Measurements of the effectiveness of dust control on cut-off saws used in the construction industry

ABSTRACT

Materials used in the construction industry frequently contain large quantities of silica. When they are cut or shaped with power tools considerable respirable dust can be produced. Three dust control systems for use with cut-off saws have been evaluated on site: wet dust suppression using mains water, the same system using water from a portable water tank, and local exhaust ventilation. The efficiency of water suppression on cut-off saws has been precisely quantified in controlled laboratory conditions by means of measurements with and without dust control. When dust control was used on-site, the mean concentrations of airborne silica were reduced by a factor of between three and seven, the accuracy being limited by the relatively high limit of detection for silica. All controls systems generally reduced respirable dust levels by at least 90%. Although the effectiveness of dust suppression did not depend on blade type, a diamond blade was more effective than a resin-bonded blade with the pressurised water system; cutting a slab with this type of blade could be completed before the water tank required repressurisation. In laboratory tests, the application of water reduced the dust concentration to <4% of its value without control. The method for monitoring the dust concentration was sufficiently sensitive to measure a difference in concentration produced during cutting in different directions. It is important, however, that the pressure in supply reservoirs is properly maintained, that the water is correctly applied and that it is used at the correct rate. If this is done effective dust control can be achieved.

ON-SITE WORK

Hand held cut-off saws used on construction sites are often powered by small capacity two-stroke combustion engines or by 110 V electric motors.

Normally the saw blades are 305 mm (12 in.) or 230 mm (9 in.) in diameter, and they may be either diamond tip or thermal resin-glass fibre composition. Both types of saw and blade were considered in this survey.

Two dust control systems involving wet dust suppression and one using local exhaust ventilation (LEV) have been examined. The wet systems are similar in principle as they both rely upon water applied to the rotating cutting disc to reduce dust emission. However, one system utilises water provided by a portable pressurised tank as shown in Fig. 2, and the other requires water supplied from the mains as shown in Fig. 1. The LEV system involves air extraction, and is made more effective by the substantial containment provided by the receptor guard attachment as shown in Fig. 3.

The three systems were assessed independently, between November 1995-January 1996, at three separate sites. Cutting of kerbs and slabs with a silica content ranging between 12±40% was carried out by experienced ground-work sub-contractors. Unfortunately it was not possible to standardise on the material, and the silica content of the materials
cut could not be measured until after the trials. A single operator was employed for each system, to eliminate the effects of variation in work practices, such as operator position. Cutting with and without controls was carried out on the same day to minimise the effects of the variable weather conditions. Cutting time was approx. 15 min, reflecting typical work and exposure patterns within the industry, with and without dust control systems fitted.

Personal protective equipment (PPE) worn by operators during cutting work included particulate respirators, either FFP3 filtering facepieces or orinasal respirators with P3 filters, safety glasses/goggles, ear defenders and head protection. Airborne personal sampling for respirable dust was carried out as described in MDHS 14 (1997), which is believed to give a fair representation of personal exposure, using Casella cyclone-operated samplers with a sampling rate of 1.9 l min-1 on operators' right and left lapels, and all samples were taken in duplicate. The mean concentration provided a better estimate of silica dust levels in the operators' breathing zone. Both gravimetric analysis and i.r. spectrophotometry were carried out on the samples.

Portable tank (wet) system

This equipment is supplied by a number of plant-hire companies and major manufacturers of cut-off saws. The system used in the survey consisted of a polypropylene bottle containing approx. 8.5 litres of water and pressurised by hand. The bottle was connected by narrow plastic tubing to two brass heads attached to the same side of the guard. An on/off valve controlled the water supply, which was filtered to prevent blockage of the heads. These were adjustable, and they supplied a fine jet of water to the disc face. The saw had a two-stroke engine with a capacity of 70 cm3 , giving a nominal speed of 5000 rpm. It could be used with a 20 segment 305 mm diamond blade or a reinforceable silicon (medium grain) resin blade.

Mains water (wet) system

The dust suppression equipment used to put this method into effect is essentially the same as the tank system, except that the water comes from mains supply through a hose, to two water jets located on each side of the guard. The saw was driven by two-stroke engine with a capacity of 65 cm3 and a nominal speed of 5100 rpm. It could be used with a 305 mm diamond blade or a composite resin blade of the same size.

LEV system

LEV is a well established technique, but the system used in our tests took a sophisticated form. It was specifically designed for the saw, and takes into account whether the user is left- or right-handed. More importantly, it adjusts to the depth of the cut by means of an inner sleeve contained with the guard, which can be adjusted to ensure maximum containment during cutting. Alternative systems are "add-on" guard attachments, which often provide only limited enclosure. The system tested was driven by a 2400 W electric motor with a nominal speed of 6500 rpm. It was used with either a diamond blade of diameter 230 mm or a composite resin blade of the same diameter. The dust extraction was provided by a vacuum cleaner, the filter grade of which was not known.

Method

The portable pressurised water tank system was used first with a diamond blade to cut paving slabs 600 X 900 mm, ten cuts in all giving a total cut length of 9 m. This was followed by cuts on similar slabs using a resin blade, three full cuts and a fraction, in all giving a total length of 3.15 m. The silica content of the slab cut was 20%.

The mains water system was used with a diamond blade to cut paving slabs 600 X 900 mm, twelve cuts giving a total length of 10.8 m, plus a single kerb 250 X 125 X 900 mm, to give a total cut length of approx. 12 m. A second set of tests using a resin blade was applied to similar paving slabs, with five cuts followed by one cut on a similar kerb giving a total cut length of approx. 5.7 m. The silica content of the paving slabs was 40%.

The LEV system was used to cut 600 X 900 mm paving slabs with a diamond blade, twelve cuts in all giving a total length of 10.8 m. This was followed by similar slabs cut with a resin blade three cuts in all giving a total cut length of 2.7 m. The silica content of these paving slabs was only 12%.

Results

The results shown in Tables 1-3 are based on a cutting time of approx. 15 min. The figures were calculated from the time of each cut, but the 8-h TWAs were calculated from the actual time/volume of sampling. For this reason there may appear to be discrepancies between the two sets of values.

The major factors affecting dust exposure are the operator's technique and posture during cutting, and the speed and direction of the wind. The results show differences resulting from sampler locations, i.e. left or right shoulder (cf. Vaughan et al., 1990). The mean levels measured provide information on typical dust concentrations likely to occur during this type of work.

The mean concentrations of silica show reduction by a factor of 3 to 7 when dust control was used on-site, the accuracy being limited by the relatively high limit of detection for silica. However, the three control systems generally reduced respirable dust levels to less than one tenth. For example, use of the control measures reduced mean respirable dust concentrations from 21.2 to 1.3 mg m-3 , 14.4 to 0.6 mg m -3 and 8.0 to 0.7 mg m-3 , with the diamond blade and the pressure tank system, mains water and air extraction respectively. Levels were reduced from 12.0 to 6.4 mg m-3 (the poor control here was due to loss of pressure in the water tank, as described below), 58.0 to 1.9 mg m-3 and 13.3 to 0.2 mg m-3 respectively when resin blades were used.

Although the effectiveness of dust suppression did not depend on blade type, a diamond blade was more effective with the pressurised water system overall, because its use enabled a slab to be cut in approximately one minute using a single pressurisation of the tank, whereas when the resin disc was used the tank needed to be re-pressurised a number of times during cutting.
LABORATORY TESTS

On-site tests provide realistic data but they are susceptible to variation as a result of external forces such as wind speed and direction, which are outside the control of the experimenter. This means that results are not sufficiently accurate to enable subtle distinctions of the sort needed for optimisation, or for discrimination between alternative systems to be made. Accurate data require laboratory experiments, which can be carried out in conditions where external effects can be controlled.

A cut-off saw with provision for the application of water to suppress dust production was sought for the tests. Petrol-driven saws had the provision but indoor tests were not possible because of the production of carbon monoxide during use. The electrically-powered saws examined did not have the provision, but a compressed air-driven saw was available with an optional water feed kit comprising two nozzles which are fitted either side of the blade onto the saw hood, along with a hand-driven water pump and an 8 litre capacity tank. Laboratory tests were carried out on this saw, which was powered by a petrol-driven compressor, which stood outside the laboratory.

Precise measurements required that the saw be tested at a constant cut speed, and that the resultant airborne dust concentration be measured in a steady state air flow (Regnier et al., 1988). The dust concentration will depend on this flow rate, which may not be typical of conditions of use, and so absolute values have no real meaning. However, the constancy of the conditions means that relative values are extremely reliable, and therefore the effectiveness of control systems can be specified accurately by this means. A test rig previously used to test hand sanders (Thorpe and Brown, 1994) inside a large recirculating dust tunnel (Blackford and Heighington, 1986) was used. The apparatus, shown schematically in Fig. 4, comprises an inner (I) and an outer framework (E) made from strong modular sections of extruded aluminium. The material to be cut (H) is mounted on top of the inner frame. This frame can be raised or lowered by an electrically operated telescopic pillar (J), and the height can be set using a limit switch. A linear module (A) with a movable platform driven by a dc motor (B) is attached underneath the top of the outer frame. A single control box regulates the speed and direction of the linear module, and a switch connected to the moving platform controls the direction of movement.

The saw, fitted with a diamond tipped blade (C), was securely attached directly to the platform using four separate brackets (D), necessary because of the saw's complex shape. It was fixed, as shown in Fig. 5, at an angle similar to that observed during a site visit and illustrated in the instruction manual, and the blade was held vertical throughout the tests. The rate of water supply to the saw was monitored by two floating ball type liquid flow indicators with ranges of 0.07-0.55 and 0.2-2 l min-1, along with a liquid flow sensor with ranges of 0.05-1.5 and 0.2- 9 l min-1, both attached to the side of the tunnel. The output of the latter meter consists of pulses at a frequency proportional to the liquid flow rate, and the flow can be logged remotely. Water was applied to the saw, either from the hand pump supplied with the system or from an electric pump, which provided a uniform supply.

The mean dust concentration inside the tunnel was measured with a gravimetric sampler consisting of an open faced sampling head containing a 47 mm diameter high efficiency glass fibre filter, and a Rotheroe and Mitchell high volume pump. The variation of dust concentration with time was measured with a Hand-held Aerosol Monitor (HAM), a direct reading instrument. This instrument is calibrated with Arizona road dust with particle sizes in the respirable range, i.e. <=7 µm in diameter, and so can only be used to make comparative measurements with other types of dust. The two sampling instruments were placed alongside each other in the centre of the tunnel, approx. 3 m downstream of the test rig. The output from the HAM was connected to a data logger module manufactured by Digitron Ltd which plugs into a Psion series 2 organiser.

In order to ensure safe use, the saw's compressed air supply was first passed through a mains operated solenoid valve, interfaced with magnetic contact switches on the doors of the tunnel so that if the doors were opened the power would be cut. Water settling on the floor of the tunnel produces a slip hazard, and so drip trays were placed in the vicinity of the saw and downstream, and an area of carborundum-impregnated non-slip floor covering was securely attached to the floor of the tunnel just downstream of the drip tray. During the tests the operators wore polyurethane rubber or soft rubber soled shoes to ensure a good grip. The paving slabs inside an enclosed space are dangerous to handle, especially when they are partly cut through, and so a small fork lift was used to transport the slabs to the tunnel and to put them on and off the platform of the test rig.

Method

The saw was designed to operate at a pressure of 7 bar and an air consumption of 2-2.4 m 3 min-1 . Its rotational speed, measured with a stroboscope, was adjusted to the specified value of 5100 rpm. The air flow through the tunnel was set to 1 m s-1  and the speed at which the saw moves, its cutting speed, was adjusted to about 0.6 m min-1 .

Cuts were made on 600 X 600 X 48 mm paving slabs placed onto a wooden platform between two wooden stops and fastened on top of the inner frame of the test rig using G clamps, with the cut depth set to approx. 10 mm. At the end of each cut the slab was moved along 2 cm using a length of wooden dowel inserted through the side of the tunnel, so that a new cut could be made in the opposite direction. In practice cuts are normally made with the saw moving away from the operator, but it was easier to return the saw to its original position cutting than not, and it will be seen below that an interesting result emerged from this exercise.

After nine cuts the linear module, saw, logger and sampling pump were switched off. The gravimetric sampling filter was removed and weighed, and the logged data were transferred to a computer spreadsheet program. The drip trays were emptied if necessary. After each test the average cut depth was determined from measurements at four positions along each cut.

When the effects of water on dust suppression were investigated, the water tank was pumped up to maximum pressure, usually with about 26 pump actions, at which a safety valve would begin to open. The initial flow was then set to either 1 or 2 l min-1 and the decrease in flow rate with time was measured. This was then repeated with an initial tank pressure of approximately half the maximum.

Results

The results of all the tests are shown in Table 4 and Figs. 6-8. Dust concentrations are calculated from measurements made during the cutting cycles and not the time between cuts, and are normalised to a standard 10 mm cut depth. The HAM measurements were normalised with respect to the gravimetric results using pooled results from all of the experiments.

When no water was applied to the saw blade the dust concentration was 184 mg m-3 measured gravimetrically, the normalised HAM result being 154.4 mg m-3 . Figure 6 shows that the levels of dust produced in this instance did not vary with the direction of cut.

When the test was repeated with water supplied by the hand pump, at an average flow rate of 0.12 l min-1 and an initial flow rate of 0.2 l min-1 , the concentration was reduced by 55%. Water at a constant flow rate of 0.2 l min-1 , supplied by the electric pump, resulted in a 73% drop in concentration, to 49 mg m-3 , though this is still high. However, increasing the flow rate to 0.5 l min -1 had a dramatic effect, reducing the dust concentration by about 97%, to 6.12 mg m -3 measured gravimetrically, the normalised HAM result being 2.64 mg m-3 . These results are summarised in Fig. 7, and indicate that further increase in water flow to 1 l min -1 does not cause further significant reduction. Figure 8 shows that when water dust suppression was used, the levels of dust were higher on the forward stroke, i.e. with the saw blade spinning in the direction of a wheel moving along the cut, by between 55 and 70%. The reason for this is not known, but the effect was repeatable.

The water flow produced using the hand pump is illustrated in Fig. 9. An 8 litre capacity tank should be capable of supplying water at a flow rate of 0.5 l min-1 for 16 min. However, even the initial pressure inside the tank at its maximum, the minimum adequate flow of 0.5 l m-1 can be sustained only for just over 4 min. The flow is not constant and the initial flows of 2 or 1 l min-1 result in waste of water without any improvement in dust suppression.

The flow of water to each nozzle was observed to be uneven to an extent that depended on the rate of water flow. At total flows of <1 l min-1 the water failed to reach the far nozzle, which was supplied by a length of tube that passed over the saw hood. Smaller diameter nozzles operating at a higher pressure would probably resolve this problem, since the resistance to flow introduced by the nozzles would dominate, reducing the effect of the resistance of any connecting tubes. The manufacturers recommended that a "trickle" of water would be sufficient to control dust levels, but they could not quantify this (private communication). Since there is no water flow indicator supplied with the kit the operator would have no idea whether the flow rate is higher than the 0.5 l min-1 required to give good dust suppression, other than by observing the dust generated.
