Waste Heat Recovery System Research

[nickbike]

Hello RE.com!
I've been lurking around this forum for awhile reading and trying to learn from the immense depth of knowledge that exists here.
I’m a graduate student at the Colorado State University Engines and Energy Conversion Lab and we are working on installing a waste heat recovery system to do some research on.
The quick run down of the system is:
Working Fluid: R245fa
Hot Side: 90C Propylene Glycol/Water from internal combustion engine exchanged through a plate HEX
Cold Side: Forced Draft Dry Condenser
Expander: Tesla Turbine

I am starting this thread to keep you updated on the progress of installing and commissioning the system, and in return, I would like to pose some questions that are refrigeration specific.

Feel free to ask any questions about the project. I'll gladly answer.

I've attached some pictures of our major components (condenser, plate HEX, and turbine).

[mad fridgie]

Hi nickbike, this subject is something I have been looking at, (more to do with application than manufacture)
I presume you are looking at trying to convert your thermal energy into electrical energy, what level of conversion are you looking for (normally just over 10% I think.)
for example 500Kw heat converted to 50Kw electrical. Does this conversions include the extra power required to drive the condensing fans.

[nickbike]

We were wanting to hit a 6-10% cycle efficiency. This does include the power required for the pump, but does NOT include the condenser fans.

Our efficiencies will be higher in the winter and lower in the summer due to the difference of the ambient air temperature. In the summer, we will almost certainly have to run the fans on the condenser, which as of right now will negate the power we would be generating. This could be solved by using a natural draft condenser but I'm well aware of the the cost increase for a larger condenser and the need for more refrigerant.

This system is just a starting point for our lab. Once we get this infrastructure in, we can test different types of expanders, refrigerants etc.

[nickbike]

Hi nickbike, this subject is something I have been looking at, (more to do with application than manufacture)
I presume you are looking at trying to convert your thermal energy into electrical energy, what level of conversion are you looking for (normally just over 10% I think.)
for example 500Kw heat converted to 50Kw electrical. Does this conversions include the extra power required to drive the condensing fans.

Also wanted to say that there are some waste heat recovery systems already out there. I would have loved to have bought the vaporizer side of an Electratherm system and piped in our own turbine and condenser, but, alas, they are ridiculously expensive compared to our budget. FYI, they are using R245fa in their system.

[mad fridgie]

It would seem that by using a secondary fluid, you are loosing potential efficiency. Is it possible that you can over feed your refrigerant directly into the engine (I understand pressure limitations on engine components)
Maybe you need to look at at evaporative cooling for condensor. This would reduce fan power and reduce condensing temps.
On your vapouriser (evap in refrig terms) have your measured the actual pressure drop across this heat exchanger (in many case is higher than theory predicts)

[nickbike]

It would seem that by using a secondary fluid, you are loosing potential efficiency. Is it possible that you can over feed your refrigerant directly into the engine (I understand pressure limitations on engine components)


mad fridgie. I like the way you think! The problem here is that you are now running refrigerant through the water jacket of your engine and as you mentioned, your engine would need to handle much higher pressures. Engine coolant systems top out at about 15 PSI whereas our refrigerant is running at 130 PSI. We've actually thought of this and it might be part of some future research on engine water jackets and other working fluids.



Maybe you need to look at at evaporative cooling for condensor. This would reduce fan power and reduce condensing temps.
We looked at evaporative cooling for this project and decided against it because the usage of water for cooling is becoming a very hot topic especially in places like Southern California, Arizona and other dry climates. Using dry cooling for this project actually became one of our main research goals. We want to answer the question, "What can we get out of a dry cooling?"



On your vapouriser (evap in refrig terms) have your measured the actual pressure drop across this heat exchanger (in many case is higher than theory predicts)
No we haven't measure it yet. All we have to go on is what the datasheet says. At least its new, so we won't have an unknown amount of fouling to deal with. I'll be sure to keep that in mind as we start commissioning.

[nickbike]

Here's a simple visio diagram of what we've got going on.

[mad fridgie]

Natural convection, the condensor is going to be as big as a football pitch. It is my understanding that the areas you are talking about the air temp is really high in summer.
One would presume that acommercial item would cost a few dollars, therefore look at you condensing approx in different manor, "ground source" Without knowing details of the areas, I suspect the ground 6 feet down will be somewhat cooler than the air temp. Another way of keeping the diff pressure up.

[SteinarN]

Interesting project.
Have you considered reclaiming heat from the exhaust? You should be able to get significantly higher efficiency from the exhaust due to higher temperature. You do not nesesarily need to have a saturated state, superheating the wapour out of the heat exchanger also increases the efficiency.

[nickbike]

Interesting project.
Have you considered reclaiming heat from the exhaust? You should be able to get significantly higher efficiency from the exhaust due to higher temperature. You do not nesesarily need to have a saturated state, superheating the wapour out of the heat exchanger also increases the efficiency.

We have looked at it and we might go that way for some future research. Right now we are sticking with the jacket water only for simplicity.

There are a few things that you must consider when you are using exhaust heat. Lets go with the old rule of thumb that about 33% of the energy of burnt fuel goes into creating power, 33% into the jacket water and 33% into the exhaust heat.

So there's just as much energy in the jacket water as there is in the exhaust. The jacket water has a much higher heat capacity but at a much lower temperature.

You might ask, "why not use the jacket water for preheat and then the exhaust for superheat." We've looked at it and the numbers just don't work out. You gain very very little from superheating most refrigerants (especially R245fa). And the cost of that extra heat exchanger in the exhaust makes the ROI not worth it.

There ARE ways to utilize both heat streams however and these mostly utilize binary systems with ammonia and water (like a Kalina cycle) and start getting really complicated.

The other problem with using exhaust heat with refrigerants is the char temperature for them. They break down at relatively low temps.

Someday we might get to the point where we can add the complexity of exhaust heat to our system.

[mad fridgie]

My theory, is not as good as it once was, what is the maximum theoretical conversion from heat energy to electrical power (presume 30C condensing)?

[SteinarN]

You are quite right about the decomposition of the refrigerant at only slightly elevated temperatures.
I have not calculated the effect of superheating this particular refrigerant. I know this effect is highly variable dependent on the actual refrigerant used.

Water would probably be an efficient refrigerant in an exhaust waste heat recovery system. At close to full load the exhaust temperature is something like 500 to 600 degres Celsius, probably slightly less after the turbo if one is fitted. With water and those temperatures you could maybe recover 20 percent of the energy in the exhaust, probably closer to 25 or 30 percent in a bit more advanced system like two or three stage turbine with superheater between the stages and high pressures like 40 to 50 bar and a large condenser at the final exhaust lowering the exhaust pressure down to 0.5 bar vacuum.

But as you say, to keep it simple for now......

[nickbike]

I've got a really simple question for the RE forums.

Piping: Of course we're on a tight budget, it's a university research project, is there any other type of budget? I know copper is the way to go. But, can we get away with steel (not stainless, that costs almost as much as copper) with a desiccant filter or some other dryer installed? This system will not be running 24/7 and doesn't need to be as robust as a building HVAC system. My main concern is ruining the refrigerant with contamination from the steel pipe.

A quick back of envelope calculation says copper is about $2,500 and steel is $800.

[mad fridgie]

I've got a really simple question for the RE forums.

Piping: Of course we're on a tight budget, it's a university research project, is there any other type of budget? I know copper is the way to go. But, can we get away with steel (not stainless, that costs almost as much as copper) with a desiccant filter or some other dryer installed? This system will not be running 24/7 and doesn't need to be as robust as a building HVAC system. My main concern is ruining the refrigerant with contamination from the steel pipe.

A quick back of envelope calculation says copper is about $2,500 and steel is $800.

Yes you can use steel, just keep it oxygen + water free

[US Iceman]

I have not read the entire thread yet, but just browsed it quickly. A Tesla turbine? For expanding liquid or pumping liquid as in a Rankine cycle?

[nickbike]

I have not read the entire thread yet, but just browsed it quickly. A Tesla turbine? For expanding liquid or pumping liquid as in a Rankine cycle?

The Tesla turbine will be used to expand 100% quality R245fa refrigerant coming out of a plate heat exchanger to generate electricity.

The pump in the ORC system is a Blackmer vane pump. It's downstream of the condenser so it's pumping liquid refrigerant.

[novoross74]

2 Nickbike

Have you measured already an isentropic efficiency of your Tesla turbine? Based on common experience it can be on quite low level - 30-50% maximum. Not so bad for a small system of waste heat recovery but much less than INTERNET hype.

[nickbike]

2 Nickbike

Have you measured already an isentropic efficiency of your Tesla turbine? Based on common experience it can be on quite low level - 30-50% maximum. Not so bad for a small system of waste heat recovery but much less than INTERNET hype.

The turbine is coming from an engineer that's outside of the project and I haven't done any testing on the turbine myself. In fact, we are working on the paperwork right now that will allow them to lend us the turbine, but not take it apart or share any information publicly about it. It's kind of a black box to us.

With that said, the engineer has the turbine running on steam and getting 60% efficiency from turbine inlet to the generator cables. Not bad considering that takes into account the efficiency of the generator along with the turbine.

However, running the turbine on refrigerant is going to be a whole different set of operating conditions. The engineer has designed the best he can around the properties of the refrigerant. Most likely, he'll have to do some tweaking of the nozzles in the turbine to get it right.

You are correct. The internet has hyped the Tesla turbine a lot, but none of the scientific literature has been successful in producing high efficiencies.

[novoross74]

With that said, the engineer has the turbine running on steam and getting 60% efficiency from turbine inlet to the generator cables. Not bad considering that takes into account the efficiency of the generator along with the turbine.


Hope you have a proper guy to lease turbine from.

If it will be helpful for you - we are performing a similar project about Tesla turbine and discussing it here:

h*ttp//:talks.guns.ru/forummessage/151/609841.h*tml

(discussion is in Russian, but you can use Google Translate to read it, delete asteriks in address).

There we tried to collect all available data about Tesla turbines that is on-line and put all valuable comments all together.

The best experimental data for Tesla turbine was 49% isentropic (Sherstuk, Naumov, 1980), best math model shows 55% isentropic.

Our own results we collect not for scientific purposes, but it shows also about 50% isentropic.

Other options for small ORC systems you can consider as back-up for Tesla turbine are screw compressors and spiral copressors in reverse mode. They are also have 50-60% isentropic efficiency in 1-10 kW size.

[nickbike]

Signs of progress. Condenser being placed on the roof this morning.

[novoross74]

Nickbike,
Look at this report (page 6). There is a comparision of working fluids for ORC machines.
h*ttp://w*ww.scansims.org/sims2005/SIMS2005_77.pdf

For 90C hot source you stipulated above - looks like R245fa is not the best liquid, but worst.

[garethjones10]

Hi

It would seem that by using a secondary fluid, you are loosing potential efficiency. Is it possible that you can over feed your refrigerant directly into the engine (I understand pressure limitations on engine components)
Maybe you need to look at at evaporative cooling for condensor. This would reduce fan power and reduce condensing temps.

Regards,

Gareth
CNM ONLINE

[nickbike]

Nickbike,
Look at this report (page 6). There is a comparision of working fluids for ORC machines.
h*ttp://w*ww.scansims.org/sims2005/SIMS2005_77.pdf

For 90C hot source you stipulated above - looks like R245fa is not the best liquid, but worst.

After going through the paper I think I can see why R245fa is not as efficient at 90C in their graph. It has to do with their particular expander and the model of it. With the Tesla turbine we are assuming the isentropic efficiency is constant at any temperature and pressure. The scroll expander they are modeling starts to lose efficiency as the vapor pressure rises.

I think it's the difference in the equipment they are using that makes their efficiencies different than ours.

Electratherm uses R245fa in their systems with a screw expander.

[nickbike]

Hi

It would seem that by using a secondary fluid, you are loosing potential efficiency. Is it possible that you can over feed your refrigerant directly into the engine (I understand pressure limitations on engine components)
Maybe you need to look at at evaporative cooling for condensor. This would reduce fan power and reduce condensing temps.

Regards,

Gareth
CNM ONLINE

Gareth,
Good thought on using the refrigerant directly in the jacket of the engine. We've thought of this and would eventually want to spend more time researching this possibility. It would certainly increase efficiency and reduce the cost of the system, 1 less heat exchanger. I know that research won't happen on my watch unless I want to be here for a few more years.

The pressure is probably the number 1 concern, the other thing to consider is material compatibility (I know R245fa eats lots of types of rubber). Overfeeding would get you the capacity you needed to pull the heat away from the engine, but I don't know if the engine could actually handle the flow rate that would be required. A specially designed block would have to be made.

As for the evapoartive cooling: It would certainly bump our efficiency, but then you need a water source to operate the system. One goal of the whole project was to NOT need water to operate.

Thanks for the input! Keep the ideas and discussion going.

[garethjones10]

Hi nickbike, this subject is something I have been looking at, (more to do with application than manufacture)
I presume you are looking at trying to convert your thermal energy into electrical energy, what level of conversion are you looking for (normally just over 10% I think.)
for example 500Kw heat converted to 50Kw electrical. Does this conversions include the extra power required to drive the condensing fans.

Regards,

Gareth
CNM ONLINE

[mizun]

Hi,
It is good to meet you here. I need some help. I on my own have designed and manufactured a turbine to run on R245fa.

Honestly, I have no idea whatsoever on refrigerant gas. I have been informed that R245fa need to be heated to 120-140C and a pressure of 4-6 bar.

The heating media we have chosen is solar collectors. Can you help me with few technical solutions?

1. How much R245fa is required as a first fill for a 10KW turbine/

2. what should be the temperature of glycol at inlet of heat exchanger to heat R245fa to 140C.

3. How much glycol should be circulated per minute for heat transfer.

I have embarked on this project after successfully designing and commissioning "VAMAN-The Midget" range of Micro Steam turbines ranging from 1KW to 50KW operatin on 10 bar steam/190-260C.

I can be contacted on mizunorc@ymail.com.

I will very much appreciate your advice on R245fa system for my new project. Thank you in advance

Best wishes,
Mohan Prabhu

[novoross74]

After going through the paper I think I can see why R245fa is not as efficient at 90C in their graph. It has to do with their particular expander and the model of it. With the Tesla turbine we are assuming the isentropic efficiency is constant at any temperature and pressure. The scroll expander they are modeling starts to lose efficiency as the vapor pressure rises.

I think it's the difference in the equipment they are using that makes their efficiencies different than ours.

Electratherm uses R245fa in their systems with a screw expander.

I spoke last year with Electratherm - and I have more questions about their device, that before I started my communication with them.

As far as I know - for low temperature source in Chena hot springs UTC power replace R245fa to R134a, which is similar to above mentioned report.

Look at their works here:
h*ttp://w*ww.chenahotsprings.com/geothermal-power/

Also, as far as I understand - less you overheat the working flued - less problems you have with heat exchangers.
Since T-s curve for fluid is isentropic or dry - overheat is not needed.
Look here:

h*ttp://w*ww.ehakem.com/index.php/IJoT/article/viewArticle/217

PDF is not available now on this site, but I have a copy if you need it.

[nickbike]

Here's some progress pictures. We're waiting for the turbine to be reworked, it has a sealing problem. The rest of the system is getting very close to being ready to collect data.


656165626563

[nick2ono]

Any update on this wonderful project?? keep it up guys!!

[chempro]

wonderful project?? keep it up guys!!

[nickbike]

We got it running under reduced load a few weeks ago. The system is designed to pull 250kW of heat off 93C jacket water from a 1.6MW natural gas engine. Running the engine is expensive, so we've been doing some reduced input heat testing using our building boiler system. Here's a video of the first turns. This is running at about 300 RPM.

http://www.youtube.com/watch?v=8V-m7WE2wLQ

[kayakalp]

Absolutely friend.

[nickbike]

We got the turbine up and running while using the Caterpillar Engine as the heat source. We ran the motor up as fast as it would go to apply load to the turbine. The system was designed around an expected turbine speed of 3600 RPM upon recommendation of the turbine designer, but that turned out to be far too slow. We removed the belt to see what the turbine "wanted" to do unloaded. Here's the video of the turbine running at almost 8,000 RPM unloaded at about 3/4 of it's designed flow rate.

http://www.youtube.com/watch?v=ZA4rdEwaQxA

We are in the process of sourcing higher speed bearings for the output shaft and new pulleys to gear down to a 3:1 ratio so we don't overspeed the motor.

[desA]

That turbine definitely sounds 'happy'. What are the critical speeds for this turbine?

[nickbike]

Critical speeds as in where it runs most efficiently? That's what we're researching and regearing for. Hopefully can get back to you soon on that. Critical as in where it detonates? About 16,000 RPM.

[mad fridgie]

[QUOTE=nickbike;238450]Critical speeds as in where it runs most efficiently? That's what we're researching and regearing for. Hopefully can get back to you soon on that. Critical as in where it detonates? About 16,000 RPM

[desA]

Critical speed/s : rotational speed/s at which resonance occurs.

Please be advised that there are likely to be more than one critical speed. Play carefully if the developer has not already determined this/these. These could occur before the 16,000 rpm value you site - or then again, not.

You could probably simulate the machine rotating portions & determine target critical speed/s. Your vibration prof can assist you here - make a class project of it.

Alternatively, track the run-ups with an accelerometer & plot the vibration magnitude response. Also plot phase angle - when it flips 180' prepare to launch upwards a little faster through the critical speed into safer territory.