“ Customized fire detection for challenging applications”

1. “Customized fire detection for challenging applications” Rick Jeffress CET, CFPS

2. Agenda Challenges of large volumes Signifire Technology Overview Fire Videos Listings and approvals Performance Based Design SpyderGuard Overview More videos Live Fire Demo

3. Customers Royal New Zealand Navy

4. Intro to Video Image Detection Based on Machine Vision (video analytics) technology Utilized in: manufacturing facial recognition inventory management product identification

5. True early detection technology is here…. Old Technology New Technology

6. Solution The most effective fire detection system is the human being, using visual identification of fire or smoke.

7. Solution Signifire uses visual imaging and artificial intelligence (machine vision) to emulate human fire recognition. “Trained” with 1000’s of real and false fires. Analyzes each frame of 480X640 pixels @15 frames/sec Scans 4.5 million datapoints (luma value)/second analyzing the image for signs of combustion 24/7/365.

8. Challenge Detection device challenges for high ceilings: Spot Smoke/Heat Detectors: Beam detectors Aspirating Systems Linear Heat Smoke must interact with detection device Installation, inspection and maintenance on high ladder, lift or scaffolding Signifire can be installed lower than ceiling level and brings detection to the fire. 65 82

9. Challenge Conventional detection devices are not suitable for many applications having large volume or high ceilings Stratification issues Sprinkler systems designed for building protection, require a large fire to activate First responders require early, “actionable intelligence” to avoid costly fire damage and injury. Cause of fires unknown – can rarely make changes to avoid future incidents

10. Customized Fire Detection Detection Algorithms Smoke Flame Reflected Fire Light Motion All included on camera Individually enabled and adjusted Detect or blocking zones Sensitivities Time delay (Alarm verification) Relay output On/Off schedules Targeted Detection Reduce False Alarms

11. Signifire Video Smoke and Fire Detection Uses visible spectrum video images from stationary cameras Analytic process identifies characteristics of flames and/or smoke Outputs alarms and video to user interface

12. Reflected Fire Light Filter Filter removes residual light, leaving only light that is originated from flames.

13. RFL Implementation

14. Distribution Center Fire

15. Magnesium Plant

16. Excitation/Aging Filter • Smoke Detection Algorithms Smoke Aging (red-orange older, blue newer) for movement Smoke Signature Recognition

17. Distribution Center Smoke

18. Cole Field House Univ of MD –Concession Smoke – 300 ft

19. Dominion Energy Smoke

20. Motion Detection Motion detection is a byproduct of analysis Generally not as sophisticated as security analytics Can be tied to image location and schedule

21. IP Network Camera

22. System Architecture CAT 5E cable Addressable Modules SLC (24vdc for NFPA) Fire Alarm Control Panel Camera Connections Terminal strip 1,2: 12 – 24vdc 3,4,5: Programmable Alarm 6,7,8: Trouble 9,10,11: Programmable Alarm 12,13: Unused Network Video Recorder Options 4, 18 or 32 camera NVR

23. Enclosures • Environment Dictates Type of Housing • Explosive • Dust or dirt • NEMA 4x • Cold - Heated • Hot - Cooled

24. Coverage Comparison It takes approximately 20 Spot-type smoke detectors to cover the area of ONE SigniFire camera. Very cost effective in ceilings with beam construction. Area of coverage based on 100’ = approx. 7800 sq ft Area of coverage based on 150’ = approx. 17,600 sq ft

25. FM Global and UL • Recognizes Use of Flame & Smoke VID Systems • Flame VID Classified as Radiant-Energy Sensing Detector • UL – 268B Video smoke detection • FM – Uses FM 3260 (new FM 3232) and UL 268. • FM 5-48 (Jan. 2011) Recommends video smoke detection • Property/Business Continuity vs. Life Safety

26. FM Global Update – January 2011

27. NFPA Journal Aug/Jul 2011

28. NFPA 72 • 5.7.6.1 VID and all components thereof, including hardware and software, shall be listed for the purpose of smoke detection • 5.7.6.2 VID shall comply with all of the applicable requirements of Chapters 1, 4, 5, 6, and 10 of the code • Chapter 4 power supply criteria • Communications integrity • Listed for the purpose • Alarm signal to FACP • 5.7.6.2.1 System Shall be designed in accordance with the performance based design requirements of 5.3 • 5.7.6.2.2 Location and spacing of VID shall comply with 5.10.5 • Prevent tampering and ensure integrity of mission of system.

29. Visible vs IR Illumination • VID systems need illumination either IR or Visible Light • LED’s most common due to long life expectancy and low power consumption • NFPA 101 requires emergency lighting (1 Foot candle) • Switching between IR and white light can result in a focus shift (competition) • Blinding of competitor CCD chip with IR • IR is not visible to humans • White light provides safety and security

30. Common Applications • Large area detection • High ceilings • Large volume spaces • Irregular ceiling, beam pockets • Special hazards • Local application detection • High fire hazard equipment • Telecom/Server rooms • Outdoorhazards • Process areas • Flammable storage • Wireless • Portable fire watch

31. SpyderGuardTM Functions with standard PC Provides video surveillance, smoke, flame and motion detection View and record the condition that caused the alarm View and manage system via internet Receive email alerts Audible voice warning Locate the event on building/site plans Coordinate response efforts

32. SpyderGuardTM Flexibility to reduce false alarms or provide targeted detection Mask or section areas of camera view to: Disable smoke or flame detection Enable motion detection Scheduler allows programming to times of day and days of week System alarms when smoke or flame crosses plane, out of masked area Detection zones target specific equipment Masking or Detect Zones Potential false alarm Smoke block RFL block Flame Detect Zone

33. SpyderGuardTM Timeline Screen • System records 24/7/365 • History timeline of detected events • View and download video from any time period • Easily identify and download video of events • Used in commissioning – eliminate false alarms

34. Commissioning and Maintenance • Audit Report Generation • Commissioning report • System tests • Propane torch • Smoke emitters • Canned smoke • Motion zones • User initiated alarms • Confirm original field of view

35. Video Detection Scenario SpyderGuard Visual ID Email message Improved response time and situational awareness Fast Detection Coordinate response with live video IP Dial 911 Dry-contact

36. Hydroelectric facility

37. Sikorsky Helicopter

38. Satellite Assembly

39. Electrical Rooms

40. National Academy of Science (DC Mall)

41. Customizable Fire Detection One or more detection algorithm(s) Detect or mask areas Zone on-off schedules Sensitivity settings Cost savings A. Less wire and conduit/scaffolding B. Fewer devices Minimum operational disruption Quick system modification “Custom Fire Detection Solutions For Challenging Applications”