Search

1-D codes are still much utilized regarding the preliminary design and the design phase of a single component. In particular, combustors is one of the most critical component in gas turbine engine and its design mainly affects all other components, starting from the turbine. During the initial phases of the design process, parameters are known approximately; during this phase, critical for the definition of a starting set of design parameters, there is little point in doing high fidelity Computational Fluid Dynamics (CFD) analyses. On the contrary, the exploration of the whole space is extremely important to better understand the behaviour of the system and to focus on the design objectives. Uncertainty quantification (UQ), mainly developed in recent years and applied in many fields, can be really helpful in the preliminary design phase and also as a support during the whole design process. This work comes with the idea to estimate the main source of geometrical uncertainties in the design phase of a combustor. The test case is based on a full annular lean burn combustor, tested at Central Institute of Aviation Motors (CIAM) during the LEMCOTEC (Low EMissions COre-engine TEChnologies) European project. Among the test points investigated in the experimental campaign, the Approach condition is analysed. The inner liner is considered to analyse the metal temperature. Therm-1D, a 1-D in-house simulation code, is used to model the combustor. After having built the baseline case of the combustor, several uncertainty analyses are investigated. In particular a classical Monte Carlo approach is compared with 4 innovative polynomial-chaos approaches for each group: Gauss quadrature, total order with Latin Hypercube sampling (LHS), probabilistic collocation and stochastic collocation. The analyses proved how the last two methods give the same results with a sensible lower amount of simulation (depending on the number of input variables). An additional sensitivity analysis to the order of the polynomial is conducted and results show how, with a 3rd order polynomial approximation, results can be very similar to those ones of the Monte Carlo simulation. Lastly, results are compared with experimental data to achieve a better understanding of the most relevant input parameters and the propagation of their uncertainty on the results. [ABSTRACT FROM AUTHOR]