Overpressure Protection

[dhenderson2014]

Hi,

We have recently completed a LOPA study of a gas (predominantly methane) system and have found that one of the knockout drums is not adequately protected from over-pressure, based on our tolerability criteria. We have concluded that the drum requires an order of magnitude risk reduction for the scenario in which the split-range control loop fails to danger i.e. the inlet PCV fails in the wide-open position.

I have attached a sketch of the system in question for your understanding.

Upstream pressure = 37barg
KO Drum design pressure = 7barg
KO Drum normal operating pressure = 4barg

Some conclusions thus far:

1) The split-range control loop is not a valid barrier as it is assumed to have failed to danger.
2) The high pressure alarm is not a valid barrier as there is insufficient process safety time. I have tried to take credit for this by installing an RO in the supply line to the KO drum. However, this is not feasible as, under normal operation, this would limit the flowrate to below the fuel gas demand of the downstream furnace.
3) Upstream pressure cannot be limited to below the design pressure of the KO drum as the DP across the inlet PCV under normal operation would not provide the required flow for the furnace.

At present, the only recommendation that we can think of is to install a SIL 1 independent high pressure trip on the inlet to the KO. I was wondering if anyone had any other 'clever' ideas for risk reduction before we implement the high pressure trip.

Thanks!

 

[RVAmeche]

Depending on your criteria/risk analysis, you could install a second regulator in series so failure would require double jeopardy.

Have you looked at the Cv of the PCV and confirmed it'll result in overpressure during normal operation, or are you assuming a worst case where it's failed open and pressure has reached steady state (ie, no flow from upstream) to 37 barg?

 

[dhenderson2014]

Thanks for your reply RVAmeche.

I'll certainly look into the potential second regulator in series.

Yes, the scenario we are considering is that the inlet PCV fails in the wide open position with 37barg upstream and the KO drum at its normal operating pressure. We have essentially assumed that everything is 'frozen'. Based on the Cv of the PCV, it will most definitely result in overpressure.

I should have mentioned that the PSV is set at 7barg and is adequately sized for this scenario. However, the PSV has an associated PFD and therefore we require another independent layer of protection against overpressure.

 

[Latexman]

A second PSV in parallel?

Good Luck,
Latexman

 

[TiCl4]

A second PCV in series may not be acceptable if the failure mode is common (i.e. fail open valve and loss of plant air, which would hit both valves, or a BPCS failure which would hit both valves).

Latexman's suggestion of a parallel PSV would work, but since this is a BPCS failure for the initiating event, you could also just upgrade the integrity of the pressure control loop to decrease the initiating event probability. If you have an existing SIS, this may be another option. You'd have to evaluate what integrity level is required(are parallel sensors or another final control element needed, or both) via PFD calcs on the loop.

It's likely you need another sensing element that routes through the SIS, at minimum.

 

[georgeverghese]

Agreed, a PSH SIL 1 trip would be standard approach to risk reduction and would be a suitable penultimate protection device. You cannot take credit for the control loop, whether there be one or two of these in parallel or in series, since we cannot be sure of the controller's response time to a downstream trip at the furnace air - fuel gas ratio controls unit. If process safety time is not adequate, add a quick closing ESDV on the inlet line.

 

[XL83NL]

Does this system/assembly need to comply with (the essential safety requirements of) PED 2014/68/EU?

Huub

 

[TiCl4]

Just another thought to throw out:
What is your normal operating position on the inlet PCV? Do you ever need 100%? 80%? If that number is fairly low, you may be able to reduce the PCV size or install a mechanical stop at X% open and still achieve sufficient gas flow. This would not eliminate the overpressure event, but it may increase the length of time it takes for the vessel to overpressure. If you didn't take credit for the outlet PCV opening to relieve the pressure due to the quick overpressure, this may increase the time available for response.

 

[Iomcube]

A self-regulating PRV after inlet control valve? Purely independent mechanical driven PRV not requiring instrument air

 

[RVAmeche]

An individual unit (self contained PRV) could still be considered a single point of failure if it would fail open.