Experimental and computational analysis of purge systems for radiation pyrometers

Abstract

Maximizing the turbine entry temperature (TET) is fundamental to increase engine efficiency and reducing fuel consumption. Nonetheless, safety and reliability requirements have to be fulfilled. The life of gas turbine blades is strictly connected to their temperature through the creep deformation process. For this reason temperature monitoring is an essential requirement. Commonly this is achieved by means of devices such as thermocouples which are placed in the bulk flow. The usefulness of these devices as the means of supplying turbine blade temperature information is limited given their slow response time and the fact that the blade temperature is inferred from that of the surrounding gas rather than measured directly. This in turn means that critical blades parts (e.g., trailing edge) or the presence of hot spots are not identified in a discrete manner. These drawbacks can be addressed by using instead a radiation pyrometer, which is characterized by a fast response time, high accuracy, and by being contactless. The pyrometer optical front-end is a lens which collects the radiation emitted by a spot on the turbine blades. However, since the lens is exposed to the harsh engine environment, contaminants entrapped by the turbine flow can therefore be easily deposited on the lens thus filtering the radiation and resulting in an under-estimation of the actual blade temperature. The fouling of the lens is generally tackled by using a purge air system that employs air bled from the compressor to divert those particles whose trajectory is directed towards the lens. Currently the employment of optical pyrometry is often confined to military applications due to the fact that their turbine entry temperatures are higher than in civil applications. Besides, the maintenance schedule established for military engines is far more frequent than what is practiced in airline engines. Therefore, the design of current purge air systems reflects these facts. Before optical pyrometers can be commonly used for civil applications more research is required since some of the fundamentals of the fouling mechanisms remain to be clarified. This is then the knowledge gap the present research sought to fill. Its aim was to conduct a comprehensive investigation of the phenomena that underpins the lens fouling process in order to provide a set of guidelines for optimising the design of purge air systems. The initial part of the research was dedicated to the study of the purge flow inside a given pyrometer configuration. The scope was to identify the main flow structure that determines the fouling process and at the same time to validate the results obtained via computational fluid dynamics (CFD) analyses conducted in a second phase of the research. Given the reduced dimensions of the pyrometer purge system, it was not possible to gain the appropriate optical access to take flow measurements. Consequently, a large scaled experiment was performed, employing the Particle Image Velocimetry (PIV) technique for the acquisition of experimental data of the flow field. The distortion of the image and light reflection introduced by the presence of curved glass surfaces was investigated by means of a feasibility experiment. The experimental study highlighted the presence of a large recirculation zone that can trap contaminants and direct them towards the pyrometer lens. The experimental data were in agreement with computational fluid dynamics results obtained by using two different turbulence models. In a second instance, attention was focused on the particle deposition as seen from a fluid dynamics perspective. A computational fluid dynamics analysis aimed at reproducing the flow field of an existing pyrometer purge system enabled the identification of those features that can significantly impact on the lens fouling process. It was found that the geometry of the air curtain configuration plays a fundamental role. However, given the high speeds involved, the main force governing the contaminants deposition is the drag. Additionally, particles with high inertia hit the purge tube wall and then bounce towards the pyrometer lens, while contaminants with low inertia can be trapped by a large recirculation zone and subsequently directed towards the lens. In a third phase of the research, the impacts between the contaminant particles and the lens were investigated through a finite element analysis (FEA) aimed at identifying the most important factors that contribute to the lens fouling process. Particles moving at low speed can be deposited on the lens by means of electrostatic and Van der Waals forces. Conversely, particles with very high velocity can be deposited on the lens through the same mechanisms involved in the cold spraying process, which is a technique commonly used for coating deposition. A local melting can occur at the interface between the lens and the contaminants due to the high stresses created by the asperities and high sliding velocity of the particles. As a result, while large particles bounce back, debris remains bonded to the lens surface. Last but not least, the findings of the several steps of the present research have been brought together in order to produce guidelines to be followed by engineers engaged in the redesign of more efficient pyrometer purge systems.