Measuring flue particulates and other parameters
Feb 11, 2019 20:08:45 GMT -8
Vortex and serg247 like this
Post by hallinen on Feb 11, 2019 20:08:45 GMT -8
I've been digging in the archives, reading articles and roaming the web looking for ideas on how to measure a rocket's particulate matter output. I learned about a Testo 380, which looks expensive and reliable, but may not answer some of the questions we should have, like what are the distribution of the various particulates, in particular the 0.3 micron, 0.5 micron and 1 micron ranges. These tiny, tiny particulates are bad news for our bodies. I think the Testo 380, which is used by German chimney sweeps, measures the PM 2.5 and PM 10, but I'm not sure if they are totaled together in their calculation or if they are separated out. I spoke with a salesperson from Testo, he didn't know any of the details about the device. I'll let you all know when I hear back from him. The Testo 380 looks super useful though, it measures oxygen, carbon monoxide, pressure, temperature, and calculates carbon dioxide. I bet it runs around $3K.
Another way to measure PM is via a laser particle counter. The sensors that do this trick are cheap, and people have been making air pollution monitors for less than $50. I have access to a commercial version of an air pollution monitor that uses the same sensors that you get buy for $24 and hook up to an Arduino if you want to make your own. The commercial version that I have access to is called PurpleAir, and runs between $179-$259. It isn't as useful as a homemade one, but it is helping me define my quest.
Some questions I'm working to answer include...
1. Can a PM be measured using syringeful of flue gas diluted into a container of room air? The European standard, for 2020, is 40 grams/cubic meter, (M3), for a cooker and closed front heater (https://www.unece.org/fileadmin/DAM/env/documents/2018/Air/WGSR/Roald_Wolters.pdf) Peterburg reported that a batch box stove hit 20 mg/m3. The limits on the PMS5003 laser particle counters, which are used in PurpleAir and are available for ~$24 U.S., are 1000 micrograms/M3 for the 2.5 micron particulates, and has a temperature limit of 60 C. This means we can't just stick the device in the flue gas, the gas has to cool and we have to dilute it so the result is in the counter's range. I did a little test today using smoking wooden matches in a sealed mug. I extracted various volumes of fumes using a 20 ml plastic syringe. I then injected it into a sealed glass jar with the purple air inside that had an approximate volume of 1.6 liter. You can look at these pretty meaningless results at purpleair.com, click on maps and do a search for Hartland, Michigan. My eventual goal is to figure out a calibration technique that a solid fuel appliance owner could use to test real world conditions of their appliance. The sensors can report in microgram/m3 or a raw particle count.
2. Will a glass/metal syringe be significantly better than cheap plastic syringes and regular metal needles for the flue gas sampling? Would a longer needle that can suck from the middle of the flue pipe be better than a cheap 1.5 inch needle? Would pumping the syringe a couple of times be useful or is the dead space in the needle not significant.
3. Does the PM count drop over time when it sits in a syringe? Could you collect a sample every 5 minutes during the burn and batch them up or would you have to run them right after collection? What is the best way to evacuate the dilution container between measurements. How long does it take for the sensor to pump the particulates around and settle into a "steady state"?
4. What is the decay in PM count after dilution? Does the internal surface of the dilution container make a difference in count, such as with rough versus smooth walls, perlite on the floor? The device has a little fan so the gasses are moving around in the container, would this accurately reflect what is happening in a bell? Other thoughts on bell design that could be tested?
From what I've been reading, making the sensor/arduino devices isn't super hard, publiclab.org/notes/cfastie/11-28-2018/nano-particle-monitoring. The parts can be ordered off ebay, but they are shipped from Asia so they take awhile to show up. The devices can be made with wifi so the data can be uploaded for free to thingspeak.com/, or the date can be stored on a SD card which can be then inserted into a computer. I think you can hook the devices up to a computer to get real time data, but I'm not sure how that would work yet. I don't know what the cheapest/easiest/best way to do it yet. The PurpleAir website was slow to refresh and was a little frustrating today. It did eventually get the date uploaded, but I can't figure out if there is a way to easily save the output. With the homemade devices, you can plug in a LCD display device so you can get real time PM 2.5 numbers in addition to what gets sent to the web and/or sd card.
While researching this project, I came across really cheap CO sensors, as in less than $10. The range is 10-1000 ppm CO with a max operating temperature of 85 C. For folks who can't justify spending big bucks on a Testo, maybe these will be helpful. Maybe not. I noticed that in Peterburg's graphs, a good run has a CO sitting less than 1000 ppm, and the Testo tops out at 5000 PPM. As a bonus I guess you can configure the sensor to read the amount of alcohol in your breath. I don't know if these CO sensors would be reliable enough to get useful information from or not. To get a picture of how a run went, you would probably have to do a 1:3 dilution with room air with repeated measurements of every couple minutes. I plan on ordering a few of these to test. I think the humidity matters in CO calculations, fortunately humidity and temperature sensors are cheap.
Oxygen sensors that can measure less than 25% O2 run about $90, at least that is what I've seen so far. I have a feeling all of these sensors have a finite lifetime, the oxygen one is said to last 3 years, the CO less than 2 years. I'm probably not going to be testing the Oxygen sensor anytime soon. I'll try to dig around on the web to see if anyone has been using them in a similar fashion. I don't know if the particulates would bother the sensor. They have CO2 sensors that are around $20, I've not read much about them yet. Would measuring the CO2 add any useful information if you do not have the oxygen level? I have a barely basic understanding of flue gas analysis.
As far as how to interpret the PM data, there are a couple of considerations. If the goal is to maximize the configuration of the bells, adjust primary or secondary draft, pop on an electrostatic precipitator, insert a filter, or use extra insulation in the firebox, a device that gives reproducible numbers would be good enough. Ideally, the data would be compared with a Testo 380 or even a fancy government type tests. PurpleAir is undergoing testing at the University of Utah, I've not yet read the results. I might have some connections at the University of Michigan air pollution group, I will explore these when I get a protocol figured out.
MathewWalker's comments about his experience at the wood stove decathlon in Washington made me wonder what the best way is to compare amazing rockets with what other wood burning stoves do. Comparing apples to oranges by measuring PM at discrete points in time or by weighing the total amount of wood and expressing emissions as mg/MJ by extrapolating the total PM over the entire burn. I think if people want to compare stoves then the amount, moisture content and type of wood would be important.
Speaking of electrostatic precipitators, there is one that runs about 2K available in the U.S., the OekoTube, which is getting tested over at Aprovecho. Sounds like it works. Another option is a filter, which seems really complicated. I spent too much time trying to figure out if they would work in a bell. They sell Merv 13 that are rated to 500 F, (260C), but I suspect they would require a draft assist device and if your bell hits 500 F, you'd have a mess. The merv 13 is suppose to filter out 80% of the 2.5 PM. It would be an interesting experiment in any case. There are higher temperature filters that they use in factories, but I was overwhelmed looking at that possibility.
I decided to post this thread before I've gone too deep to see what you all think. I also am wondering if anyone would be interested in borrowing (or buying if I can make them cheap) a device to see what their personal stove's configuration is putting out. Logically, there is a direct correlation with the efficiency of a burn and the emissions output, but I'm curious about what can be done to make a more perfect stove. As Peterburg frequently mentions, you need numbers! (I paraphrased that, sorry)
If there are a couple people with testable stoves in the USA or Canada, I'd be happy to mail a sensor device in about 3 months (assuming China sends me the stuff by then). For folks who do not live in Canada or US, it would be cheaper to make one than mail it. When I get the parts I'll make a youtube video and detailed instructions on the how to make and use them. I'm still sourcing the components, I'll probably end up trying a couple of different configurations. I think the particulates will be doable, no promises on the CO.
Another way to measure PM is via a laser particle counter. The sensors that do this trick are cheap, and people have been making air pollution monitors for less than $50. I have access to a commercial version of an air pollution monitor that uses the same sensors that you get buy for $24 and hook up to an Arduino if you want to make your own. The commercial version that I have access to is called PurpleAir, and runs between $179-$259. It isn't as useful as a homemade one, but it is helping me define my quest.
Some questions I'm working to answer include...
1. Can a PM be measured using syringeful of flue gas diluted into a container of room air? The European standard, for 2020, is 40 grams/cubic meter, (M3), for a cooker and closed front heater (https://www.unece.org/fileadmin/DAM/env/documents/2018/Air/WGSR/Roald_Wolters.pdf) Peterburg reported that a batch box stove hit 20 mg/m3. The limits on the PMS5003 laser particle counters, which are used in PurpleAir and are available for ~$24 U.S., are 1000 micrograms/M3 for the 2.5 micron particulates, and has a temperature limit of 60 C. This means we can't just stick the device in the flue gas, the gas has to cool and we have to dilute it so the result is in the counter's range. I did a little test today using smoking wooden matches in a sealed mug. I extracted various volumes of fumes using a 20 ml plastic syringe. I then injected it into a sealed glass jar with the purple air inside that had an approximate volume of 1.6 liter. You can look at these pretty meaningless results at purpleair.com, click on maps and do a search for Hartland, Michigan. My eventual goal is to figure out a calibration technique that a solid fuel appliance owner could use to test real world conditions of their appliance. The sensors can report in microgram/m3 or a raw particle count.
2. Will a glass/metal syringe be significantly better than cheap plastic syringes and regular metal needles for the flue gas sampling? Would a longer needle that can suck from the middle of the flue pipe be better than a cheap 1.5 inch needle? Would pumping the syringe a couple of times be useful or is the dead space in the needle not significant.
3. Does the PM count drop over time when it sits in a syringe? Could you collect a sample every 5 minutes during the burn and batch them up or would you have to run them right after collection? What is the best way to evacuate the dilution container between measurements. How long does it take for the sensor to pump the particulates around and settle into a "steady state"?
4. What is the decay in PM count after dilution? Does the internal surface of the dilution container make a difference in count, such as with rough versus smooth walls, perlite on the floor? The device has a little fan so the gasses are moving around in the container, would this accurately reflect what is happening in a bell? Other thoughts on bell design that could be tested?
From what I've been reading, making the sensor/arduino devices isn't super hard, publiclab.org/notes/cfastie/11-28-2018/nano-particle-monitoring. The parts can be ordered off ebay, but they are shipped from Asia so they take awhile to show up. The devices can be made with wifi so the data can be uploaded for free to thingspeak.com/, or the date can be stored on a SD card which can be then inserted into a computer. I think you can hook the devices up to a computer to get real time data, but I'm not sure how that would work yet. I don't know what the cheapest/easiest/best way to do it yet. The PurpleAir website was slow to refresh and was a little frustrating today. It did eventually get the date uploaded, but I can't figure out if there is a way to easily save the output. With the homemade devices, you can plug in a LCD display device so you can get real time PM 2.5 numbers in addition to what gets sent to the web and/or sd card.
While researching this project, I came across really cheap CO sensors, as in less than $10. The range is 10-1000 ppm CO with a max operating temperature of 85 C. For folks who can't justify spending big bucks on a Testo, maybe these will be helpful. Maybe not. I noticed that in Peterburg's graphs, a good run has a CO sitting less than 1000 ppm, and the Testo tops out at 5000 PPM. As a bonus I guess you can configure the sensor to read the amount of alcohol in your breath. I don't know if these CO sensors would be reliable enough to get useful information from or not. To get a picture of how a run went, you would probably have to do a 1:3 dilution with room air with repeated measurements of every couple minutes. I plan on ordering a few of these to test. I think the humidity matters in CO calculations, fortunately humidity and temperature sensors are cheap.
Oxygen sensors that can measure less than 25% O2 run about $90, at least that is what I've seen so far. I have a feeling all of these sensors have a finite lifetime, the oxygen one is said to last 3 years, the CO less than 2 years. I'm probably not going to be testing the Oxygen sensor anytime soon. I'll try to dig around on the web to see if anyone has been using them in a similar fashion. I don't know if the particulates would bother the sensor. They have CO2 sensors that are around $20, I've not read much about them yet. Would measuring the CO2 add any useful information if you do not have the oxygen level? I have a barely basic understanding of flue gas analysis.
As far as how to interpret the PM data, there are a couple of considerations. If the goal is to maximize the configuration of the bells, adjust primary or secondary draft, pop on an electrostatic precipitator, insert a filter, or use extra insulation in the firebox, a device that gives reproducible numbers would be good enough. Ideally, the data would be compared with a Testo 380 or even a fancy government type tests. PurpleAir is undergoing testing at the University of Utah, I've not yet read the results. I might have some connections at the University of Michigan air pollution group, I will explore these when I get a protocol figured out.
MathewWalker's comments about his experience at the wood stove decathlon in Washington made me wonder what the best way is to compare amazing rockets with what other wood burning stoves do. Comparing apples to oranges by measuring PM at discrete points in time or by weighing the total amount of wood and expressing emissions as mg/MJ by extrapolating the total PM over the entire burn. I think if people want to compare stoves then the amount, moisture content and type of wood would be important.
Speaking of electrostatic precipitators, there is one that runs about 2K available in the U.S., the OekoTube, which is getting tested over at Aprovecho. Sounds like it works. Another option is a filter, which seems really complicated. I spent too much time trying to figure out if they would work in a bell. They sell Merv 13 that are rated to 500 F, (260C), but I suspect they would require a draft assist device and if your bell hits 500 F, you'd have a mess. The merv 13 is suppose to filter out 80% of the 2.5 PM. It would be an interesting experiment in any case. There are higher temperature filters that they use in factories, but I was overwhelmed looking at that possibility.
I decided to post this thread before I've gone too deep to see what you all think. I also am wondering if anyone would be interested in borrowing (or buying if I can make them cheap) a device to see what their personal stove's configuration is putting out. Logically, there is a direct correlation with the efficiency of a burn and the emissions output, but I'm curious about what can be done to make a more perfect stove. As Peterburg frequently mentions, you need numbers! (I paraphrased that, sorry)
If there are a couple people with testable stoves in the USA or Canada, I'd be happy to mail a sensor device in about 3 months (assuming China sends me the stuff by then). For folks who do not live in Canada or US, it would be cheaper to make one than mail it. When I get the parts I'll make a youtube video and detailed instructions on the how to make and use them. I'm still sourcing the components, I'll probably end up trying a couple of different configurations. I think the particulates will be doable, no promises on the CO.