Thermal Oil Heater Piping system 2

[Rafleonard]

Hello everyone,

I'm currently designing a thermal oil system.
I have sized up the heater (16MBTU/hr) with 1,200 GPM flow rate.

I'm having issues calculating the pressure loss (due to friction) in my piping system. I'm confused when it comes to how to calculate this.
I know I have to add pressure losses due to friction in pipes, fittings, valves, heat exchangers, etc to get my total pressure loss.

I'm going to try my best to describe our piping system.

I have a 6" header coming from the heater, that pipe goes across our building (width), from that 6" header, I have (4) 3" sub-headers that are going to each production line, each 3" sub-header has (14) 1" service lines, at the end of each 1" line, there is a heating coil inside a tank.

Each 3" sub-header should have a flow of around 300 GPM (1200/4), then that 300 GPM is divided into (14) 1" lines, if all the tanks are heating, I should have an even flow rate (at least close) on each 1" pipe (300/14) of around 21 GPM.

should I only use the 1" lines (pipe + fittings + valves) to calculate my pressure loss?
or, since the 3" line has to fill up 1st. before going into the 1" lines, do I need to add the pressure loss on the 3" pipe as well?

I appreciate any help with this.

Regards.

 

[jari001]

A sketch with the lengths of the segments and the main components would help us out a lot. Is this a loop configuration where oil is being constantly circulated from the main heating tank through the building and back or is it some sort of in-line heating whenever oil is called for?

I would look at max flow, the expected normal flow, and the minimum flow cases. You need to calculate losses at all diameters from whatever you count as the start of the distribution. Do you know the pressure of the oil after the heater element and that's the available driving force you have to work with?

 

[danschwind]

or, since the 3" line has to fill up 1st. before going into the 1" lines, do I need to add the pressure loss on the 3" pipe as well?

"Fill up"?


Post a sketch please.

Daniel
Rio de Janeiro - Brazil

 

[LittleInch]

Is the return line the same set up but in reverse?

You need to look at the entire system including 6",3 and 1"lines, feed and return.

With that many off takes you will also need some sort of flow control on each one or you will get wildly varying flows.

The worst case is the one with the longest total run.

Getting equal flow along branched systems without some sort of active or variable flow control is very difficult


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[EdStainless]

Start by drawing the system to scale.
Lable with min and max flows.
For a first pass I would ignore the 6", it will be a very small portion of the pressure drop.
You will likely need a restriction where each heater flows back into the return header.
This will be only way to keep flows roughly balanced.
You normally have to tune these in service.
Start with everything wide open and then gradually close the valves slightly on the ones with higher flow until the variations are acceptable.
Then check everything with various combinations.
You may need a VFD on the pump, depending on how much of this system can be shut off at a time.
You may actually need multiple pumps.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[georgeverghese]

Check that all hot oil service and return piping, including hot oil heating coils, have no high point pockets where air may get trapped.
What is the minimum heating demand - is it one out of 14 tanks? If so, that would be beyond the turndown capability of a single burner main hot oil heater, which is about 1:5. So below 20% (or 3 tanks, each running on on-off temp control, say), you'd need a smaller auxiliary burner.

 

[Latexman]

I recommend you buy Crane Technical Paper No. 410 and study it.

Good Luck,
Latexman

 

[Rafleonard]

Thank you all for taking the time to answer.
I will answer everyone on this post.


@Jari001:
I have done a quick P&ID

It doesn't have lengths or all the components but I hope that makes it clearer.
The image shows 4 systems (production lines)

These are the total pipe lengths:

1" pipe (service lines) = 1,700 ft, longest service line is 747'
3" pipe (Sub-header) = 606ft
These numbers are just for 1 line (we have a total of 4)

6" pipe (header) = 300 ft.

This is a closed system, thermal oil (therminol 66) will be recirculating.

I want to have 20 GPM on each service line (1"), I'm planning on installing flow meters and glove valves to try to balance flow on each line.
I know the pressure drop across each coil (heating element).

Do I need to calculate the pressure loss on the main header (6") before it branches, then add the pressure drop on the longest service line for each production line? or do I need to calculate the pressure drop on each service line and add them together?

danschwind:

Daniel, please see the image uploaded above.

LittleInch:
I will pair a flow meter with a globe valve to try to balance the flow across each tank.
I have included an image of our system.

Do I need to account for just the longest run? or do I need to account for the pressure loss on each individual tank?


EdStainless:

Please take a look at the image I have uploaded, it shows our P&ID, it doesn't have all the components but it should give you a better idea of the system.

Im planning on paring a flow meter with a globe valve to try to balance the flow across each tank.
But when calculating pressure drop across the system, do I need to calculate the pressure loss on each individual tank and then add them together or just the longest run? I'm confused with that part.


georgeverghese:
Im adding bleeding valves in our system to avoid having air.

Minimum demand is around 12 MBTU/hr.


Latexman:

Thank you for the advice.

 

[LittleInch]

God, you like making it complicated don't you.

So not only are you having a set control to max 20GPM, you now have 60 individual temperature controllers as well. And then your 3" header runs both ways, not equal from the bigger header...

You can do this two ways.

The complex way whereby you need to segregate your pressure drop in each header between each offtake to account for the assumed equal flow into each branch.

So your 6" header is into 4 parts and your 3" header into 12 parts to get to your worst case location of the tank on the bottom right.

Or you can just oversize the header such that the worst case pressure drop right up to the 1" offtake is within 5% of the best case offtake.

So your 6" header should be probably 10" and your 3" header s at least 6". Then each offtake will be more or less at the same DP.

However you now have apparently quite different lengths of 1" tube? So you will need to size on the worst case 1" tube and throttle back the rest.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Latexman]

You'll need to calculate pressure drop from the discharge nozzle of your pump(s) to the suction nozzle of the same pump(s). Include all pipe, fittings, oil heater, coils in the tanks, etc. Any filters/strainers in case the oil cokes?

I hope all 4 lines are identical, like the P&ID, but I bet they are not.

Good Luck,
Latexman

 

[EdStainless]

I am with LI, you need to upsize the headers so they are not a factor.
I think that you could get away with going to 8" and 4", as long as we are talking light gage piping.
What is the pressure limit on your heat exchangers and other components? Can you do this in sch10 pipe?
A lot less hangers, easier to handle and faster to install, and it costs less.

What possible combinations can these be run? Can they actually use any combination at random?
If so then you need a lot more than just a pump.
It would not be out of reality to see this system with 4x4 MBTU/hr heaters and four pumps.
That should give you an operating range of 5%-100%.

I can't make out the details on the P&ID, but I presume that each HX has some form of temperature control and a shut off valve?
Your trim valves should be on the outlet side of the HX, and your flow meters can be anywhere between the header and the trim valve in the 1".

For pressure drop I would put together some of the most likely high use cases.
Use larger headers and ignore them.
If they can ever run everything at once then work with that.
Lump each small group together. If one bank of machines has 1" runs ranging from 32' to 46' just call it 39' average and work with that.
One concern is what your delta T will be, will you need to account for change in viscosity from the hot to the cold side?

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[Rafleonard]

I want to thank all of you for helping me with this.

LittleInch:

Each production line has 14 tanks, so yes, we are going to have 56 tanks that will need to be heated.
Each tank will have its own temperature transmitter and controller to adjust the flow going inside each coil.

I was planning on changing my pipe size from 6" to 8 and from 3" to 4", to get the velocity recommended by the heater manufacturer.

Ideally, I would've preferred to split my sub-header (4") evenly but I cant.

Yes, I have 4 different pipe lengths (1") before going into the heating coils.

1 - 494" Total of 4
2 - 203" Total of 3
3 - 747" total of 1
4 - 644" Total of 6.

The longest 1" pipe is service line #7 (left to right), the last service line is 203" (bottom left).

wouldn't the worse case be the farthest right tank on the top system? (take a look at the attachment).


Latexman:

If I had software I could run a simulation to find out my pump head, but I don't have it.

Probably they are not, but that's why I'm using a glove valve with a flowmeter to balance it.


EdStainless:

Material for the heat exchanger (coils) is Sch 40 316 SS
We are planning on using Sch 40 on all our piping, the thermal heater has a designed pressure of 225 PSI, I believe once in operation, it wont be running at that pressure, maybe 100 PSI.

If you are talking about our production lines combinations, most of the time all 3 lines will be running, the worst-case scenario will be maybe 2 lines down and 2 running.

You are correct, we have a thermocouple on each tank, that thermocouple is sending a signal to a temperature controller and that is controlling our 3-way valves.

So, regarding the pressure drop, I should get the pressure drop across a single 1" line (longest or average) instead of all 14?

Thanks

 

[Compositepro]

I have never seen anything like this. Hot oil is a poor choice for transferring energy any significant distance. The only reason for using oil as a heat source, rather than direct electric, is for even heating and to avoid any possibility of hot spots degrading heat sensitive materials like plastics. It also allows for the ability for cooling.

Electricity is far better for distributing power through-out a plant. Small, packaged oil heaters are then located next to the process equipment. There certainly are huge oil heaters out there, but they are used for huge single processes.

 

[EdStainless]

The heat exchanger coils should probably be sch5 and all of the piping sch10.
Look at Victaulic PressFit or similar systems.

Yes Comp, I couldn't figure out why they weren't using 56 small electric heaters.
Simple to turn off, simple to control.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[georgeverghese]

Agreed, your 3inch supply and return subheaders are too small - they should be about 8inch to keep dp in these headers small.
Also, why do you have only 3 tanks on one side of the subheader, and the remaining 11 on the other side? It should be 7-8 on one side and 7-8 on the other. Then on each side with 7tanks, the supply and return could be 6inch (8inch reduce to 6inch).
The main header laterals into the 4 subheaders does not look right. If you keep the supply laterals the way they are now ( with the main supply coming in from the bottom of the page), then the return main header should go out of the page from the top, not at the bottom of the page as you have it now.
Similar maldistribution exists on the current 3inch collection lateral...
With this suggested arrangement, the single current hot to cold manual bypass valve location may not be convenient. Suggest a hot a hot to cold bypass for each of the 4nos of supply and return subheaders.

 

[Rafleonard]

Compositepro:

You hit it right on the head, that's the main reason we cannot use an electric heater, the liquid being heated is too sensitive, we have used electric heaters in the past and it will burn the product.
Certainly, not all tanks have this issue and we could have a combination of electric and hot oil, but most of the tanks are heat sensitive.

Plus, the price for mmBTU/hr is around $9, which will cost us around $160/hr and the price for kWh is around $0.09/kWh, I need close to 4,700 kW.

EdStainless:

Thermail oil heater manufacturers recommend using Sch 40 in all piping systems. Im not really sure the Sch of the coils (those were purchase) but I can find out, they are probably Sch 10, but i cant be sure.


Georgeverghese:

If I make my subheaders 8", then my main header will have to be around 16" (4 - 8" lines will fit into a single 16" line to maintain flow velocity).
Why run the subheader that big?

There are several reasons for this, but the main reason is that I can cross all 4 machines from that location, if I move to the center, then I do not have enough headroom to run my main header across all 4 machines.

The main header has to return to the thermal heater (loop) so I can re-heat the oil again, Im not sure what do you mean when you say that the return main header should go out of the page from the top.

I can eliminate the three-way valve and install a two-way valve (open and close), if I do that, then I might need to install a back-pressure valve on each sub-header.
Would you recommend that option? if so, why?

 

[georgeverghese]

Okay, you can still make things work if you must have 3tanks on one side and the rest on the other provided lines are sized to enable similar dp per unit length. Here is a rough sketch and indicative line sizes.
A single FIC-FCV on the main supply header will be better as an alternative to the 56nos of 3way valves. Since you say heat turndown is 2:1, this is well within the turndown capability of a standard trim min flow control valve. Backpressure control is less preferred to minimum flow control.
If product is heat sensitive, then thermal design of heating coil should confirm the tube o.d. film temp at the hot end of the coil is less than thermal degradation temp of the product at the max normal hot oil supply temp.

 

[Compositepro]

There are a number of ways to do anything. This could, theoretically work. But, to me, this looks like it will become a nightmare to operate and maintain. Steam is a far more common option, but is becoming less popular, due to having to employ licensed operators for large boilers. Oil heating has its disadvantages as well as advantages. It is messy, smells, and every btu of heat supplied by oil results in a temperature drop in the oil. Thus you cannot supply a large heat load and maintain a uniform temperature.

 

[georgeverghese]

Ways to reduce heating coil tube od surface temp to be less than thermal degradation temp of the product:
a) Use externally finned tubing for the heating coil, provided the product cannot corrode the fins and coil tubing, and product is not dirt laden.
b) Reduce hot oil supply temp and increase the coil surface area

With 56tanks in operation, it seems clear that corrosion resistant coils would be preferred, else it may be a nightmare to find which tank is leaking hot oil. If product must in no way be contaminated with hot oil through coil leaks, then corrosion resistant coils would be a must.

 

[Rafleonard]

georgeverghese:

There are a couple of question regarding your sketch:

1. Why 6" sub-headers and not 4"? I know you are trying to keep a similar dP per unit length, but using a 6" will increase the cost, and dp on a 4" line is not that terrible.
2. Why is a backpressure control less preferred?
3. Why is the main header return double? the main return is looping back, why not just a single pipe coming back?
4. I understand the concept of the FCV, but why is the FT on the supply side? wouldn't you want to check the flow at the return to control the flow valve? I know you need to know how much you are supplying, but shouldn't you need another FT on the return side to make sure both supply and return are equal?

Corrosion-resistant coils are being used.

Thank you for your help on this.