ANSWERS:
I FIRST addressed this question in about 1980 when a Du Pont colleague was investigating reliability and safety shutdown systems. His work eventually led in to the effort in ISA SP84. My conclusion then, and I still believe this is valid, is that there is no way to accurately calculate or predict solenoid valve failure in the general sense.
Most failures I saw during my years on the plants (20) were electrical. Some of these were very interesting with whole sets of SVs failing at one time during a unit start up. Rare failures have been seen with dirt blocking either flow or, more likely, switching. Water frozen in air lines can prevent operation, I suspect this problem is very rare in Saudi Arabia! Dirt and water are enemies of reliable operation of any pneumatic systems.
Another sort of failure experienced was a failure to switch because the DC voltage applied was not quite sufficient to overcome the magnetic structure in the valve. Note that the required pull-in voltage is higher than the hold-in voltage. These different values might not be listed in the catalogue specifications. I have been a party to opening up valves and adjusting the internal magnetic spacing.
Electrical failures may be linked to voltage spikes, or the reverse, low voltage heating the coils but insufficient to cause armature motion which would reduce current to a lower value in AC applications. Exposure to voltage/current spikes is heavily dependent on installation practices and even local earth conductivity. Voltage suppressing diodes are sometimes used. These may raise questions of possible of their failure to a shorted condition. Long wire runs will increase exposure to lightning induced spikes. There is a large difference in coil temperature between valves normally energized and those normally deenergized. It is generally believed that heat contributes to coil failure. My conclusion was that some brands and model numbers seemed more likely to fail. I have no way to prove this, and would never name names.
One approach supported by SP84 is that proof of reliability can be achieved after extended experience in exactly the same service. This is not much help until enough service time has passed to develop useful data. Proper record keeping is required for this, but not often seen.
The only recommendations that I could make is to select valves from a proven reputable manufacturer, that they are used well within their ratings, and installed in a manner to reduce exposure to voltage spikes and excessive heat. If very high reliability is required, then redundancy in the system design is required to reduce the impact of any single failure. That being said, I reject some of the quoted valve failure numbers which make solenoid valves look terrible, this is not justified by experience. It all comes down to intelligent engineering of the entire system and application.
Cullen Langford
RESPONSE:
I WOULD like to wish you happy new year, and also thank my teacher Dr. Liptak for his excellent support to give me this occasion to talk to an experienced person. In relation to your feedback, I have the following:- The reliability of solenoid valves has been always under the microscope. We have in one plant many Plants S/D report, which was attributed to Field components failures.
- Reliability for instrument can be discussed as in a qualitative way, but this will lead to a certain set of wide recommendations.
- My opinion is Reliability for instrument must be evaluated based on quantitative Data, so that we can an objective list of recommendations.
- As good practice, I have been informed by some of my colleague that in most Fertilizer chemical plants, they replace critical solenoid valves every T/A. The T/A is scheduled every 3 years.
- In Nuclear facility of U.S., there were some investigation reports for nuclear accidents that was caused by solenoid valves failures. In one of these reports, it was found that the life of the SOV in the 2,5 years, and the failure has happened after 6 months. Therefore, a lesson like this, will send a proactive measures such as: (a) replace critical SOV between 2~3 years of service. (b) improve on-line testing methods, (c) re-sort to new intelligent SOV with diagnostic check.
- In SAFCO, we would like to have an expert to come and help us to evaluate life for some of our critical instruments. The environment we have is dirty with high humidity rate that reaches to 100% across many days in the year. Thus, this environment is really difficult top control spurious trips. But, as PM is improved, the plant S/D occasion, can be reduced to minimum.
- Can you suggest, if you work with us as a consultant to action point 6.
- Is the Reliability Centered Maintenance concept, has been deployed fully to the Instrumentation systems?
Mufeed Al-Ghumgham
MY COLLEAGUES of process control professionals and myself are answering many more questions than the few, which get published in Control Magazine. These combinations of questions and answers you can find on our web page under Ask the Experts at www.ControlGlobal.com.
Thank you for describing Indian practices on solenoid valve (SOV in your letter) maintenance, life expectation and turn-around (T/A in your letter). I am sure that Cullen Langford will comment on them and inform you of or industry specific - American practices. I also hope that Cullen will visit with you, so that you might benefit from his experience.
