Determination of the energetic characteristics of commercial explosives using the cylinder expansion test technique

(1)

Determination of the energetic

characteristics of commercial explosives using the cylinder expansion test technique

Bestämning av civila sprängämnens arbets- förmåga med cylinderexpansionsprovet

Sedat Esen Ulf Nyberg Hiroyuki Arai Finn Ouchterlony

Swebrec - Swedish Blasting Research Centre Luleå University of Technology

Department of Civil and Environmental Engineering • Division of Rock Engineering

(2)

(3)

Sammanfattning

Cylinderexpansionsprovet har använts som en viktig metod i SveBeFos och Swebrecs forskning för att bestämma civila sprängämnens arbetsförmåga. Swebrecs huvudmannaprojekt Cylinderexpansionsprov och verkansberäkningar har som mål att utnyttja sprängämnesenergin optimalt genom att styra energitransmissionen från sprängämne till berg och att kvantifiera energiförlusterna i närzonen utanför laddningen.

Den del som rapporteras här har mer specifikt inriktats på att utveckla metoden, dvs. att förbättra försökstekniken och standardisera utvärderingen av äldre resultat utvärdera spränggasernas expansionsegenskaper med s.k. JWL-teknik undersöka effekten av inblandning av aluminium i emulsionssprängämne

jämföra resultaten för arbetsförmåga med data från undervattensprov och fältförsök och bedöma metodens användbarhet för att bestämma civila sprängämnens förmåga att producera

arbete och andra effekter inom bergsprängning.

I rapportern summeras resultaten från 58 äldre cylinderprov som utförts mellan januari 2002 och januari 2005. Proven omfattar 11 civila sprängämnen i kopparrör med innerdiametrar mellan 40 och 100 mm. Sprängämnena täcker det spektrum från ren ANFO och emulsionssprängämne till blandningar som utgör den absoluta majoriteten av den civila sprängämnesmarknaden. Till detta kommer 9 cylinderprov som sköts augusti-november 2005.

Huvudresultaten från arbetet är

Då flera forskare haft ansvar för arbetet genom åren gicks gamla data igenom i detalj med avseende på kvalitén i rådata, datalagring, utvärderingsrutiner för Gurneyenergi och JWL- expansionsdata. En enhetlig procedur för arbetet har tagits fram. Tveksamma data har eliminerats och övriga har analyserats med statistiska metoder.

Passningen av JWL-tillståndsekvationen till expansionsdata har medfört att såväl sprängämnets detonationsegenskaper (detonationstryck, specifik volym och inre energi) som spränggasernas hela expansionskurva (tryck och energi) har kunnat bestämmas. Detta innebär en väsentlig generalisering av den tidigare Gurneytekniken som enbart ger ett begränsat värde på arbetsförmågan.

Den ingående analysen av äldre data för 100 mm laddningsdiameter visar att

Tillsats av ANFO i ett rent emulsionssprängämne (E682) sänker till att börja med energiöverföringen till omgivningen. Senare, vid ”cut-off”-trycket 20 MPa, är

Determination of the energetic characteristics of commercial explosives….Swebrec Report 2005:1

(4)

energiöverföringen lika stor som för den rena emulsionen. Då densiteten är högre så väntas emellertid arbetsförmågan för blandningen var större på volymsbas.

Tillsats av 3,5 % finkornigt aluminiumpulver till den rena emulsionen har initialt ingen effekt på energiöverföringen. Senare, vid 20 MPa, har tillsatsen en positiv effekt. Tillsats av samma mängd grovkronigt aluminiumpulver verkar inte höja emulsionens arbetsförmåga, vilket alltså det finkorniga pulvret verkar göra.

Resultat från de nya försöken

Två satser med sprängämne, E682 och E682 med 6 % finkornigt aluminiumpulver, tillverkades vid Dyno Nobels nya pilotanläggning i juni 2005. De testades i augusti i syfte att undersöka effekten av aluminiumtillsatsen. Den planerade densiteten om 1170 kg/m³nåddes med god noggrannhet.

De utförda cylinderproven visar att Gurneyenergin för E682, ren och med 6 % aluminium, blev 1,92±0,08 MJ/kg och 2,16±0,10 MJ/kg. Skillnaden är signifikant på 99 %-nivån.

Tillsatsen av aluminiumpulvret höjer arbetsförmågan med 12,5 %, vilket är mindre än motsvarande effekt i undervattensprovet som är 19 % (Hagfors, 2005). Det entydiga resultatet gör att tidigare resultat bedöms som relevant, en tillsats av 3,5 % fint pulver ger en höjning av arbetsförmågan med 5,4 %.

Onoggrannheten i resultaten bedöms ha minskat från ca 7 % till 4,5-5 %.

Resultaten från JWL-analysen stämmer i allt väsentligt med resultaten för Gurneyenergin.

JWL-data visar att aluminiumtillsatsen ger en lägre energiöverföring i början av expansionsförloppet (inom ett par ggr ursprungsvolymen) och en högre mot slutet (ner mot cut-off trycket 20 MPa). Minskningen i detonationstryck bedöms vara 3,9 % vid tillsats av 6

% fint aluminium medan arbetsförmågan ökar med 12,2 %.

Några skott med friliggande laddning med E682 med aluminium gjordes också. Mer data krävs emellertid för att kunna göra en detaljerad detonationsanalys. Detonationsegenskaperna som erhållits med DeNE-programmet (Esen, 2004) stämmer någorlunda med dem som räknats fram ur cylinderprovsdata. Däremot gav DeNE sämre resultat för själva expansionsförloppet, särskilt för E682 med aluminium. Orsaken kan vara begränsningar i programmet eller de antaganden som gjorts i det. Cylinderprovets förmåga att beskriva spränggasernas hela expansionsförlopp är lovande och detta bör följas upp för komplicerade sprängämnestyper såsom emulsioner med inblandning av aluminium.

(5)

Effekten av tillsats av ANFO i emulsion

Äldre cylinderprovsdata räckte inte för att påstå att en tillsats av ANFO höjer emulsions- sprängämnens arbetsförmåga på massbas. Datorprogram för ideal detonation och undervattensprov pekar emellertid i den riktningen. För att bekräfta detta bör de tidigare försöken med ANFO-tillsats upprepas när metoden nu är noggrannare än förut.

Effekten av tillsats av aluminium i emulsion

Datorprogram för ideal detonation, undervattensprov och cylinderprov pekar alla i riktning mot att en aluminiumtillsats har en positiv effekt på en emulsions arbetsförmåga. En större ökning av arbetsförmågan erhölls med 6 % aluminium (typ A60 – medelstorlek 62 µm) än med 3,5 % aluminium i cylinderprovet. En 3,5 % tillsats av typ A20 (medelstorlek 250 µm) hade emellertid ingen nämnvärd effekt.

Undervattensprovet jämfört med cylinderprovet

Tjugonio data från undervattensprov hittades i litteraturen. De pekar mot att en tillsats av ANFO och/eller aluminium påverkar ett emulsionssprängämnes arbetsförmåga positivt. En jämförelse av proven visar att cylinderprovet ger ett bättre mått på ett sprängämnes arbetsförmåga, dvs.

energiöverföring till berget runt ett borrhål.

Fältprov, effekten av aluminiumtillsats till emulsions- eller ANFO-sprängämne

De fyra fall som hittats i litteraturen pekar mot att en aluminiumtillsats ger såväl bättre fragmentering som bättre kast. Undersökningarna är emellertid dåligt dokumenterade och proven upprepades inte så en definitiv slutsats om effekten är inte möjlig att dra. Nya fältprov är troligen nödvändiga för att visa om aluminium bidrar till fragmentering eller kast eller till båda och om detta beror på bergförhållandena.

Genomgång av fullskaleprov i fält, en jämförelse av sprängämnen

Fyra fall där olika sprängämnen undersöktes i olika bergförhållanden har analyserats. I Chiappettas (1998) undersökning gav en inblandning av ANFO i emulsion (blend) 20 % högre kasthastighet än den rena emulsionen. Stagg m.fl. (1989) hävdar att effekten av sprängämnet på styckefallet är försumbar i täkterna i Waterloo och Granite Falls där bergmassan var relativt uppsprucken. Stagg och

(6)

Otterness (1995) rapporterade senare att en tillsats av 30 % ANFO gav finare styckefall än ANFO i Maniwotoc-brottet där bergmassan var relativt sprickfri.

Allmänt gäller att väldokumenterade fältprov med försöksupprepning och data från bergkartering, funktionskontroll- och styckefallsmätningar är sällsynta. Därför behövs en testserie där sådana mätningar görs för att säkerställa sprängämnets inverkan på styckefall och kast under olika bergförhållanden.

Cylinderprovets användbarhet i samband med bergsprängning

Den nuvarande cylinderprovsmetoden har följande användningsområden:

Metoden kan användas för att bestämma såväl detonationsegenskaperna (detonationstryck, specifik volym och inre energi) som spränggasernas hela expansionskurva (tryck och energi) under en relativt realistisk inneslutning ner till ett bestämt cut-off tryck.

Metoden kan använda för att jämföra eller gradera olika sprängämnen, t.ex. nya recept eller blandningar. Effekten av tillsats av aluminium eller andra fasta ämnen som ANFO till emulsioner eller andra receptförändringar kan utvärderas.

Det är den enda metod där spränggasernas expansionsarbete, dvs. energiöverföringen till omgivande material, kan plottas under realistiska, dvs. icke-ideala detonationsförhållanden.

Motsvarande modeller för icke-ideala detonationer behöver dock utvecklas mera.

Cylinderprovsmetoden har följande begränsningar:

Den mäter inte direkta sprängningsresultat (styckefall, kast sprängskada) men det gör ingen annan indirekt provmetod heller.

Resultaten gäller strängt taget för 100 mm laddningar inneslutna i 5 mm tjocka kopparrör.

Detonationsreaktionen kan bli mer fullständig i större laddningar av höggradigt icke-ideala sprängämnen som ANFO och ANFO med aluminiumtillsats.

Förslag till fortsatt arbete

Efter en diskussion om cylindermetodens fördelar och nackdelar beslöt Swebrecs projektgrupp att förslå en fortsättning på projektet där det kortsiktiga arbetet går ut på att prova

Emulsionssprängämne med tillsats av 4 och 6 % grovt aluminiumpulver Emulsionssprängämne med tillsats av 35 % AN

Emulsionssprängämne med tillsats av 35 % ANFO

En blandning 65/35 % av emulsion med 6 % fint aluminium och emulsion med 35 % AN.

(7)

På längre sikt vill gruppen också validera cylinderprovsresultaten i fältprov i full skala med särskilt tonvikt på hur tillsatsen av aluminium och AN(FO) påverkar styckefall och kast. Det långsiktiga målet skall vara att utveckla riktlinjer för rätt val av sprängämne beroende på bergförhållandena.

Huvudmannaprojektets långsiktiga mål är att utnyttja sprängämnesenergin optimalt genom att styra energitransmissionen från sprängämne till berg och att kvantifiera energiförlusterna i närzonen utanför laddningen. De arbeten som gjorts hittills stämmer väl med dessa mål. Arbeten som kvantifierar energiförlusterna i närzonen runt borrhålet saknas emellertid ännu.

(8)

(9)

Executive Summary

The cylinder expansion test technique has been a principal testing technique for the determination of energetics of commercial explosives at Swebrec. The aims of this project are to evaluate the work capacity of commercial explosives, compare the products, judge the addition of Aluminum into ANFO and emulsion explosives, assess the addition of ANFO into emulsion explosives, compare the test results with the underwater test and full-scale results as well as determine the usefulness of the cylinder test in the mining and quarrying industry.

This report presents the comprehensive cylinder expansion tests collected to date. 58 cylinder tests were carried out during the period from January 2002 to January 2005. A total of 11 commercial type explosives have been tested during this period at charge diameters between 40 and 100 mm in copper pipes. The explosives ranged from pure ANFO to pure emulsions and covered the majority of the commercial explosive types used in today’s production blasting operations. In addition to the previous in-house data, 9 more cylinder tests were conducted between August 2005 and November 2005.

The major conclusions drawn from this project are as follows:

Because of the issues (mainly experimental) with the old data, a detailed analysis was undertaken in this study by analyzing the data for each explosive type; charge diameter; raw cylinder expansion, kinetic and JWL expansion data; making consistent analysis; doing JWL fitting for each experimental data; eliminating faulty data (due to e.g. moisture, insufficient pin data, etc); and including elaborate statistics for data analysis. All data collected/analysed to date are placed in a consistent electronic format.

JWL fitting for each experimental data resulted in the determination of the full expansion (pressure and energy) curve as well as the detonics properties (detonation pressure, specific volume, internal energy) of the explosive. This was a key extension of the previous Gurney analysis technique, which gives only a limited value.

The detailed analysis of previous in-house data (100 mm diameter cylinder test) suggested the following:

o It appears that ANFO addition into emulsion decreases the early energy release into the surroundings but the late energy values (at cut-off pressure of 20 MPa) are almost the same as for the pure emulsion. As the densities of the blends are higher than pure emulsions, their bulk energies (in MJ/l) are expected to be higher though.

o It is shown that the early energy releases for both fine aluminized (3.5%) and pure emulsions are similar; however, fine aluminized emulsion delivers more energy at later stages (at the cut-off pressure level). The addition of coarse aluminum (3.5%)

(10)

does not appear to increase the work capacity whereas the addition of the fine Aluminum might be interpreted to do so.

Results from new trials in 2005:

Two explosives (E682 and fine aluminized E682 emulsion with 6% Al) were manufactured in June 2005 and tested in August 2005 in order to investigate whether the Al addition is beneficial or not. These products were manufactured with good mixing conditions at the Dyno Nobel’s laboratories and the targeted densities of 1.17 g/cm³ were achieved for both products.

Cylinder test results suggest that maximum Gurney energies for E682 and 6% aluminized products are 1.92±0.08 MJ/kg and 2.16±0.10 MJ/kg respectively. The statistical analyses confirm that the fine aluminized emulsion delivers more energy (99.9% significance level).

The increase in the maximum Gurney energy is 12.5% with the addition of fine Al into emulsion. A somewhat larger increase (19%) was found by Hagfors (2005) in underwater tests. In summary, cylinder tests now show that 3.5 and 6% fine Al addition into emulsion increase the maximum Gurney energies by 5.4 and 12.5% respectively.

The inherent error in the testing technique was found to be in the range of 4.5-5.0%. Previous trials had larger variations (7.0±5.4%) due to the experimental conditions.

The results of the JWL analysis are in-line with Gurney analysis: higher energies with the aluminized product. It is shown that aluminized product has less early (within a few volume expansion ratios) energy release but delivers more late energies (at cut-off pressure level of 20 MPa). The decrease in the detonation pressure is 3.9% when 6% Al is added into E682.

However, JWL energy at the 20 MPa cut-off point is increased by 12.2% with 6% fine Al addition.

Limited unconfined VoD tests were conducted with the aluminized product. More data is required though for a detailed modeling study. Detonics properties obtained from the DeNE code (Esen, 2004) and cylinder tests were found to be reasonably in agreement. However, when it comes to the prediction of the expansion, the code was found to be limited particularly with aluminized products due to the limitations and/or assumptions made in the code. The capability of cylinder test in providing the full expansion is promising and should be followed up for particularly complex compositions like 6% fine aluminized emulsion.

The effect of ANFO addition into emulsion:

Previous cylinder tests did not allow the conclusion of any benefit with the addition of ANFO into emulsions on a mass basis; however, ideal detonation codes and underwater tests show improvement

(11)

in explosive performance when AN/ANFO is added into the emulsion. Our old cylinder tests with blends should be repeated for confirmation.

The effect of Aluminum addition into emulsion:

It is shown that ideal detonation code, underwater and cylinder test results suggest benefit from the addition of Aluminum into emulsion. Higher expansion energies were obtained with a 6% fine aluminized (type A80 – average particle size 62µm) emulsion than the one with a 3.5% fine aluminized emulsion using cylinder tests. Our previous cylinder tests show that the addition of 3.5%

coarse aluminum (type A20 – average particle size 250µm) into emulsion does not appear to affect the energetics of the explosive.

Revisiting the underwater tests and comparison with cylinder tests:

Twenty-nine underwater test data were collected to date from the literature. The data suggests that aluminum addition into both ANFO and emulsions is beneficial to the explosive performance. A comparison between underwater and cylinder testing techniques was also made and it was found that the cylinder test method is a more suitable measure of energy transfer to rock in a blasthole than the underwater test is.

Review of full-scale blasts – The effect of Al addition into the explosive (emulsion and ANFO):

Aluminized products appear to lead to better fragmentation and heave based on four case studies collected to date. None of them were documented well and the repetitions were not carried out. Thus, it is difficult to reach a definite conclusion. A few case studies may be needed to investigate whether Al contributes to breakage or heave or both and whether the benefits from Al depend on the rock mass conditions.

Review of full-scale blasts – Comparison of explosive types:

Four case studies were collected to date in which blasting/geotechnical conditions varied. The face velocity of the blend was found be 20% higher than for pure emulsion in Chiappetta’s (1998) study.

Stagg et al (1989) suggest that the effect of explosive on fragmentation is small (considered negligible) in the Waterloo and Granite Falls quarries where rock mass was moderately/highly fractured. Stagg and Otterness (1995) later reported that a blend with 30% ANFO resulted in finer fragmentation than ANFO in the Manitowoc Quarry where the rock mass was competent.

(12)

In general, data is limited in the literature where one tries to find well-documented, repeated case studies with sophisticated blast monitoring systems including reliable fragmentation and rock mass data. Therefore, a series of case studies are needed in the future considering these with a view to make the explosive selection depending on the specific blasting conditions.

The usefulness of the cylinder test in blasting:

The following uses were identified for the current cylinder testing technique which includes the new JWL evaluation;

It can be used to determine the detonics properties (VoD, detonation pressure and specific volume) of explosives, and the pressure-volume and energy expansion curves to a pre-defined cut-off pressure level under a reasonably realistic confinement type.

The test can be used to compare or rate explosive formulations and new explosive products.

The effect of aluminum or any other solid energetic ingredients, ANFO addition into emulsion, changes in the emulsion formulation, etc can be assessed.

This is the only technique, which plots the full expansion energy (work done to the surroundings) curve under real, i.e. non-ideal conditions. Non-ideal detonation modeling is also promising; however, it has not been developed satisfactorily.

The cylinder test has the following limitations:

The results do not measure any direct blasting results (fragmentation, heave, damage) but nor does any other indirect tests.

The results are valid for the charge diameter (100 mm) and confinement (copper) conditions.

Reactions may become more complete at larger diameters for some highly non-ideal explosives such as ANFO or aluminized ANFO.

Future work:

Having discussed the potentials and limitations of the cylinder expansion testing technique, the project group decided to continue with the project with the following cylinder test trials in the short term:

emulsion + coarse Al (4 and 6%).

emulsion + AN (35%).

emulsion + ANFO (35%).

fine Aluminized emulsion (65%) + AN (35%). Fine Al proportion in the emulsion part would be 6%.

(13)

In the long term, the project group is keen on validating the cylinder test results in full-scale and focus on the effects of Al/ANFO addition into emulsions in terms of fragmentation and heave as well as developing guidelines for explosive selection in different geotechnical conditions. These are in discussion currently.

We also revisited the original project target which was the “optimum use of explosive energy”. It was initially aimed to investigate the control of energy transmission from explosive to rock and quantify energy losses in the near-zone around the charge. It appears that the proposed ideas are in-line with these goals. However, more work is required to quantify the energy losses in the crushed-zone around the blasthole in the long-run.

(14)

(15)

List of Contents

Sammanfattning ... i

Executive Summary ... vii

List of Figures ... xv

List of Tables... xvii

List of Symbols ... xix

1 Introduction ... 1

1.1 Cylinder test method as a measure of explosive work capacity... 1

1.2 Project background ... 2

2 Cylinder test technique... 5

2.1 Detonation process of cylindrical commercial explosives... 5

2.2 The Gurney model of explosive output... 5

2.3 Cylinder test set-up ... 9

3 Preliminary analysis of the previous in-house data ... 13

3.1 Explosive types tested... 13

3.2 Data analysis – 100 mm charges... 14

3.3 Data analysis – Diameter effect ... 16

4 Detailed analysis of the in-house published data ... 19

4.1 Brief description of the JWL Equation of State and its relevance to the analysis ... 19

4.2 Detailed analysis of φ100 mm cylinder test data ... 22

4.3 The effect of ANFO addition into pure emulsion explosives ... 26

4.4 The effect of Aluminum addition into pure emulsion explosives... 28

4.5 The density effect in pure emulsion and ANFO type explosives... 29

4.6 Comparison of the energetics of commercial explosives... 30

4.7 Detailed analysis of diameter effect data ... 31

5 New Cylinder Test Trials ... 37

5.1 Manufacturing of the pure and aluminized emulsion explosives for the cylinder test project ………...37

5.2 Loading and preparation of the charges for the cylinder test trials... 40

5.3 Data analysis ... 42

5.4 More detailed analysis ... 45

5.5 New data set for the 40 mm aluminized (6%) E682 ... 51

5.6 Non-ideal detonation modelling... 53

(16)

6 The review of published underwater and full-scale blast results and the usefulness of the

cylinder test in rock blasting ... 57

6.1 Underwater test results... 57

6.2 Review of full-scale blasts relevant to the project ... 60

6.3 Comparison of the performance of explosives using different techniques ... 62

6.4 The usefulness of the cylinder test in blasting ... 63

7 Conclusions... 67

References ... 73

APPENDIX A ... 77

APPENDIX B ... 109

APPENDIX C ... 117

APPENDIX D ... 121

APPENDIX E... 125

APPENDIX F... 129

APPENDIX G ... 145

APPENDIX H ... 153

APPENDIX I... 157

APPENDIX J... 167

(17)

List of Figures

Figure 1. Non-ideal (two-dimensional) detonation representation (Byers Brown, 2002)... 5

Figure 2. The positions of the radii. Size scale is not proportional (Arvanitidis et al., 2004)... 7

Figure 3. Contact signals vs time during cylinder expansion (Arvanitidis et al., 2004). ... 7

Figure 4. An example of an expansion history (radial expansion, ∆r_m vs time, t). ... 8

Figure 5. The disposable rig with copper tube and contact pins (Nie, 2001)... 10

Figure 6. Mounted contact pins with gauge jig (Nie, 2001)... 11

Figure 7. Gurney energies in MJ/kg... 15

Figure 8. VoD (in copper) vs charge diameter effect curve. ... 17

Figure 9. JWL (a) pressure versus volume (b) expansion energy vs volume and (c) expansion energy versus pressure curves for Nagolita 2. Detonation products are allowed to expand down to 1 atm pressure. Nagolita 2’s density and heat of reaction are 0.92 g/cm³ and 3.841MJ/kg respectively. ... 20

Figure 10. Experimental expansion energy data and JWL fitting result for Nagolita 2... 21

Figure 11. Expansion energy versus pressure curves of the pure emulsion (test 160), blends including 20 and 30% ANFO (149 and 186 respectively). Note that detonation products are allowed to expand down to cut-off pressure of 20 MPa... 27

Figure 12. Expansion energy versus pressure curves of the Titan 6000 and 6080... 27

Figure 13. Expansion energy versus pressure curves of the pure emulsion and aluminized (fine and coarse Al) emulsion products. ... 28

Figure 14. Inverse diameter-unconfined VoD relations for the pure and 3.5% aluminized (fine) emulsion (Nie, 2000; Arvanitidis, 2004). ... 29

Figure 15. The effect of the density on the explosive performance (a) pure emulsion and ... 30

Figure 16. The expansion curves (expansion energy versus pressure) for the selected commercial explosives tested. ... 30

Figure 17. Diameter effect for E682... 32

Figure 18. Diameter effect for the blend with 20 and 30% ANFO. ... 34

Figure 19. The expansion energy vs v/v_o for the coarse and fine aluminized emulsion explosives... 35

Figure 20. The relationship between density and microballoon content... 39

Figure 21. Loading of the copper pipes... 40

Figure 22. Double pin set-up... 41

Figure 23. A cylinder test with double pin sets prior to the blast... 41

Figure 24. Comparison of the maximum Gurney energies of the tested explosives. ... 43

Figure 25. Comparison of the Al addition into E682 explosive during current and previous campaigns. ... 44

(18)

Figure 26. Raw expansion data rm(t) for the pure emulsion explosive... 46

Figure 27. Raw expansion data r_m(t) for the aluminized emulsion explosive. ... 46

Figure 28. Radial kinetic energy data Ek(v/vo) for the pure emulsion explosive... 47

Figure 29. Radial kinetic energy data E_k(v/v_o) for the aluminized emulsion explosive... 47

Figure 30. The JWL curves (Ed vs v/vo) for the cylinder tests with the pure emulsion explosive. The curves have been terminated at the cut-off pressure of 20 MPa. ... 48

Figure 31. The JWL curves (Ed vs v/vo) for the cylinder tests with the aluminized emulsion explosive. The curves have been terminated at the cut-off pressure of 20 MPa... 49

Figure 32. Comparison of the average JWL curves Ed(v/vo)of the pure and aluminized (6%) product. The curves have been terminated at the cut-off pressure of 20 MPa... 50

Figure 33. Radial kinetic energy data E_k(v/v_o) for all E682 shots including test 207. ... 51

Figure 34. Radial kinetic energy data Ek(v/vo) at 2.2, 4.4 and 7.2. The data is connected for clarity. .. 52

Figure 35. The JWL curves (Ed vs v/v_o) for the cylinder tests with the aluminized emulsion of 40 mm diameter. The curves have been terminated at the cut-off pressure of 20 MPa... 52

Figure 36. Taylor wave modeling results: p (v/v_CJ). ... 56

(19)

List of Tables

Table 1. Methods for the determination of explosive performance (Esen, 2004). ... 1

Table 2. Explosives tested during the previous campaigns... 13

Table 3. Density statistics of the explosives tested. ... 14

Table 4. VoD statistics of the explosives tested. ... 15

Table 5. Ranking of the Gurney energies of explosives on mass and volume basis for the first six explosives. ... 16

Table 6. Descriptive statistics for Gurney energy at different diameters. ... 17

Table 7. Density and detonation pressures for the E682 tests. ... 22

Table 8. Density and detonation pressures for the E682+3.5%coarse Al tests. ... 23

Table 9. Density and detonation pressures for the E682+3.5%fine Al tests. ... 23

Table 10. Density and detonation pressures for the Blend (E682+20/30%ANFO) tests. ... 24

Table 11. Density and detonation pressures for the ANFO tests... 25

Table 12. Density and detonation pressures for the Titan 6000 tests. ... 25

Table 13. Density and detonation pressures for the Titan 6080 tests. ... 26

Table 14. Ranking of the expansion energies in MJ/l. ... 31

Table 15. Density and VoD for the E682 tests... 32

Table 16. Density and VoD for the blend with 20% ANFO. ... 33

Table 17. Density and VoD for the blend with 30% ANFO. ... 33

Table 18. Density and VoD for the blend aluminized emulsion with coarse Al. ... 34

Table 19. Density and VoD for the blend aluminized emulsion with fine Al... 35

Table 20. Composition of the matrix... 38

Table 21. Composition of the pure (E682) and aluminized emulsion explosives. ... 39

Table 22. Preliminary cylinder test results for the pure emulsion... 42

Table 23. Preliminary cylinder test results for the aluminized emulsion. ... 42

Table 24. Previous E682 data... 44

Table 25. Recently collected E682 data. ... 44

Table 26. Summary of the test 207... 50

Table 27. Preliminary cylinder test results for the aluminized emulsion of 40 mm diameter... 51

Table 28. Comparison of the 40 and 100 mm cylinder test for the aluminized emulsion... 53

Table 29. Input parameters for the DeNE. ... 54

Table 30. Unconfined VoD versus charge diameter data for pure and aluminized emulsion. ... 54

Table 31. DeNE and cylinder test results. ... 55

Table 32. Underwater test results from the literature. ... 58

(20)

Table 33. Comparison of the underwater and cylinder test results. ... 59 Table 34. Comparison of the explosive performances using different techniques. “Yes” and “No”

suggest benefit (improved explosive performance) and no benefit from the explosive

respectively... 63

(21)

List of Symbols Upper-case letters

A JWL constant

A_o Total underwater expansion energy

B JWL constant

C JWL constant

D Detonation velocity

D_c Confined VoD DCJ Ideal VoD

D_u Unconfined VoD E Energy

E_c Energy of compression E_d Detonation energy EG Gurney energy

E_k Radial kinetic expansion energy

E20 Expansion energy at the 20 MPa cut-off pressure level E_o Chemical energy

Es Specific internal energy

E682 A generic emulsion explosive designed by Swebrec M/C Metal mass to charge mass ratio

P Pressure

P50 50% fragmentation passing size R₁ JWL constant

R₂ JWL constant U_s Detonation velocity U_L Gurney velocity

Lower-case letters

p Pressure rm Centre radius of the tube q Heat of reaction

t Time

tpin Copper wall thickness at the pin sector

v Specific volume

v/vo Volume expansion ratio

(22)

vp P-wave velocity of the intact rock x_CJ Reaction zone (DDZ) length

Greek letters

Γ Grüneisen coefficient γ Adiabatic gamma coefficient λ Extent of chemical reaction

ρ Density

ρ_L Liquid density of the unreacted explosive ρo Unreacted explosive porous density ρ_r Density of the intact rock

ρ_S Solid density of the unreacted explosive

ω JWL constant

θ metal deflection angle

Superscripts and subscripts

c Confined

CJ Chapman – Jouguet state

o Initial state

u Unconfined

Abbreviations

AEL African Explosives Limited

Alnafo Aluminized ANFO

ANFO Ammonium nitrate/fuel oil explosive CJ Pertaining to Chapman – Jouguet DDZ Detonation driving zone

DeNE Detonics of Non-ideal Explosives. A non-ideal detonation code developed by Esen (2004).

EoS Equation of state

Gamma CJ CJ gamma Gamma R Ideal gamma

JWL Jones-Wilkins-Lee Equation of State MFL Mass fraction of liquid

Swebrec Swedish Blasting Research Centre

Vixen-i An ideal detonation code developed by African Explosives Limited (AEL VoD Velocity of detonation

(23)

1 Introduction

1.1 Cylinder test method as a measure of explosive work capacity

The determination of the explosive work capacity in rock blasting is an active research field and there is no established method of predicting this. In general, researchers and/or practitioners in rock blasting have used the simple calculations and/or field tests to extremely complex computations (Table 1).

Table 1. Methods for the determination of explosive performance (Esen, 2004).

A) Early performance test

methods (Persson et al., 1993)

- Ballistic mortar - Grade strength - Brisance

- Trauzl lead block test - Plate dent test - Cylinder test

- Underwater detonation test - Crater test

- Langefors’ weight strength B) Theoretical studies (Peugeot

and Sharp, 2002) - Ideal detonation modelling - Non-ideal detonation modelling

With the widespread use of blasting agents, which do not detonate in small laboratory sample sizes, the usefulness of small scale laboratory test methods (Table 1) for explosive strength has become very limited (Persson et al., 1993). The weight strength method is traditionally associated with the strength markings of different dynamite grades, and has little correlation with the effectiveness of an explosive in blasting. It has no meaningful relation to modern commercial products such as ANFO products, water gels, and emulsions. Therefore, weight strength method is considered not a useful parameter for the rating of explosive performance (Tosun, 1991; Persson et al., 1993; Hopler, 1998). In this project, the cylinder test is preferred over the other experimental test methods and the comparison of this test with the underwater test is discussed in proceeding sections. The energy transported into the surroundings in succeeding stages of the expansion process can at present only be determined experimentally by these two methods (i.e. cylinder and underwater tests).

The use of detonation theories to compute the detonation properties of explosives is potentially an effective method of predicting the performance of explosives and offers valuable insight into the realistic performance of commercial explosives. The detonation codes namely Vixen-i (an ideal detonation code developed by African Explosives Limited, see Cunningham (2001)) and DeNE (a non-ideal detonation code developed by Esen (2004)) were made available during the project. The

(24)

relevance of these codes and the discussion on the modeling and experimental analyses were also given in later sections.

The cylinder test has long been the principal method for obtaining the Equation of State (EoS) parameters for military (high) explosives (e.g. Souers and Haselman, 1994; Trzcinski, 2001) and recently for commercial explosives (Davis and Hill, 2001; Trzcinski and Cudzilo, 2001). The tube size is usually Ø = 25.4 (1”) or 50.8 mm by 0.3 m long. Sometimes 4” and 8” tubes have been used with non-ideal explosives (Davis and Hill, 2001). The expansion adiabat (pressure versus volume curve) and detonics properties (detonation pressure and volume at the end of the detonation driving zone;

detonation velocity using a continuous VoD system) as well as expansion work of explosives can all be determined using this testing technique. It is believed that cylinder test is the only experimental technique which gives these explosive characteristics.

1.2 Project background

Swebrec has chosen the cylinder expansion test as the most suitable test to measure the initial work capacity of civil explosives in a realistic geometry and charge size. This goes back to SveBeFo (Nie, 2001) and the Less Fines project where emulsion E682, Titan 6000&6080 bulk emulsion and ANFO (with or without aluminum) were tested (Nyberg et al., 2002). The work capacities of different explosives based on the Gurney energies were compared in the Less Fines project.

During 2003-2004, the cylinder tests continued under Ioannis Arvanitidis. After he left, the test series was finished by Ulf Nyberg, Hiroyuki Arai and Lars Granlund. The 2004 work investigated the effect of adding 30 % ANFO or coarse or fine grained aluminum to the E682 emulsion, in Ø 100 mm copper tubes. A few Ø 40 mm tubes were shot to investigate the size (diameter) effect (Arvanitidis et al., 2004).

The resulting differences in work capacity between formulations now became much smaller and were sometimes masked by the scatter in the data. A special point was that the addition of 3.5 % aluminum didn’t seem to increase the work capacity of E682 more than marginally, contrary to expectations.

The 2004 work had revealed that the copper tubing didn’t meet the manufacturer’s specifications for the wall thickness. The tubing’s basically circular walls were not concentric so the wall thickness varied considerable. Therefore, a special set of Ø 100 mm tubes was ordered and used during the last test campaign.

(25)

A test with two pin sets revealed that different wall velocities and hence Gurney energies were obtained when the expansion was measured at the thin and thick sectors of the copper tube. The computed difference decreased when the actual wall thickness was used in the evaluation.

The wall thickness at the measuring point had not been measured at the beginning, which introduces a scatter in the results. Furthermore, the in-tube density of the explosives showed inconsistent variations and the previous work had indicated that the work capacity increases linearly with density. In addition, the evaluation procedures had changed between different operators.

Because of these difficulties, the program board decided on February 23^rd to re evaluate the cylinder data obtained to date using a standardized evaluation procedure and also specify the test’s usefulness for testing commercial explosives with respect to work capacity or other rock blasting results.

The project group (Dyno Nobel, Kimit and Swebrec) met on April 13^th and decided to suggest the following:

1. Re-evaluate the in-house cylinder test data obtained to date with a standard evaluation procedure. Do the Vixen-i calculations of the recipes to obtain the ideal detonation characteristics of the formulations tested during the testing campaign.

2. Do the JWL EoS fitting to obtain the EoS parameters of the explosives and compare these with Gurney energies.

3. Collect underwater test data, in-house and external, plus field test data for explosives. Make a special comparison of the effect aluminum when added to ANFO and emulsion.

4. Do 8 new cylinder tests, 4 each of E682 and of E682 with aluminum.

▪ Choose as much aluminum as can be put in without changing the salt solution (ratio AN/SN/water).

▪ Design micro sphere content to give same density in both cases.

▪ Use double sets of pins to evaluate effect of tube wall thickness

▪ Make diameter effect curves for explosives.

5. Specify the test’s usefulness for testing commercial explosives with respect to work capacity or other rock blasting results.

These tasks were carried out in 2005. The pure emulsion (E682) and aluminized E682 (6% fine Al of A80 type) explosives were manufactured in July 2005 and the tests were conducted in August 2005.

The manufacturing of the explosives involved Linda Lindberg and Hans Perlid of Dyno Nobel and Sedat Esen of Swebrec (Swedish Blasting Research Centre). Iver Hauknes and Lars Granlund of Dyno

(26)

Nobel; Hans Karlström of Kimit; Ulf Nyberg and Sedat Esen of Swebrec were all involved in the loading and testing of the explosives.

In this report, the re-analysis of the published in-house data, detailed analysis of the new trials, ideal and non-ideal detonation modeling results, unconfined diameter effect data, published underwater tests and comparison with the cylinder test, published full-scale data, the usefulness of the test in the mining industry as well as discussions for the way forward and recommendations are given in detail.

(27)

2 Cylinder test technique

2.1 Detonation process of cylindrical commercial explosives

The cylinder expansion tests conducted in this project use cylindrical charges which are commercial explosives showing non-ideal detonation behavior. It is also widely recognized that performance of commercial explosives is dependent on the charge diameter and confinement type. In a non-ideal situation, as represented in Figure 1, the shock front is always curved, the flow of the reaction products diverges and reaction is never complete in the detonation zone. The detonation driving zone (DDZ) terminates at the sonic line and contributes to the support of the detonation process (Byers Brown, 2002). In this case, the detonation velocity may approach, but never exceeds, the ideal detonation velocity (Persson et al, 1993).

Figure 1. Non-ideal (two-dimensional) detonation representation (Byers Brown, 2002).

he flow behind the sonic line is supersonic, so that perturbations such as compression and rarefaction

urney model of explosive output

The cylinder expansion tests determine the kinetic data that can be used to evaluate the work capacity f an explosive. The work capacity has been usually expressed by the so-called Gurney energy T

waves, which move at the local sound speed, can never catch up with the DDZ, and therefore cannot contribute to or diminish the speed of the detonation wave. The rarefaction (Taylor wave) which occurs in the supersonic, still reactive, the flow between the sonic line and the end of the reaction zone is of importance, especially with commercial explosives which may have slow energy-releasing reactions and can contribute substantially to the blast (Byers Brown, 2002).

2.2 The G

o .

(28)

Gurney energy is the final kinetic energy of driven metal and expanding detonation products per unit ass of the explosive. The well-known Gurney equation (Kennedy, 1998) gives the relation between m

the work capacity or Gurney energy, EG (MJ/kg) and the final tube wall velocity (Gurney velocity), UL

(km/s):

⎥⎦⎤

⎢⎣⎡ +

= 2

1 C M 2

E_G U^L² (1)

is the result of applying the mass, momentum and energy conservation equations both the detonation product gases and the metal confinement (Kennedy, 1998). The radial velocity distribution inside th

are considered negligible compared with the kinetic energy of explosive and metal. The Gurney elocity is the final velocity during the later stage of tube wall expansion.

he radial wall expansion history and velocity of detonation (VOD) are measured during the test.

sing the relations given in Equations 2 nd 3 respectively.

where M (kg) is the metal mass and C (kg) is the explosive charge mass.

The Gurney equation to

e gases is assumed to be linear. The energies due to heat, deformation and friction

v

T

Therefore, metal deflection angle (θ) and UL are determined u a

⎟⎠

⎜ ⎞

⎝

= ⎛

VOD arctan U

θ ^m (2)

⎟⎠

⎜ ⎞

⎝

⋅ ⎛

⋅

=2 VOD sin U_L

2

θ (3)

true radial motion is about 7° off from the perpendicular measured radial change. Also, the acceleration is calcu

function of time.

direction of the copper tube (Hornberg where U_m (km/s) is the velocity perpendicular to the tube axis.

The experimental results from the cylinder expansion tests are presented as the radial change of the confining copper cylinder wall as function of time. The radial change is perpendicular to the length axis of the copper tube. The

lated from the perpendicular radial change and presented as

In order to evaluate the kinetic data of the expanding cylinder, the velocity of the copper should be calculated at the centre radius of the tube, r_m, which could be estimated under the assumption of incompressible deformation and no material flow in the length

(29)

and Volk, 1989). Under these conditions, the center or mid-wall radius, rm, is calculated from the half cross sectional surface area:

( ) ( ) (

i²

)

2 y 2

i 2 m 2

m 2

y π R R

2 r 1 r π r r

π − = − = − (4)

where ry and ri are the outer and inner radii, Ri and Ry are the initial inner and outer radii (Figure 2). ry

is measured as function of time using contact pin in our case. An example is given in Figure 3. The radial change of the centre radius, ∆rm, is given in Equation 5.

∆rm = rm – Rm =

2 R R R

2 R r R

2 y 2 2 i y 2

y 2 2 i y

+ −

− −

+ (5)

r_m r_i

r_y

Figure 2. The positions of the radii. Size scale is not proportional (Arvanitidis et al., 2004).

-20 -10 0 10 20 30 40 50 60 70 8

Time (µs)

0

Voltage

Figure 3. Contact signals vs time during cylinder expansion (Arvanitidis et al., 2004).

(30)

The calculated expansion curves were obtained by fitting a combined linear and exponential function to the experimental data (Hornberg and Volk, 1989). This function is given below:

( )

_⎥⎦^⎤

⎣

meters. The shift t = t_exp- t₀ makes comparable expansion d data is shown in Figure 4.

⎢⎡ − − ⋅ −

⋅

=

∆ _exp ₀ 1 1 ⁻b^⋅⁽t^exp⁻t⁰⁾

m e

t b t a

r (6)

where a, b and t₀ are the curve fitting para

curves start at t= 0, ∆rm = 0. An example for experimental expansion data as well as fitte

0 10 20

0 10 20 30 40 50 60 70 80 90

Time (µs)

Ra ³⁰

40 50 100

dial

60 70 80 90

expansion (mm)

Measured Fitted

Figure 4. An example of an expansion history (radial expansion, ∆rm vs time, t).

The tube wall velocity perpendicular to the tube axis, U_m,is obtained by differentiating Equation 6 with respect to time:

[

^bt

]

m

m m a e

dt

U d∆dtr =dr = 1− ₋

= (7)

U_m is then converted to the true wall velocity U_Lvia Equations 2-4 and to E_Gvia Equation 1. a represents the asymptotic expansion velocity perpendicular to the tube axis, U_m, in Equation 2. 1/b describes the length of acceleration phase.

The volume expansion ratio ν/ν_oof the detonation product gases are approximately given by:

(31)

( )

2 2 2 2

2

2 2

1

i y

i rm R R

r − −

= ν =

i i

o R R (8)

wed as a testing method which determines the early (initial) work capacity of explosives in a realistic size and geometry.

The Gurney energy should not be confused with other energy characteristics of explosives such as heat of detonation and expansion work. The heat of detonation is the heat of decomposition of the explosive into the detonation products, the chemical equilibrium of which is assumed to occur at the Chapman-Jouguet point. Ideal detonation codes (e.g. Cheetah, Vixen-i) give this parameter. The expansion work of detonation products is the parameter of explosives, which characterizes the degree of conversion of the chemical energy into the mechanical work, ie, the expansion work determines the capacity of explosive to do work (Trzcinski and Cudzilo, 2001). The maximum theoretical adiabatic expansion work is sometimes named detonation energy. Underwater and cylinder tests are able to measure the energy transferred to the surrounding medium (water and copper respectively). The initial work capacity of detonation products is expressed by the so-called Gurney energy. The Gurney energy is the final kinetic energy of a driven metal envelope and expanding products per unit mass of

xplosive (Trzcinski and Cudzilo, 2001; Kennedy, 1997).

.3 Cylinder test set-up

ll thicknesses have been used: one is 1/20^th of the inner diameter (half wall) and the ther is 1/10^th the inner diameter (full wall). The latter has been used widely.

ν

Souers and Kury (1993) states that one of the strengths of the cylinder test is that the actual relative volume of the product gases is close to the geometrical relative volume as determined by the expansion of the cylinder, at least for 2<ν/νo<7. Thus, the cylinder test should be vie

e

2

The cylinder test was originally developed at Lawrence Livermore National Laboratory (LLNL) (Kury et al., 1965). The test uses the streak camera (e.g. Kury, et al, 1965; Bjarnholt, 1976; Hornberg and Volk, 1989; Lan et al, 1993; Souers and Haselman, 1994) or fabry interferometer (e.g. Souers and Haselman, 1994) or impulse X-ray photography (Trzcinski and Cudzilo, 2001) or contact pins (Nie, 2001). Two wa

o

The set-up used at the Swebrec is the one adopted by FOI, the Swedish Defence Research (Helte et al., 1999). Soft annealed oxygen free copper tubes with a fixed ratio (half wall) between the inner diameter and the wall thickness are mounted in disposable rigs as shown in Figure 5.