Document EqKE32w34Rzk0g7q7gkKKpgzx

HSE Health & Safety Executive HSE CONTRACT RESEARCH REPORT No. 40/1992 AN EVALUATION OF THE ROACHES 'DUST PARTICLE APPARATUS' DUSTINESS TESTING EQUIPMENT C P Lyons and D Mark Warren Spring Laboratory Stevenage Price 20.00 HSE Health & Safety Executive HSE CONTRACT RESEARCH REPORT No. 40/1992 AN EVALUATION OF THE ROACHES 'DUST PARTICLE APPARATUS' DUSTINESS TESTING EQUIPMENT C P Lyons and D Mark Warren Spring Laboratory Stevenage This report describes a laboratory evaluation of the Roaches `Dust Particle Apparatus' (DPA) dustiness tester. The performance of the device was evaluated, in comparison with the Warren Spring Laboratory rotating drum dustiness tester (WSL), whilst measuring the dustiness of 30 powder samples recently employed in the British Occupational Hygiene Society (BOHS) round-robin evaluation of dustiness testers. Generally, the dustiness results from the DPA followed the same trends as those measured by the rotating drum. However, the DPA device collected fewer coarse particles than the WSL and may therefore underestimate the dust dispersed in some material handling processes. Nevertheless, the DPA tester was easy to operate and clean, and with some modifications, may be used successfully to determine the rate of dust released when a material is dropped onto a surface in calm air conditions. The results for the WSL device agree well with those obtained previously at this laboratory during the BOHS round-robin exercise. * Copyright Controller HMSO 1992 This report and the work it describes were funded by the Health and Safety Executive. Its contents, including any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy. No part of this publication may be photocopied or otherwise reproduced without the prior permission in writing of the Health and Safety Executive. Enquiries should be addressed to: The Director Warren Spring Laboratory Gunnels Wood Road Stevenage Hertfordshire SGI 2BX Tel: (0438)741122 Telex: 82250 WSLDOIG Fax: (0438) 360858 . DTI Environmental Enquiry Point: 0800 585794 Warren Spring Laboratory is the Environmental Technology Executive Agency of the Department of Trade and Industry AN EVALUATION OF THE ROACHES 'DUST PARTICLE APPARATUS' DUSTINESS TESTING EQUIPMENT C.P.Lyons and D.Mark. SUMMARY This report describes a laboratory evaluation of the Roaches 'Dust Particle Apparatus' (DPA) a new device for measuring the dustiness of powders during handling operations. The performance of the device, a version of the single-drop type of dustiness testers, was evaluated whilst measuring the dustiness of the 30 samples recently employed in the 60HS round-robin evaluation of dustiness testers. For comparative purposes, the dustiness of a further 30 samples of the same materials were measured at the same time using the Warren Spring Laboratory rotating drum tester (WSL). The results showed that generally the dustiness of the material samples, as measured by the DPA dustiness tester, followed the same trends as those measured by the WSL rotating drum. However, the DPA device collected fewer coarse particles than the WSL device, and may therefore underestimate the dust dispersed in some materia! handling processes. The variability of results obtained with the two testers were usually similar, with coefficients of variation generally between 10 to 30%. However, when insufficient dust was produced the variability was high for both testers. With very dusty materials, sampled particles could fall off the downwards-facing filter of the DPA device. The results for the WSL device agree well with those obtained previously at this laboratory tor the same device during the BOHS round-robin exercise. The DPA tester was easy to use and to clean. It could possibly be of use to determine the rate of dust released when a material is dropped once onto a surface in caim air, provided that certain operational inadequacies are improved. The broad impressions from the work demonstrate that considerable research and development work is required before a reliable method of measuring the dustiness of a material is realised. 1 CONTENTS 1. INTRODUCTION 2. AIMS 3. DESCRIPTION OF THE EQUIPMENT 3.1 Roaches Dust Particle Apparatus (DPA) 3.2 Warren Spring Laboratory Rotating Drum (WSL) 4. THE TEST POWDERS 5. EXPERIMENTAL PROCEDURES 6. OPERATIONAL ASPECTS 6.1 The DPA tester 6.2 The WSL rotating drum 7. RESULTS 8. DISCUSSION OF RESULTS ' 9. GENERAL REMARKS 10. CONCLUSIONS 11. ACKNOWLEDGEMENTS . 12. REFERENCES Tables _ Figures 2 1. INTRODUCTION The dust/ness of a powder is its propensity to emit dust during handling operations. This is a matter of great practical importance due to the hazards to health, and nuisance that can be caused in the workplace when high concentrations of airborne dust are produced during materials handling. There is, at present, no standard method for determining the dustiness of a material. A wide variety of different methods have been used to assess dustiness. A preliminary assessment of some of these methods has been carried out by members of the British Occupational Hygiene Society (BOHS) Technology Working Party on Dustiness Estimation, which was published as BOHS Technical Guide No 4, 1985. Subsequent work by this group has involved a round-robin exercise in which samples of ten different powders were analyzed by a number of different laboratories using their own methods. These methods included the impact, or single drop test, a method incorporating the use of a fluidised bed, and the rotating drum method. The data has not yet been fully analyzed, and so no firm conclusions made as to the validity and reliable of the various methods. Members of the Ecological and Toxicological Association of the Dyestuffs Manufacturing Industry have recently reported a single-drop method for assessing the dustiness of reactive dyestuffs (Berger-Schuun et al, 1989). It is manufactured in Britain by Roaches Engineering Ltd. (Dust Particle Apparatus (DPA)), and is currently being used by six international chemical companies. Unfortunately, this method was reported too late to be included in the BOHS-organised studies. The Health and Safety Executive has, as part of its programme, a requirement to recommend a standard test method for dustiness measurement, and so asked Warren Spring Laboratory (WSL) to carry out an assessment of the performance of the new device. This report describes this assessment. 2. AIMS The assessment had three main aims. These were:- 1) to determine the dustiness of the 10 test materials involved in the BOHS round-robin exercise, using the Roaches DPA single drop dustiness tester; 2) to carry out similar measurements with the WSL rotating drum method and compare the two sets of results; and, 3) to assess the performance of the DPA, both in terms of its reproducibility and its ability to rank materials for their dustiness, and in terms of the operational aspects such as ease of use and versatility. 3 Contents 3. DESCRIPTION OF THE EQUIPMENT 3.1 Roaches Dust Particle Apparatus (DPA) The DPA tester is of the impact, or single-drop type of dustiness tester. A schematic diagram of the apparatus is given in Figure 1. It comprises a cylindrical tower 0.6 m long leading to a sealed cylindrical chamber of size 0.2 m high and 0.22 m diameter. This gives a drop height of 0.8 m. A sampling filter holder is fixed to the roof of the chamber, which is mounted onto a baseplate together with a pump and control system. All components in contact with the material under test are constructed from stainless steel. The apparatus, which is automatically controlled, has the following operating cycle. 10 g of test material are placed in the conical hopper at the top of the tower, which is fitted with a trap door. On pressing the start button, the trap door opens and the material falls to the bottom of the chamber. 5 seconds after the trap door opens the vacuum pump starts and a sample of the dust cloud, which is generated in the chamber, is collected, at a flowrate of 15 l/min, onto a pre-weighed glass fibre filter for a period of 2 minutes. After sampling, the filter is taken out and weighed and the amount of dust collected used as a measure of the dustiness of the material tested. In practice, a dummy run. involving the use of the hopper empty, is carried out to enable the flowrate through the filter to be set to the required value. ' 3.2 Warren Spring Laboratory Rotating Drum (WSL) The WSL rotating drum tester was used as a reference tester against which the results from the new DPA tester could be compared. A diagram of the device is given in Figure 2, and a full description can be found both in Taylor (1985) and Higman (1984). It consists of a large drum (300 mm diameter, 460 mm long) with conical ends. It is fitted internally with eight vanes attached to the walls to lift the sample as the drum rotates at 30 rpm. Any dust made airborne during this process is drawn though a modified Andersen size selective sampler using three stages. The preseparator stage, which has a 50% cut point (d50) at an aerodynamic diameter of 10 //m is retained to minimise overloading of the impaction stage which has a d50 of 9 ^/m. The backup filter collects all particles which penetrate the impaction stage. Three measures of material dustiness are therefore obtained; a) for particles <9 //m, b) for particles >9 //m, and c) for all particles. This apparatus, which is automatically controlled, has the following operating cycle. 100 g of the test material is spread evenly over the bottom of the drum. The system is started and the drum rotates. Five seconds later, the vacuum pump starts drawing air through the sampler (and the drum) at 28.3 Ipm (1 cfm) and continues to sample for 1 min. At the end of the test the pump and drum are switched off, and the sampling system is dismantled. The size selecting stages are weighed and the amounts of dust in the three stages calculated. The drum and sampler stages are then cleaned and prepared for the next sample. 4 Contents 4. THE TEST POWDERS The test powders were very kindly supplied to us by Dr Bernard Wells of Unilever Environmental Safety Laboratory. They comprised samples of the same materials that were distributed in the BOHS round-robin assessment of dustiness testers. These were initially selected to give a wide range of dustiness values, but with minimum complications in terms of particle size, homogeneity, low toxicity, and cost.They included; calcium carbonate (precipitated, heavy); calcium orthophosphate (precipitated); charcoal (activated, decolorising); charcoal (animal, technical powder); kaolin (light); magnesium oxide (heavy); magnesium oxide (light); sodium chloride (GPR); sulphur (sublimed, flowers); and talc (fine powder). Three samples of the ten different materials were supplied for each tester. Samples of approximately 10 g, and 100 g in weight were supplied for the DPA and the WSL devices respectively. The samples were randomly ordered and sent to us in 60 numbered containers, 30 for the WSL apparatus and 30 for the DPA tester, but to enable direct comparisons of the results of each test, the order of materials in each set was the same. According to Dr Wells, the samples have been tested once already by single drop methods, but he was confident that this should have no effect on the dust yield in further tests. Consequently, it was considered that, besides providing a means of evaluating the performance of the DPA tester, the new results obtained in this project would (subject to HSE's approval) make a useful addition to the BOHS data set. For the purposes of this evaluation it has been assumed that the three samples of each material are identical. 5. EXPERIMENTAL PROCEDURES In order to minimise additional variability due to variations in temperature and humidity, tests on each sample for both the DPA and the WSL testers were performed (as closely as possible) at the same time . Before testing, the insides of both devices were cleaned thoroughly by a combination of brushing, wiping and vacuum cleaning. Water was not included in the cleaning process to prevent raising the humidity inside the testers. After cleaning, the impaction plate in the WSL tester was greased to minimise particle bounce, 8s recommended by Vaughan (1989). The filters and impaction plate were then weighed on a 5-place balance, the preseparator on a 3-place balance, and the components inserted into their respective testers. The chosen material samples were weighed in their containers, the sample powder was then emptied into the appropriate test apparatus and the testing cycle started, whilst the empty container was reweighed. This enabled the final values of dustiness to be adjusted for variations in the amount of material actually tested. With the DPA, at the end of the tests, the filter was removed from the filter holder, taking care to avoid any loss of material from accidentally knocking the filter surface, and the weight of dust collected on the filter determined. This value was then used to calculate the dustiness of the material. For the WSL tester, the sampler complete with the connecting right angle bend were removed from the rotating drum, and any particles deposited in the bend added to those collected in the preseparator stage. The sampler was dismantled and the weights of dust collected on the three stages determined. These values were then used to calculate the dustiness of the material. During each test measurements of both temperature and relative humidity were made. The tests were carried out without knowledge of which of the samples came from the same material so that subjective bias was eliminated. The information allowing the samples Contents to be grouped into ten sets of three samples was supplied by Dr Wells after the measurements had been made, and this was used to compare the results of the two testers for the same materials. Information identifying the composition material of each set of ten results was not supplied. 6. OPERATIONAL ASPECTS 6.1 The DPA tester Vibration is the main problem with the DPA. The internal pump of the DPA runs continuously throughout the test cycle and a solenoid valve in the line controls the period that dust laden air is drawn through the filter. However, as the pump does not have vibration damping, and is not isolated from the test chamber, the filter is continuously vibrated during the test cycle. Whilst the pump is drawing air through the filter, particle losses due to this vibration should be small. However, when the test is over, air is no longer drawn through the filter, the pump continues to run, and loose particles may be shaken off the downwards-facing filter. With very dusty, non-cohesive materials, losses of this sort could be high, especially if the filter is left in the machine for a long time before the power is switched off. Removal of the filter has to be done with great care as it has a tendency to fall away from the filter holder and again, sampled dust may be lost. The DPA also loses sample when the trap door snaps open, and with some very light and fine materials substantial clouds of the powder are dispersed into the laboratory atmosphere. How much sample is lost in this way will depend on the material and its condition, but as only 10 g are used in the test, the loss could be important. A simple double-trap system would overcome these problems. The DPA, however, has one major advantage over the WSL device - that of cleaning. The entire cleaning process, including dismantling and cleaning the trap door mechanism, takes 15 min at most. 6.2 The WSL rotating drum . The WSL system, despite being in use for some time, still has a number of operational difficulties which may contribute to the overall variability of results. The major problems are to be found in the size separating system. The wall losses in the pipe to the adapted Andersen sampler have not been calibrated and for very dusty materials these losses can be quite large, although in these tests, powder deposited here was added to the dust collected in the pre-separator. Because of the tare weight of the pre-separator it could only be measured to three decimal places whereas the impactor plate and the filter were measured to five decimal places. So an error in the last digit of the pre-separator will have a large effect on measurements of the total dust given off from the least dusty materials. Also, if the pre-separator becomes swamped, as happens with very dusty materials, there is a risk that powder caught here may be re-entrained, causing a degradation of size separation. Another problem with the WSL tester occurs when very dusty materials are measured. The grease layer on the impaction plate becomes overloaded with particles and any new particles may bounce off and be re-entrained to be collected by the filter. So for these 6 Contents materials the < 9 //m fraction will be overestimated. This problem arises partly from the large amount of material used in th.e test (100 g) and partly from the small amount of dust that can be held on a grease layer. In research applications this is not a problem as the test can be repeated with less dust until the plate is no longer overloaded but if it is to be used as a standard test method either the total amount of dust used will have to be reduced for all the samples, possibly reducing the sensitivity of the test. One advantage of the larger sample however, is that representative samples of lumpy or agglomerated materials can be tested. Cleaning the WSL apparatus can be a relatively long job, taking for one material 60 min, but more usually 15 to 30 mins. The nooks and crannies in the drum can become filled with fine sticky particles and these have to be cleaned out to avoid contaminating the next test material. This will need to be improved if the tester is to be used routinely without the risk of contamination from previous test materials. One main operational advantage that the WSL device has over the DPA is that of containment of test sample. No leaks of dust were obvious during the tests. 7. RESULTS The basic measurements for each sample are given in Table 1, from which it is reassuring to note that the conditions of temperature and relative humidity changed very little during the programme of tests, ranging from 14.9 to 18.4 C, and from 39.1 to 47 % RH. Dustiness, or rate of dust emission, is expressed as a fraction of the weight of material used, and the sampling time, using units of g kg'1 min'1. Plots of this data are given in Figures 3 and 4, where the results from the DPA are plotted on the vertical axis, and the horizontal axis comprises either the < 9//m results, or the total dustiness results from the WSL tester. The measurements from the three samples of each material are identified by separate symbols, each set being assigned an arbitrarily chosen letter. Lines of equality (1:1) are drawn on each graph together with a simple linear regression analysis. The relationships from the regression are as follows:- For the WSL < 9 //m results (Figure 3), the correlation coefficient is 0.54, and y = 0.240 x, + 1.177 (1) where y, and x, are the dustiness results from the DPA and the WSL devices respectively. For the WSL total results (Figure 4). the correlation coefficient is 0.62, and y = 0.0275 x2 + 1.174 (2) where y, and x2 are the dustiness results from the DPA and the WSL devices respectively. In order to compare the results from the two testers for the same individual materials, the mean values of the three repeat measurements for each material, the associated standard deviation (<7) and coefficient of variation (cv) were calculated and are given in Table 2 for each tester, together with the ranking in order of dustiness from 1 (the most dusty) to 10 (the least dusty). The results are also presented graphically in Figure 5 for the < 9 //m WSL results, and in Figure 6 for the total dustiness results. Log axes are used so that the 7 Contents characteristics of each material set can be clearly seen. The size of the error bars on each axis of both figures represent + /-1 a about the mean value, but it should be remembered that these are also plotted on the log scale and are therefore of unequal linear size on either side of the mean value. Two further graphs are included in the results to compare the new data obtained with the WSL rotating drum tester with those obtained previously in the BOHS round-robin exercise. The mean values and standard deviations for each material are plotted in Figures 5 and 6 for the < 9//m and the total dustiness, respectively. / Further statistical treatment is beyond the scope of this small project, and as mentioned above, may be carried out by Dr Wells in the BOHS round-robin exercise. 8. DISCUSSION OF THE RESULTS In a number of previously reported studies (for example, BOHS, 1985, Higman et al, 1984) the position of each material in a ranking order of dustiness has been used to compare the relative performance of dustiness testers. Whilst this may be useful when there are a large number of testers to compare, in this study, with just two machines to compare, more accurate comparisons are gained by considering both the variability of the results as well as the mean values. The first point that becomes clear from the results, as displayed in Figures 3 and 4, is that the dustiness values obtained with the DPA tester are nearly always larger than the < 9/ym values obtained with the WSL tester, but almost always smaller than the WSL total dustiness values. Secondly, it is clear from both the coefficients of the regression (correlation coefficients (r) of 0.54, and 0.62 respectively) and the poor fit of the regression lines to the plotted points, that the relationship between the results from the two testers is non linear. Closer inspection of the two graphs may lead to the conclusion that the regression analyses are biassed to a large extent by the results from material B. However, analyses carried out with these results omitted did not improve the relationships greatly. For the < 9 pm dustiness results, the value of r improves to 0.66, and for the total dustiness results, r drops to 0.48. These findings are to be expected if the mechanisms involved in dust dispersion and sampling in the two testers are compared. The DPA device relies on a single drop of 0.8 m and impact on a hard surface to disperse dust. In this process some fine particles are stripped from the edge of the stream of material by aerodynamic shear but most of the dust is produced by 'splash' producing a disturbed airflow in the test chamber caused by compaction of bulk material when the fall is arrested. Whilst this latter mechanism may disperse large particles into the air, they may not be sampled because of elutriation before reaching the filter surface in the roof of the chamber. Nevertheless, this method may be useful for determining the quantity of dust dispersed from a given material when it is dropped once onto a surface (e.g. in conveyor transfer points). In the WSL tester, material is continuously lifted and dropped, and so dust is dispersed by both the same mechanisms as the DPA device and also by attrition as the particles mix together. Sampling takes place from the axis of the drum, and so elutriation effects are relatively small. In any case, to allow for this, particles deposited on the internal walls of the connecting pipe between the drum and the sampler are added to the preseparator catch. It is expected therefore, that the WSL device would disperse, and 8 Contents Table 1. Dustiness data for the 30 samples tested. Sample Mass tested (flm) WSL DPA 1 101.6 11.6 2 101.8 11.7 3 108.7 11.1 4 100.1 11.2 5 103.0 11.2 6 102.5 11.5 7 102.3 15.0 8 103.7 11.3 9 103.5 11.1 10 101.2 11.4 11 103.5 12.0 12 103.4 10.6 13 100.9 12.3 14 106.3 11.5 15 100.5 11.9 Temp (C) 15.7 15.9 15.9 15.9 16.2 16.6 16.3 15.6 15.7 18.0 18.2 17.7 18.2 18.3 16.9 RH (%) 45.7 45.9 46.4 46.9 47.0 46.5 46.3 46.7 46.8 48.2 48.2 48.2 48.2 48.0 39.1 Normalised Dustiness (gm kg'1 min'1) < 9//m WSL > 9/sm Total DPA 0.16 12.20 0.02 0.053 0.75 1.29 0.65 97.86 0.03 2.00 7.16 21.70 0.81 110.06 0.05 2.05 7.91 22.99 0.27 5.00 0.005 0.11 3.65 2.64 0.87 10.10 10.97 3.47 0.13 0.37 0.21 0.56 0.34 0.93 1.73 0.33 12.35 1.72 101.61 8.70 113.96 10.42 1.41 1.67 0.22 30.70 30.92 0.92 0.16 0.36 0.52 1.61 0.02 0.03 0.05 0.003 14.97 54.85 69.82 5.56 Contents Table 1 (cont). Sample No Mass tested (gm) WSL DPA 16 102.2 13.1 17 103.8 11.8 18 102.5 11.8 19 102.1 10.5 20 102.8 10.8 21 101.6 11.5 22 101.2 10.6 23 104.5 10.3 24 103.8 12.2 25 100.9 11.5 26 101.0 9.4 27 103.8 11.5 28 103.0 11.9 29 101.5 10.6 30 98.0 11.6 Temp (C) 18.5 18.4 17.8 15.3 16.6 16.2 16.7 15.5 16.3 16.2 16.2 16.1 14.9 16.0 15.7 RH (%) 41.1 41.0 39.4 40.1 40.0 40.5 39.4 38.9 39.9 42.4 43.0 42.6 42.9 42.3 42.5 Normalised Dustiness (gm kg'1 min'1) <9//m WSL ' > 9pm Total DPA 0.21 1.45 0.62 1.70 8.90 1.30 1.91 10.35 1.92 1.26 3.84 0.78 0.16 1.57 0.58 0.03 1.68 0.17 1.28 0.16 0.04 0.18 1.33 0.06 0.82 8.10 1.27 1.21 7.70 0.69 22.65 0.32 0.08 10.17 24.93 1.50 0.98 9.67 1.85 1.24 9.38 0.86 23.93 0.48 0.12 10.35 26.26 1.56 0.28 1.62 0.74 0.12 2.90 0.31 1.46 2.27 0.001 1.85 2.58 0.10 Contents Table 2. Moan and standard deviations for the dustiness of oach material tasted elonQ with tfie coefficients of variation and ranking in order from the most to the least dusty. IIII 4 ,2 20,3 0 I||I 6 ,2 5F,2 9 1 11,20,23 M CO to cn co to vO>D x Vo to O hi o> - m sg aO Vo -si O tn w> to A O Oo O o to O b> in io o o o o 0) to in o KJ in CO *4 a 'o o O o o o o o o oo to o> o M o M CO o o to sj oto o * CO 10 to CO O to O Co to O to o> Kt in CB CO A to O - sj ? p* CD to sj o * O to o hi to o CO to SJ to in Sj O o 4k at A *J Sj O sj to K> O p o o o to O 4k CB u 4k o -j co 0> M CO io to b to to oSj 03 9 cn co co to CO 0> A to O SJ tn N> in o (0 in *o CO to A o> O - CD A co <0 CO CO in -in4 o A A to A Cat CO 10 Vj A to o o CO sj CO A O CD to to CD p usj O a o> O hi io sr io to o co CO o 6 to co io in o b h Of Jk M M e> A Kot *to KJ O to W A sg CO to <n o> 3 - 09 Mo to CO o O CO O CO * o o> b M is si to CO bcn 8 to io co to CO .. oo oO oooo O u00 iun to O Vo CO in cn o 8 to 4k b H hCDi hi to to Kt hi A * CO os>j A 03 sj o> * si <n to to CO O 09 Cft If eZ ?* ft f s M a o i < A f a31 . * ? a3 CL o 2 E. "< 3J a * 5 w r ac. a S' 5' V co 1 32 7T 0- <ft 1s 3 _r a g. a 5' ca 2aa 3 a o 2. ~< Ja S- -el 2aa3 ft Q. o i *" < 3ft5 s- o*0 > ac a* ft 1! -< 2| 3-S 'a E. a s- Contents Drum cross-section Figure 2. The Warren Spring Laboratory Rotating Drum Tester (WSL) Contents w uiuj ^ B>i 6) sssui^sno BB6L ISM 100 Figure 7. Graph showing the mean < 9 //m fraction for each material measured using the WSL in 1988 compared with the present values. The error bars show the standard deviation on each measurement and the straight line corresponds to a 1:1 relation between the tw o measurements. O* o Contents /o--* --o ( uau B>| 6) ... ssaui^sna gg6l ISM * l- 1000 Figure 8. Graph showing the mean total dustiness of each material measured using the WSL in 1988 compared w ith the present values. The error bars show the standard deviation on each measurement and the straight line corresponds to a 1:1 relation between the two measurements. o -o-o' 6 Contents