Relief valve sizing for Heat Exchanger

See the following book by Crosby for "Safety Valve Engineering Handbook": Safety_Valve_Engineering_Handbook

Have a look into API 520/521. I am sure I did something like this a couple of years ago and there is a procedure in those standards particularly for two phase flow relieving conditions.

Dear Walt48,

RV for heat exchanger normally we consider two relief scenarios.

1. Thermal expansion - this occurs when both valve on the water side closed or blocked

2. Tube rupture - I believe in your case tube rupture is the dominant case

There is no way we can calculate the flowrate precisely. We can only estimate. The most conservative method is to use Bernoulli equation since in reality we need to consider pressure drop due to sudden contraction and this will reduce the flowrate. What I meant is that actual flowrate is lower than the one estimated by Bernoulli's.

For the pressure drop, we need to take the difference between steam pressure and the set pressure of the RV, not the water operating pressure.

However, we need to check whether the flow is under choked conditions or not. As a rule of thumb, if the set pressure is about 1/2 of steam pressure or lower, the system is choked. that is if the set pressure is 125 psig or less, most likely it is under choked flow conditions. in this case we need to determine the choked pressure and use that to determine the pressure drop for the flow calculation.

Bernoilli equation gives you the velocity. U can convert this into flowrate by multiplying velocity with x-sectional area of the tube.

Good Luck

Tube rupture takes in the shell and water box is not affected. Water box relief valve is only thermal expansion relief valve. The valve provided would be adequate.

Dear Walt48

What is the design pressure of the tube?

If the design pressure is higher than 2/3 of the shell design pressure (rule of thumb) then RV for tube rupture emergency scenario is not necessary. Then the existing RV most likely for thermal relief.

No one has yet considered the fact that- when steam contacts relatively cool water, the steam will condense and create a local vacuum.

Consequently, IF/WHEN a tube ruptures, the steam that enters (due to higher pressure) will immediately become much smaller volume of condensate and the steam flow will essentially stop or continue to become more condensate mixing with the hot water.

This condition will cause the water to become contaminated by the chemicals in the steam, so a bigger issue might be to protect whoever/whatever will be on the using side of the hot water flow. Ideally- that would be with a pressure activated 3-way valve set to "dump" the hot water if triggered. The trigger could be the drop in steam pressure in the HXU shell or the increase in pressure at the water side.

Hello Walt

Is this a multi-tube exchanger? For 785 usgpm need something like 80 0.824" ID tubes in parallel to get a reasonable velocity,.

Heat power needed for the water comes to about 31.4*10⁶ BTU/hr (9216 kW). Your steam flow agrees roughly with this, but looks a bit light. To get the enthalpy change needed for your 34,200 lbs/h I estimate the condensate would be at 115°F i.e. right down at the water inlet temperature.

There's a reference to 250 psig steam - is that the max pressure?

For steam pipe size - I haven't got data with typical velocities to hand, but assuming 220 ft/s it comes to about 4".

To get back to your question - it depends on what maximum pressure the water side system can take. If it's above max steam pressure (250 psig?) you could set the relief valve at just above that and everything's OK, in fact you probably don't need a relief valve (depends on the water side details).

But this seems unlikely, and if not, the valve set pressure must suit the water side. To size the valve you need to take account of the steam supply pipe length and diameter. When relieving, the steam flow is determined by the difference between max supply pressure and relief valve setting, and supply pipe details. That would be a lot of steam, and might be overdoing it as it assumes total failure of the tubes. If you can make an assumption/reach agreement about the size of the leak (equivalent orifice diameter) it might be more realistic. Leak likely to start off small, and when you see relieving is happening you can investigate. In either case need valve data and do an iterative calculation.

Of course the relief valve must be specified for steam duty.

Cheers.........Codey

Now got my steam data and for saturated steam over 50 psig it says 100-120 ft/s. On that basis the steam supply pipe needs to be 6" dia.

Cheers........Codey

Dear Walt

I am wondering why you are not responding to any of the comments?

If I understand you question correctly, you are interested to determine the steam flowrate under the situation of tube rupture, so that you can check the RV capacity.

I have mentioned that we can use Bernoulli's equation and most flow calculation or pressure drop calculation is derived from Bernoulli. The equation to be used depends on the accuracy that you are interested in. For more accurate computation, you need to consider friction losses, adiabatic expansion factor, heat loss etc, which is a very tedious calculation and requires iteration.

However, we can simplify the computation by assuming an orifice of ¾" diameter (full bore rupture of one tube).

The general equation for orifice:

W = C a √ [2 g ρ (P₁ – P₂)]

Where W in lb/s,

C = constant,

a = x-sectional area (ft²)

g = 32.2 ft/s²

ρ = lb/ft³

P = pressure in lb/ft², (1 & 2 refer to upstream and downstream respectively), if the pressure in psi then multiply by 144

Under real situation, C < 1, but for conservative estimate use C = 1

For choked flow, P₂ is the choked pressure, otherwise use the set pressure of the RV.

I hope by now you have some idea on how to determine the flowrate.

Good luck

Sorry for not getting back to you. Tubes are specified for design of 200 psig on water side and 250 on shell side.

Walt48

From my own experience in sizing RV for exchanger, I shall not consider tube rupture anymore since the tube design pressure is more than 2/3rd of shell design pressure. The tubeside can stand up to 300 psig i.e. 1.5 x design pressure provided it is hydrotested at 1.5 design pressure.

The only consideration would be for thermal expansion. The worst case emergency scenario is when inlet and outlet valve on the water side closed inadvertently which may occur during start-up or shutdown due to operator not following procedure.

Good luck

I personally enjoy process design

Another way of doing it is to give the process details to the local friendly Valves Distributor, and get a quotation for the valve from that organisation.