article_024_tnail
article_024_tnail
article_024_tnail
article_024_tnail
article_024_tnail

Flame detector stops false alarms

Jan. 22, 2006
Multi-spectrum IR sensing and neural network processing is combined in one device to give automation professionals a potent new tool in protecting people and plants from dangerous flame sources.
THERE ARE five primary optical flame sensing technologies in use today, all of them based on either UV spectrum or IR spectrum sensing. Depending on the application, the designer selects a specific technology. While these devices meet basic flame detection needs, they are prone to false alarms caused by reflected sunlight, arc welding, hot piping, machinery, and other false-alarm sources. 

“False alarms are more than a nuisance. They are both a productivity and cost issue,” says Ardem Antabian, product manager at General Monitors. “From the perspective of productivity, a false alarm incident in most process plants requires a system shutdown. Turning off and then restarting a process line may take anywhere from an hour to a whole shift.”

    

FL4000 FLAME DETECTOR

The FL4000 gives automation professionals a potent new tool in protecting people and plants from dangerous flame sources.

Stopping spurious alarms is critical for both the proper functioning and the credibility of the flame detector’s outputs. Typically, flame detection is based on multiple optical flame detectors, each set at a different wavelength, and each integrated with expert signal processing. This type of system relies on an expert to design a fixed set of conditions. As long as these fixed conditions obtain, the device works reasonably reliably.

The FL4000 Flame Detection System from General Monitors is designed with four multi-spectral infrared sensor systems and a unique neural network with a recursive training algorithm. In other words, the flame detector itself is trained to discriminate between background radiation and a valid infrared target.

In operation, each of the four sensors’ signals are sent to an A/D converter and then into the neural network’s processor. Once the neural network has done its discrimination, the signal is routed to three different output paths. The first path is a 0-20 mADC analog output. Second is a logic signal to LED indicators and built-in 8A relays. The third path is Modbus for complete addressability.

FL4000 provides a detection depth approximately 10% longer than competitive devices, with a detection range of 230 feet, compared to the nearest competitor’s 210 feet. At the same time, FL4000’s horizontal field of view is almost 10% wider at maximum distance than its nearest competitor, according to the company.

Built-in diagnostics check both window cleanliness and the detector’s electronic circuitry every two minutes. Also, 128 units (247 with repeaters) can be serially daisy chained to a host computer using the Modbus RTU protocol. The communication registers provide alarm status, fault and other information for operation, troubleshooting or programming the unit. FL4000’s electronics are integral in an explosion-proof (Class I, Div. 1, Groups B, C and D, includes Zones 1 and 2; Class II, Div 1, Groups E, F, and G; Class III; Type 6P, IP67) stainless-steel enclosure.

Accurately and reliably detecting flames has long been a serious challenge. No process plant can afford the costs of false alarms in this competitive environment. Losing a day’s production can mean the difference between a profitable plant and one that’s not. FL4000 is a new weapon in the fight for profitability in process manufacturing.

  About this Product Exclusive
For more information about the FL4000 flame detection system, contact General Monitors at 949/581-4464, or visit the company's web site atwww.generalmonitors.com.