HX claning schedules 2

[billbill20002000]

Hello Group,

I am trying to set up a spread sheet to help us schedule cleaning of our HX's. I found an article on the CEP magazine in which a technique using EUAW (equal Uniform Annual Worth) is used. After Google for 2 days I haven't been able to find how to do this. The article describes using a first order model to predict the change in the U of the exchanger byt the following equation:

U=Uo*e^-kt
Where k is the rate of fouling calculated from operation data
t is the time (yr o month)
Uo=U when the exchanger is clean

Now once I have this data I will create table in which the actual value of U from year 0 (when clean) to year 1,2, etc. I will tabulate the value of U, the diference between Uo and U for each year. Now Here is the part where the article stops at describing the calulation. In the table you have an incremental hot and cold utility cost.

The data for the example is:

MAR (minimum aceptable Return)=20 DCF (any idea what this means?)
Tax rate 34%
Hot utility cost $5/MMBtu
cold Utility $0.5/MMBtu
Operation time 8760 stream hr/yr (100% factor)

now in the table you get:

year Process duty Trim Duty Incremntal Incremental
MMBtu/h MMBtu/h hot Utility Cold Utility
$/yr $yr
0 10 0 0 0
1 9.93 0.07 3269 327
2 9.76 0.24 10415 1042
etc

1. I am stuck on how they are calculating the incremental utility cost
2. Once they have this, the obtained plots of EUAW vs time for different values of k and total cleaning costs, How do you think they went from what you have in the table to EUAW's? all they say is that:

"Using the decline of U over time, one can calculate the duty of the process exchanger from the end of year 0 to the end of years 1, 2, 3, and so on, by solving Eqs. 1–3 (which Are:
Q=Mh*Cph*(T1-T2)
Q=Mc*Cpc*(t1-t2)
Q=A*U*LMTD)

simultaneously. Because energy not recovered in the process exchanger must be made up in the trim exchangers, ncremental increases in utility costs that result for each year of the analysis can also be calculated. Cleaning is a one- time expense, which is assumed to return U to its original (i.e., clean exchanger) value. Figure 2 plots U vs. time for various cleaning schedules.Cleaning changes the project life, or analysis period, to the number of years between cleanings. A cash flow analysis is used
to find incremental increases in operating expenses until the year of cleaning. Cleaning is treated as a year zero cost for a new analysis period. Discounted cash flow is used to calculate a net present value for the cleaning and subsequent incremental operating costs. From the net present value, an EUAW can be calculated based upon the life of the project between cleanings. Using EUAW allows a direct comparison of results between projects with varying life spans. Plotting EUAW vs. cleaning schedule reveals the maximum annual worth or optimum cleaning schedule. "

I havent been able to figure out or reach the author of this, there is a second example without a tax rate and still they get a plot of EUAW but no description on the method. Any comments, references, experiences on scheduling or articles will be appreciated it.

Cheers

William

 

[EdStainless]

Does this model even apply to your case? Do you expect a linear fouling progression of is it seasonal/cyclic?

Either case, the way that I have done them is to set up the cost of wasted energy as a cash flow stream. When the NPV at some future date equals the cost of cleaning you clean at that date.

In some power plants they have to look at the loss in generating capacity vs the spot power price at that time of year. I know of a power plant that cleans twice a year, in mid-July and early August. Loosing a little capacity the rest of the year when power is cheap doesn't offset the cleaning costs.

= = = = = = = = = = = = = = = = = = = =
Corrosion, every where, all the time.
Manage it or it will manage you.

http://www.trent-tube.com/contact/Tech_Assist.cfm

 

[liberoSimulation]

The way I was doing to assist as a process engineer the plant manger is that calculating the U factor based on the duty (Q) and LMTD where the area is constant.
U = Q / (LMTD * A)
The duty is found by the cold stream or the hot stream
Q = m Cp DeltaT whichever has on line readings for flow and temperatures.
Note that three on line temperature readings are usually installed around a heat exchanger as the fourth one can be calculated by:
Qh (duty of hot stream) = Qc (duty of cold stream)
Qh = m,h Cp,h DeltaT,h
Qc = m,c Cp,c DeltaT,c
If flow reading not given on-line, material balance can be done to find out the flow from other on-line flow readings.
Cp is usually assumed constant at the average temperature.

Spreadsheet can be developed as follow:

Time m,h Th,1 Th,2 Qh m,c Tc,1 Tc,2 LMTD U
Plot time vs U and check the trend and the exchanger profile or performance. By doing analysis of the plot with knowing the history of the HE you will see when the HE can be taken offline and cleaned.

I have done this technique and I was most of the time successful in making the decision of the HE cleaning.

Regards

 

[billbill20002000]

Thanks a lot guys,

Quick question for Stainless, can you give a short example for the cash flow? I have tried to understand the way I should apply NPV for my case.

Is it like:

calculate U
Calculate heat duty required with a reduced U due to fouling
Calculate cost for that trim duty
now this is my pickle, if this is the U for the day do you use these values to set an equation that fits the process data and use to predict U in the future and based on this calculate NPV for these values and once they are equal to your cleaning cost schedule cleaning?

Thanks again

Will

PS I am using the first order model to fit my data, for what i see your method is less restrictive rigth?

 

[epoisses]

With all due respect, what if you counted the energy that you are spending yourself when you're working out these calcs? I think you can simplify things considerably without throwing much money out of the window.

For exchangers that have a long cleaning frequency, like once every unit shutdown or less, you just rely on the inspection reports. Was it "very dirty", "dirty", "almost clean" etc - you then clean it when necessary and don't open it between shutdowns.

Exchangers that cost you throughput when they foul, don't spend time calculating but clean them as soon as you notice you're losing performance. Meanwhile you should of course solve the fouling problem.

Any exchangers between those two extremes are worth a closer look IF you can open and clean them on the run in the first place and IF they are actually costing you money, for steam heaters that is if you're not venting steam or letting down steam from one pressure level to the other for example (there are strong seasonal effects in here).
If they cost money, then you may want to assume that the loss of heat transfer is linear, which is more or less true during the first stage of fouling. I guess it's improbable that you can let an exchanger foul well into the exponential part of the curve without it having you cost significant money, it would be largely overdesigned otherwise. Balancing cleaning cost versus operating cost will then be easier (and you can spend the rest of the afternoon on something else).