Virginia Tech
Application of a Non-intrusive Optical Non-spherical Particle Sizing Sensor at Turboshaft Engine Inlet
Abstract
dc:description.abstractThe aviation industry struggles with particle ingestion as it is highly corrosive and damaging. More importantly, though, the issue of engine particle ingestion has more questions than answers. There is a large industry need to understand the effect that particle ingestion has on turbine engines as they alter the life-cycle of parts and raise safety concerns. For those reasons, our team with the Advanced Propulsion and Power Lab developed a non-intrusive an optical sensor that is able to measure particle size and mass flow rate in flow. This sensor utilizes scattering and extinction theories and neural-network machine learning accompanied by a scattering and extinction library to measure the particle characteristics captured using optics at the test section. This sensor has been demonstrated in various applications; however, the most recent experiment was completed at Virginia Tech's M250 engine inlet. This was the first time the sensor was fully attached to an engine, putting it under operational conditions. The particulate used at this demonstration was CSPEC sand, also known as Mil-E-5007 is a military specification used for turbomachinery. CSPEC is 99% crushed quartz, which is non-spherical and ranges widely in shape and size distributions. Utilizing the scattering light measurement taken using a laser sheet and cameras, as well as an intensity drop measurement from an extinction setup, the machine learning model returns the average aspect ratio, distribution width, and Sauter mean diameter of the particle. Those measurements are then used to calculate the average mass flow rate of ingestion. The most recent inlet demonstration measured the average mass flow rate with a root-square-mean error of 0.1 g/min. The success and accuracy of this non-spherical measurement method allows for versatility and robustness in different applications with various seeding particulates. The scattering and extinction library used to train the machine learning models can be expanded to allow for a number of seeding materials to be identified. This method was also expanded to inform local particle size distributions. To do this, the scattering information captured by the cameras was first processed through a binning code. Those binned images were then run through the characteristic prediction code that was modified to accept the smaller scattering data. The code worked to find the particle size distribution averaged over the smaller test section area scaled by the number of bins used. The mass flow rate for each bin was calculated defining a local distribution of the mass flow rate. The result of the experiment described above as well as the local adaptation results are presented in this thesis.
Degree
thesis:*- Name thesis:degree_name
- Master of Science
- Level thesis:degree_level
- masters
- Discipline thesis:degree_discipline
- Mechanical Engineering
- Department dc:contributor.department
- Mechanical Engineering
- Grantor dc:publisher
- Virginia Tech
- Year dc:date.issued
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Antous, Brittney Louise
- Chairs dc:contributor.committeechair
-
- Lowe, K. Todd
- Ng, Wing Fai
- Committee member dc:contributor.committeemember
-
- Son, Chang Min
Subjects
dc:subject × 4Rights
dc:rights- Statement dc:rights
-
- In Copyright
- Licence dc:rights.uri
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Dc Identifier Other
- vt_gsexam:36674