15 In-flight Icing: Modeling, Prediction, and Uncertainty
501
The European regulations are found to be analogous to the FAA ones presented in
the Sect. 15.4.3.1. The icing envelopes are presented in appendices C, D, and O and
the regulation and certification it is described in the sections 25.1419, 25.1420, and
25.1093.
15.5 Concluding Remarks
An overview of the in-flight icing phenomena has been presented, including its
physics, research methods, and ice protection systems. Within the research methods
the replication of ice shapes, whether experimental, numerical, or analytical is subject to uncertainties. Input data for models gathered in-flight presents uncertainties
related to the accuracy of the measuring devices. For instance, the error of the King
probe, which measures the cloud mean volume diameter, can reach up to 30%.
Although the technology of the measuring devices has evolved, some published data
was gathered in field campaigns several years ago. This data still needs to be used
since, in some cases, it is the only data available. It has been used for validation
purposes to replicate icing shapes, as model inputs or to elaborate regulation to
derive possible existing icing envelopes. The uncertainties need to be addressed, to
measure their individual and overall effect on the discrepancies between models and
real phenomena.
Physical reproduction of in-flight icing phenomena in icing tunnel tests presents
also uncertainties related to the replication of actual airflows and cloud properties.
The cloud-related uncertainties are also applicable to in-flight tests in which the
studied aircraft is preceded by a water tank that sprays water droplets. It was found in
literature that the replicability of the same icing test could reach up to 30% in some
cases, when ideally it should be 0%. Some of these uncertainties can be reduced
by further testing, regular calibration of the devices and retrofitting facilities and
devices, however, this implies further efforts.
There has been a huge progress on the numerical modeling of in-flight icing as
it has been presented in the part Sect. 15.3. Some of the numerical models present
really good agreement with experimental results. However, additional studies are
required to capture intricate ice shapes. Further work would entail the quantification
of the uncertainties in model inputs and in analytical and numerical models
themselves to assess the global uncertainty of the results. Then, the uncertainties
could be rated depending on their importance in the overall uncertainty of the results
and further efforts could be conducted in the reduction of the largest epistemic
uncertainties. In addition, input and model uncertainties could be propagated
through the models to predict the uncertainty of the modeling results and elaborate
more robust predictions. Improved capture of the in-flight icing phenomena can lead
to the design of more robust ice protection systems, which can ensure safety still in
a more efficient manner.
501
The European regulations are found to be analogous to the FAA ones presented in
the Sect. 15.4.3.1. The icing envelopes are presented in appendices C, D, and O and
the regulation and certification it is described in the sections 25.1419, 25.1420, and
25.1093.
15.5 Concluding Remarks
An overview of the in-flight icing phenomena has been presented, including its
physics, research methods, and ice protection systems. Within the research methods
the replication of ice shapes, whether experimental, numerical, or analytical is subject to uncertainties. Input data for models gathered in-flight presents uncertainties
related to the accuracy of the measuring devices. For instance, the error of the King
probe, which measures the cloud mean volume diameter, can reach up to 30%.
Although the technology of the measuring devices has evolved, some published data
was gathered in field campaigns several years ago. This data still needs to be used
since, in some cases, it is the only data available. It has been used for validation
purposes to replicate icing shapes, as model inputs or to elaborate regulation to
derive possible existing icing envelopes. The uncertainties need to be addressed, to
measure their individual and overall effect on the discrepancies between models and
real phenomena.
Physical reproduction of in-flight icing phenomena in icing tunnel tests presents
also uncertainties related to the replication of actual airflows and cloud properties.
The cloud-related uncertainties are also applicable to in-flight tests in which the
studied aircraft is preceded by a water tank that sprays water droplets. It was found in
literature that the replicability of the same icing test could reach up to 30% in some
cases, when ideally it should be 0%. Some of these uncertainties can be reduced
by further testing, regular calibration of the devices and retrofitting facilities and
devices, however, this implies further efforts.
There has been a huge progress on the numerical modeling of in-flight icing as
it has been presented in the part Sect. 15.3. Some of the numerical models present
really good agreement with experimental results. However, additional studies are
required to capture intricate ice shapes. Further work would entail the quantification
of the uncertainties in model inputs and in analytical and numerical models
themselves to assess the global uncertainty of the results. Then, the uncertainties
could be rated depending on their importance in the overall uncertainty of the results
and further efforts could be conducted in the reduction of the largest epistemic
uncertainties. In addition, input and model uncertainties could be propagated
through the models to predict the uncertainty of the modeling results and elaborate
more robust predictions. Improved capture of the in-flight icing phenomena can lead
to the design of more robust ice protection systems, which can ensure safety still in
a more efficient manner.