Remaining life evaluation of coke drum using strain gauge

[Anonymous]

We have heard that some coker plant installed strain gauges to monitor the stress condition of their coke drums.
 
We would like to know if it really can predict the life time of a coke drum.
 
How many coker plants in USA adopt this technology in their coke drums?
 
 
 

[Anonymous]

I also want to know it.
can anybody give us some details?

[Richard BoswellParticipant]

Thanks for asking this great question. We are asked this a lot. Our company has installed these gages and intrinsically safe systems on more than 50 coke drums and have recorded and viewed over 5,000 cycles of operation. The short answers to your question are:

1) It is difficult to accurately predict the life of a coke drum with strain gages.
2) Yes, you can reasonably estimate the drum life if gages are installed when drum is new.
3) The Equipment Health Monitoring System provides you feedback to modify your process, so that you can minimize the damage per cycle and maximize the life of your drums.
4) I call them my flashlights in the dark, because they let me find the switch to turn the light on.

Let me provide you a more detailed explanation of what the Health Monitoring System does and its benefits. First, the strain gage is a relative measurement. When the strain gage is installed and a zero set, it has no knowledge of the past or current stress history. Coke drums are often corrugated, internal flows create random stress conditions from self constraints, the drum can have a tilting performance (banana effect), and does not exactly repeat itself from cycle to cycle. This is why a number of cycles (30-50) need to be recorded to gain a representative statistical snapshot of stress range per cycle. From this we can calculate an estimated fatigue life and yes, the associated cycles.

High Temperature Strain Gages on skirts have the most accurate predictions because they are more consistently responding to the push and pull conflict between the skirt and the drum. Gages on the shell show a wider spread of response during quenching due to random channel flows and bulge interactions.

Extrapolation into the future or the past assumes that the drum has and will always operate the same as a limited data sample. This is a big assumption because cycle time changes, feedstock changes, and corrugations change. If gages are installed on new drums, then a more accurate prediction can be constructed. Often, gages are put on drums near the end of economic life when damage is accumulating in a non-linear increasing pace regarding bulges and crack growth. We recommend deployment early in the drum life, so that benefit can be maximized, and operating procedures adjusted to have the least damage while maintaining production. This early installation can be done before bulges can create amplification of the stress, and the results are easier to interpret.

We have developed a procedure to focus more on low cycle fatigue damage accumulation per cycle rather than per a large collection of cycles. Although the net result may be the same, the reading of fatigue damage is a more useful and timely metric than any mention of cycles. When cycles are mentioned, most people will misuse the answer and try to estimate the number of cycles remaining in the life of the drum. Cycles have an inverse and nonlinear relationship with fatigue damage.

Low cycle fatigue includes metal deterioration as well as crack growth leading to a through wall failure and leakage. End of fatigue life does not always mean end of economic life. We can monitor only a few locations on the drum surface which continues to change as the bulge corrugations grow. It is not economically feasible to have the entire surface covered, or every seam, or every bulge peak and valley. This is where engineering judgments come into play, because you have no alternative but to extrapolate and provide evaluations for the drum and not just a certain location. Economic end of life means you can not economically repair the defects as they become known while maintaining production goals.

Unfortunately, many drum sites are in this dilemma because replacements are not a trivial task and there is a lead time before this can be done.

If we were to tell you the drum has a cyclic life of 3600 cycles, for example, this should not be interpreted as no cracks before then and on cycle 3601 watch out because it will be a leaker! . This means you should be making plans to have maintenance procedures, expenses and delivery schedules of similar time frame. More importantly, the owner/user should seek out less damaging procedures to extend the drum cyclic life and reduce the outages and repair cost.
The best way to seek this path is to have a good dashboard for feedback on how fast you are going, how much tread is left on the tires, working headlamps, and some flashlights in the dark. The odometer is not the most useful gage to tell you how many miles you have left to go.

As you can see, this issue is a bit more than a forum reply can do justice to. Contact us for more information. We will come explain it to you if that helps.

Coke Drum Team
Stress Engineering Services

[Anonymous]

We actually produce bespoke strain gauges, have a look at our website for info: http://www.sensorsolutions.biz

Cheers

[Author]

Posts