David W. Spitzer
David W. Spitzer

RELATED: Read Part I in the  "Troubleshooting Operational Issues" Series

RELATED: Read Part II in the "Troubleshooting Operational Issues" Series

RELATED: Read Part III in the "Troubleshooting Operational Issues" Series

The reactor scram (shutdown) system was activated at about 2 p.m. on a warm June day because the liquid ammonia flowmeter feeding the reactor measured low flow. The flowmeter appeared to be properly installed, and its flow control valve located downstream was caked in ice. The reactor was restarted without incident about an hour after the scram, and I went back to my office to think about what really caused the scram.
   
A control valve caked in ice is not a common occurrence. Investigation revealed that the vapor pressure of ammonia at warm ambient temperatures can be as high as 10 to 12 bar. This means that if the ammonia was in an open cup, the ammonia liquid would boil (vigorously) because the atmospheric pressure above the ammonia is lower than its vapor pressure at this temperature.
 
The design pressure upstream of the control valve is above the ammonia vapor pressure. However, the pressure downstream of the control valve is below the vapor pressure, as is the pressure in the reactor. Therefore, the ammonia downstream of the control valve is partially or completely in the gaseous state. Either way, the downstream ammonia will be cooler within and downstream of the control valve due to the refrigeration effect of boiling ammonia.
 
Operating this control valve under these conditions on a warm day with high humidity will cause water to condense on the outside of the valve. Further, ice will form if the surface temperature of the valve is below zero degrees Celsius.  

Although it is not a common occurrence to operate a control valve that is caked with ice, the presence of ice was deemed to be normal under the operating conditions in this application. The control valve was used to start up and operate the reactor, so it was unlikely that the control valve had any role in causing the scram.  
We’ll investigate the flowmeter next month.

RELATED: Read Part V in the "Troubleshooting Operational Issues" Series

RELATED: Read Part VI in the "Troubleshooting Operational Issues" Series

RELATED: Read Part VII in the "Troubleshooting Operational Issues" Series

RELATED: Read Part VIII in the "Troubleshooting Operational Isssues" Series

RELATED: Read Part IX in the "Troubleshooting Operational Isssues" Series

RELATED: Read Part X in the "Troubleshooting Operational Isssues" Series

David W. Spitzer is a regular contributor to Flow Control magazine and a principal in Spitzer and Boyes, LLC offering engineering, seminars, strategic, marketing consulting, distribution consulting and expert witness services for manufacturing and automation companies. Spitzer and Boyes is also the publisher of the Industrial Automation Insider. Mr. Spitzer has more than 35 years of experience and has written over 10 books and 300 articles about flow measurement, instrumentation and process control. 

Mr. Spitzer can be reached at 845 623-1830 or www.spitzerandboyes.com. Click on the “Products” tab to find his “Consumer Guides” to various flow and level measurement technologies.