9 replies to this topic

[jac_lh]

Hi All,

I have a question about a tube rupture relief scenario.

Both sides of the exchanger are liquid. My high pressure tube side fluid is being pumped through the heat exchanger at a fairly low flow rate. When I look at the tube rupture case for PSV sizing, can I say that the relieving rate is limited to the pumping rate? Although initially the amount of fluid flowing through the tube break will be high, immediately afterwards the flow can only match the pump rate.


I have been advised by a senior engineer to go with the pump rate, but the client has rejected the calculation on this basis. It significantly changes the required PSV size.

Thanks in advance

[ankur2061]

jac_lh,

No problem in considering the pump flow rate but even there consider the flow rate at the end of the pump curve by refering the pump curve. This will provide you a conservative flow rate (higher value) compared to the rated flow rate of the pump.

Regards,
Ankur.

[jac_lh]

Ankur,

Thanks for your quick response.

[Bobby Strain]

You need to consider what is on the tube outlet, too.
Bobby

[fallah]

jac_lh,

You fail to upload a simple sketch of your system in order to get more precise response . Also you fail to say if client comment will result in higher (or lower) relief load than pump's rated flow. It is not specified if there would be a phase change through tube break during tube rupture. Also, Isn't the flow choked across the tube break?. Anyway considering the flowrate corresponding to the end of curve without having above data may lead to an oversized PSV that, in turn will result in PSV chattering.

Fallah

[Art Montemayor]

I agree with Fallah’s logic. Once the rupture occurs, if the PSV capacity is related to end-of-curve rate, the chattering will be inherent and possibly damaging.

I have always justified and employed rupture discs or rupture (buckling) pins for this type of application because of the fact that if this is a process heat exchanger, the entire process (and unit) are, in effect, totally shut down as a result of the tube rupture. If that is, in fact, the case then nothing is lost in using a rupture device. In fact, a lot is gained with its simplicity and cost – both capital and maintenance.

[jac_lh]

Thanks all,

Some more information:

The client has said we cannot use a bursting disc although this would be our normal procedure.

The pump is a diaphragm type, and the rated flow is less than the flow through the tube rupture (using flow through a pipe end plus flow through an orifice). The PSV size based on flow through the tube rupture gives a J orifice, whereas if limited to pump flow it is a D.

There is no phase change through the rupture. I have allowed credit for the minimum operating outlet flow through the exchanger discharge line on the shell side as per client instructions and as per API 521.

[kkala]

Tube rupture PSV was once specified at local work, http://www.cheresources.com/invision/topic/10441-exchanger-tube-rupture. Shifted to the site I could not follow details, but old procedure (not well understood) for maximum flow into low pressure (LP) side was: high pressure (HP) liquid flow Q from broken tube = twice that from a single tube section, upstream pressure = normal operating pressure of HP liquid, downstream pressure=1.1*design pressure of LP liquid, discharge coefficient =0.7 (no flashing around broken tube). If thus calculated flow was higher than normal HP liquid flow, the latter had to be taken into account.
Suppose that a pump feeds the HP side of the exchanger and this pump does not feed anything else. According to above, Q shall be equal to pump rated flow or less, depending on the results of mentioned procedure, whether the pump is centrifugal or positive displacement (e.g. diaphragm, rotary, plunger).
In case of a positive displacement pump, adopting Q=rated flow covers also the case when downstream HP valves are closed during rupture. Consequently this adoption (Q=rated pump flow) is adequate (no need for more flow).

It is noted that mentioned procedure contains assumptions. Indication of tube rapture complexity can be seen in http://www.equityeng.com/consulting-services/pressure-relieving-systems/heat-exchanger-tube-rupture-analysis, as well as in http://www.cheresources.com/invision/topic/4246-heat-exchanger-tube-rupture/ along with some clarifications. Transient analysis articles (seen but not studied) looks quite complicated.

Since relevant PSV will face events from (serious) tube leak to tube bore rupture, PSV may pop up, get deteriorated and replaced after the "accident". But main aim is to protect exchanger. Rupture disks could be better concerning this issue.

Edited by kkala, 23 August 2012 - 07:30 AM.

[Art Montemayor]

Jacqueline:

Thank you. Finally, you give us the REAL basic data - the pump type. This should have been in your query's first paragraph and is the basis of the flow rate - which is positive displacement. This pump - I have to assume - is not on a variable speed drive and, therefore, sets the relief rate. That is probably why your supervisors advised you to set the relief based on the pump rate - which is constant (albeit, pulsating) and steady.

This is why I have ranted and raved through the years that our members should always qualify WHAT type of pump and compressor they are writing about - dynamic or positive displacement. It does make a difference. The answer becomes very simple: use the positive displacement rate of the pump.

[jac_lh]

Thanks for all your responses.