Flare design for simultaneous failure of 2 switchboards?

[Guidoo]

For checking of a flare header, we have to decide which partial power failures can be disregarded as uncredible.
The electrical system of the refinery consists of two main substations, located in two separate buildings. Each substation consists of two switchboards, each distributing the power to different parts of the refinery.
Question is what the possibility is that two switchboards in a single substation fail simultaneously. Does anyone have figures, or know where to find figures in relation to damage (e.g. car crash, lightning), or if one board fails the other board will be damaged as well.
Also more qualitative considerations (e.g. experiences in your own plant/refinery) are wellcome.

 

[owg]

I think it would depend on how indepent the "switchboards" motor control centers, or substations are. If they both are fed from the same utility (third party power supplier), then you would have to consider failure of that utility.

 

[MortenA]

switchboards may not fail simultaneously but how about a general power failure?

Best Regards

Morten

 

[Shortstub]

Responding to Guido's query:

Yes, simultaneous failure is possible! I have investigated at least 3 such incidents in two refineries and one chemical plant. All three were initiated on one side of a LV (400-Volt) secondary-selective substation. One incident was initiated in the incoming-breaker cubicle. The ensuing arc-products and gasses were blown thru openings in the tie-breaker compartment, and eventually enveloped the entire switchgear lineup.

The root cause was inadequate protection to detect the arcing-fault. A secondary cause was, that, even though the main bus was compartmented, the bus was bare, allowing the arc to travel from one side of the lineup to the other. Remember, an arc will move in a direction away from the source of power.

The resultant rms, was too low to be detected by relaying. Thus the fault persisted until upstream relaying, on the HV side of the supply transformers cleared the fault.

An aside! A traveling arc, and its resultant plasma was the basis for the Star-Wars space-gun!

 

[22082002]

Looking overall, I feel simultaneous power failure across the whole plant(i.e. Substaion 1 & 2) should be considered for design of flare header.I think this is by default consideration, involving CW/air coolers loss/ reflux, etc. Check it !

If this is considered then partial failure should not pose any problem with regards to flare header.

Regards,
SAA

 

[SooCS]

There are a few sources you might want to check. EG

1. Offshore Reliability Data
2. E&P Forum Report 11.8/250
3. Mil Std No: 721 (This is especially good source for electrical components).

In any case, I would recommend a HIPP (High integrity pressure protection)Study for your system. Download, Norsok Std P-001, Process Design for more info about HIPP.

I have been done a HIPP Study for a flare system for one of the gas production facility for the similar concern with the risk of multiple failures that can lead to overloading the flare system.

Best of luck.

 

[Guidoo]

The study for which I asked the question has been finished some time ago. Basically my conclusions were that:
- a more detailed system analysis is required to estimate a failure figure (analysis can be done according to standard IEEE 494 ("gold book&quot)
- it is important that feeds to the switchboards are separated, if they are all fed from same utility power supplier, all switchboards fail when this utility fails

About reply from SooCS: a HIPPS system is used specifically in a gas production facility, to protect the facility against the well pressures.
In the refinery, flare load mitigation is applied where high integrity trips are used that reduce probability that flare load exceeds the design load. The trips are initiated e.g. by local high pressure sensors and/or by global sensors that detect power failure or cooling water failure. The trips cut off heat supply to reboilers, trip compressors etc. and thereby reduce flare load. Of course there is always the possibility that one or more of these trips fail on demand. Therefore, also mechanical devices (safety valves) are provided, and statistical calculations need to be done to show that probability that total flare load exceeds flare design load is small (e.g. probability < 1E-4) so that the resulting risk is tolerable.
Anyway, thanks for the replies to my query!