Hi guys,

I'm searching for a Through Beam Sensor to use in a boiler room.
The beam can not be affected by open flame and smoke.
I'm having trouble identifying such sensor, did any of you guys ever use something like this?

 

What are you wanting to detect?

 

Solid fuel being loaded into the furnace.
The sensor would be mounted behind glass, outside of the furnace.
Flame and smoke could sometimes get in the way of the beam, and I'm wondering what type of sensor would be best suited for this application.

 

This is a tricky application. Any digital through-beam sensor is going to have a threshold that will trigger a state change and with all the sensors I have experience with, this is set imprecisely, either by a potentiometer on the receiver, or by a "teach mode" button on the receiver. I don't think either type will perform to your requirements. I think you need something "smarter" than that. Immediately I think of a computer vision camera, but that may be overkill. I think you can probably pull it off with a thru-beam sensor if you use an analog model (ex Banner Analog Omni-Beam) and use the signal to do a more advanced detection than simple threshold detection. You will be looking for certain events; solid fuel passing through the beam should look differently than smoke/flame passing through the beam, when the sensor's analog output is printed out on a strip chart or plotted on an oscilloscope time graph. You would need to monitor this output and characterize these events, and write an algorithm which is effective at detecting the events. Solid fuel passing through I would expect to manifest as brief (but predictable/consistent timespan) total losses of signal with sharp rising/falling edges, while flame/smoke I would expect to look more organic, smooth, drops to low (but not completely gone) signal strength, on a variable timespan. There are at least 3 characteristics there i have identified as potential parameters for the algorithm: signal level, timespan, rising/falling time, but there may be more.

 

An ultrasonic beam might work. I would expect solids to attenuate it more than flames or smoke would (but I haven't researched that).

 

Hi Alec.
If the beam break is near a furnace, hot air draughts caused by the heat could deflect the ultrasound path.?
E

 

Quite possibly.

 

This is a situation where we need more information!! While detecting fuel feed may be the goal, there is clearly more to it than that. So there is more to the story than we are told. Probably a "smart camera" vision system would be the solution. Even though I do not yet understand the whole situation.

 

An ultrasonic beam might work. I would expect solids to attenuate it more than flames or smoke would (but I haven't researched that).

Most of the sensors I've looked up have a maximum working temperature less than 100deg C.
I wouldn't be able to place an Ultrasonic sensor behind a safety glass.

 

This is a tricky application. Any digital through-beam sensor is going to have a threshold that will trigger a state change and with all the sensors I have experience with, this is set imprecisely, either by a potentiometer on the receiver, or by a "teach mode" button on the receiver. I don't think either type will perform to your requirements. I think you need something "smarter" than that. Immediately I think of a computer vision camera, but that may be overkill. I think you can probably pull it off with a thru-beam sensor if you use an analog model (ex Banner Analog Omni-Beam) and use the signal to do a more advanced detection than simple threshold detection. You will be looking for certain events; solid fuel passing through the beam should look differently than smoke/flame passing through the beam, when the sensor's analog output is printed out on a strip chart or plotted on an oscilloscope time graph. You would need to monitor this output and characterize these events, and write an algorithm which is effective at detecting the events. Solid fuel passing through I would expect to manifest as brief (but predictable/consistent timespan) total losses of signal with sharp rising/falling edges, while flame/smoke I would expect to look more organic, smooth, drops to low (but not completely gone) signal strength, on a variable timespan. There are at least 3 characteristics there i have identified as potential parameters for the algorithm: signal level, timespan, rising/falling time, but there may be more.

Thank you for your input.
I like this approach, I think I will go ahead with an Infrared Thru-Beam sensor, and then write a custom bit of code to analyze its output, rather than depending solely on the output itself.

 

There is still a whole lot of needed information missing. That includes how the fuel is fed in, what sort of stream of fuel, is it a stoker feed or a person shoveling in coal? Or contraband, or wood chips, or chopped sugar cane?? Each type of fuel would have one kind of sensor working best. The last stationary boilers I had anything to do with were gas fired and the control had to verify adequate pilot flame before switching on the BIG (4 inch) gas valve.
So it mght actually be that a vision system works best. It is a challenge to sense something whan we don't know what it is!

 

There is still a whole lot of needed information missing. That includes how the fuel is fed in, what sort of stream of fuel, is it a stoker feed or a person shoveling in coal? Or contraband, or wood chips, or chopped sugar cane?? Each type of fuel would have one kind of sensor working best. The last stationary boilers I had anything to do with were gas fired and the control had to verify adequate pilot flame before switching on the BIG (4 inch) gas valve.
So it mght actually be that a vision system works best. It is a challenge to sense something whan we don't know what it is!

Thank you for your input.
The fuel is in form of wood pellets, and they are being fed in from bags by an operator.
The sensor would be mounted on the bottom of a vertical drop, where the pellets meet the surface of the conveyor belt.

 

OK, so now we have a lot more information. So the pellets could either be sensed falling into the belt, or traveling along the belt after landing. Or even sensed after they leave the belt and travel into the firebox, since I am guessing that the conveyor belt does not enter the fire box.

 

That's a correct guess.
The idea is that the pellets would be detected after they have landed on the conveyor and are traveling along.

 

OK, then the sensor system needs to detect the pellet motion, not just the presence.
And now I am guessing that at the end of the conveyor the pellets fall down a chute of some kind into the firebox. Or possibly a fire-tube, since most boilers have a tube structure, unless they are rather small.
So now a question about how wide is the conveyor that the pellets are traveling on? What sort of gap will a sensor be looking across? There are all sorts of conveyors, from roller-chain driven sweepers along a steel platform to rubber/textile belts on large drum-rollers. And some conveyors move fast enough to throw the chips a foot or more into the fire-pot.