Lambda sensor question

[Eric Johnson]

Seeing the boiler with the lambda sensor in Sweden raises a few questions in my mind that I hope our resident techies like nofossil and TCaldwell can answer. To wit:

The lambda sensor measures the amount of oxygen in the exhaust. I gather that is a way of measuring combustion efficiency. Is the goal to have the lower O2 levels? If so, how would the sensor regulate that--by changing the secondary air mixture?

You guys seem concerned about the durability of the sensor over time, if located directly in the exhaust stream. Is this a concern related to the heat in the exhaust, or the particulates and other nasties, especially during idle and startup?

If so, would it work to somehow filter out particulates and/or locate the sensor out of the direct line of hot exhaust gas, say by diverting a small, cooled "sample" to the lambda?

 

[steam man]

I'll give my take on O2 sensors since I deal with them on an industrial scale albeit liquid/gas boilers. Typically in the"old" days for a boiler we looked for about 15% excess air to achieve relatively good efficiency by ensuring all the fuel was burned (max CO2-no CO) with minimal stack loss. Keep in mind the biggest heat loss in a boiler is up the stack. Note that condensing boilers try to eliminate this problem. I think it is about 3.5% O2 will equate to about that 15% excess air limit. With an O2 sensor in the stack we drive down the excess air to about 2% which is I believe about 5% excess air. Again we want to minimize CO and stack heat loss.

The maintenance of an O2 sensor system is an issue, I would think more so with a solid fuel device. I don't know much about the Lambda sensor but the ones I encounter don't have a long life. Typically <2 years. They are expensive. They also need some sort of calibration periodically to ensure accuracy. They do take a beating in any stack enviroment. I have built shields around some. Typically an air line is piped in but just for calibration purposes. Some sort of known span gas is used for the calibration on the low end, say .5% O2 with air used at the high end 20%+.

Again, I don't know the Lambda sensor but if it is used strictly in a gasifier it may work. I would think that some comprises would be made for wood burning. The sensor would use some type of control scheme to modulate the air. I would think these upscale systems are going to be pricey.

 

[Nofossil]

I don't know the chemistry of what happens inside an O2 sensor, but even small amounts of some chemicals in the exhaust can kill them pretty quickly. That's why you don't have lead in your gasoline any more.

They are expensive and finicky, but the ability to know how much excess oxygen you have tells you a lot about your combustion. In a gasifier, you want a little bit of excess oxygen. Too little and you're wasting heat from too much unburned combustion byproducts. Too much and you're sending extra heated air up the chimney.

 

[deerefanatic]

Anyone thought of building a circuit around an automotive O2 sensor? They're fairly cheap, even if they are short lived.. But I think they should fair alright, after all, gasoline engines put out some pretty noxious chemicals due to the emissions controls themselves....

 

[Redox]

Anyone thought of building a circuit around an automotive O2 sensor? They're fairly cheap, even if they are short lived.. But I think they should fair alright, after all, gasoline engines put out some pretty noxious chemicals due to the emissions controls themselves....

I don't think there is anything too noxious coming out of a modern car engine any more. The controls have cleaned them up quite well over the last decade or so. Replacement used to be a regular maintenance item, but nowadays I just wait for the computer to tell me that one is bad. Most of mine (12 last count?) seem to be working good and I have only had to replace one in the last few years. It did have about 240k on it, though. I think fuel injection was the breakthrough that made them work.

The contamination issue is obviously the problem and a workaround probably wouldn't be too hard to pull off. I'm thinking some type of remote sampling device that periodically sniffs the stack and makes the adjustment might be the easiest way to do that. You don't need the split second accuracy that an automotive system needs to make. If there was a way to tell the computer when your gasification was happening, you could enable the sampler and fine tune it in "closed loop mode".

Obviously going to have to mandate unleaded firewood, but it's a price we'll all just have to pay...

Chris

 

[TCaldwell]

I was told by a oxygen analizer/trim kit designer that the design of a automotive zirconia based sensor would take some peripherials to operate in a wood or for that mater in any boiler, the sensor works by combusting on its sensor surface the ammt of oxygen that passes over it in a flue gas stream and outputs a milivolt signal. for the minute combustion to take ,the exhaust gas needs to be atleast 1500deg f . as this is not condusive on a consistent basis to either car or boiler, each industry is willing to accomodate. the cars sensor is located to accept the hottest exhaust stream possible then the computer is able to somehow compensate and scale the sub temps to a signal. in a boiler the sensor has a surrounding heater to make the difference up from the sensed flue temp. if combustion effiency is stated from the monitor it will use the exit flue temp at the back of the appliance for calculation. and yes this manufacturer said sensor head would probabally last 2 years at a cost of $160 and in a high particulate enviornment would benefit from a sensor pre filter.Most manufacturers offer a severe duty sensor, of course it costs more. from my search a dependable unit sensor/monitor,that outputs a scaleable /variable speed drive compatable signal costs $4000/9000. And yes no matter small or large fire this technology will modulate fan to target a pre-set o2 level post combustion with flue temp monitored to real time display combustion effiency. I am not sure how the lambda sensor and associated computer fits in with all this but if you can buy a euro effiency boiler, lambda equipted for under 20k , usd, and that technology is widely there, i would think the component list would be at relatively mainstream?, for alot less money than the above. the engineer said that more than half of the o2 kits installed, ambient air is used to calibrate. also statad that a less efficient way to controll combustion, but was the way all boilers were controlled before first co2 and 02 sensors was simple flue draft... consider that a given draft pressure correlates to ammount of post combustion air volume. at the start of burn there is usally more combustibles than available air so the fire is starved , o2 deprived, each boiler has its way telling you,the garn will puff . the draft pressure will sway from the pos. to neg. in a violent cycle. mid burn when things appear to have the highest combustion rate the draft measured will be at a optimum value and your set point. at the end of burn we require less air to maintain the optimum draft rate setpoint because there are less are less combustibles in the firebox. so maintaining a optimum draft setpoint throughtout the burn crudely correlates to a o2 level in flue gas. by using a differential pressure transmitter with the sample tube in the primary fire box and through test burns concluding the optimal draft ,scaling the optimal output to a variable speed fan in theory should be a improvement over a static air volume controll....what do ya think

 

[steam man]

TCaldwell-I'll have to think about this. The liquid fossil fuel boilers I have always operated measure the differential pressure between the windbox and the firebox. This pressure drop is used to infer airflow to the boiler. One firebox sensing port will not work if say a stack is plugged. Pressure does not necessarily mean there is flow. Simply put the air would change based on load demand or pressure change. In our case we would be reacting to temperature. That makes me think that maybe just using a temperature sensing for load change could be used. Possibly even a relationship between airflow and temperature could be established to define limits to compensate for the amount of fuel available. The one big difference is in a liquid fossil fuel boiler the fuel would be adjusted for the amount of air called for. I suppose there would be times such as start up an dwindling fuel supply that the modulation would have a problem. In a system with an O2 sensor controlling the air, they typically don't respond as well during large load changes and can be turned off and the differentilal pressure is used. I have a wood/coal boiler that I am thinking of installing (yes-I have thought about playing around on the "dark side") that has on/off forced draft that I have thought about trying to modulate. You've made me think that maybe by measuring the temperature of the boiler and the stack some sort of modulation could be achieved. A timer could be used based on a typical burn time for the size of the firebox that would let the boiler revert back to a fixed airflow possibly even shutdown (as some have done on this forum). I have some Moore 353 PLC's I could try this with. One caveat is that each boiler has a specific recommended draft, say 1" max so modulating around that would be interesting. Even operating a damper verses a variable speed fan could be done. I have friend that operates a large biomass boiler. I'll check with him to see how they modulate that type of boiler.

One other thing about the stack O2 sensors-industrial ones use zirconia type and they do have the heaters built in. Based on that once they are turned one they really need to be kept on to stay stable. They can't just be turned on/off on short cycles.

One easy thing anyone could do is to install a video camera on their stack to watch for smoke to add just enough excess air. Let's see how many people on this forum do this one................

 

[Nofossil]

I've thought about tracking combustion temperature - if there's not enough secondary air, then secondary combustion temp will increase as you add more. At some point, it will level off. Perhaps the ideal ratio could be inferred from the slope of the temperature vs. secondary air flow curve....

 

[TCaldwell]

google KBM systems, AFS lambda display system and TM technomatic, all info on lambda monitoring and control