|David W. Spitzer|
The following article is the latest in a series based on the troubleshooting of a reactor unit shutdown due to flowmeter performance problems.
The appearance and operation of a control valve encased in ice was evaluated last month (Flow Control, June 2014, page 12), and it was determined that the control valve operated properly and was not a likely source of the reactor scram (shutdown). Now let’s focus on the liquid ammonia flow measurement that appeared to cause the reactor to shut down.
As I mentioned a few months ago (Flow Control, May 2014, page 14), the technician (whom I met only a few days prior) immediately pleaded with me not to have him check the flowmeter calibration and its functionality. He knew that it would be a waste of time because he had seen the same chain of events occur a number of times over a number of years.
The technician elaborated, saying that the scrams seemed to start every year in about the same month (June) at about the same time (early afternoon) under the same weather conditions (warm and sunny). The spurious scrams stopped when the weather got cooler (until the next year).
Failure to keep the ultrasonic flowmeter full of liquid ammonia would break the ultrasonic circuit between the transducers and cause the flowmeter to suddenly measure zero flow—even when flow is actually present. Such an event would be consistent with the chain of events leading to the scram. However, observation of the installation revealed that the ultrasonic flowmeter was properly installed in a submerged leg so it should be full of liquid.
Examination of the upstream piping revealed about 3 meters of jacketed pipe that was cooled with glycol. The jacketed pipe was clearly an afterthought in this well-designed unit. When questioned, the technician told me that the scrams occurred more frequently before the cooling jacket was installed.
Is it possible that we are getting closer to the source of the problem? Could it be that the liquid ammonia was entering the unit with bubbles of gas that caused the flowmeter to incorrectly measure zero flow and, in turn, cause the reactor to scram? Cooling the liquid ammonia feed reinforces this hypothesis because it reduced the number of scrams.
Stay tuned … we’ll journey across the plant to discover more information about this application next month.
David W. Spitzer is a regular contributor to Flow Control magazine and a principal in Spitzer and Boyes, LLC. Spitzer and Boyes is also the publisher of the Industrial Automation Insider. David has more than 35 years of experience and has written over 10 books and 300 articles about flow measurement, instrumentation and process control.