle size measurement After measuring the particle size distributions we performed preliminary electron microscope tests to find a proper leading range on membrane filters and to confirm the adequacy of the procedures selected for electron micro- scope analysis The RBC for all braking cycles was used for this purpose We found that the loading range of 0.1 - 0.2 mg was adequate for fiber counting by electron microscopic analysis Finally various RBC's defined in Table 1 were run to collect airborne and deposited particulates Usually one test was run for each RBC For each test several mem- brane and depth filters were collected by applying an appropriate number of braking cycles to get a proper loading on a membrane filter by trial and error The sampling box and brake system were then cleaned to find deposited particulates The details of tests and samples collected are summarized in Reference 6. The weight ratios of samples on membrane and depth filters were close to the air flow ratio of 1:30 The emission rate per brake application and percentage of airborne particulates to total wear were calculated for each cycle and are summarized in Table 3. First the averages of the airborne and deposited particulates per brake test application were calculated for each If there were more than one test for an RBC the average was used to represent it Average emission rates of airborne and deposited particulates were estimated for all RBC's by weighting the frequency distribution listed in Table 1. The total wear per brake application and percentage of airborne particulates were calculated directly from emission rates of airborne particulates and deposited particulates shown in Table 3 Generally particulate emissions increased for higher speeds which was expected considering the mechanical work done during a braking cycle The percentage of airborne particulates also increased with respect to speed These increases were probably caused by agitation by the rotor at a higher speed but a change in particle size distribution with speed may also be a contributing factor The weighted average of particulate emissions from a front disc brake was 2.43 brake application Airborne particulates accounted for about 31 or equivalently 0.75 brake application These results are comparable with those reported by Williams and Muhlbaier which showed the total wear and airborne par- ticulate emissions to be 3.3 stop and 1.6 stop respectively Table 3. Particulate Emissions Application and Percent of Airborne Particulates for Various Representative Braking Cycles RBC No. Airborne Particulates Brake Application mg Deposited Particulates Brake Application mg Total Wear Brake Application mg Airborne Particulates % ] : 3 4 0.105 0.210 0.210 a8 5 6 7 B 0.378 0.875 0.856 0.756 9 0.297 022 0.831 022 0.474 0.474 12 wc 2525 2.830 2525 4.592 2525 soe 2525 -- 17 18 19 20 21 22 23 24 2.098 2.903 3.266 4.196 1.324 3.410 3.202 2.648 Weighted Average 0.746 See Table 1 for RBC numbers 0.258 0.516 0.516 alin 0.896 2.052 2 628 1.792 1.270 2.849 1.812 ane 3.604 4.600 ae ote 3.300 7.363 5.936 6.624 1.925 5.304 2.391 3.849 1.682 0.363 0.726 0.726 oe 1.274 2.927 3.484 2.548 1.567 3.680 2.286 ote 6.434 9.192 434 ---- 5.399 10.266 9.202 10.820 3.254 8.714 5.593 6.497 2.428 28.93 28.93 28 93 eos 26.67 29 89 24.57 26.67 18.95 18.95 22.58 20 73 orn 43.99 4996 4996 oo oe 38.86 38.86 28.28 35.49 38.78 40.76 39.13 57.25 40.76 30.72 No test was done due to small percentage occurrence of braking cycles Instead emission rates were estimated from a braking cycle having the same initial speed v and speed reduction AV Emission rates of a long cycle duration > 7.5 s were assumed to be twice more than those of the shortest cycle duration 0 - 7.5 s Emission rates of the shortest cycle were taken as a half of the longest cycle Using the average brake application 5 05 times and the results in Table 3 we calculated the average emissions per mile The total particulate emissions were 11.84 mile the airborne particulates emissions from a disc brake were 3.78 mile while the deposited particulates were generated at a rate of 8.06 mile To find the significance of background particulates we sampled the air that had passed through the absolute filter three times on membrane filters during the tests The re- sults indicated that the contribution of back- ground particulates in filtered air was only 0.017 g The airborne particulate concen- tration from the disc brake had an value greater than 1.8 gduring average a test MEASUREMENT OF ASBESTOS FIBERS AND ESTIMATION OF THEIR EMISSIONS After measuring the various particulate emission rates of the RBC's we analyzed sam- ples with proper loading on a transmission electron microscope TEM to count the number of asbestos fibers and find their sizes Asbestos concentrations were estimated by cal- culating the asbestos fiber volumes This section presents the procedures followed and results found 14 ELECTRON MICROSCOPE ANALYSIS - Trans- mission electron microscopy often used with selected area electron diffraction has to date been the most reliable technique used to identify and characterize asbestos fibers in ambient air and water samples Other analytical methods such as ray diffraction and thermogravimetric analysis are less sensitive than electron microscopy In this study a fiber is defined as a particle having a diameter ratio of at least 3 a definition that has been used internationally by occupational hygienists when monitoring dust Characterization of asbestos fiber levels is detailed in Reference 8 and summarized below First a quarter section of the mm filter was cut and placed on a slide The particulate sample on the quarter section was secured by carbon depositing in a carbon evaporator The carbon sample was cut into 2 x 2 mm squares and transferred onto electron microscope grids coated with Formvar film The layers of the polycarbonate filter and Formvar film were dissolved in a Jaffe Washer with chloroform transfer was completed 30 for about 10 min as a solvent Sample by drying in an oven at The prepared samples were examined with an optical microscope to check for adequacy of preparation and to count fibers longer than 5 u m The magnification factor for phase contrast was 400x The samples were then mounted on a TEM and fibers were observed to measure length and width at 10,000 to 22,000x magnification When fibers were detected only chrysotile fibers were sorted out by identifying elemental composition and crystal type with energydispersive ray spectroscopy and selected area electron diffraction Major elements appearing in ray analysis were silicon magnesium Fiber and iron number and sizes were determined for fiber bundles whenever possible Identi fication and size determination of chrysotile fibers were in most cases straightforward In some cases interfering foreign materials or condensed hydrocarbons partially covered the fibers and we approximated the fiber size we believed the effect of the approximation on measurement error was minimal Photographs in Figures 7 and 8 show three typical topographies of various fibers appearing during TEM analysis ESTIMATION OF ASBESTOS FIBER EMISSIONS - Based on the fiber counting and size measure- ment general statistics related asbestos fibers were estimated to chrysotile First we computed the individual volumes of fibers and bundles By using the scanned area during fiber counting and the specific gravity of chrysotile asbestos 2.56 we calculated the asbestos fiber weight per unit filter area The asbestos concentration was calculated by dividing this number by the particulate weight per unit filter area for the corresponding sample The number of single fibers and the number of fiber bundles per unit mass were estimated similarly During these estimations the average size of single fibers average length of fiber bundles and average number of~ single fibers in a bundle were also calculated These results are summarized in Reference 6 as well as information about the samples used for TEM analysis The results indicate that there is no strong correlation between mechanical work done during a cycle and number of fibers per unit mass particulates or their sizes Since the airborne particulate emissions per brake application were known for each RBC Table 3 we can calculate the number of airborne single fibers and bundles emitted per brake application as well as airborne asbestos emissions per brake application These calculated results and asbestos concentrations of airborne particulates are shown in Table 4 for all RBC's The weighted averages were calculated similarly to those in Table 3 from the frequency distribution of RBC's in Table 1 In Table 4 the asbestos concentration of airborne particulates does not show any noticeable correlation with mechanical work done during a cycle but shows a slight increase for higher speed Asbestos emissions or the number of fibers emitted per brake application increased clearly together with mechanical work This arose from the higher particulate emissions rather than from higher asbestos fiber concentrations The average asbestos concentration is about 0.02 for downtown driving The weighted average of airborne asbestos emissions was 0.24 brake application which contained approximately 3.87 million single fibers and 1.98 million fiber bundles The weighted averages for single fiber size bundle length and number of fibers in a bundle were also estimated using the frequency distribution of RBC's The average single fiber length was 0.54 pm with 8.2 aspect ratio while bundles were 0.68 pm in average length containing about 3.4 bundle For an average brake application of 5.05 times the estimated asbestos emissions were 1.13 mile Single fiber emissions were 1.95 x 107 mile and fiber bundle emissions were 9.98 x 106 mile The asbestos fiber level in the background particulates was also analyzed The results indicate that the background asbestos concentration was less than % of the weighted average concentration in Table 4 and the max~-mum contribution of background asbestos was less than % DISCUSSION In this section the test results from this study are compared with those from previous investigations The asbestos contri- bution from automobile brakes to the ambient level is also discussed 15 Figure 7. Typical topography of fibers fiber bundle 48,000x top single fiber 80,000x bottom Table 4. Airborne Asbestos Fiber Emissions Application and Other Statistics for Various Representative Braking Cycles RBC No. Asbestos Concentration of Airborne Particulates % Asbestos Emissions Brake Application g No. of Single Fibers Brake Application No. of Fiber Bundles Brake Application 1 2 3 3.73 x 10-4 3.73 x 1074 3.73 ^ 10-4 4 wan 3.91 ^ 1004 1.16 1074 1.16 x 10 wae 1.84 x 105 3.67 x 105 3.67 x 10 wae 6.12 x 10 1.22 x 105 1.22 x 105 wee 5 6 7 8 1.36 x 10-3 6.19 ^ 10-2 1.18 x 10 1.18 ^ 10-2 5.15 x 10 5.42 x 10-1 1.01 x 10 8.92 x 10-3 3.02 x 10 2.48 x 10 2.90 x 106 2.56 x 106 4.53 x 106 2.57 x 106 9.25 x 105 8.17 x 10 9 72 1.93 ^ 10 11 4.87 ^ 10-2 2.89 x 10 1 61 x 10 2.31 x 10 34 ^ 106 6.86 x 10 1 65 x 108 1.05 x 105 2.83 x 105 2.06 x 106 4.06 x 10-2 14 27 ^ 10 16 an 179 1.78 ^ 10-2 18 1.78 ^ 10-2 4.69 x 1075 20 1.02 x 10-2 1.15 5.81 x 10 -- 72 x 1071 5.15 x 10 1.53 x 1003 4.26 x 10 1.20 x 10 1.81 x 107 wee 5.90 x 106 8.16 x 106 6.88 x 106 9.48 x 200 16 x 106 1.60 x 106 a 8.90 x 105 1.23 x 106 1.25 x 106 6.34 X 105 continued 18 Table 4. continued RBC No. Asbestos Concentration of Airborne Particulates % Asbestos Emissions Brake Application g No. of Single Fibers Brake Application No. of Fiber Bundles Brake Application 21 22 23 24 7.97 x 10 7.97 x 10-2 1.42 x 1071 3.10 x 10 1.06 2.72 4.56 8.21 1.62 x 106 4.17 x 106 3.53 x 107 1.72 x 107 5.40 x 105 1.39 x 106 2.00 x 106 2.65 x 106 Weighted -2 Average 1.74 x 10 10-1 2.24 x 10-1 106 3.87 x 106 1.98 x 106 * See Table 1 for RBC numbers wee No TEM analysis was done due to small percentage occurrence of braking cycles Instead asbestos concentration and number of fibers or number of bundles per unit mass of airborne particulates were assumed to have the values of the v closest braking reduction Av cycle number with the same initial speed and speed of 11 and sion rates 2.1 and - 3.3 other ... Generally measured values the for emisair- borne and agreement Muhlbaier deposited particulates with those reported by 9 show good Williams and The estimation of particulate emissions is believed to be fairly accurate compared with the estimation of asbestos concentration by fiber counting with an electron microscope even though the brake wear is not highly repeatable The primary uncertainty in asbestos fiber measurements was due to a few large fibers present in a sample which enormously change the concentration and the intrinsic problem in scanning only a tiny portion of a sample to measure fibers with an electron microscope The asbestos concentra- tions of airborne particulates ranged from 0.14 to 0.5 x 104 in this study with the average at 1.8 x 10 % Jacko et al 2 showed a somewhat greater value of 0.23 but the result by Muhlbaier and Williams 3 0.03 agrees well with the current study In the three investigations compared above the RBC's and their weighting factors were different from each other Furthermore temperature thermal wear occurred during some of the tests by Jacko et al 2 But if only the Detroit traffic cycle is considered which included no thermal wear and represented city driving the three results can be compared For the Detroit traffic cycle the asbestos emission rate from automobile brakes was 21.8 car mile of which 0.852 car mile was airborne Muhlbaier and Williams 3 estimated 2.2 g airborne asbestos emissions per brake application By multiplying this value by 2.0 mile the airborne asbestos emission rate was 4.4 car mile They also indicated that there appeared to be no significant difference between the asbestos content of particulates from disc and drum brakes Based on this and the airborne particulate emission between drum and disc brakes the ratio of airborne 69 asbestos emission rate from a car was estimated the same way as the particulate emissions The result was an emission rate of 3.82 car mile The study comparison showed a fairly good agreement even though the lower emission rate reported by Jacko et al 2 was believed to be due to the sampling system generating a lower percentage of airborne particles Muhlbaler and Williams 3 estimated the contribution of asbestos emissions from auto- mobile brakes to the environment by using the registered automobiles in the United States in 1977 and the total amount of asbestos disposed or emitted to the environment in 1974. They concluded that the asbestos emitted from auto- mobile brakes accounts for approximately 0.0019 of the total United States asbestos emissions or 0.23 of airborne asbestos emissions These calculations were based on total 9.2 g asbestos emissions mile or 4.4 g airborne asbestos emissions car mile Since about 32 of total particulates emitted in our study were airborne the total asbestos emissions can be approximated from the airborne asbestos emission rate of 3.82 car mile resulting in 11.9 car mile Considering these close results we can infer that the asbestos emissions from automobiles contribute only a small portion to the total national level The total number of airborne asbestos fibers emitted in this study was 3.0 x 107 fibers mile the fibers averaged 0.59 m in length There were about 26,000 fibers of asbestos Muhlbaier and Williams 3 reported compatible values showing a median fiber length of 0.5 pm and fiber density of 90,000 fibers of asbestos The results of these three typical investigations are summarized in Table 5. Other investigations also used an electron microscope for asbestos analysis Anderson et al 13 used a brake dynamometer and concluded that particulates emitted from automobile brakes contained only 0.005 asbestos Rowson 14 collected airborne particles from a shrouded brake assembly mounted on an inertia dynamometer and measured an asbestos concentration less than 0.5 in wear debris All these reports show the asbestos concentration to be less than 0.5 CONTRIBUTION OF BRAKE ASBESTOS IN URBAN AIR - The RBC's and the frequency distribution used to measure particulate and asbestos emis- sions from automobile brakes in the current study were based on inner driving Because most brake emissions are expected in urban areas it would be appropriate to estimate automobile brake contributions to the asbestos level in urban air by using meteorological models or surrogate models These models can determine the maximum possible contribution of brake emissions to the ambient asbestos level from which the corresponding health effect can be inferred Bradow 15 used dispersion models devel- oped as a result of the U.S. Environmental Protection Agency's Regional Air Pollution Study to find automobile exhaust particulate contributions for the city of St. Louis MO In the calculations the central city area had a maximum annual average concentration of 13 gfor the vehicle of 0.195 mile For particulate emission the same models the rate contribution of asbestos emissions from auto- mobiles in the current study is a maximum of 0.25 mfor the emission rate of 3.8 car mile Based on the measured value for the number of fibers per unit mass 26,000 fibers the contribution is equivalent to 6500 fibers The automobile asbestos emissions in St. Louis can be similiar to those for other large cities in the United States From the data in Reference 1 the ambient asbestos concentrations in large cities Dayton OH and Frankfort KY are excluded vary from mately 20 1.6 to ng 200 ng averaging approx~-- Therefore the maximum con- tribution of asbestos from automobiles from 0.13 to 16 for large cities the maximum contribution is about 1.3 ranges average Williams and Muhlbaier 4 estimated the contribution of brake wear asbestos in New York City by using a lead tracer model In this model the vehicle emission was assumed to contribute to the ambient air at the same ratio as lead lead emission emissions rate from a In their study a car was found to be 108 mile of which 60 remains airborne with a typical air concentration of 2 -car Using the same ratios we found an asbestos contribution of 0.12 mfor an emission rate of 3.8 car mile This value is close to that obtained by Williams and Muhlbaier 4 Since the asbestos concentration of ambient air in New m1 York City ranges the contribution from from 8 to 41 automobile brakes represents only 0.3 to 1.5 and Considering the results of the dispersion lead trace models the maximum asbestos contribution from automobiles to the ambient air in cities is 0.25 mand ranges from < 1 to 16 SUMMARY AND CONCLUSIONS RBC's and their frequency distribution were found by statistically analyzing downtown city driving in Raleigh NC The particulate as well as asbestos emissions from automobile brakes were characterized by simulating these RBC's on a controlled dynamometer Average total particulate emissions deposited plus airborne were approximately 12 brake mile of which 3.8 mg or 32 remained airborne The average aerodynamic diameter of airborne particulates was 3.7 pm Based on the 1.6 wear ratio between disc and drum brakes the estimated airborne particulate emission rate was 12.8 car mile Airborne particulates contained 0.018 asbestos fibers resulting in an asbestos emission rate of 1.1 brake mile or 3.8 car mile for the same disc to drum brake wear ratio The average length of single fibers and fiber bundles was about 0.6 ...m From dispersion models for the city of St. Louis MO the maximum contribution of asbestos emissions from automobiles to the ambient air was 0.25 mor 6500 fibers Assuming other big cities have similar dispersion and emission factors the contribution of automobile brakes varied from 0.13 to 16 averaging 1.3 with Results from this those of Williams investigation and Muhlbaier agree 4 well Judging from the current and previous investigations using electron microscope analysis asbestos emissions from automobile brakes con- tribute little to the ambient asbestos level compared to other sources Table 5. Comparison of Test Results with Those of Previous Investigations Parameter Jacko et al 2 Williams and Muhlbaier 3,9 Current Study Test Facility RBC's Percent of Airborne Particulates Airborne Particulate Emission Rate car mile Asbestos Concentration % Analysis Method for Asbestos Fibers Airborne Asbestos Emission Rate car mile Real car Brake dynamometer 6 brake 8 performance test cycles and Detroit traffic cycle 2 22 55 oe 10.8 0.23 TEM 1.05 0.03 TEM 4.4 Brake dynamometer 15 32 12.8 0.018 TEM 3.8 TEM = transmission electron microscope REFERENCES 1. Suta B.E. and Levine R.J. 1979. NonOccupational Asbestos Emissions and Exposures In Asbestos Properties Applications and Hazards Volume I Michaels L. and Chissick S.S. editor p 179 New York John Wiley and Sons Jacko M.G. Ducharme R.T. and Somers J. 1973. Brake and Clutch Emissions Generated During Vehicle Operation 730548. New York NY motive Engineers Society of Auto- Muhlbaier J.L. and Williams R.L. 1980 Characterization of Asbestos Emissions from Brakes 3435 Warren MI General Motors Research Laboratories Williams R.L. and Muhlbaier J.L. Contribution of Brake Lining Wear to Urban Airborne Asbestos 3422 Warren MI General Motors Research Laboratories 1980 5 Society of Automotive Engineers 1978 Brake System Road Test Passenger Car and Light Duty Truck J843d In SAE Handbook Part II pp 31.69-31.73 Warrendale PA Society of Automotive Engineers 6. Cha S. and Carter P. 1982. Estima- tion of Urban Automobile Brake Emissions through Simulation of Wear Dynamics 82-08 Research Triangle Park NC Northrop Services Inc. 7 Preston J.D. and Forthofer R.J. 1971 Correlation of Vehicle Dynamometer and Other Laboratory Tests for Brake Friction Materials 71050. New York NY Society of Automotive Engineers B. Samudra A.V. Bock F.C. Harwood C.F. and Stockham J.D. 1978. Evaluating and Optimizing Electron Microscope Methods Characterizing Airborne Asbestos EPA 2-78-038 Research Triangle Park for NC U.S. Environmental Protection Agency 21 Williams R.L. and Muhlbaier J.L. 1980 Gas and Particulate Emission Rates from Asbestos Linings 3434 Warren MI General Motors Research Laboratories 10 Williams R.L. 1980. Design and Construction of a Test Facility to Charac- terize Brake Wear Emissions 3477 Warren MI General Motors Research Laboratories 11 Preston J.D. 1973. Inertia Dynamometer Evaluation of Brake Lining Materials 730192. New York NY Society of Automotive Engineers 12 Jacko M.G. and Ducharme R.T. 1973 Simulation and Characterization of Used Brake Friction Material and Rotors 730191. New York NY Society of Automotive Engineers 13 Anderson A.E. Gealer R.L. McCune R.L. and Sprys J.W. 1973. Asbestos Emissions from Brake Dynamometer Tests 730549. New York NY Society of Automotive Engineers Rowson M. 1978. The Chrysotile Content of the Wear Debris of Brake Linings Wear 315-321 Bradow L Emissions 797-811 1980. Diesel Particulate Bull N.Y. Acad Med This paper is subject to revision Statements and opinions ad- vanced in papers or discussion are the author's and are his responsibility not SAE's however the paper has been edited by SAE for uniform styling and format Discussion will be printed with the paper if it is published in SAE Transactions For permission to publish this paper in full or in part contact the SAE Publications Division Persons wishing to submit papers to be considered for presentation or publication through SAE should send the manu script or a 300 word abstract of a proposed manuscript to Secretary Engineering Activity Board SAE Printed in U.S.A.