14 replies to this topic

[cat]

I am currently surveying/auditing all the heat exchangers in our plant utilizing our closed loop cooling water system. There is this certain exchanger wherein I can get the inlet and outlet temperatures as well as the flow rates of both streams. I was able to compute the Q, LMTD and U (using known A). I compared it with the information given in the technical data sheet (provided by the EPCM who built the plant). The U I calculated was higher than the specified. What could be the implication of this?

The following are the data: given as T, A where

(T) - technical book data or preliminary design data
(A) - actual data (averaged values)

1-1 shell, tube pass

Water inlet: 33, 47.33
water outlet: 43, 53.46
HC inlet: 97.1, 105.83
HC outlet: 37.3, 48.25
Flowrate (CW) [TPH]: 1320, 1567.89
U: 690.8, 836.35

The deviations from the specified was due to recycling part of the water outflow, which is done to prevent too much "cooling" which could have adverse effect on the reactor (the HC is fed to the reactor).

I know that the difference in the temperatures have an effect on the calculated values. However, there is another heat exchanger (with identical specifications) utilized in another portion in the plant. The details are the following:

Water inlet: 33, 40.75
water outlet: 43, 47.48
HC inlet: 97.1, 103.37
HC outlet: 37.3, 43.17
Flowrate (CW) [TPH]: 1320, 1527.07
U: 690.8, 602.82

Certainly, the second condition show that actual plant conditions are different from design data. However, the calculated U is less than the specified. I wonder what could this imply..Is "U" for an existing heat exchager constant even though fluid conditions differ (but, I think not. though Im very confused now with the results).

Thanks a lot.

[Root]

Hi,

What is your design margin when you calculated "U"?

Have you taken into account the heat transfer coefficient correlation for the calculation of heat loss due to system surroundings?

Re-calculate "U" with heat transfer correlation and then add design margin for fouling factor.

Toor

[SSWBoy]

As Toor says the contractor will often state two U values on the datasheet, Clean and Service. Please be clear which you are quoting in your post.

What service are the two exchangers in? I am unclear what you mean by same specification but another location in the plant, are these exchangers in two different trains?

[cat]

Hi:

[Toor] I did not took into consideration Q lost due to surroundings.

However, I tried simulating via Chemcad (Heat exchanger rating mode), and added the "fouling factor"..I got nearly the same results as I have posted.

[SSWBoy]

There is only 1 given U in the technical data book (Heading given as Overall Heat Transfer Rate). I believe that this is service U

Both exchangers: water on the shell side; HC gas (mainly ethylene and nitrogen) on the tube side.

Yes, the exchangers are in two different trains. Basically, we have two identical Polyethylene plants (same in everything).

[fatimah]

Hi there

for anybody who does not know;
clean - without fouling
service - fouling added

usually if you design a HE with a fouling factor, the U value-service will increase. i believe you use value as in the preliminary data sheet.

try to check the baffle type. single segmental type baffle can a result in higher U-service value than double segmental.

also, please check the internal tube. any installation of twisted tape can make the u value differs.

[cat]

I computed the U based on the actual temperatures I got, and tried it as well using Chemcad. Under the rating case, I inputted the geometry specs of the heat exchanger supplied by the vendor. I compared it with the preliminary U (which I double checked with chemcad as well).

I calculated and compared values since it was noted in the article "Energy Performance Assesment of Heat Exchangers" published by the Bureau of Energy Efficiency (available on the web), that an actual lower value of U (compared to technical specifications) is an indication of fouling. However, nothing was mentioned what could be the indications of high U.

[fatimah]

Hi there

i'm interested with your statement that you use actual temperature for calculation. for designer, the calculation shall be based on worst case and not actual value. The worst case applies on flowrate, temperature and pressure drop.

[cat]

Hi,

Uhm, I am not designing a heat exchanger. The exchangers exist physically already and working. I am currently "auditing" them with regard to their "performance." As I mentioned, I've read that comparing the actual vs design heat transfer coefficient is a way of assessing heat exchangers. For me to compute the actual "U", I used the actual temperature readings.

[katmar]

What all this data means is that the first exchanger is still clean, but the second one is fouled. The purpose of doing these audits is to determine exactly this. Many plants have this monitoring built into the DCS system to give a continuous indication of exchanger performance and condition. You do not want to waste time and money cleaning clean exchangers, so you do the performance audits to determine which are dirty, and clean them when necessary.

Collecting all the data that you have is the correct way to do it, but a quick indication can also be derived from the approach of the outgoing HC temperature to the incoming cold water. The first exchanger has an approach of 0.92 deg C (48.25-47.33) while the second one has an approach of 2.42 deg C (43.17-40.75). Obviously this approach will also be affected by the flow rates, but in a steady running plant it gives a good guide to exchanger condition.

It is not unusual for a clean exchanger to have a significantly higher HTC than shown on the spec sheet as the "Service HTC", because the fouling factors are usually a high proportion of the overall resistance to heat transfer. If this were not so you would be cleaning your exchangers every week.

[cat]

Thanks a lot!

[Zauberberg]

Assuming that the temperature values are given in degC, the cooling water inlet/outlet temperatures seem to be too high. What is the reason for that? Normally one should avoid having CW temperatures exceeding 45 degC (some authors "allow" for 50-55 degC based on experience) due to accelerated fouling and corrosion since the carbonate salts/hardness deposits will start precipitating from water.

For heat exchangers where hot process streams are cooled with water, verify the tube wall temperature. In cases where calcium carbonate may deposit, heat transfer surface temperatures above 60 degC should be avoided. As a rule of thumb, this may occur if the inlet process temperature is above 90-95 degC for light hydrocarbon liquids and 150-200 degC for heavy hydrocarbons (source: Hewitt, Gulley). For calculating tube wall temperature, refer to attached spreadsheet or read about this procedure in Kern's "Process Heat Transfer".

You can easily setup a spreadsheet by pulling process data from DCS/PI, and having your actual U value calculated on a regular basis. This can give you a feeling whether something is happening with exchanger efficiency.

Attached Files

•  Tube-Wall Temperature.xls   278.5KB   150 downloads

Edited by Zauberberg, 13 August 2010 - 05:20 AM.

[cat]

Temperatures are in deg C.

Inlet temperatures were high since part of the outlet water is recycled and mixed with inlet cooling water. This operation is controlled via a temperature control valve, which is cascaded to the reactor temperature. This is done to prevent "over cooling" of the process fluid (hydrocarbon mixture mainly nitrogen and ethylene) and may affect the operation, as there is an optimum temperature that needs to be maintained. In cases wherein runaway reactions were experienced, the control valve fully opens to hasten cooling of the reactor.

[Zauberberg]

A three-way valve could have been placed on the process stream side, i.e. maintaining process temperature set point with reactor feed stream bypassing the cooling water exchanger. However, there might be some other reasons why this configuration was avoided.

Perhaps fouling on the cooling water side is not significant in your case, particularly if the quality of water is high, and always maintained as such. I was surprised to see design temperatures (inlet or outlet) higher than 50 degC. This is the first time I encounter such high values, but there is always a first time for every thing in life.

Best regards,

[S.R.Shah]

Dear All,

If the Overall heat transfer Coefficiient get Higher than design,

More Heat transfer can occur and result in Higher temperature in Cold side Outlet and Lower temperature
on Hot side Outlet stream.

Reasons considering same process parameters maintained as in design:
1. As discussed ;difference between Service and Dirty Overall Co-efficient
2. More excess area is taken by designer

Warm Regards

[Ranga v]

Hi,
Other than soft ware is their a formula to calculate the healthiness of exchangers??

We have a multipass(4pass) air(shell)- water (tube)after cooler which has developed leak in tubes due to corrosion.

Opt cond shell side: Air,In temp:128 c ,out temp: 35 c,flow :9330 nm3 /hr

tube side: water ,in temp: 31 c, out temp: 36 c

total surface area: 38 m2

how to calculate the U?





What is your design margin when you calculated "U"?

Have you taken into account the heat transfer coefficient correlation for the calculation of heat loss due to system surroundings?

Re-calculate "U" with heat transfer correlation and then add design margin for fouling factor.

Toor