Tilman Rasche BE MSc Senior Inspector of Mines, DEEDI

1. A Regulator’s View of Automation, Proximity Detection and Collision Avoidance Systems Tilman Rasche BE MScSenior Inspector of Mines, DEEDI

2. Link between Proximity Detection and Automation Triumph of partial and full automation will depend on the success of proximity detection and collision avoidance technologies and approaches. Therefore …. We can learn from our current experience (successes and mistakes)

3. Assumptions….. Two blondes living in Brisbane were sitting on a parkbench talking........ and one blonde says to the other, "Which do you think is further away..........Melbourne or the moon?" The other blonde turns and says "Helloooooooooo, can you see Melbourne ...?????"

4. Dictionary Assumptions are beliefs or ideas that we hold to be true — often with little or no evidence required. We make assumptions every day of our lives… …. blondes have more fun…. …. the earth is flat….. …. as a driver on the highway, I assume that other drivers will obey traffic signals, so that when I go through an intersection with a green light, I assume that the cross traffic will stop at its red light.

5. Stats from the US - perspective 11% 4.5 in 1000, 1 in 222 http://lifesavers.ky.gov/lifesavers_2006/session17-allred.ppt#671,2,Red Light Running

6. Assumptions ‘solve’ a lot of problems we would otherwise have to investigate and answer for

7. ~ 3 out of 4 = ~ 75% - vehicle related Scientific investigation into large scale accidents has allowed us to come a long way

8. Why do we see what we see… Increasing dependence on mine vehicles – OC & UG More & more vehicles, increase in size & speed • Visibility and space around machine • Operator & worker behaviour • Recurring accidents • Same picture worldwide Key mining hazard but there are solutions • Need to clarify assumptions…… Its up to all of us to solve the problem….

9. HPIs as of March 2011 Source : DEEDI - Serious accidents and high potential incidents Mining and Quarrying Compilation of reports for March 2011

10. Almost Likely Possible Unlikely Rare Almost Likely Possible Unlikely Rare Likelihood certain certain (has (has (heard of it (heard of it (not likely to (not likely to (almost (almost (common / happened) happening) happen) impossible) (common / happened) happening) happen) impossible) Consequence repeating) repeating) Catastrophic Catastrophic 1 2 4 7 11 1 2 4 7 11 (eg fatality) (eg fatality) Major Major 3 5 8 12 16 3 5 8 12 16 (eg permanent (eg permanent disability) disability) Moderate Moderate 6 9 13 17 20 6 9 13 17 20 (eg medical / (eg medical / hospital treat) hospital treat) Minor Minor 10 14 18 21 23 10 14 18 21 23 (eg first aid) (eg first aid) Insignificant Insignificant 15 19 22 24 25 15 19 22 24 25 (eg (eg no injury) no injury) ‘Pegging’ the Risk - Rank Assumptions?? Maximum reasonable consequence? Most likely consequence? Likelihood?

11. Hazards • Moving equipment (‘tons’ of kinetic energy - CAT 797F, payload 360 ton GVM 623 tons, 3,800 HP, top speed 67 km/h, 14.8m long, ~10m wide) • Boeing 747-400 178 tons 3.5* • LTA visibility • LTA hazard awareness

12. Automation NOW CAS

13. Basic design assumptions What is the basis for …. • Road widths - 3.5 rule • Separation distances – vehicles in motion require ??? m to stop • Maximum allowable speeds to allow safe braking • Queuing and parkup distances • Detection envelopes for proximity detection technology – cloverleaf-shape??? • Where is the empirical data to substantiate the above??? • Braking distances • Operator reaction times • How will people react? We too often assume we know what the problem is

14. Engineering Assumptions Q: What is the braking distance of xyz haultruck? (50 km/h, level ground, dry road, loaded) Q: What is the minimum length of runway for the safe landing of a Boeing 747? A: Google http://www.boeing.com/commercial/airports/acaps/747_4.pdf

16. V2V V2P slow speed e.g. Parkup areas V2V – overtaking collision V2V – high speed rear end collision V2V - rear end collision V2V – slow speed rear end collision V2V – head on collision V2I or V2P forward collision V2V collision or reversing over dump V2V collision – fast –slow moving vehicles V2V collision – mining face V2V V2P V2I reversing collisions V2V collision - intersection What is the problem?

17. Continuous Miner Shuttle Car Typical Underground Scenarios, there are many more….. No Go-Zones !

18. Legend Victim location X Operating the remote Not operating the remote A moving RCCM collided with another at an intersection, causing the stationary RCCM to pivot and crush the other Op Maintenance activity Example – design decision partially responsible for mining deaths MSHA - UG Fatalities - Continuous Miner 29 fatalities or 72% of victims were operating the remote at the time of the accident. Most/all could be avoided if Proximity Systems had been available and installed

19. Stopping distanceAcknowledged braking capabilityBehavioural expectation

20. ‘There are only so many ways to kill people, and we know them all’

21. ? ? Design Visibility = Opportunity to identify a hazard & react in time

22. Shuttlecar or Truck Miner

23. Person vs. Machine 1 Human Reliability - People are inherently unreliable Mental processing time • Sensory – perception/recognition • Interpretation – what does this mean – friend or foe? (Car stopped in the middle of the road) • Response Selection – what happens next? Expectation • Expected to brake – 0.7 secs = 0.5 secs perception 0.2 secs movement • Unexpected – 1.25 secs = 1.05 secs perception 0.2 secs movement • Surprise – 1.5 secs =1.2 secs perception 0.3 secs movement Movement time • Brake engagement time – foot movement, on pedal, depress, mechanical delays

24. Person vs. Machine 2 Other factors • Urgency – time to collision • Cognitive load – ‘non driving’ matters – music, mobile phone, autopilot • Age ~ lower levels of fitness ~ lower response capability • Gender • Nature of signal – can it be seen? Is it distinct? Is the vehicle in front accelerating/decelerating? Aspect – frontal/from side • Visibility vs recognition • Reaction time at night – visual contrast (amber/yellow brown shooting glasses) • People will make mistakes (wrong decisions)

26. HR RA Tools Risk 1. High Level QRA, Engineering type analysis, FTA, BTA, FMEA, HAZOP WRAC 1. Major Hazard Baseline RA Elimination ‘Specialists’ 2. Project/Change Issue RA eg. Coll. Awareness Substitution 2. FTA, BTA, FMEA, HAZOP, WRAC 3. Routine & Non Routine Task Planning RA Engineering Control 3. WRAC Degree of Difficulty Reliance on People to control risk ‘Generalists’ Administrative Control - Procedure 4. Individual ‘continuous’ Face RA 4. JSA Mine workers PPE Human behaviour Risk Tilmans ‘Meaning of Life’ Ver.1.0 Risk & Issue People to administer risk Control Options RA Tools Req’d HR Resources

27. Some CAS Technologies • RFID – tags and readers • Radar • ‘Magnetic bubble’ • Laser scanning • GPS – surface only • Cameras • Combination of the above • Opencut and Underground metalliferous - available now • Underground Coal – requires IS certification – mid 2011 • Cost from $5k per vehicle

28. Fit for purpose equipment - Selection of the equipment • Sites to review all risk assessments based on local and published collision scenarios • Change management – change of system functions • Verify that selected proximity detection system is in fact able to mitigate the collision scenarios • Need explicit underlying assumptions (speed, distance etc.) • Polar diagrams’ showing the actual detection envelope of their systems, not assumed envelopes - ‘clover leaf’ vs actual pattern. • Clarify CAS - ‘collision awareness/avoidance system’ • Physics - mine sites must understand what the chosen system can and cannot do - Manufacturers to declare the capabilities of their systems.

29. Fit for purpose equipment - Selection of the equipment • Manufacturers to declare if their systems are ‘collision awareness’ or ‘avoidance systems’, and provide sound, logical and unambiguous evidence for their judgement. • Sites to check inference with other radio frequencies • Hardware – veiling (reflection), clarity of display etc. • Placement of screens/ alarming units – should be in field of vision

30. Fit for purpose equipment - Selection of the equipment • Combination of screens and method of alarming – intuitive exception based alarming based on criticality? • Mine sites to review the systems ability - future proofing • Review breaking distances for all vehicles for the range of operating speeds and conditions - verify that current site vehicle separation distances are sufficient

31. Competent People • Review ‘Operations rules’ i.e. if systems deemed safety critical, then operational procedures must ensure consistency of approach. • Site champions - effective acceptance and utilisation of proximity detection systems – link between site – operations and maintenance and the proximity detection manufacturer. • Dedicated maintenance personnel to ensure a successful commissioning and implementation of the system. • Who is going to maintain proximity detection and automation systems – specialised skill • Proximity detection issues - simulator training • Maintainability – easy and safe access to all external hardware must be achieved.

32. Safe Work Practises • Sites to review and update relevant site procedures incl. prestart checks to ensure proximity detection systems and their importance as a safety control is assured and recognised. • Proximity detection system requirements for contractor vehicles operating at different sites. Commonality of approach and rule-set. • Design and roll out a comprehensive training program that outlines how to use the system effectively. Incorporate a section that explains what the system can do and what it cannot do. • ‘Nuisance’ alarms or conditions may be in fact real alarms due to the systems design and capabilities (physics)

33. Controlled Working Environment • Proximity detection equipment must not be considered as the primary solution to mitigate collision risks. • Must also consider their pit design & layout – intersection, haulroad, dump designs, road separation human behaviours etc. • Inherently safer operating environment

34. Market Place • Collaboration - several prox detection OEMs integrating their systems into ‘one’ • Combined systems better than sum of all • Some machinery OEMs allowing prox system to manage their machine – eg. braking • GPS ( high speed) plus radar (slow speed) – opencut • ‘Magnetic bubble’ plus … - underground • Ability to create non-detection envelopes on the equipment

35. In Summary • ‘What is the problem, then look for solutions’ - effectiveness • (V2V, V2P V2I) Accidents are preventable • Free lessons from proximity detection systems implementation towards automation • Proximity detection systems are not the complete answer but are an essential part of the solution • Must also look at human factors – human ‘unreliability’ • Proximity Detection Technology is available or rapidly becoming available • Need a side by side integrated combination of approaches • Must be embraced – life saving technology • Key to making automation a safe reality

36. Every miner home safe and healthy every day