Eckhoff RK — Dust Explosions In The Process Industries, 3 Ed

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying,recording, or otherwise, without the prior written permission of the publisher.
Recognizing the importance of preserving what has been written, Elsevier Science prints its books on acid-free paper whenever possible.

Library of Congress Cataloging-in-PublicationData Eckhoff, Rolf K.
Dust explosions in the process industries / Rolf Eckhoff.-3rd ed.
p. cm.
Includes bibliographical references and index.
ISBN 0-7506-7602-7 (HC : alk. paper) 1. Dust explosions. 2. Fire prevention. 3. Industrial accidents. I. Title.
TH9446.D86E34 2003 604.7-dc2 1

2003040799

British Library Cataloguing-in-PublicationData A catalogue record for this book is available from the British Library.
The publisher offers special discounts on bulk orders of this book.
For information, please contact:
Manager of Special Sales Elsevier Science 200 Wheeler Road Burlington, MA 01803 Tel: 781-313-4700 Fax: 781-313-4882 For information on all Elsevier Science publications available, contact our World Wide Web home page at: http://www.elsevier.com 10 9 8 7 6 5 4 3 2 I

Printed in the United States of America

With deep gratitude for their love and support, I dedicate this book to my wife Astrid and our children Kristian, Ragnar, Solveig, and Jorunn, and their families, and to my mother and the memory of my father. The words in Isaiah 42.16 also gave me hope and courage.

Gonte nts
Foreword
Prefaces

1 Dust Explosions-Origin, Propagation, Prevention, and Mitigation:
An Overview 1.1 The nature of dust explosions 1.1.1 The phenomenon 1.1.2 Materials that can cause dust explosions 1.1.3 Explosible range of dust concentrations-primary and secondary explosions 1.1.4 Ignition sources 1.2 Significanceof the dust explosion hazard: statistical records 1.2.1 Recording dust explosions, an activity of long traditions 1.2.2 Dust explosions in the United States, 1900-1956 1.2.3 Dust explosions in the Federal Republic of Germany, 1965-1985 1.2.4 Recent statistics of grain dust explosions in the United States 1.3 Dust and dust cloud properties that influence ignitability and explosion violence 1.3.1 Dust chemistry, including moisture 1.3.2 Particle size or specific surface area 1.3.3 Degree of dust dispersion effective particle size 1.3.4 Dust concentration 1.3.5 Turbulence 1.3.6 Oxygen content of oxidizer gas 1.3.7 Initial temperature of the dust cloud 1.3.8 Initial pressure of a dust cloud 1.3.9 Combustible gas or vapor mixed with a dust cloud (“Hybrid” mixtures) 1.3.10 Inerting by mixing inert dust with combustible dust 1.3.11 Concluding remarks 1.4. Means for preventing and mitigating dust explosions 1.4.1 The means available: an overview 1.4.2 Preventing ignition sources .4.3 Preventing explosible dust clouds 1.4.4 Preventing explosion transfer between process units via pipes and ducts: explosion isolation I .4.5 Explosion-pressure-resistantequipment 1.4.6 Explosion venting

xvii
xix

n
i

1
5

7
10
20
20
21
22
25
25
25
29
32
34
36
39
44
47

50
55
56
57
57
57
67
73
81
84

viii

Dust Explosions in the Process industries

1.4.7 Automatic suppression of dust explosions 1.4.8 Control and interlocking systems to prevent and mitigate dust explosions in integrated process plants 1.4.9 Prevention and removal of dust accumulationsoutside process equipment: good housekeeping 1.4.10 Dust control by the addition of liquid 1.4.11 Construction and layout of buildings 1.4.12 The “human factors” 1.5 Selecting appropriate means for preventing and mitigating dust explosions 1.5.1 Basic philosophy, cost estimation, and risk analysis 1.5.2 Selection scheme suggestedby Noha for the chemical process industry 1.5.3 Special aspects for some specific groups of powders and dusts: a brief literature survey 1.5.4 Standards, recommendations, and guidelines

101
109
112
113
116
118
121
121
127
133
141

2 Case Histories

157

2.1 Introduction
2.2 The explosion in a flour warehouse in Turin on December 14, 1785 2.3 Grain dust explosions in Norway 2.3.1 Wheat grain dust, Stavanger Port Silo, June 1970 2.3.2 Wheat grain dust, new part of Stavanger Port Silo, October 1988 2.3.3 Grain dust (barley/oats),head house of the silo plant at Kambo, June 1976 2.3.4 Malted barley dust, Oslo Port Silo, July 1976 2.3.5 Malted barley dust, Oslo Port Silo, June 1987 2.4 Four grain dust explosions in United States, 1980-1981 (Source: Kauffman and Hubbard, 1984) 2.4.1 Inland grain terminal at St. Joseph, Missouri, April 1980 2.4.2 River grain terminal at St. Paul, Minnesota, June 10, 1980 2.4.3 Train-loading country grain terminal at Fonda, Iowa, July 15, 1980 2.4.4 Large export grain silo plant at Corpus Christi, Texas, April 1981 2.5 A dust explosion in a fish meal factory in Norway in 1975 2.6 Smoldering gas explosion in a silo plant in Stavanger,Norway, in November 1985 2.7 Smoldering gas explosions in a large storage facility for grain and feedstuffs in Tomylovo, Knibyshev Region, USSR 2.8 Smoldering gas explosion and subsequent successful extinction of smoldering combustion in pelletized wheat bran in a silo cell at Nord Mills, Malmo, Sweden, in 1989 2.9 Linen flax dust explosion in Harbin Linen Textile Plant, Peoples Republic of China, in March 1987

157
157
160
160
162
163
165
165
166
166
168
169 171 173 177 178 181 182

Contents

2.9.1 General outline
2.9.2 Explosion initiation and development, Scenario 1 2.9.3 Explosion initiation and development, Scenario 2 2.9.4 Additional remarks 2.10 Fires and explosions in coal dust plants 2.10.1 Methane explosion in 17,000m3coal silo at Elkford, British Columbia, Canada, in 1982 2.10.2 Methane/coal dust explosion in a coal storage silo at a cement works in San Bernardino County, California 2.10.3 Gas and dust explosion in a pulverized coal productiodcombustion plant in a cement factory in Lagerdorf, Federal Republic of Germany, in October 1980 2.10.4 Further explosion and fire incidents involving coal 2.11 Dust explosion in a silicon powder grinding plant at Bremanger, Norway, in 1972 2.12 Two devastating aluminum dust explosions 2.12.1 Mixing section of premix plant of slurry explosive factory at Gullaug, Norway, in 1973 2.12.2 Atomized aluminum powder production plant at Anglesey, United Kingdom, in 1983

3 Generation of Explosible Dust Clouds by Reentrainment and Redispersion of Deposited Dust in Air 3.1 Background 3.2 Structure of the problem 3.3 Attraction forces between particles in powder or dust deposits 3.3.1 Van der Waals forces 3.3.2 Electrostatic forces 3.3.3 Interparticle forces due to liquids 3.4 Relationship between interparticle attraction forces and strength of bulk powder 3.4.1 Theories 3.4.2 Measurement of the mechanical strength of cohesive bulk powders and dusts 3.5 Dynamics of particles suspended in a gas 3.5.1 Terminal settling velocity of a particle in the gravitational field 3.5.2 Drag on a particle in general 3.5.3 Movement of a particle in an arbitrary flow 3.5.4 Speed of sound in a dust cloud 3.5.5 Propagation of large-amplitudepressure waves in dust clouds 3.6 Dislodgement of dust particles from a dust or powder deposit by interaction with an airflow 3.6.1 Airflow parallel to a monolayer of particles on a plane, smooth surface 3.6.2 Airflow parallel to the surface of a powder or dust deposit 3.6.3 Entrainment of particles by an upward airflow through a particle bed

ix

182
184
186
187
187
187
188

B 89
189
190
192
192

P 95
199
199
200
202
202
203
204
206
206
208
213 213 215 218 219 22 1 221 22 1 224 230

x

Dust Explosions in the Process Industries

3.7 Dispersion of agglomerates of cohesive particles suspended in a gas by flow through a narrow nozzle 3.8 Diffusion of dust particles in a turbulent gas flow 3.9 Methods for generating experimental dust clouds for dust explosion research 3.9.1 Background 3.9.2 Transient dust clouds generated by a short air blast 3.9.3 Stationary dust cloud in a closed circulation system 3.9.4 Stationary dust cloud in an open-circuit system 3.9.5 Conclusion

232
234
239
239
240
244
244
246

4 Propagation of Flames in Dust Clouds

251

4.1 Ignition and combustion of single particles 4.1.1 Aluminum 4.1.2 Magnesium 4.1.3 Zirconium 4.1.4 Carbon and coal 4.1.5 Wood 4.2 Laminar dust flames 4.2.1 Laminar flame propagation in premixed, quiescent gases 4.2.2 Differences between flames in premixed gas and in dust clouds 4.2.3 Experimental burning velocities, flame thicknesses, quenching distances, and temperatures of laminar dust flames 4.2.4 Theories of one-dimensionallaminar flame propagation in dust clouds 4.2.5 Theories of laminar flame propagation in closed vessels 4.2.6 Minimum and maximum explosible dust concentrations 4.3 Nonlaminar dust flame propagation phenomena in vertical ducts 4.4 Turbulent flame propagation 4.4.1 Turbulence and turbulence models 4.4.2 Turbulent dust flames: an introductory overview 4.4.3 Experimental studies of turbulent dust flames in closed vessels 4.4.4 Turbulent flame propagation in partly or fully unconfined geometries 4.4.5 Systematic comparative studies of turbulent gas and dust explosions 4.4.6 Maximum experimental safe gap for dust clouds 4.4.7 Acceleration of turbulent dust explosions in enclosures of large LID (ducts, pipes, galleries, and the like) 4.4.8 Theories of flame propagation in turbulent dust clouds:
computer models 4.5 Detonations in dust clouds in air 4.5.1 Qualitative description of detonation 4.5.2 Experimental evidence of detonations in dust clouds in air 4.5.3 Theories of detonation

251
25 1
253
255
256
264
265
265
268
270
283
294 303 318 325 325 330 332 341 344 346 35 1 363 367 367 367 370

Contents

5 Ignition of Dust Clouds and Dust Deposits: Further Consideration of Some SelectedAspects 5.1 What is ignition?
5.2 Self-heating and self-ignition in powder deposits 5.2.1 Overviews 5.2.2 Some experimental investigations 5.2.3 Further theoretical work 5.2.4 Applications to different powder/dust types: a brief literature survey 5.3 Ignition of dust clouds by electric spark dischargesbetween two metal electrodes 5.3.1 Historical background 5.3.2 The ohmic resistance of a spark channel between two metal electrodes 5.3.3 Influence of spark discharge duration on the minimum electric spark ignition energy for dust clouds 5.3.4 Influence of some further parameters on the minimum ignition energy of dust clouds 5.3.5 Theories of electric spark ignition of dust clouds 5.4 Ignition of dust clouds by heat from mechanical rubbing, grinding, or impact between solid bodies 5.4.1 Background 5.4.2 Sparks and hot-spots from rubbing, grinding, and multiple impacts 5.4.3 Sparks, hot spots, and flashes from single accidental impacts 5.5 Ignition of dust clouds by hot surfaces 5.5.1 Experimental study of the influence of size of the hot surface 5.5.2 Theories for predicting the minimum ignition temperatures of dust clouds

6 Sizing of Dust Explosion Vents in the Process Industries:
Further Consideration of Some Important Aspects 6.1 Some vent sizing methods used in Europe and the United States 6.1.1 Vent ratio method 6.1.2 “Nomograph’method 6.1.3 The Swedish method 6.1.4 The Norwegian method (modified Donat method) 6.1.5 The Radandt scaling law for vented silo explosions 6.1.6 Other vent sizing methods 6.2 Comparison of data from recent realistic full-scale vented dust explosion experiments with predictions by various vent sizing methods 6.2.1 Experiments in large silos of LID I 4 6.2.2 Experiments in slender silos of LID = 6 6.2.3 Pneumatic pipeline injection experiments in vessels of small LID

xi

385
385
388
388
389
397
40 1

404
404
404
407
415
414
419
419
420
421
423
423 424

431
431
431
431
433
434
434
434
435
435
438
443

xii

Dust Explosions in the Process Industries

6.2.4 Experiments in Japan on venting of dust explosions in a 0.32 m3cyclone 6.2.5 Realistic experiments in bag filters 6.2.6 Other large-scale experiments relevant to industrial practice 6.3 Vent sizing procedures for the present and near future 6.3.1 Basic approach and limitations 6.3.2 Large empty enclosures of LID < 4 6.3.3 Large, slender enclosures (Silos) of LID > 4 6.3.4 Smaller, slender enclosures of L/D > 4 6.3.5 Intermediate (10-25 m3)enclosures of small LID 6.3.6 Cyclones 6.3.7 Bag filters 6.3.8 Mills 6.3.9 Elongated enclosures of very large LID 6.3.10 Scaling of vent areas to other enclosure volumes and shapes and to other Predand dusts 6.3.11 Concluding remarks 6.4 Influence of actual turbulence intensity of the burning dust cloud on the maximum pressure in a vented dust explosion 6.5 Theories of dust explosion venting 6.5.1 Introductory outline 6.5.2 Theory by Maisey 6.5.3 Theory by Heinrich and Kowall 6.5.4 Theory by Rust 6.5.5 Theory by Nomura and Tanaka 6.5.6 Theoretical analysis by Nagy and Verakis 6.5.7 Theory by Gruber et al.
6.5.8 Theory by Swift 6.5.9 Theory by Ural 6.5.10 Concluding remarks 6.6 Probabilistic nature of the practical vent sizing problem 6.6.1 Basic philosophy 6.6.2 The “worst credible explosion” 7 Assessment of Ignitability, Explosibility, and Related Properties of Dusts by Laboratory-Scale Tests 7.1 Historical background 7.2 A philosophy of testing the ignitability and explosibility of dusts:
the relationship between test results and the real industrial hazard 7.3 Sampling of dusts for testing 7.4 Measurement of physical characteristics of dusts related to their ignitability and explosibility 7.4.1 Particle size distribution and specific surface area 7.4.2 Dispersibility 7.4.3 Powder mechanical properties 7.4.4 Moisture content 7.4.5 Electrical resistivity

444
447
451
452
452
453
453
453
454
454
454
455 455 455 456 457 458 458 460 460 462 462 463 463 464 464 465 465 465 468

473
473
475
477
479
479
480
485
485
487

Contents xiii

7.5 Can clouds of the dust produce explosions at all? Yes/No screening tests 7.6 Can hazardous quantities of explosible gases evolve from the dust during heating?
7.6.1 The industrial situation 7.6.2 Laboratory tests 7.7 Ignition of dust deposits and layers by self-heating or hot surfaces 7.7.1 The industrial situation 7.7.2 Laboratory tests 7.8 Minimum ignition temperature of dust clouds 7.8.1 The industrial situation 7.8.2 Laboratory tests 7.9 Minimum electric spark ignition energy of dust layers 7.9.1 The industrial situation 7.9.2 Laboratory tests 7. IO Minimum electric spark ignition energy of dust clouds 7.10.1 The industrial situation 7.10.2 Laboratory tests 7.11 Sensitivity of dust layers to mechanical impact and friction 7.11.1 The industrial situation 7.11.2 Laboratory tests 7.12 Sensitivity of dust clouds to ignition by metal sparks, hot spots, or thermite flashes from accidental mechanical impact 7.12.1 The industrial situation 7.12.2 Laboratory tests 7..13 Minimum explosible dust concentration 7.13.1 The industrial situation 7.13.2 Laboratory tests 7.14 Maximum explosion pressure at constant volume 7.14.1 The industrial situation 7.14.2 Laboratory tests 7.15 Maximum rate of rise of explosion pressure at a constant volume (explosion violence) 7.15.1 The industrial situation 7.15.2 Laboratory tests 7.15.3 Further development of adequate test methods for dust explosion violence assessment 7.16 Efficacy of explosion suppression systems 7.17 Maximum explosion pressure and explosion violence of hybrid mixtures of dust and gas in air 7.18 Tests of dust clouds at initial pressures and temperatures other than normal atmospheric conditions 7.19 Influence of oxygen content in the oxidizing gas on the ignitability and explosibility of dust clouds 7.19.1 The industrial situation 7.19.2 Laboratory tests

488
490
490
49 1
493
493
493
499
499
499
505
505 505 508 508 508 513 513 513 515 515 516 518 518 519 525 525 526 533 533 534 535 536 539 540 540 540 541

xiv

Dust Explosions in the Process Industries

7.20 Influence of adding inert dust to the combustible dust on the ignitability and explosibility of dust clouds 7.21 Hazard classification of explosible dusts

542
543

8 Electrical Apparatuses for Areas Containing Combustible Dusts

549

8.1 Introduction
8.1.1 Background and objectives of chapter 8.1.2 Basic similarities and differences between dusts and gases 8.1.3 The “Atex 1OOa” directive gives unclear signals with regard to dusts 8.1.4 Scope of the IEC standards on dusts 8.2 Classification of areas containing combustible dusts 8.2.1 What is area classification?
8.2.2 Definition of zones according to the three-zone concept 8.2.3 International standards 8.2.4 Need to revise area classification standards to include dust fires as a hazard in its own right 8.3 Why different electrical apparatus design criteria are needed for areas with combustible dust and explosive gas atmospheres 8.3.1 Influence of inertial forces on the movement of dust particles 8.3.2 Thermal hazards associated with accumulation of dust layers inside electrical apparatus enclosures 8.3.3 Thermal hazards associated with accumulation of dust layers on external surfaces of electrical apparatus enclosures 8.4 Enclosing potential ignition sources to prevent hazardous ingress of dust 8.4.1 The IP code for prevention of dust ingress 8.4.2 Design of apparatus to prevent ignition of dust clouds and dust layers by hot enclosure surfaces 8.4.3 Other requirements to IP enclosures 8.4.4 International standards for design of enclosures for electrical apparatuses for areas containing combustible dusts:
an overview 8.4.5 The IEC standard for pressurized electrical equipment enclosures for areas containing combustible dust 8.4.6 Encapsulation by molding 8.4.7 Why the concept of flameproof enclosures is not relevant for combustible dusts 8.5 Intrinsically safe electrical apparatuses 8.5.1 The original concept for gases and vapors 8.5.2 The situation with dusts 8.5.3 The new IEC Ex“iD” standard for dusts 8.5.4 Minimum ignition energy, a universal ignition sensitivity parameter for the design of electrical apparatuses that are intrinsically safe in the presence of explosive dust clouds 8.6 Summary and conclusions 8.6.1 “Atex 1OOa” directive 8.6.2 Area classification

549
549
550
55 1
552
553
553
553
554
557
557
557 558 559 560 560 562 564 564 565 567 567 568 568 568 570 57 1 574 574 574

Contents xv

8.6.3 Protection by enclosures
8.6.4 Intrinsic safety

575
576

9 Research and Development, 1990-2002

580

9. I Introduction
9.1.1 Background and objective of chapter 9.1.2 Books and conference proceedings published after 1990 9.2 Status and outstandingproblems in fundamental research on dust explosions 9.2.1 The main topics covered 9.2.2 Generation of primary dust clouds and resulting dust cloud structures 9.2.3 Ignition and combustion of single particles and dust deposits:
ignition of dust clouds 9.2.4 Flame propagation processes in dust clouds 9.2.5 Blast waves generated by burning dust clouds 9.3 Status and outstandingproblems in preventing and mitigating dust explosions in industry 9.3.1 The role of fundamental knowledge in assessing hazards in practice 9.3.2 Inherently safe process design 9.3.3 Papers covering several methods and specific applications 9.3.4 Generation and properties of explosive dust clouds in industry 9.3.5 Preventing ignition sources 9.3.6 Preventing explosive dust clouds 9.3.7 Protective and mitigatory measures 9.3.8 Risk, safety, and hazards analysis 9.3.9 Human and administrativerisk and hazard factors:
risk and safety management 9.3.10 Costs of explosion prevention and mitigation 9.3.11 New European Union legislation to prevent and mitigate accidental explosions 9.4 Status and outstandingproblems in testing dust ignitability and explosibility 9.4.1 Historical background and introduction 9.4.2 Two approaches for achieving differentiation 9.4.3 New test methods 9.4.4 Determining the limits of flame propagation: a problem of the scale of the experiment 9.4.5 Miscellaneous 9.5 Dust explosion statistics and case histories 9.6 Expert systems: friends or enemies?
9.7 Joint research efforts in Europe 9.8 Research and development in the Peoples Republic of China 9.9 Conclusions

580
580
580
582
582
583
587
595
609
610
610
611 611 613 613 619 622 634 634 635 635 636 636 637 638 639 640 641 643 645 646 646

xvi

Dust Explosions in the Process Industries

Appendix: Ignitability and Explosibility Data for Dusts from Laboratory Tests

681

A.l Tables A.l, A.2, and A.3 and comments from the BIA (1987) A. 1.1 Limitations to the applicability of the data A.1.2 Comments on the various items in Table A.1 A.2 Applicability of earlier USBM test data A.2.1 Background A.2.2 Minimum ignition temperature of the dust cloud A.2.3 Minimum ignition temperature of the dust layer A.2.4 Minimum ignition energy of the dust cloud (MIE) A.2.5 Minimum explosible dust concentration A.2.6 Maximum explosion pressure A.2.7 Maximum rate of pressure rise A.2.8 Maximum permissible O2 concentration for inerting

68 1
68 1
682
685
685
685
702
702
702
702
703
703

Index

705

t-oreword
Experience has shown all too clearly that ignition and explosion can occur wherever combustible dusts are handled or permitted to accumulate as a by-product of related activities. Despite reasonable precautions, accidents can and do happen; recognition of this universal hazard and the potential means for its control is widespread, as evidenced by the many individuals and groups worldwideperforming research and developing codes and regulations.
The primary means of controlling and minimizing this recognized hazard are study, regulation, and education;to accomplish this, specific knowledge must be generated and disseminatedfor the benefit of all interested people. Rolf Eckhoff has, in my estimation, prepared an outstanding book. It presents a detailed and comprehensive critique of all the significant phases relating to the hazard and control of a dust explosion and offers anLup-to-date evaluation of prevalent activities, testing methods, design measures, and safe operating techniques.
The author is in an outstanding position to write this text, having spent a lifetime in relsearch on dust and gas explosions. He assimilates information from worldwide contacts while retaining his independence of thought and the ability to see clearly through prloblems.His clear and concise language and thorough approach will benefit his fellow workers and all who read his book. His presentation of the mathematics,tables, and figures is clear and striking. The inclusion of a comprehensivebibliography indicates not only his own thoroughness but also the widespread nature of research into dust explosions throughout the world.
To my knowledge this book is the most complete compilation to date of the state of the art on industrial dust explosions.
John Nag3 Finleyville, PA (Formerly of the U.S. Bureau of Mines)

Prefaces

PREFACE TO THIRD EDITION
As with the second edition, I have not revised the entire book, which, although desirable, would have been an impossible task for me to undertake. However, the entire original body of the book (Chapters 1-7) had, for various reasons, to be reproduced for this new edition. This has given me the opportunity, when reading through the new production, to make minor adjustments in the original text, where this was considered necessary in the light of more recent evidence.
The major revision has been limited to the final review chapter of the second edition, which has been expanded and rewritten to cover the whole span 1990-2002. Nearly 400 new literature references have been added. I am indebted to all the colleagues, over the whole world, who have kindly provided reprints and reports of their valuable work.
An entirely new chapter on electrical equipment for areas containing explosible dusts has been added to the book. I became acquainted with this special field over the last 12 years and gradually realized that a chapter devoted to this topic would be useful. I am greatly indebted to Thore Andersen, secretary of the Norwegian national Ex committee, NIK 3 1, for many valuable discussions, help with retrieving printed information, proofreading the new chapter, and arranging for me to take part in some of the meetings of the IEC working group 31H WG3. Sincere thanks are due also to Dr. Gerold KlotzEngmann, Endress + Hauser, Germany, for valuable advice and help during the preparation of the section on intrinsic safety.
My special thanks goes to my outstanding research student Trygve Skjold, for numerous stimulating discussions and bringing to my attention many important papers. He also kindly proofread both Chapters 8 and 9.
In the original part of the book, the adjective expEosible is used both in connection with dust and dust cloud, hence explosible dust and explosible dust cloud. In the new Chapters 8 and 9, I distinguish between explosible dust and explosive dust clouds. The reasoning is that a combustible dust as such is only potentially explosive, whereas a dust cloud that can propagate a flame is explosive in the same sense as a premixed gas cloud.

PREFACE TO SECOND EDITION
The present book was first published in August 1991 as a hardcover version, which was out of print by spring 1994. The publisher then decided to produce a new paperback version, which was essentially the original book with some minor adjustments. This second version was out of print by mid-1996.
In 1992 I was asked to give a review lecture on the state of the art of research on dust explosion prevention and mitigation, at an international summer school. This provided

xx Dust Explosions in the Process Industries

an opportunity to pick up from where I had to stop reviewing the most recent literature for inclusion in the original book manuscript. The summer school was repeated both in 1993and 1994,which encouragedme to update the review accordingly.It gradually came clear to me that the review would only need to be modified slightly to form a useful new Chapter 8 of my book. The publisher agreed to this idea and decided that such a chapter, covering material published after 1990,be included in the second addition of the book to appear in 1996/1997.I therefore continued to incorporate new material right up to the submission of the final manuscript.
After having worked for more than 30 years at CMR (CMI up to 1992), performing contract research and consultancy work for industry, I started, from 1996, a new challenging career as a full-time professor in process safety technology at the University o f Bergen. It is my hope that my students will find the present book, with the new Chapter 8, a helpful guide into one of the important facets of process safety.

PREFACE TO FIRST EDITION
The ambitious objective of this book is to provide an overview of the present state of the art. However, the amount of published information on dust explosions worldwide is vast, and a substantial amount of additional work was never printed in retrievable literature.While I feel that I may have been able to cover some of the EnglisWAmericanand German open literature fairly well, most of the valuable research published in other languages had to be left out simply because of the language barrier.
Although I have tried to give a reasonably balanced account,the book also reflects my personal research background. To me the important role of powder mechanics in understanding dust explosions is evident. I have, therefore, included a separate chapter on the mechanics of dust particles and dust deposits.The book also reflects that most of my dust explosion research has been related to ignition, venting, and testing.
The confrontation with the early research carried out by the pioneers in the United Kingdom, Germany, United States, and other countries creates deep humility and admiration for the outstanding work performed by these people. Lack of sophisticateddiagnostics did not prevent them from penetrating the logical structuresof the problems and to draw long-lasting conclusions from their observations. It is a pity that much of this work seems to be forgotten in more recent research. Too often humankind “reinventsthe wheel.” This also applies to dust explosion research.
I am indebted to professor emeritus H. E. Rose for bringing the existence of the phenomenon of dust explosions to my attention for the first time and for giving me the opportunity to become acquainted with the subject, during my two years of post-graduate studies at King’s College, London, 1966-1968.
I am also indebted to Mr. Alv Astad and Mr. Helge Aas for their encouragementand active participation when dust explosion research, sponsoredby Norwegian industry,was initiated at Chr. Michelsen Institute, Bergen, Norway, about 1970.
The Royal Norwegian Council for Scientificand Industrial Research (NTNF) has given valuable financialsupportto CMI’sdust (and gas) explosionresearch from 1972until today, and also allocated a generous special grant for the writing of their book. An additional valuable grant for the work with the book was given by the SwedishFire ResearchBoard (Brandforsk).

Prefaces

xxi

I am also most grateful to all the industrial companies, research institutions,and COIleagues in many countries, who made available to me and allowed me to make use of their photographs and other illustrations.A specialthanks to BerufsgenossenschaftlichesInstitut fiir Arbeitssicherheit (BIA), in Germany, for permission to translate and publish the tables in the Appendix.
I also wish to express my gratitude to those who have kindly read through sections of the draft manuscript and/or given constructive criticism and advice: John Nagy, Derek Bradley, Geoffrey Lunn, Bj@rnHjertager, Gisle Enstad, Dag Bjerketvedt, Ivar @. Sand and Claus Donat should be mentioned specifically.
Also my indebtedness goes to Chr, Michelsen Institute, Department of Science and Technology, with its director Dr. Jan A. Andersen, which in its spirit of intellectual freedom coupled with responsibility, offered me the opportunity to establish dust explosion research as an explicit activity of CMI. The institute also gave high priority to and allocated additional resources for the writing of this book, for which I am also most grateful.
This short Preface does not allow me to mention all the good people with whom I had the privilege to work during my 20 years of dust explosion research at CMI and who deservemy sincere thanks. The exception is Kjell Fuhre, who worked with me from 1970 to 1988. I wish to thank him specially for having devoted his exceptional engineering talent to our experimental dust and gas explosion research, in laboratory scale as well as in full-size industrial equipment.
Finally, I wish to express a special thanks to Mrs. Aaslaug Mikalsen, who, aided by more than 20 years’ experiencein interpreting my handwriting, was able to transform the untidy handwritten manuscript to a most presentable format on CMI’s word-processing system. Many thanks also go to Mr. Per-GunnarLunde, CMI, for having traced the majority of the drawings in the book.
Rolf K. Eckhoff

Rust Explosions
in the Process Industries

apter 1
Dust Explosions-Origin, Propagation,
Prevention, and Mitigation: An Overview 1.
THE NATURE OF DUST EXPLOSIONS 1 .1 .1 THE PHENOMENON 1 .I .I .I What Is an Explosion?

The concept of explosion is not unambiguous. Encyclopedias give varying definitions that mainly fall in two categories. The first focuses on the noise or “bang” due to the sudden release of a strong pressure wave, or blast wave. The origin of this pressure wa.ve, whether a chemical or mechanical energy release, is of secondary concern. This definition of an explosion is in accordance with the basic meaning of the word (“sudden outburst”).
The second category of definitions is confined to explosions caused by the sudden release of chemical energy. This includes explosions of gases and dusts and solid explosives. The emphasis is then often put on the chemical energy release itself, and explo~ i o nis defined accordingly. One possible definition could then be “An explosion is an exothermal chemical process that, when occurring at constant volume, gives rise to a sudden and significant pressure rise.” In this text, the definition of an explosion shifts pragmatically between the two alternatives, focusing on either cause or effect, depending on the context.
1 .I .I .2 What Is a Dust Explosion?

The phenomenon named dust explosions is in fact quite simple and easy to envisage in terms of daily life experience. Any solid material that can burn in air will do so with a violence and speed that increases with increasing degree of subdivision of the material.
Figure l.l(a) illustrates how a piece of wood, once ignited, burns slowly, releasing its heat over a long period of time. When cut in small pieces, as illustrated in Figure I.l(b), the combustion rate increases, because the total contact surface area between wood and air has increased. Also, ignition of the wood has become easier. I f the subdivision is continued right down to the level of small particles of sizes on the order of 0.1 mm or less and the particles are suspended in a sufficiently large volume of air to give each particle

2 Dust Explosions in the Process Industries

(a1 SLOW COMBUSTION

(b) FAST COMBUSTION

Id EXPLOSION

Illustration of how the combustion rate of a given mass of combustible solid increases with increasing subdivision.

Figure 1 .I

enough space for its unrestrictedburning, the combustionrate is very fast and the energy required for ignition very small. Such a burning dust cloud is a dust explosion. In general, the dust cloud is easier to ignite and burn more violently the smaller the dust particles are, down to some limiting particle size that depends on the type of dust material.
If such an explosive combustion of a dust cloud takes place inside process equipment or work rooms, the pressure in the fully or partly enclosed explosion space may rise rapidly; the process equipment or building may burst; and life, limb, and property can be lost.
1.I .1 .3 Specific Surface Area-A

Convenient Measure of Dust Fineness

The degree of subdivision of the solid can be expressed in terms of either a characteristic particle size or the total surface area per unit volume or unit mass of the solid. The latter characteristic is called the specijic surface area of the subdivided solid.
Figure 1.2illustrates the relationshipbetween the particle size and the specific surface area. After subdivision of the original cube to the left into eight cubes of half the linear dimensionof the original cube, the total surface area has increased by a factor of 2, which indicates that the specificsurface area is simply proportional to the reciprocal of the linear dimension of the cube. This can be confirmed by simply expressing the specific surface area S as the ratio between surface area and volume of one single cube of edge length x.
One then finds

This is also the specific surface area of a powder or dust consisting of monosized cubes of edge length x.
The same result applies to spheres of diameter x,because S=

m2

6
--

( n / 6 ) x 3- x

Dust Explosions: An Overview

3

Figure 1.2 lllustralion of the increase of the specific area of a solid with increasing subdivision (From Hammond and Kaye, 1963).

For flake-shaped particles, for which the thickness x is much smaller than the characteristic flake diameter, one has S = -2 X

If a spherical particle of diameter 5 pm, for example, is compressed and deformed plastically to a thin flake of thickness, for example, 0.2 pm (flake diameter about 20 pm), equations (1.2) and (1.3) show that the specific surface area increases by a factor of 8.3.
This effect is utilized when producing highly reactive aluminum flakes from atomized (spherical) aluminum particles (see Section 1.3.2).
If the “particles”are fibers of large length-to-diameterratio and the diameter is x,one gets

S = -4
X

Fibrous dusts may either be natural (for example, cellulose) or synthetic (such as flock materials). Ignitability and explosibility of synthetic flock materials were discussed by Schenk (1984).
In the case o f polysized cubes or spheres, the specific surface area equals S = 6 CxfniI C x:ni

(1.5)

where niis the number of particles in the size category xiin the sample considered.
As the particles get smaller, the interparticleforcesplay an increasinglyimportantrole compared with gravity forces; and in a given practical situation,the dust in a dust cloud may not necessarily be dispersed into the small individual primary particles but rather into larger agglomerates,or lumps. The effective particle size therefore is larger and the effective specific surface area smaller than if the primary particles had been completely dispersed. This important aspect is discussed in Section 1.3.3 and in depth in Chapter 3, see also Section 7.4.2.

4

Dust Explosions in the Process Industries

1.I .I .4

Factors Influencing Ignition Sensitivity and ExplosionViolence of Dust Clouds Particle sizehpecific surface area of the dust is a central factor. However, there are other important factors too, and the comprehensive list may look as follows:
1. Chemical composition of the dust, including its moisture content.
2. Chemical composition and initial pressure and temperature of the gas phase.
3. Distributions of particle sizes and shapes in the dust, determiningthe specificsurface area of the dust in the fully dispersed state.
4. Degree of dispersion, or agglomeration, of dust particles, determining the effective specific surface area availableto the combustionprocess in the dust cloud in an actual industrial situation.
5. Distribution of dust concentration in an actual cloud.
6. Distribution of initial turbulence in an actual cloud.
7. Possibility of generation of explosion-induced turbulence in the still unburned part of the cloud. (Location of ignition source important parameter.) 8. Possibility of flame front distortion by mechanisms other than turbulence.
9. Possibility of significant radiative heat transfer (highly dependent on flame temperature, which in turn depends on particle chemistry).
Factors 3 and 4 have already been mentioned. These and other factors are discussed in more detail in the subsequent sections. Factors 1,2,3, and 9 can be regarded as basic parameters of the explosible dust cloud. Factors 4 to 8 are, however, influenced by the actual industrial dust cloud generationprocess and explosion development.These, in turn, depend on the nature of the industrial process (flow rates, etc.) and geometry of the systemin which the dust cloud bums. The location of the ignition point is another parameter that can play an important role in deciding the course of the explosion.
In view of the wide spectrum of dust cloud concentrations,degrees of dust dispersion and turbulence, and locations of potential ignition sources in industry, a correspondingly wide spectrumof possible dust cloud ignition sensitivitiesand combustionrates must be expected.
This complex reality of the process industry is also shared by laboratory experimentation and represents a constant challenge in the design of adequateexperiments and interpretation of experimental results.
1 .I .I .5

Previous Books on the Dust Explosion Problem
Over the years, several textbooks on the dust explosionhazard have been produced. One of the first ones, by Beyersdorfer (1925),was published in Germany; he mentioned that his motivation for writing the book arose from three questions. The first, asked by most people, was, “Are dust explosions really existing?” The second question, asked by the plant engineer, was, “Why are we having so many dust explosions?”The final question was asked by the researcher, “Why are we not having many more dust explosions?” Although out of date on some points, Beyersdorfer’s pioneering book is still fascinating reading.
Almosthalf a century elapsed from the publication of Beyersdorfer’stext until the next comprehensive book on dust explosions appeared. It should be mentioned though, that

Dust Explosions: An Overview

5

in the meantime some valuable summaries were published as parts of other books or as reports. Examples are the reports by Verein deutscherIngenieure (1957) and Brown and Jaimes (1962) and the sections on dust explosions in the handbook on room explosions in general, edited by Freytag (1965). In his book on hazards due to static electricity,Haase (1972)paid attentionto the dust explosionproblem as well. However, Palmer (1973a)produced the long-desiredupdated,comprehensiveaccount of work in the Western world up to about 1970.In Eastern Europe, a book on the prevention of accidental dust explosions, edited by Nedin (1971), was issued in the USSR two years earlier. Cybulski’s comprehensive account of coal dust explosionsappearedin Polish in 1973,that is, simultaneously with the publication of Palmer’s book, and the English translation came two years later (Cybulski, 1975).In the Federal Republic of Germany, Bartknecht had conducted extensive research and testing related to dust explosions in coal mines as well as in the chemical process industries.This work was summarized in a book (Bartknecht, 197&),which was subsequentlytranslated to English. The book by Bodurtha (1980) on industrialexplosion prevention and protection also contains a chapter on dust explosions.
Two years later, two further books were published, one by Field (1982a) and one by Cross and Farrer (1982), each quite comprehensive but emphasizing different aspects of the dust explosion problem. In the next year, a book by Nagy and Verakis (1983) was published, in which they summarized and analyzed some of the extensive experimental and theoretical work conducted by the US. Bureau of Mines up to 1980 on the initiation, propagation, and venting of dust explosions. Three years later, a book by Korol’chenko (1986) was issued in the USSR, reviewing work on dust explosionspublished in both the West and Eastern Europe. The next year, Bartknecht’s (1987) second book was published, describinghis extensive, more recent research and testing, at CibaGeigy AG, related to dust explosion problems. The Institution of Chemical Engineers in the United Kingdom published a useful series of booklets reviewing the status of various aspects of the dust explosion problem (Lunn, 1984, 1988; Schofield, 1984; and Schofield and Abbott, 1988). The comprehensive book by Glor (1988) on electrostatic hazards in powder handling should also be specificallymentioned at this point. Valuable information on the same subject is also included in the book by Liittgens and Glor (1989).
The proceedings of the international symposium on dust explosions, in Shenyang, Peoples Republic of China, published by North East University of Technology (19871, contains survey p a p r s and special contributions from researchers from both Asia, America, and Europe. EuropEx (1990) produced a collection of references to publications related to accidental explosions in general, including dust explosions.The collection is updated at intervals and contains references to standards,guidelines, and directives as well as to books and papers. Finally, attention is drawn to the proceedings of three conferences on dust explosions, in Nurnberg, published by the Verein deutscher Ingenieure (VDI) in 1978, 1984, and 1989 (listed under the publisher’s name in the References).
Section9.1.2in Chapter 9 reviews books and conferenceproceedings published after 1990.

1.1.2
MA,TERIALSTHAT CAN CAUSE DUST EXPLOSIONS Dust explosions generally arise from rapid release of heat due to the chemical reaction Fuel + oxygen + oxides + heat (1-6)

6

Dust Explosions in the Process Industries

In some special cases, metal dusts can also react exothermallywith nitrogen or carbon dioxide, but most often oxidationby oxygen is the heat-generatingprocess in a dust explosion. This means that only materials not already stableoxides can give rise to dust explosions. This excludes substances such as silicates, sulphates, nitrates, carbonates, and phosphates and therefore dust clouds of Portland cement, sand, limestone, and the like cannot produce dust explosions.
The materials that can cause dust explosions include Natural organic materials (grain, linen, sugar, etc.).
Synthetic organic materials (plastics, organic pigments, pesticides, etc.).
Coal and peat.
Metals (aluminum, magnesium, zinc, iron, etc.).
The heat of combustion of the material is an important parameter,because it determines the amount of heat that can be liberatedin the explosion. However, when comparing the various materials in terms of their potential hazard, it is useful to relate the heat of combustion to the amount of oxygen consumed.This is because the gas in a given volume of dust cloud contains a limited amount of oxygen, which determines how much heat can be released in an explosionper unit volume of dust cloud. Table 1.1 lists the heat of combustion of various substances,per mole of oxygen consumed. Ca and Mg top the list, with A1 closely behind. Si is also fairly high up on the list, with a heat of combustionper mole of oxygen about twice the value of typical natural and synthetic organic substances and coals. Table 1.1 is in accordance with the experiencethat the temperatures of flames of dust of metals like A1 and Si are very high compared with those of flames of organic dust and coal.
Table 1 .I

Ca

Heats of combustion (oxidation)of various substances per mole O2 consumed

CaO

I

1270

The equation of state for ideal gases describes the mutual interdependenceof the various parameters influencing the explosion pressure:
p = -TnR V

(1.7)

Dusr Explosions: An Overview

7

Here P is the pressure of the gas, Tits temperature, Vthe volume in question,n the number of gas molecules in this volume, and R the universal gas constant. For constant volume, P is proportionalto T and n. Normally, the increase of T due to the heat developed in the burning dust cloud has the deciding influence on P , whereas the change in n plays only a minor role.
Combustion of metal dust can cause the maximum possible relative reduction of n, by consuming all the oxygen in the formation of condensed metal oxides. If the gas is ab-and all the oxygen is consumed and all the nitrogen is left, n is reduced by about 20%.
In the case of organic dust and coal, assuming that CO, (gas) and H,O (gas) are the reaction products, the number of gas molecules per unit mass of dust cloud increases somewhat during combustion.This is because two H20molecules are generated per 0, molecule consumed.Furthermore, in organic substancescontaining oxygen, some H,O and CO, are generated by decompositionof the solid material itself, without a supply of oxygen from the air.
Consider as an example a starch of composition (C6H1005)* suspended in air at the dust concentration that just consumes all the oxygen in the air to be completely transformed to GO2 and H,O (= stoichiometric concentration); 1 m3of air at normal ambient conditions contains about 8.7 moles of 0, and 32.9 moles of N2. When the starch is oxidized, all the 0, is spent on transforming the carbon to CO,, whereas the hydrogen and the oxygen in the starch are in just the right proportions to form H20 by themselves. The 8.7 moles of 0, is then capable of oxidizing 8.7/6 = 1.45 moles of (C6HI0Oj),that is, ablout 235 g, which is the stoichiometric dust mass per m3 of air at normal conditions.
The reaction products then are 8.7 moles of CO, and 7.3 moles of H,O. The total number of 41.6 moles of gas in the original 1 m3 of dust cloud has therefore been transformed to 48.9 moles, that is, an increase by 17.5%. In an explosion, this contributes to increasing the adiabatic constant-volumeexplosion pressure correspondingly.
13must be emphasized,however, that this formal considerationis not fully valid if the combustion of the organic particles also results in the formation of CO and char particles. This is discussed in greater detail in Chapter 4.

1 .I .3
EXPLQSIBLE RANGE OF DUST CONCENTRATIONSRJMARY AND SECONDARY EXPLOSIONS Thle explosive combustion of dust clouds, illustrated in Figure l.l(c), cannot take place unless the dust concentration (i.e., the mass of dust per unit volume of dust clouds) is within certain limits. This is analogous to combustion of homogeneous mixtures of gas,eousfuels and air, for which the upper and lower flammability limits are well established. Figure 1.3 shows the explosiblerange for a typical natural organic material, such as corn starch, in air at normal temperature and atmosphericpressure.
The explosible range is quite narrow, extending over less than two orders of magnitude, from 50-100 g/m3 on the lean side to 2-3 kg/m3 on the rich one. As discussed in greater detail in Chapter 4, the explosibility limits differ somewhat for the various dust materials. For example, zinc powder has a minimum explosible concentration in air of about 500 g/m3.Explosible dust clouds have a high optical density, even at the lower

Dust Explosions in the Process Industries

8

[ €NE I
I

10,:

::j ; / :,: ,:.,;I

I...'.<........... I
1'' :,:......
,:,;...;,
..".....'..[ : ..:..- I .;:1.
...:.,*: ..;

I
I

Figure 1.3 The range of explosible dust concentrationsin air at normal temperature and atmospheric pressure for a typical natural organic dust (corn starch), compared with typical range of maximum permissible dust concentrations in the context of industrial hygiene, and a typical density of deposits of natural organic dusts. (Note that self-sustainedflame propagation may also occur in such deposits.)

explosible limit. This is illustrated by the fact that the range of maximum permissible dust concentrations specified in the context of industrial hygiene in working atmospheres are three to four orders of magnitude lower than minimum explosible dust concentrations.This means that the unpleasant dust concentrationlevels that can sometimes occur in the general working atmosphereof a factory, and calls on the attention of industrial hygiene authorities, are far below the concentration levels that can propagate dust flames.Therefore, the minimum explosible concentration corresponds to dust clouds of high optical densities, which are unlikely to occur regularly in work rooms of factories.
A visual impression of the density of explosible dust clouds is provided in Figure 1.4, which illustrates a cubical arrangement of cubical particles.

L
Figure 1.4 Cubical dust particles of edge x
arranged in a cubical pattern, with interparticle center-to-center distances of L.

On average, there is one cubicalparticle of volume 2 per cube of dust cloud of volume L3. If the particle density is pp, the dust concentration equals c = p,(x/L)3

(1.8)

Dust Explosions: An Overview 9

or, in a rearranged form,
Lllx = (pp/c)1’3

For particles of density 1 g/cm3,that is, lo6g/m3,a dust concentrationof 50 g/m3corresponds to L/x = 27. For 500 g/m3,which is a typical worst-case explosible concentration, L/x = 13.The actual density shown in Figure 1.4,of L/x = 4, corresponds to a very high dust concentration, 16 kg/m3,which is well above the maximum explosible concentration for organic dust (2-3 kg/m3).
It is important to note that the absolute interparticle distance correspondingto a given dust concentrationdecreases proportionallywith the particle size. For example, at a dust colncentrationof 500 g/m3and a particle density of 1g/cm3,L equals 1.3 mm for 100pm particles, whereas it is only 13 pm for 1 pm particles.
Zehr (1965) quoted a rule of thumb by Intelmann,saying that, if a glowing 25-W lightbulb is observed through 2 m of dust cloud, the bulb cannot be seen at dust concentratictns exceeding 40 g/m3.This is illustratedin Figure 1.5.It followsfrom this that the dust clouds in which dust explosions are primarily initiated are normally found inside process equipment, such as mills, mixers, screens,dryers, cyclones, filters,bucket elevators,hoppers, silos, aspiration ducts, and pipes for pneumatic transport of powders. Such explosions,initiatedby some ignition source(see Section 1.1.4) are calledprimary explosions.

Figure 1.5 A cloud of 40 g/m3 of coal dust in air is so dense that a glowing 25-W lightbulb can hardly be seen through a dust of cloud 2 m thick.

This reveals an important difference between primay dust and gas explosions. In the case of gases, the process equipment normally contains fuel only, with no air, and under such circumstances,gas explosionsinside process equipment are impossible.Therefore, most primary gas explosions occur outside process equipment, where gas from accidental leaks is mixed with air and explosible atmospheres generated.
An important objective of dust explosion control (see Section 1.4) is to limit primary explosions in process equipment to the process units in which they initiate.A central concern is then to avoid secondary explosions due to entrainment of dust layers by the blast wave from the primary explosion. Figure 1.3 shows that there is a gap of two orders of magnitudebetween the maximum explosible dust concentration and the bulk density of dust layers and heaps. The consequence of this is illustrated in Figure 1.6. This figure

10 Dust Explosions in the Process Industries

1 mm LAYER OF DUST
OF BULK DENSITY

Figure 1.6 The potential hazard of even thin dust layers. A I-mm layer of a dust of bulk density 500 kg/m3 (a) generates a cloud of average concentration 100 g/m3 if dispersed in a room of 5 m height (b).Partial dispersion up to only 7 m gives 500 g/m3 (c).

shows that the simple relationship between the bulk density of the dust layer, Pbulk; the layer thickness, h; the height, H , of the dust cloud produced from the layer; and the dust concentration, c, is h (1.10) = Pbuk If a dust layer of thickness h on the internal wall of a cylindrical duct of diameter D is dispersed homogeneously over the whole tube cross section, one has = Pbulk

4h
5

(1.11)

In the case of a tube diameter of 0.2 m, typical of many dust extraction ducts in industry, a layer thickness of only 0.1 mm is sufficient for generating a dust concentration of 1000 g/m3with a dust of bulk density 500 kg/m3.
In general, dispersibledust layers in process plants represent a potential hazard of extensive secondarydust explosions,which must be reduced to the extent possible. Figure 1.7 illustrates how secondary explosions in workrooms can be generated if preventive precautions are inadequate.

1 .I .4

IGNITION SOURCES
1.I .4.1

Background

A combustible dust cloud will not start to burn unless it becomes ignited by a source of heat of sufficient strength. The most common ignition sources are

Smoldering or burning dust.
Open flames (welding, cutting, matches, etc.).
Hot surfaces (hot bearings, dryers, heaters, etc.).
Heat from mechanical impact.
Electrical discharges and arcs.

Dust Explosions: An Overview

17

PRIMARY EXPLOSION
DUST LAYER IS
ENTRAINED AND
DUST CLOUD FORMED

EXTENSIVE SECO
EXPLOSION CAN

ibl

Figure 1.7 How the blast wave from a primary explosion (a) entrains and disperses a dust layer, which (b) is subsequently ignited by the primary dust flame.

Some of these sources are discussed more extensively in Chapter 5, and only a brief outline is given here.
There is considerable variation in the ignition sensitivity of various types of dusts. This is discussed in Section 1.3. To quantify the ignition sensitivity of dust clouds and dust deposits when exposed to various kinds of ignition sources, a range of laboratory-scale test methods have been developed, which are described in Chapter 7.
1 .I .4.2

Smoldering or Burning Dust
Experience has shown that combustible dust, when deposited in heaps or layers, may under certain circumstances develop internal combustion and high temperatures. This is due to the porous structure of dust deposits, which gives oxygen access to the particle surface throughout the deposit and makes the heat conductivity of the deposit low.
Coinsequently, heat developed due to comparatively slow initial oxidation at moderate temperatures inside the dust deposit may not be conducted into the surroundings sufficieiitly fast to prevent rising temperature in the reaction zone. As long as oxygen is available, the increased temperature increases the rate of oxidation, and the temperature inside the dust deposit increases even further. Depending on the permeability of the dust deposit and geometrical boundary conditions, the density difference between the hot combustion gases and the air of ambient temperature may create a draft that supplies fresh oxygen to the reaction zone and enhances the combustion process.
If a dust deposit containing such a hot reaction zone, often called a smoldering nest, is disturbed and dispersed by an air blast or a mechanical action, the burning dust can easily initiate a dust explosion if brought in contact with a combustible dust cloud.
Sometimes,the dust in the deposit that has not yet burned, forms the explosible dust cloud.

12

Dust fxplosions in the Process Industries

The initial oxidation inside the deposit may sometimes be due to the dust or powder being deposited having a higher temperaturethan planned. However, some natural vegetablematerials may developinitial spontaneouscombustioneven at normal ambienttemperatures due to biochemical activity, if the content of fat or moisture is high.
In other cases, the dust deposit or layer rests on a heated surface, which supplies the heat needed to trigger self-ignition in the dust. Such surfaces can be overheated bearings, heaters in workrooms, lightbulbs, walls in dryers, and the like. If the surface is not intended to be covered with dust, the dust deposit may prevent normal coolingby forming an insulating layer. This may give rise to an undesirable temperature rise in the surface, which further increases the probability of ignition of the dust. In general, the minimum temperature of the hot surface for the dust layer to self-ignite decreases with increasing thickness of the dust layer.
Figures 1.8, 1.9, and 1.10 illustrate various ways in which smoldering combustion in dust deposits can initiate dust explosions. The critical conditions for the generation of smoldering nests are discussed in Chapter 5, and test methods assessing the self-heating tendency of various dusts are described in Chapter 7.

DUST DEPOSIT

SMOLDERING
NEST
GROWING
SMOLDERING
NEST

Figure 1.8 A smoldering nest in a dust or
powder deposit in a silo can initiate a dust explosion if the nest is discharged into an explosible dust cloud.

Figure 1.9 Complex ignition sequence via gas
explosion: Due to limited supply of oxygen, the smoldering nest develops CO and other combustible gases and creates an explosible mixture above the dust deposit. When the edge of the smoldering nest penetrates the top surface of the dust deposit, the gas ignites and the gas explosion blows up the silo roof, Dust deposits in the room above the silo are dispersed and a major secondary dust explosion results.

Dust Explosions: An Overview

13

Figure 1.1 0 A hidden dust deposit inside a
duct can be brought to ignition by heat supplied to the duct wall from the outside.

It should be mentioned that van Laar (1981) found that burning cigarettes and cigars may give rise to smoldering fires in tapioca and soybean meal. Further information is given in Sections 1.4.2.5,and 9.2.3.3 and 9.3.5.2 in Chapter 9.
1 .'I .4.3 Open Flames

The flames of welding and cutting burners are sufficiently powerful to initiate explosions in any d