HX Dirty Overall Coefficient Question 4

[JoeChem]

Greetings,

It has been some time since I have specified a heat exchanger. The last time I did, there were two overall heat transfer coefficents provided on the heat exchanger specification sheet - clean and service coefficients.

Now I see a third coefficient - dirty. I am having some difficulty understanding what this is. I e-mailed the supplier I am working with and I received the following response:

"The clean rate is under a zero fouling or new condition, service is under normal operating conditions. The dirty transfer rate is in the fouled condition. ie if 100% of the fouling resistance used in the design (line 28) were present. The heat exchanger is designed with enough surface to meet the performance specified when operating in the fouled state."

Seemed to make sense until I tried some calculations based on the data the supplier provided:

Target Heat Duty = 14,000,000 BTU/hr
Surface Area/Unit = 876.7
Corrected MTD = 75.87 F

Clean Coefficient = 535.56 BTU/hr ft2 F
Service Coefficent = 210.46
Dirty Coefficient = 175.25

210.46 x 876.7 x 75.87 = 13,998,795 BTU/hr
175.25 x 876.7 x 75.87 = 11,656,794 BTU/hr

Looks to me that the exchanger is designed based on the "service" condition.

When I called the supplier and asked to explain he said the exchanger is oversized to ensure the unit would transfer 14 MM BTU/hr using the dirty coefficent and the indicated surface area. He could not explain to me what exactly was "oversized". Sounded to me like a black box software issue.

Can anyone shed some light on this? Am I missing something?

Thanks in advance,

JoeChem

 

[speco]

JoeChem,

There's obviously a mistake somewhere. Normally the dirty coefficient is higher than the service coefficient. That is, there is enough surface provided to cover the dirty coefficient and then some.

The data you have shown indicates a negative safety factor. Are you sure they are not reversed?

Speco

 

[rmw]

I am not going to crunch your numbers, (I do that for my day job) but the fouling factor is basically there to add surface area so that when the heat transfer surface becomes fouled, there will still be enough left to meet the required duty. Simplistically put it is like dividing your 'clean' result by some value less than 1.0 (say .95 for example) arriving at an area of an area 1.0526 times your required area. The divisor is determined by what you (or the Hx designer) pick as a fouling factor.

Your numbers should reflect that.

In the simple example above, until the unit becomes fouled, it will perform at a rate of 1.0526 duty in theory, simplistically put. And I say simplistically put, because more than one factor is at play here during heat transfer.

rmw

 

[TD2K]

I'm not following what the 'service' coefficient is from your vendor's reply.

I agree the exchanger is designed for the service coefficient simply by back calculating a U from the Q, area and dT.

What fouling factors did you specify?

 

[mauricestoker]

It would be helpful if the HX media was given-is this liquid-to-liquid? If so, a minimum fouling factor would be available if using ARI.

 

[speco]

JoeChem and all,

Let me take another run at this, by defining the terms here.

The Clean Coefficient is the overall heat transfer coefficient calculated based on the fluid properties and configuration of the heat exchanger, but NOT including the specified fouling resistances (fouling factors).

The Dirty Coefficient is the same coefficient, but with the specified fouling resistances added in. In this particular case the fouling resistance is quite significant, since the clean coefficient is three times the dirty coefficient (535 vs. 175) This often happens with shell and tube exchangers in the process industries, since the specified fouling factors are often a bit conservative.

The Service Coefficient is simply the heat load divided by the product of the actual surface area provided and the effective LMTD. This number is used to show some additional safety factor, usually resulting from the manufacturer going to the next larger exchanger within their manufacturing standards.

In the case shown, the Service Coefficient is higher than the Dirty Coefficient. This isn't supposed to happen, but the manufacturer may be saying that the fouling factors are too high, and this design is their recommendation. They should indicate what the effective fouling factors are, and explain why they are providing less surface than would be required with the specified fouling factors.

Regards,

Speco (

http://www.stoneprocess.com

 

[JoeChem]

Thanks for all the feedback.

Speco - I agree with what you are saying. If the information(i.e., fouling factors) I initially provided are unrealistic then the supplier should indicate so and offer their recommendation(s) on their heat exchanger specification sheet. If they are using a "dirty" coefficient to size the unit that is perfectly fine. However, I should be able to use the "dirty" U along with the A and MTD specified on the sheet and be able to calculate the required heat load (also shown on the spec sheet). I am just trying to understand how the unit will be sized. If I cannot extract this directly from the specification sheet provided then the sheet is useless (in my opinion anyway).

The fouling resistance is purposely significant in this case as we have problems with our cooling water at this particular plant. The fouling factors I used were 0.0005 (shell) and 0.002 (Tube). The data spec sheet I received back from the supplier has the same factors listed.

I posted this question here as I could not get the supplier rep to clarify the issue. When push came to shove the only response I got was "this is just the way we design exchangers". I did not get a warm and fuzzy so I am going to look up another supplier.

Thanks again.

JoeChem

 

[TD2K]

Thanks Speco for the explanation, I haven't seen a service coefficient that I recall on a data sheet but your explanation makes sense.

With a 'clean' coefficient of 535 BTU/hrft2F (if that number includes no fouling resistances) that's over a 1000 BTU/hrft2F on each side. If you are condensing steam than that's a reasonable coefficient for the steam side and your fouling factor matches what I've typically seen for steam if I remember correctly. 1000 BTU/hrft2f on the cooling water side to make the clean coefficient work seems high though.

I'd be more disturbed by their inability to explain the results. I wonder if someone is just using a black box to size the exchanger and isn't reviewing the results or doesn't know how to review the results. Or the rep doesn't want to go back to the design group for some reason.

 

[mauricestoker]

As it is liquid-liquid, is it ARI 400 certified?

 

[TD2K]

I'm going to back peddle on my comment about the CW side seeming high. I looked in a Wolverine catalogue I have and for 3/4" 16 BWG tubes, 5 feet/sec cooling water you can get inside coefficient over 1000.

Is this a competitive bid? Could they be trying to make their exchanger look better by providing less area than would seem to be required with your fouling coefficients?

Do you have any other similar exchangers and if so, what sort of Uo do you see on them? Does this exchanger provide enough ft2 area?

You said you've had problems with cooling water at this plant, what sort of tube velocities are being provided? What sort of outlet temperatures and inlet process temperatures do you have? I've seen clients limit both the outlet cooling water temperature and the inlet process temperature to cooling water exchangers to avoid too high of cooling water film temperatures resulting in their experience scaling and fouling.

 

[JoeChem]

The service in the case is water - water. The shell side is high pressure tempered water - closed loop pumped system (very clean)for cooling a batch reactor. The tube side (3/4" OD tubes)is cooling tower water.

The fluid velocity on the tube side is 4.9 ft/sec. Before I submitted the design to the vendor I rated it using aspen TASC software:

Clean U = 483 BTU/hrft2 F
Service U = 202
Dirty U = 195

Like I said earlier this is the first time I have seen three overall coefficients on a data sheet.

The design we are speaking of limits cooling tower water return to < 120 F in the summer. The make up water at this plant is terrible and they do not take care of it very well at all. Fortunately this is recognized by management and a more rigorous treatment plan is being put in place.

There are several other units drawing cooling water via supply/return headers from this system. The condition of all jackets are very poor in term of flow and pressure drop. The exchanger of interest is just one piece of a much larger upgrade project.

 

[tickle]

To digress slightly, there is an interesting article from the Chemical Engineer that discusses fouling. In summary, heat exchangers are generally over designed by using worst case conditions, conservatism in the design and added factors by engineer and designer eg +10% of area. This can result in a HE that quite easily have a surface area of 200% required.

The excess surface area the HE effectively fouls itself to the design point.

 

[jcazenave]

Hi

the result means that the exchanger doesnt have enough surface for the specified fouling.

The service results are based on using the "required" surface with the fouled condition.

The dirty is using the real surface of the exchanger with the fouled condition

the clean is using the real surface with NO fouling.

Normally, the service coefficient is the lowest because the real surface is bigger than the required area (oversurface)

In you case the real surface is not enough by 17% based on the result you gave for the duties.

I hope it helps

 

[AndersE]

This is the way I am used to interpret these values:

Clean coefficient: The calculated heat transfer coefficient without fouling resistances.

Dirty coefficient: Heat transfer coefficient including fouling resistances.

Service coefficient: Dirty coefficient divided by a safety factor. This safety factor is usually specified as required oversurfacing. If the required oversurfacing is 10 %, the safety factor is 1.1.

For example:

Assume that the clean coefficient is 1 000 W/m2K and that the fouling resistances correspond to 100 % extra required surface when compared to to a clean heat exchanger. The required oversurfacing is 10 %.

Clean coefficient: 1 000 W/m2K
Dirty coefficient: 500 W/m2K
Service coefficient: 454.5 W/m2K

 

[aftamath77]

I know this question may seem quite remedial, but I was curious if the amount of excess surface area designed in to the exchanger can be calculated using these transfer rates, and if so, what is the calculation? Much appreciated.

 

[speco]

Aftamath77,

The excess surface is very easy to calculate. If you divide the product of the surface, LMTD, and dirty coefficient by the actual heat load, you should normally get a number greater than 1.0. Something like, maybe 1.067,indicating a 6.7% excess surface.

If a service coefficient is given this would be the same as the relationship between the dirty coefficient and the service coefficient.

Regards,

Speco (

http://www.stoneprocess.com)

 

[Christine74]

The percent excess surface area is typically something like 2-3 times greater than the percent excess heat transfer (more surface area means a closer temperature approach, which reduces the LMTD and the amount of heat transfer per unit of additional surface area).

-Christine