Appropriate pump for ORC coupled with AC system 1

[akr768]

We have a small-scale ORC system (less than 3 kWe)that is powering the AC loop where the working fluid in both loops are R134a. The ORC is powered by electrical water heater. Upon testing the system, our hypothesis for very low thermal output (or system efficiency about 2%) is low mass flow rate (about 0.02 kg/s or 160 lb/hr) in the ORC loop. We are currently using a magnetic drive rotary vane pump (Make: Fluid-O-Tech, Model: THOT1001A) driven by a 1/13 HP motor at 975 RPM. Since the pump characteristics were given at 1450 or 1725 RPM, we decided to change the motor first. In order to increase the flow rate, we are changing the motor to a 3/4 HP (1725 RPM) and planning to run it at lower speed for the desired flow rate (about 400 lb/hr). My question is: Is changing the motor capacity and controlling RPM the right way to increase the flow rate? or is there another approach to it?

Please let me know what you think.

Thanks.
Arun
P.S.: Here system efficiency is the ratio of cooling output over heat input from hot water. The Carnot efficiency for this system should be about 13%.

 

[georgeverghese]

Increasing rpm and changing the motor HP would be one way, provided it doesnt result in a built up backpressure at the pump discharge which exceeds say 90% of process design pressure - this design pressure may most likely be the same as the setting of the PSV on the pump discharge. Presume your rational for increasing mass rate is to enable higher dp across the temperature / expansion valve? Can we see a process flow diagram of this 2 loop system ?
All applications I've come across for ORC use some kind of waste heat that would otherwise be be dumped to the environment.

 

[akr768]

Thanks for your response. I was afraid of the backpressure too due to increased mass flow. Yes, primary motive is to have higher deltaP and thereby higher shaft work output from the cycle. I have attached a process flow diagram for our system. We ran the system once after changing the motor, but regardless of RPM change (10% to max), deltaP across the pump was about 5 psi. Our theory was that R134a low viscosity was the cause. We tested the same motor and pump combination with water with bucket of water, and it was working fine. Any suggestions why this could be?

 

[georgeverghese]

Intriguing concept.
Apart from viscosity , can only suspect NPSHa for R134a. Probably needs several feet of elevation delta between condensor / accumulator and pump suction to avoid gas from filling up the pumping chamber. If this is the case, then is there some way you could subcool the R134a before the pump? Some refrigeration cycles have built in subcoolers to do this.

 

[akr768]

Good point. We have a flowmeter with sight glass before the pump and that indicates liquid. Also, analyzing test data on the T-s diagram of R134a is suggesting that its completely saturated liquid at the pump suction. We are running the system again with a different pump (but 3/4 hp motor) that was known to work based on our low viscosity theory.

 

[georgeverghese]

It must be boiling up at the entrance to the pumping chamber close to the rotating members. From memory, you need in excess of 10ft of NPSH at the pump suction for pd pumps (gear,vane etc) in volatile fluid applications.

 

[akr768]

Okay. Thank you. I will check with the manufacturer for this information. Appreciate your inputs.

 

[georgeverghese]

Strange, the pump datasheets for this Flo Tech dont have NPSHa curves. However, if I was to take a stab at this, the operating manual says not to operate at above 80degC in the case of hot water. If this limitation is NPSHa related, from this I gather NPSHa is about 5m for water. For pumping R134a, referring to fig 10-25 in Perry 7th edn and the related footnotes( and assuming R134a approximates propane in this regard),we get NPSHa to be 2.5m = 8ft of R134a.

Beware vendors often quote NPSHa for water, even when the service fluid is not water. Fig 10-25 in Perry is abstracted from the Hydraulic Institute standards. To be on the safe side, see if you can find a recent version of Fig 10-25 from the Hyd Inst. which should call out the deduct for R134a.

 

[akr768]

I calculated NPSHa to be about 18 ft compared to NPSHr from Perry's about 8ft. I used the corrections from Propane as well. This suggests that we are safe in terms of NPSH. What do you think?
We are contacting the pump manufacturer for clarification too.

 

[georgeverghese]

Maybe wait to hear from Flotech; if they dont have NPSHr specifically for R134a, then see if you can find this Hyd Inst graph with R134a correction. Note that NPSHa is net head available at pump suction, so all frictional losses in piping, exit nozzles from condensor,strainer etc should also be deducted from static head differential.
If I were to look closer at how this Hyd. Inst graph is constructed in Perry, it does look like you dont really need a dashed line trend for R134a. All that is required is vapor pressure and operating temp, and extrapolate to get the deduct. So all we will then is NPSHr data from Flotech, presumably referencing water. Then get the deduct from fig 10-25 in Perry. Given this is volatile liquid service, I would add a safety margin of 1m to the final NPSHr.

 

[akr768]

We ran the system with the new motor (higher capacity) and same vane pump, system was working fine without any issues. So, we are ruling out any pump issues.