Historical Filter Test Rig Designs

For decades filter designs have routinely been tested on laboratory facilities, however the evaluation process in regard to particulate filters for the mining industry is relatively recent. One of the most detailed reports of a test rig specifically constructed to evaluate diesel particulate traps (exhaust filters for over-the-road vehicles) was by the 3M company (Brunner 1995). When developing their system the 3M company determined the primary requirements of a suitable sampling system are that it should have:

• Mass-based measurements

• Ability to sample both sides of the trap simultaneously or in rapid succession

• Minimal time for efficiency test • Excellent repeatability

• Good resolution

• Simplicity of use and maintenance • Minimal cost and space investment

• Fluorocarbon (TFE) coated borosilicate glass filter media, and • Ability to measure the volatile fraction of the particulate

deposited on sample filters on a post-test basis

This resulted in the development of four systems which had the following characteristics (Table 4.4).

Table 4.4

Characteristics of Test Rigs Developed By 3M Co Inc Systems Characteristics 47 mm Raw Gas Sampling 90 mm Raw Gas

Sampling Dilution Partial Sampling Smoke

Mass-based Measurement Yes Yes Yes No Simultaneous Upstream/Downstream

Sampling Yes Yes No Rapid Sequence Minimal Time for Efficiency Test Yes Yes No Yes

Excellent Repeatability No Yes Yes Yes Good Resolution No Yes Yes Yes Simplicity of Use and Maintenance No Yes Yes Yes Minimal Cost/Space Investment Yes/Yes Yes/Yes No/Yes No TFE-coated Borosilicate Glass Sample

Media Yes Yes Yes No

Ability to Measure Volatile Fraction Yes Yes Yes No

Operational experience with these four systems was varied, with the 47 mm raw gas sampling system being discontinued due to the following reasons:

• The small size of the sample tubing resulted in high particulate attenuation between the inlet of the sample probe and the filter holder.

• Unacceptable weighing errors were experienced, particularly on the downstream filter.

• The use of separate vacuum pumps, desiccant columns and gas clocks resulted in high variations in sample flow between upstream and downstream sample trains.

• High maintenance items such as the desiccant resulted in leaks frequently developing in the connecting tubing.

As reported by Brunner (1995), these faults were addressed in the 90 mm raw gas system subsequently constructed by the 3M company. Sample times were limited to two minutes which gave a deposit on the upstream filter of 7 – 15 mg, depending on the mass concentration of the particulate in the exhaust. One issue cited by Brunner (1995) was the difference in the volatile fraction (organic carbon) observed between the upstream and downstream filters. Brunner postulated that the common occurrence of the upstream filter always containing a lower volatile fraction than the downstream filter was due to:

• The upstream filter being at a higher temperature, allowing more of the volatile hydrocarbons to bake off during sampling. • The upstream filter, which collects more particulate mass,

develops a larger pressure drop and thereby increases the mass of hydrocarbons that are volatilised.

• The transport time between the two sample probes allows greater opportunity for particulates to adsorb hydrocarbons.

Brunner concluded that all of the above effects could be realised via conducting an efficiency test of the system (blank test) using a straight pipe in place of a particulate trap.

Such tests had indicated a negative efficiency of up to 10% which approached zero as the flowrate and gas temperature increased.

The third system developed by the 3M company was a partial flow dilution sampling system.

This system was operated in both gravimetric and real time formats by using a TEOM (Tapered Element Oscillating Microbalance) as the analysis technique. The use of the TEOM provides a means by which transient effects can be observed.

In recent years (Lomb 2002), partial flow systems to measure particulates have lost support in that they are critically dependent on the accurate measurement and control of operating parameters in order to give the desired dilutions. A full flow system depends on the operation of critical orifices (whose performance is consistent) and does not require the same level of calibration, measurement or control as a partial flow system. Major diesel research projects within the Australian transport sector (Anyon et al 2000) have focused on full flow systems for these reasons.

Over the past 20 years there have been many variations of the above approaches used to evaluate the performance of particulate traps for use on over-the-road diesel vehicles. In respect to the mining/tunnelling sector, several organisations have conducted the majority of research on diesel particulate traps using engines mounted on dynamometers. These are:

• Canada Centre for Mineral and Energy Technology

(CANMET).

• Mines Safety and Health Administration (MSHA) Approval and Certification Centre at Triadelphia, West Virginia.

• Verminderung der Emissionen von Realmaschinen in

Tunnelban (VERT), a Swiss joint project to curtail the emissions from engines at tunnel sites.

The first reported testing of diesel exhaust control technologies using dynamometer based systems was most probably undertaken by CANMET (Mogan and Dainty 1987) in which a prototype venturi scrubber system was evaluated.

Procedures were developed for the testing of exhaust treatment devices as part of the approval process for use in Canadian non-coal, non-gassy mines (CANMET 1995). These procedures involved isokinetically sampling the exhaust emissions followed by gravimetric analysis on 47 mm fibre of glass filters.

Since 1996 CANMET has been intimately involved in the Diesel Emission Evaluation Programme (DEEP), a joint venture of operators, labour, regulators, research agencies and OEMs. The DEEP programme has funded a number of major research projects of ceramic filter trap evaluation but given the very low number of underground coal mines in Canada there has been no research on DDEFs.

The test system used by MSHA was developed by the US Bureau of Mines (Anderson et al 1992) and involves the use of a partial flow dilution tunnel with the collection of diesel particulate matter on Teflon fluorocarbon polymer-coated glass fibre filters. Gases are also evaluated via direct reading instrumentation and a vapour phase sampler (XAD-2 resin) is installed downstream from the particulate filter.

MSHA (Ambs and Setren 1995) have used the above system to conduct a safety evaluation of DDEF, however the focus was on establishing the levels of emissions generated when the filters were subjected to temperatures above their operational range. No data appears to have been published by MSHA in regard to the efficiencies of DDEFs against diesel particulate. One reason for this is the system in Triadelphia is not fitted with an exhaust water conditioner system (a prerequisite for the use of DDEFs). During discussions at a mining diesel emissions conference in Canada during 2002, this aspect was confirmed (Setren 2002), however the use of a water conditioner was under investigation.

In August 2003 MSHA reported on the limited testing of filters (Stackpole 2003) where samples were collected using a Sierra Instruments particulate sampling system with collected samples being sent to NIOSH in Pittsburgh for diesel particulate matter, elemental and organic carbon analysis.

To date only a total of 50 filters have been tested with a comparison being made to a standard paper filter (Donaldson P530866). A filter was judged to have failed if:

1. An equivalence of 0.97 or greater is not achieved (equivalence = efficiency candidate filter divided by efficiency of Donaldson P530866).

2. Filter shows excessive backpressure.

3. Filter is physically damaged by exposure to exhaust stream.

Using the above criteria, the Microfresh DA101 filter (200 pleat filter in USA format) was judged to have an equivalence of 1.05, the equal highest of the filters tested.

Details of the procedure used by MSHA are not readily available, however the focus appears to be to find a range of filters that meets the MSHA “standard” filter rather than assessing the characteristics of individual filters.

The basis for choosing the MSHA “standard” filter also remains unclear at this stage, which limits the usefulness of this test protocol outside the USA.

VERT commenced the evaluation of particulate traps in 1993 (Mayer 1998) using a 105 kW construction site engine (or equivalent) on a test bed (dynamometer) using ISO 8178 (1996) as the test protocol. Emission measurements for particulate are evaluated gravimetrically and a transient “soot puff” test is performed using an opacimeter. Particle size distribution is also established.

Using the above criteria and the 105 kW reference engine, VERT recommends the following filtration efficiencies for particulate trap systems on construction engines (Table 4.5)

Table 4.5

VERT Filtration Efficiency Recommendations

% EFFICIENCY

Test New After 2,000 Hrs

Total Particulates (gravimetric using ISO 8178 @ 4

test points) >80 >75

Elemental Carbon (coulometric) >90 >85 Soot Puff during Free Acceleration (opacity) <10 <10 Particulate Penetration in Size Range 10 – 500 nm <5 <10

This recommendation was the first to propose an efficiency rating of the filter system under test against elemental carbon, yet few details are provided (Mayer 1998).

As was the case in Canada, VERT has focused on ceramic type filter traps and has not reported any testing of DDEFs.