As infrastructure ages and we extend the life cycle of our power plants there has been an increased need to monitor the integrity of our boilers. In the past, monitoring repaired boilers/piping for cracking or weld creep long-term has been out of reach due to the extreme temperatures and inhospitable environment. Oxidation, thermal conductance, and low frequency response presented themselves as challenges to making true long-term testing a reality.
Thin-film sensor technology had showed promise but low power densities of sputtering and PVD deposition technology limited manufacturing time and produced a microstructure that is less than ideal for corrosion protection and sensor sensitivity. Utilizing our Syncrotron ULTRAHIPIMS deposition process we are able to coat traditionally difficult materials such as silver, nickel alloys, oxides, ceramics, etc. Opening up thin film possibilities that were either not possible or impractical with conventional means.
The test article in question was a T-bar specimen of boiler steel with a fill weld exactly like the ones that would be found in boiler repairs in the field. The T-bar was equipped with an extensometer and Suprock high temp gages then subjected to 60MPa tension and heating cycles with temps up to 600°C.
For this boiler test the thin film gage was applied as a thin film to an alloy steel shim with similar thermal expansion ratio to the test article then welded directly to the test article for maximum transfer of strain. The gages are welded at each end longitudinally so that they are being pulled in tension with
the bar. Since the gages are welded at each end, they are under tension end-to-end and read
the average strain between end points of the gage assembly. Care is taken to understand how
the gage deformation relates to its location on the weld so that real results can be correlated to
the location and expected stiffness of the different material cross sections that each gage
spans.
During the T-bar test, the weld material is expected to have a significantly lower modulus of
elasticity than the base material. This is by design and choice of the filler metal so that it can
endure with higher toughness and elasticity without creating cracks along the weld fusion
zones. However, creep damage can occur in these zones, therefore, straddling the weld fusion
zone with a strain gage allows for average strain to be measured across the creep-sensitive
region. In this test it is important to note the slightly different locations of the gages as attached
across the weld span and cross-section. Finite element analysis (FEA) is conducted to
understand the effect of gage location on relative static strain of each gage. Figure 26 shows
the anticipated asymmetrical deformation due to the weld stiffness on the weld span versus
the weld cross-section.
The results from this test indicate that the gage signal tracks in good correlation with extension
of the T-bar gage length. The response from the two operating gages on the T-bar spanning the
weld on the weld span and cross-section are likewise in agreement with the FEA that was
conducted on the T-bar and weld assembly. Figure 28 shows strain measured from both
operational gages through the first approximate 64 days of testing. The breaks in the data
represent brief interruptions in data logging on the recording laptop due to external factors
(laptop updates requiring reboot, IT issues, etc.).
The weld cross-section sensor indicates about 60% of the total strain value measured by the
extensometer, while the weld span sensor measures within 10% of the true extensometer
measurement. The latter is very encouraging. Given that the extensometer spans a larger
section than the strain gage, its measured displacement is an integration over the
measurement span. For this reason, the results of the weld span strain gage are satisfactory
considering the accuracy of local strain measurement is reasonable within some percentage of
the extensometer-measured value. Regarding the output from the weld cross-section sensor,
the physical result is in agreement with the FEA-predicted outcome of a lower strain value in
the sensor. The weld cross-section sensor sensitivity is affected by its placement and the
specific strain across less weld material as compared to the weld span sensor. Thus, it is positive
to understand that the relationship between the physical measurement locations is accounted
for.
The test ended ahead of schedule due to failure of the test article. The gages were fully operational until the substrate physically gave way. Note the remarkable condition of the oxide coated 400 series stainless gages 4,383 hours @600C. This test has been an excellent proof of concept of the survivability of Suprock high-temp strain sensors and their ability to monitor both static and dynamic strain, functioning as a type of online NDT (non-destructive testing).
There are other effects to consider that relate to systematic vibration and flow induced signals
in plant components. This particularly affects online measurement. A partial list of vibration
sources in component systems includes:
• Flow induced vibration
• Aperture related turbulence, swirl, or other valve position-related effects
• Pipe hangers, in particular thermal expansion and vibration affected by hangers
• Pipe hanger condition and how it may change or degrade over time
• Operational vibration sources such as pumps
• Component natural frequencies
We are encouraged and excited by the results of this test proving the fidelity and long-term survivability of these sensors in extreme heat conditions.