Vous êtes sur la page 1sur 14

Heat Exchangers - Course Content

Section 7 - Troubleshooting Problems

Failures and Troubleshooting


Sample Problem

EDS 2004/EXC 7-1

Table of Contents
Problem Solving Shell & Tube Exchangers
Failures and Troubleshooting

Exchanger operation sometimes deviates


from design because of:
Poor design
Fouling (excessive fouling)
Mechanical failure
How do we measure performance?
Calculate U value?
Measure duty?
Measure hot approach (or cold approach)?
Measure P?

EDS 2004/EXC 7-2

Sometimes exchangers do not performs as they should. Sometimes they do not


accomplish what they are capable of and other times they are asked to perform what
they are not capable of.

How do you measure performance. U values and duty are hard if not impossible to
measure and often difficult to calculate. Single phase temperatures are relatively
easy to measure but not two phase temperatures. Accurate pressure can be hard to
measure due to dynamics in the system.
Failures and Troubleshooting
(continued)

Deciding on performance criteria is difficult.


1. Operating conditions often different from design
(flows, compositions, temperatures, etc.)
hs vary with composition, Tlm different
from design
result => Uoperating often Udesign
2. fouling factors. Who knows how much fouling
there is?
fouling difference between operation of
CLEAN and DIRTY exchanger

EDS 2004/EXC 7-3

What criteria that can be measured accurate may not adequately define the true
performance of the exchanger.

Knowing how fouled the surface is on average is difficult. Fouling is often heavier
in some part of the exchanger than others.
Failures and Troubleshooting
(continued)

What can you decide from a single evaluation of a Heat


Exchanger that has been in service for a long time?
You CAN get:
Calculate one U value, fouled
Measure one DP for shellside and tubeside
Calculate one heat duty
You CANNOT get:
Any useful idea of fouling levels
Any useful information for maintenance
Any hints of developing problems

A single rating is no use in exchanger monitoring


EDS 2004/EXC 7-4

A single rating of an exchanger is good for getting a baseline data on its


performance but it must be done on a regular basis to define trends.
Failures and Troubleshooting
(continued)

Heat transfer calculations are accurate to about 10-15%.


For a $10 million per year fuel saved by Platforming
VCFE, if efficiency drops by 10% over 2 years ($ 1
million)
DO YOU WANT TO KNOW ABOUT IT?
Suppose one calculation after 2 years shows Heat
Exchanger under performing by 10%.
The QUESTION: Is it really under performing?
Who will rent a $7500 per day crane for turnaround
based on ONE calculation, accurate to only 10%?

EDS 2004/EXC 7-5

Since most convection correlations and overall correction factors are based on
empirical data, the heat transfer calculations for exchangers are only accurate to
about 10 to 15% when all the necessary data is known. When some data has to be
estimated, the accuracy gets worse.
Failures and Troubleshooting
(continued)

THE KEY TO EXCHANGER PERFORMANCE


MONITORING
TAKE MANY MEASUREMENTS
WATCH TRENDS
THINK ABOUT WHAT YOU SEE
Who has panel instruments accurate to 1%?
Who wants to do single-gauge surveys every week?

EDS 2004/EXC 7-6

MONITOR and check TRENDS rather than rely on a single survey to check on
performance.
Failures and Troubleshooting
(continued)

The $1 000 000 answer is:


1) When any (LARGE or EXPENSIVE) exchanger
is new:
Do single-gauge pressure and temperature survey
Personally supervise sampling of process
stream(s)
Do best possible job of energy and material
balance
Perform detailed rating of exchanger
Keep records on file
EDS 2004/EXC 7-7

Do a detailed performance evaluation of each exchanger when it is new. Repeat


after 6 months or so. Trend data in-between and afterwards.

Keep the data and the calculations.

Use Distributed Control trend logs for data collection.


Failures and Troubleshooting
(continued)

2) Track P and hot or cold approach of


exchanger on DCS trending, or graph paper if
you have to.
Mark major changes in flow rate, feed
temperatures, or feedstock on graph. Adjust
measurements for changes.
Any gradual trends not explained by changes
above mean that something is steadily getting
worse.
If the cause is not clear, get new single-gauge
survey, double check results.

EDS 2004/EXC 7-8

Trend pressure drop and hot or cold end approach temperatures besides the raw
data. Graphical data is easier to pick out trends.
Failures and Troubleshooting
(continued)

Examples:
A downward trend in hot approach indicates
fouling progression
An upward trend in P - fouling
Drastic changes in approach temperatures or
pressures can indicate leaks or blockages

EDS 2004/EXC 7-9


Failures and Troubleshooting
(continued)

PARALLEL EXCHANGERS
For exchangers in parallel, use the above
procedure but also watch trends in the differences
Example. watch trends in:
Hot end approach (A) - Hot end approach (B),
or Th,out(A) - Th,out(B)
If differences change at all, one exchanger has a
problem
Both exchangers should foul at the same rate
Performance of both exchangers should change
equally under all circumstances
EDS 2004/EXC 7-10

Parallel exchangers should also be tracked individually.

Flow distribution to parallel trains is never identical. We have seen splits come out
at 65-35% rather than 50-50%. As a result, the overall performance was not as
predicted.
Failures and Troubleshooting
(continued)

SUMMARY
If exchangers are regularly monitored in this way
You will not get many unpleasant surprises
You will have information for follow-up
investigations

EDS 2004/EXC 7-11

Monitoring will often pick up trends that can avoid failures.


Sample Problem

Reactor (Gas) Effluent Coolers


Calculate the design and operating U values from real
plant and design data:
Design Operating
Q MW 4.76 Q MW 5.27
G kg/hr 70056 G kg/hr 73899
T in C 147 T in C 172
T out C 41 T out C 65
t in C 27.2
t in C 32
t out C 35
t out C 40
Tlm C ? Tlm C ?
F - ? F - ?
UA W/C ? UA W/C ?
2
A m
2
480 A m 480
2
U
2
W/m /C ? U W/m /C ?
U/Udesign ?

EDS 2004/EXC 7-12

Actual exchanger design and operating data.

Compare the operating numbers to design.

Duty is a little higher but so is flow rate. Temperature changes on both sides are
similar (hot side and cold side T).

Calculate the LMTD and F factor for the design case. UA is Q/F/ T. Since A is
known, you get U.

Do the same for the operating case.


Sample Problem
(continued)

Solutions
Design
Q MW 4.76
G kg/hr 70056
T in C 147
T out C 41
t in C 32
t out C 40
Tlm C 39.6
F - 0.845
UA = Q/(FTlm) W/C 142250
2
A m 480
2
U W/m /C 296

EDS 2004/EXC 7-13

Compare the two U values. Operating is about half the design.

Since the hot side inlet was much higher and the cold side inlet was lower, the
LMTD was nearly double. Therefore, if the same performance was available the
duty should have been nearly double. Since the process had not really changed, the
only conclusion is that the performance is down due to fouling.

Opening the exchanger confirmed this was true.


Sample Problem
(continued)
Solutions
Design Operating
Q MW 4.76 Q MW 5.27
G kg/hr 70056 G kg/hr 73899

T in C 147 T in C 172
T out C 65
T out C 41
t in C 27.2
t in C 32
t out C 35
t out C 40
Tlm C 77
Tlm C 39.6
F - 0.97
F - 0.845
UA = Q/(FTlm) W/C 70558
UA = Q/(F Tlm) W/C 142250 A m
2
480
2
A m 480 U W/m /C
2
147
2
U W/m /C 296 U/Udesign 0.49

U is of design value even though shell and tube flows are higher
than design! The problem is severe fouling.
EDS 2004/EXC 7-14

Compare the two U values. Operating is about half the design.

Since the hot side inlet was much higher and the cold side inlet was lower, the
LMTD was nearly double. Therefore, if the same performance was available the
duty should have been nearly double. Since the process had not really changed, the
only conclusion is that the performance is down due to fouling.

Opening the exchanger confirmed this was true.

Vous aimerez peut-être aussi