Posts Tagged ‘thermowell’

Is your thermowell up to standard?

November 8, 2013

By Steve Gorvett, Temperature Product Specialist, ABB Instrumentation

Did you know that recent amends to manufacturing standards may mean that some of your thermowells aren’t up to scratch?
Thermowells are often so fundamental to plant safety that any design flaws can have disastrous effects. By way of example, in 1995 a thermowell failure in the coolant loop at the Monju fast breeder reactor in Japan closed the plant for the next 15 years. It’s for this very reason that there are tight industry manufacturing standards that govern thermowell production.

Following some catastrophic failures (including the Monju disaster) the American Society of Mechanical Engineers decided to amend the ASME PTC 19.3-1974 standard to which these thermowells were designed to.

The latest revision of the ASME PTC19.3 standard makes use of significant new knowledge about the behaviour of thermowells, compared to the criteria laid out in 1974. The standard evaluates thermowell suitability with new and improved calculations for various thermowell designs and material properties. It also takes some detailed information about the process into account.

In particular, the standard looks at the incidence of vortex shedding. This is the phenomenon where vortices formed in the wake of the thermowell move from side to side; this is what causes vibrations in the thermowell. If this vortex shedding rate matches the natural frequency of the thermowell, resonance occurs, and dynamic bending stress on the thermowell increases.

The frequency ratio is the ratio between the vortex shedding rate and the installed natural frequency. In the old standard, the frequency ratio limit was set to 0.8. The new standard stipulates that in some cases, the limit should be set to 0.4. The new possibility of having a much lower frequency ratio limit of 0.4 means tighter design constraints in many cases. As the majority of existing assets will have been designed to the 1974 standard, the new 0.4 frequency ratio means a lot of thermowells will NOT pass the new standard.

Re-evaluation and re-certification services are available. Operators will need to consider the implications when an existing thermowell fails the new calculation. If process conditions change, for example increased throughput for a part of plant, this should also be evaluated.
At a brownfield modification we recently examined for new process conditions, 29 existing thermowells were evaluated under existing and new conditions. Only six passed the new standard under existing conditions!

Bearing in mind what happened at the Monju fast breeder reactor in Japan, which did conform to the 1974 standard, isn’t it about time you reviewed your thermowell installations?