A
In our multideck offshore platform we have 3 modules - Module 1(fire zone-1), Module 2 (fier zone-2) & Module 3 (fire zone-3). There is a fire wall between Module 1 & Module 2. There is another blast wall between Module 2 & Module 3. Each module consists of 3-levels of decks. Module 1 basically contains Riser flow lines, manifold area & inlet separator area. Module 2 is gas processing area. Module 3 is utility area + LQ.
I am in slight confusion with gas detectors voting in this platform. I am habituated to do gas detectors voting in each deck and each fire zone area. In that way the voting can go maximum 2oo12 or 2oo16. But in this case Client is insisting to do the voting in entire fire zone area considering multidecks. In this case voting logic has to be 2oo36 or 2 oo40. I am not habituated to use so many detectors in the voting logic. What is other's experience?
Technically speaking, as I see it, there are two problems associated with considering huge number of detectors in the voting logic.
1. If any two detectors are in fault there will be nuisance shutdown. The probability of two detectors at fault increases with increase in no. of detectors for voting. However, we have one advantage here, in our FGS voting function block, fault of a detector is not considered as an executive action. Only voting logic changes to 2ooN-1. So in that case there will be no nuisance trip. However, if there is zero shift in the output of gas detectors and the output goes high, the shutdown may happen. This chance is very remote, because normally the IR detectors are very reliable and they will be calibrated every after 3-months.
2. Our present function block can go maximum up to 2oo12. We have to modify the function block and also have to perform independent testing.
I will be grateful if anyone can share his experience with me on this matter.
I am in slight confusion with gas detectors voting in this platform. I am habituated to do gas detectors voting in each deck and each fire zone area. In that way the voting can go maximum 2oo12 or 2oo16. But in this case Client is insisting to do the voting in entire fire zone area considering multidecks. In this case voting logic has to be 2oo36 or 2 oo40. I am not habituated to use so many detectors in the voting logic. What is other's experience?
Technically speaking, as I see it, there are two problems associated with considering huge number of detectors in the voting logic.
1. If any two detectors are in fault there will be nuisance shutdown. The probability of two detectors at fault increases with increase in no. of detectors for voting. However, we have one advantage here, in our FGS voting function block, fault of a detector is not considered as an executive action. Only voting logic changes to 2ooN-1. So in that case there will be no nuisance trip. However, if there is zero shift in the output of gas detectors and the output goes high, the shutdown may happen. This chance is very remote, because normally the IR detectors are very reliable and they will be calibrated every after 3-months.
2. Our present function block can go maximum up to 2oo12. We have to modify the function block and also have to perform independent testing.
I will be grateful if anyone can share his experience with me on this matter.
