robot65
Petroleum
- Apr 22, 2017
- 8
We have a nucleonic density gauge level measurement system on a vessel which is measuring the bulk oil level in the vessel. The source (emitter) is Cesium CS-137 and the detector is scintillation type detector both (source and detector) mounted external to the vessel i.e. non-intrusive.
The way this technology works is If there is high level in vessel, then the detector (scintillation crystal) picks up low counts thus initiating HiHi Trip. If the vessel is empty then the detector (scintillation crystal) picks up high counts thus initiating a LoLo Trip. This is commonly used in Oil and Gas industry and is considered to be extremely reliable once setup/calibrated correctly during commissioning. In our case, we only have a LoLo trip on this vessel. There is no HiHi Trip. Our system has been in service for over 2 years now and generally has been reliable except following issue:
The problem we are having is we are having unexpected nuisance process trips from this device. Looking at the trends in ESD system, the level drops below the trip setpoint for around 10 seconds and automatically recovers to healthy state again. This 10 seconds drop in level is sufficient to cause ESD system to initiate process trip. We can confidently say that it is impossible for level to drop for 10 seconds in that vessel below LoLo trip setpoint. We have also cross checked it with the trends from Guided wave radar level transmitter which is used for control purposes. This on average has happened once every two months. The way these nucleonic detectors works, it appears to me that the detector (scintillation crystal) is picking up high counts radiation from somewhere else for 10 seconds and goes back to normal. There was no radiography happening at that time so it can't be radiography.
We have replaced all the components in the loop but the problem is not going away. Both nucleonic detector manufacturer and ESD System vendor have done extensive investigation at site but we still can't figure out what is causing this issue. Does anyone has experienced any issue similar to this in the past?