Protective clothing used in the explosive industry

Following an ignition in a fireworks production facility 1 , work was done to quantify the size, duration and thermal emissive power of fireballs produced when different quantities of a range of pyrotechnic compositions are ignited 2 . This study was subsequently followed by an equivalent investigation relating to propellants 3 .

Work of this nature can be used to estimate the thermal threat to exposed workers and also to produce possible estimates of burn injuries 4,5 . However, as a result of risk assessments, steps are generally taken to reduce this type of exposure. Engineering controls (such as remote working, barriers and screens) are the preferred method since these offer the best protection. In those circumstances where engineering controls are not reasonably practicable, personal protective equipment (PPE) is used as a last resort.

Since the commercial UK explosives industry is not particularly large, little specialised PPE is available and fire protective clothing from other hazardous industries (e.g. oil, offshore) is generally utilised 6 .

Little work has been done to assess the effectiveness of this PPE against the heat outputs from burning explosives (which generally produce an intense, short duration event).

This situation was reviewed some years ago 7 and a need identified to develop appropriate test methods to complement the thermal output studies. Interim guidance on the selection of fire protective clothing was published by the Confederation of British Industry 8 and this document also recognised that work was needed to test and rank PPE when exposed to burning explosives.

As a result, HSE commissioned HSL and the British Textile Technology Group to jointly develop an instrumented manikin to evaluate PPE for use in the explosives industry. The design concept involved assessing not only frontal exposure at a simulated workbench containing burning material but also the exposure received at the rear of an escaping worker. This was achieved by having the manikin mounted on rails and capable of rotation and retraction from the source of the heat.

Details of the design and instrumentation of the manikin have been published elsewhere 9 , as have the results of commissioning trials 10 . The purpose of the present article is to report and analyse data obtained for a range of body clothing exposed to different burning explosives.


The unclothed manikin mounted on the turntable and located on the track is shown in Figure 1. Some of the 120 flush-mounted thermal sensors are clearly visible. These comprise copper-constantan thermocouples attached to matt black copper discs; flush-mounting to the body of the manikin ensures that where the clothing is in contact with the manikin it is also in contact with the sensors.

Explosive Physical Form/ Quantity Typical Duration, s Flame shape Mean diameter, m Mean Peak Surface Emissive Power, kW m -2
Firearm propellant (AS/24) Loose powder, 5 kg 4 Cylindrical 1.0 1380
Flare composition (SR 562H) Loose powder, 5 kg 10 Cylindrical 3.2 670
Star composition (SR 700) Loose powder, 5 kg 2-3 Cylindrical 7.0 1000
Magnesium Teflon Viton (MTV) Pressed pellet, 0.39 kg 2-3 Cylindrical Unknown Unknown

Sensors are located over 88% of the body, including the head, legs and arms of the manikin but not the hands, which were protected by appropriate heat resistant mitts, nor the feet. The un-sensored hands and feet therefore account for the remaining 12%. The manikin would need to be modified to enable evaluation of the fire protection offered by gloves since it currently has no fingers to the hands.

The outputs from the thermocouple responses were recorded continuously during each test using a data-logger located in the chest cavity of the manikin. Post-processing involves utilising the approach advocated by Henriques and Moritz 11 , in the draft international standard for instrumented manikins, to estimate skin burns 12 . This uses the criteria for degrees for surface skin burn developed by Stoll and Greene 13 supplemented by the work of Takata 14 for deeper burn injuries. Thus, the degrees of burn injury can be categorised as

  • 1st degree burn – burn damage to the epidermis
  • 2nd degree burn – partial thickness burn to the dermis
  • 3rd degree burn – full thickness burn to the dermis and onset of damage to the subcutaneous layers

The motion of the manikin is programmable and the time of front exposure, speed of rotation and speed of escape along the 5.5 m track can all be varied.

Fixed conditions were used for all the experiments undertaken and these were based on documented reaction times to fireball radiation and typical escape speeds 9 . Thus, for the experiments reported in this article, the manikin was stationary at a waist high (88 cm) work-bench containing the burning explosive for 2 s, after which it retracted, while still facing the source, for 1 s, and then continued to retreat while making a 180 o turn over the next second. The manikin subsequently continued along the track at a velocity of 2.5 m s-1, braking at the end.

Designation Garment Size State
A White double-layer Proban-treated cotton coverall 180/108 cm Unused and unwashed
B Proban-treated cotton jacket and trousers (Blue) Medium Used, less than 10 wash cycles
C Nomex III coverall 112 cm Removed from service; used and washed
D Proban-treated cotton coverall 108-116 cm Removed from service; used and washed
E White double-layer Proban-treated cotton coverall 180/108 cm Unused but washed for 50 cycles
F Proban-treated cotton jacket and trousers Medium Removed from service; used and washed for 40-50 cycles
G Pyrovatex treated 70% cotton/30% polyester coverall 180/108 cm Unused and unwashed
H 100% Nomex III coverall chrome leather jacket. Aluminised rayon trousers, braces and fly front. Aluminised rayon jacket. Leather belt. Removed from service
I Firotex coverall Suede leather jacket 100% cotton underwear (vest and long johns) Leather apron Removed from service
J 100% Nomex III coverall chrome leather jacket chest cool pack Aluminised rayon long coat* Removed from service

Details of the explosives used to provide the thermal sources in the trials are listed in Table 1. The clothing evaluated in this study is summarised in Table 2.

For all tests, some form of personal protective headwear covered the head of the manikin. The results obtained for the performance of this headwear, which included more specialised systems for use when working with energetic compositions such as MTV, are presented in a separate article 15 . Since the purpose of the present study was to assess and rank the performance of body clothing, the burn injury data for the head were not included in the analysis of the percentage of the body receiving burns.

Clothing System Average Predicted Burn Injury, %
3rd o Burn 2nd o Burn 1st o Burn Total ≥2nd o Burn
A 2 3 3 8 5
B 14 8 4 26 22
C 3 3 3 9 6
D 3 3 4 10 6
E 0 0 3 3 0
F 17 5 4 26 22
G 7 5 4 16 12
Clothing System Average Predicted Burn Injury, %
3rd o Burn 2nd o Burn 1st o Burn Total ≥2nd o Burn
A 1 3 2 6 4
B 11 6 4 21 17
C 5 6 3 14 11
D 29 5 5 39 34
E 3 3 3 9 6
F 5 3 4 12 8
G 7 4 4 15 11
Clothing System Average Predicted Burn Injury, %
3rd o Burn 2nd o Burn 1st o Burn Total ≥2nd o Burn
A 23 10 8 41 33
B 73 7 0 80 80
C 42 0 9 61 52
D 67 5 3 75 72
E 29 9 7 45 38
G 44 9 4 57 53

Testing with flare and star compositions and firearm propellant was carried out in a sheltered location in the open. Triplicate tests were generally done in order to provide mean data and therefore minimise the effects of localised wind on the results. No testing was done when the wind speed was greater than 2 m s -1 .

Tests with MTV were done in a large enclosed facility.

In order to compare the results of the full scale clothing evaluations presented in this article with data obtained using small scale fabric fire tests, experiments were done to test methods EN 367:1992 16 and EN 366:1993 Method B using a heat flux of 20 kW m -2 17 .


Published data 18 were used to convert the outputs from the heat sensors to the responses of human skin and, subsequently, to the degrees of burn injury at different temperatures 12 . The external face of the skin was assumed to be at a temperature of 32.5 o C in all cases and the percentages of the body of the manikin receiving burns relate to the 88% proportion of the body surface area that is sensored i.e. the maximum injury is 88%.

A typical processed result, for clothing system E exposed to the fireball produced by 5 kg of star composition, is shown in Figure 2. In this case a Proban-treated cotton anti-flash hood protected the head of the manikin.

A specific sensor located just above the head of the manikin was used to monitor the thermal threat and the other sensors on the body provided data to enable evaluation of the percentages of the surface of the manikin that would receive first, second and third degree burns. The remote head sensor was useful in evaluating the repeatability of the exposure and the effect of localised wind conditions on the fireballs. For example, in the flare composition tests with clothing system D there was a large variation in the received thermal energy, which was reflected in the burn injury predictions.

Tables 3, 4 and 5 summarise the data obtained in tests on the body clothing systems using burning firearm propellant, flare composition and star composition, respectively.

The performance of the clothing systems is ranked against total burn injuries (the summation of percentages of first, second and third degree burns) in Table 6 and against second degree and above burns (i.e. those causing very serious injury requiring hospital treatment or that could result in fatality) in Table 7.

Clothing System Ranking* vs Composition Mean ranking
Firearm Propellant (AS/24) Flare Composition (SR 562H) Star Composition (SR 700)
A 2 1 1 1
B =6 6 6 7
C =3 =4 4 3
D =3 7 5 6
E 1 2 2 2
F =6 3 5
G 5 =4 3 4

The performance of the specialised clothing used to protect against effects resulting from the ignition of MTV is summarised in Table 8.

Ranking Firearm Propellant (AS/24) Flare Composition (SR 562 H) Star Composition (SR 700)
1 E A A
2 A E E
3 C, D F C, G
4 C, D C, G C, G
5 G C, G D
6 B, F B B
7 B, F D

Table 9 presents the results obtained using the small scale fabric fire tests EN 367:1992 for convective heat transfer and EN 366:1993 for radiant heat transfer.


The results presented in Tables 3 and 4 suggest that exposure of a worker to the thermal effects produced by the ignition of 5 kg quantities of propellant or flare composition would be serious, generally requiring hospital treatment, but would be survivable. In both cases the majority of injuries in the tests were to the front of the chest region with the back remaining relatively undamaged. In relation to garment design, these sets of results highlight that better heat protection may be achieved by using double (or multiple) layers (garments A and E) rather than by changing the material. Kim and co-workers 19 have previously noted the critical role played by air gaps in multiple layer fabrics in relation to the heat transfer process.

Star composition produced a substantially larger average received radiation at the manikin (686 kW m -2 ) than the firearm propellant (397 kW m -2 ) and the flare composition (329 kW m -2 ), as evaluated from the monitoring sensor in the head. This suggests that superior protective equipment would be advisable when working with star composition and the results in Table 5 substantiate this. The extent of second and third degree burn injuries indicates that very serious or fatal burns could result even when wearing the clothing tested in this study. The tests with star composition confirm the observations made above for the tests with propellant and flare composition relating to double layer garment performance.

Table 6 provides a ranking of the performance of the PPE evaluated in this study in relation to the thermal protection provided against a range of sources. The mean ranking for the seven clothing systems provides a crude guide to selection for performance against similar thermal threats.

Table 7 examines the effectiveness of the clothing in preventing more serious (second degree and above) burn injuries. Generally, the trends are encouraging; for example, the top two rankings only contain two types of clothing, both of which were composed of double-layer fabrics. Again, a summary of this nature should help inform the selection of appropriate clothing for protection against incident thermal threats in the range ca. 300-700 kW m -2 .

Whereas, in principle at least, the manikin test enables accident scenarios to be reproduced and therefore an assessment to be made of the performance of protective body clothing against realistic thermal threats, it is recognised that small-scale test methods are not necessarily applicable to actual fire conditions and are often useful merely for ranking materials. Nevertheless, it is noted that samples of clothing systems A and E gave the longest times in both the convective and radiant heat transfer tests, Table 9. Since HTI24 and t 2 are essentially the times to second degree burns, the results indicate that, in common with the findings from the manikin test presented in Table 7, body clothing systems A and E are evaluated as having the potential to offer the best protection. The correlation of rankings between the manikin method and the two small scale tests is encouraging and may suggest that the small scale methods could have a role in screening the performance of certain fabrics.

The data obtained when assessing systems for protection against the ignition of MTV military infrared decoy flare composition, Table 8, are more limited than those discussed above. Nevertheless, some interesting observations were made. In the test with clothing J, the contaminated outer surface of the garment continued to burn for at least a minute after withdrawal of the manikin from the ignition point. The ‘cool-pack’ that forms part of the ensemble was thought to have protected the chest from burn injury because this intense burning activity was not reflected in the burn injury body map. Generally, however, this test highlights the serious effects that can result from wearing clothing contaminated with explosive compositions that are capable of rapid burning. The MTV results indicate that burn injuries were exclusively in the leg, rather than the torso, region of the body and therefore more likely to be survivable. The superior performance of clothing system I is ascribed to the presence of a leather apron that provides direct shielding for the legs.

Clothing System Proportion of manikin body surface (%) receiving burn injury (1 o , 2 o and 3 o ) Mean (%) Proportion of manikin body surface (%) receiving 2 o and higher burns (%) Mean (%)
Run 1 Run 2 Run 3 Run 1 Run 2 Run 3
H 11 5.1* 15.4 13.2 6.6 2.2* 8.1 7.4
I 1.5 0 0 0.5 1.4 0 0 0.5
J 9.5 9.5 6.6 6.6
Clothing System Convective heat transfer EN 367:1992 Radiant heat transfer EN 366:1993 Method B, 20 kW m-2
HTI 24 (s) HTI 24 -HTI 12 (s) t 2 (s) t 2 -t 1 (s)
A 9.2 3.0 17 6
B 6.6 1.3 13 5
C 6.4 2.0 13 5
D 6.1 1.8 13 5
E 9.6 3.1 18 7
F 6.7 1.5 13 5
G 5.8 1.4 13 5

One factor that also needs to be considered in relation to cleanliness is that aluminised systems, such as those used for protection against MTV ignitions, offer excellent protection against radiation when shiny. However, once they are dirty they become much more effective conductors.

The effect of repeated washing on the performance of fire protective clothing has been previously identified as an issue (see, for example, reference 7). The data obtained in the present study enable a comparison to be made between new and laundered clothing. Samples A and E and, to a lesser extent, B and F (the colours differ) are relevant: comparison of the mean rankings in Table 6 indicates that there is no significant difference in the performance of the clothing systems, suggesting that little or no protection is lost in 50 washing cycles.

Generally, washing causes all textiles to shrink by about 3% (this applies also to cotton which exhibited negligible thermal shrinkage in the tests) and the key result of this in terms of affecting performance may have been expected to be the reduction of insulating air gaps rather than loss of any intrinsic fire retardant nature of the fabric. Shrinkage on washing increases the mass per unit area of fabrics but this can be countered by the loss of fibres from the weave during the laundering process. Mass per unit area data for test samples A (outer fabric), E (outer fabric), B and F of 330, 333, 384 and 366 g m-2, respectively, indicate that these changes were probably not significant. Fabric production variability could also exert an influence.

As mentioned above, the present study has enabled a relevant assessment to be made of a range of fire protective body clothing in realistic conditions. The performance of the clothing is, however, influenced by a number of factors and the role of some of these has not been examined.

Although the present study is limited in the range of clothing systems evaluated and number of tests done, examination of the results for Proban-treated cotton coveralls (garment A) and the Proban-treated cotton jacket/trouser combination (garment B) indicates that flame ingress may have occurred in the latter case and given rise to the more extensive burn injury predictions observed in the waist region. Clearly, the likelihood of this will be dependent on the intensity and location of the fireball source but the possibility of potentially improved performance with one-piece garments may need to be considered as part of any selection process.

Fit is also of prime importance as this will affect the air insulation properties of the garment. The clothing tested was selected to be an appropriate fit to the manikin in all cases. Since the manikin has the dimensional characteristics of a six-foot tall male of athletic build there will clearly be issues relating to the degree of fit of the PPE to different body shapes and sizes, including women. This has been recognised, for example, by Kim and Whang20 who have developed a family of nine manikins to represent Korean males with ages between 30 and 50.

The main part of the study was also limited by the fact that only three primary thermal sources were examined. Additionally, the degree of thermal exposure is highly dependent on the programming of the manikin movement sequence, although this was standardised in the present study in order to enable comparisons to be made.

It has already been noted that the manikin currently does not permit evaluation of the performance of hand protection, as the hands are not sensored. Work in this area would be complex and would involve detailed instrumentation. More importantly, House21 has highlighted the importance of head protection when considering clothing designed to enhance survivability when exposed to fire or flash. Although not considered in our paper, full protection against thermal threat will involve the appropriate selection of a range of PPE covering body, head, hand and foot protection.

Initial data for a range of protective headwear are presented in Part 2 of this article15. It is recognised, however, that more detailed work is needed to extend our initial studies: this could involve the assessment of protection offered by complete ensembles against a representative range of realistic threats.

Finally, it should be recognised that the manufacture of many explosives is a lengthy and labour-intensive process and the comfort of workers will be an essential consideration when selecting PPE. The ideal solution would involve the use of the minimum weight of clothing that provides the maximum protection against any particular threat.


The results presented in this article indicate that the use of the moving manikin system provides a useful, realistic and repeatable method of assessing fire protective clothing for workers involved in certain types of explosive manufacturing.

Work of this nature could also be used to refine product design. For example, the current study on body clothing has identified the much better performance offered by double layer garments, which suggests that design improvements could be made to certain single layer coveralls. The effectiveness of these changes could then be evaluated in a full-scale simulation test using the manikin.

The responsive manikin system outlined in this article could also have applications beyond explosives workers. Examples are in the offshore oil industry, for the fire services and for the defence industry.


  • R. K. Wharton, Areas for future safety research identified by recent accidents in the pyrotechnics industry, Proc.16th International Pyrotechnics Seminar, Jönköping, Sweden 1991, 514.
  • R. K. Wharton, J. A. Harding, A. J. Barratt and R. Merrifield, Measurement of the size, duration and thermal output of fireballs produced by a range of pyrotechnics, Proc. 21st International Pyrotechnics Seminar, Moscow 1995, 916.
  • R. K. Wharton and R. Merrifield, Measurement of the size, duration and thermal output of fireballs produced by a range of propellants, Propellants, Explosives, Pyrotechnics 2000, 25, 179.
  • R. K. Wharton and R. Merrifield, Thermal hazards from the accidental ignition of pyrotechnic compositions, J. Pyro. 1997, 6, 1.
  • B. Lawton, R. Merrifield and R. K. Wharton, Thermal output of pyrotechnic compositions and evaluation of skin burns, models for aircrew safety assessment: uses, limitations and requirements, p. 32-1 to 32-12, held at Hope Hotel and Conference Center, Wright– Patterson Air Force Base, Dayton, Ohio, USA, 26-28 Oct. 1998.
  • R. Merrifield, Safe handling requirements during explosive, propellant and pyrotechnic manufacture, Specialist Inspector Reports No. 31, 1991, Health and Safety Executive.
  • N. I. Sorensen and R.K. Wharton, Fire Protective Clothing in the UK Pyrotechnics Manufacturing Industry – A critical review of current practice, Proc.16th International Pyrotechnics Seminar, Jönköping, Sweden, 1991, 747.
  • Fire Protective Clothing, CBI Explosives Industry Group, Confederation of British Industry, 1995 (ISBN 085201 513 5).
  • R. K. Wharton, A. W. Train, N. I. Sorensen, R. W. Rennell and R. I. Davies, The development of a portable manikin for assessing fire protective clothing for explosives workers, Explosives Engineering, June 1998, 18.
  • R. K. Wharton, N. I. Sorensen and A. W. Train, The commissioning of a versatile manikin for assessing fire protective clothing, “Thermal Protection of Man under Hot and Hazardous Conditions”, p265-270, 24-26 March 1999, Paris, France.
  • F. C. Henriques and A. R. Moritz, Studies of thermal injury – V. The predictability and the significance of thermally induced rate processes leading to irreversible epidermal injury, Archives of Pathology 1947, 43 489.
  • Protective clothing against heat and flame – Test method for complete garments – Prediction of burn injury using an instrumented manikin, Draft International Standard ISO/DIS 13506.2, 2000.
  • A. Stoll and L. C. Greene, Relationship between pain and tissue damage due to thermal radiation, J. Appl. Physiol. 1959, 14, 373.
  • A. N. Takata, Development of criterion for skin burns, Aerospace Medic. 1975, 45, 634.
  • R. K. Wharton, D. H. Pickering, A. W Train, N. I. Sorensen, M. T. Healey and P. M. Eaton, The application of a moving manikin for assessing and ranking the performance of fire protective clothing used in the explosives industry. Part 2. Headwear, Propellants, Explosives, Pyrotechnics, in press.
  • EN 367: 1992 Protective clothing – Protection against heat and fire-Method of test: evaluation of materials and material assemblies when exposed to a source of convective heat
  • EN 366: 1993 Protective clothing – Protection against heat and fire Method of test: evaluation of materials and material assemblies when exposed to a source of radiant heat
  • J. D. Dale, E. M. Crown, M. Y. Ackerman, E. Leung and K. B. Rigakis, Instrumented mannequin evaluation of thermal protective clothing, ASTM STP 1133, American Society for Testing and Materials, Philadelphia.
  • I. L. Kim, C. Lee, P. Li, B. D. Corner and S. Paquette, Investigation of air gaps entrapped in protective clothing systems, Fire Mater. 2002, 26, 121.
  • J. H. Kim and M. C. Whang, Technical Note. Development of a set of Korean manikins, Appl. Ergonomics 1997, 28, 407.
  • J. R. House, The level of head protection is the most important factor to consider for clothing ensembles designed to enhance survivability when exposed to fire or flash, 5th Scandanavian Symposium on Protective Clothing, 5-8 May 1997, Elsinore, Denmark 1997.
  • Acknowledgements

    The authors gratefully acknowledge the contributions made by the following organisations in donating protective clothing for testing:

    Institute of Naval Medicine, Alverstoke, Gosport, Hampshire, PO12 2DL.

    Standard Fireworks (now Black Cat Fireworks Ltd.), Standard Drive, Crossland Hill, Huddersfield, West Yorkshire, HD4 7AD.

    Chemring Countermeasures (formerly Pains Wessex Ltd), High Post, Salisbury, Wiltshire, SP4 6AS.

    Wallop Defence Systems, Craydown Lane, Middle Wallop, Nr Stockbridge, Hampshire, SO20 8DX.

    Particular thanks go to Wallop Defence Systems for the use of their indoor test facility during trials with MTV, and to the following individuals for their assistance during the project: M McGowan and S Vass (Wallop Defence Systems); B McCay and K Oliver (Chemring Countermeasures); J House (Institute of Naval Medicine).

    © Crown copyright (2008)

    For more information on Protective Clothing

    Published: 10th Jan 2008 in Health and Safety International