Simultaneous Robust Design and Tolerancing of Compressor Blades

Abstract

The manufacturing processes used to create compressor blades inevitably introduce geometric variability to the blade surface. In addition to increasing the performance variability, it has been observed that introducing geometric variability tends to reduce the mean performance of compressor blades. For example, the mean adiabatic efficiency observed in compressor blades with geometric variability are typically lower than the efficiency in the absence of variability [?]. This "mean-shift" in performance leads to increased operating costs over the life of the compressor blade. These detrimental effects can be reduced by using robust optimization techniques to optimize the blade geometry. The impact of geometric variability can also be reduced by imposing stricter tolerances, thereby directly reducing the allowable level of variability. However, imposing stricter manufacturing tolerances increases the cost of manufacturing. Thus, the blade design and tolerances must be chosen with both performance and manufacturing cost in mind.This paper presents a computational framework for performing simultaneous robust design and tolerancing of compressor blades subject to manufacturing variability. The manufacturing variability is modelled as a Gaussian random field with non-stationary variance to simulate the effects of spatially varying manufacturing tolerances. The statistical performance of the compressor blade system is evaluated using the Monte Carlo method. A gradient based optimization scheme is used to deter- * Corresponding author mine the optimal blade geometry and distribution of manufacturing tolerances. NOMENCLATUREA You may include nomenclature here. α There are two arguments for each entry of the nomemclature environment, the symbol and the definition.The spacing between abstract and the text heading is two line spaces. The primary text heading is boldface in all capitals, flushed left with the left margin. The spacing between the text and the heading is also two line spaces. INTRODUCTIONThis article illustrates preparation of ASME paper using L A T E X2 ε . The L A T E X macro asme2e.cls, the BIBT E X style file asmems4.bst, and the template asme2e.tex that create this article are available on the WWW at the URL address http://iel.ucdavis.edu/code/. To ensure compliance with the 2003 ASME MS4 style guidelines [?], you should modify neither the L A T E X macro asme2e.cls nor the BIBT E X style file asmems4.bst. By comparing the output generated by typesetting this file and the L A T E X2 ε source file, you should find everything you need to help you through the preparation of ASME paper using L A T E X2 ε . Details on

The Implications of Tolerance Optimization on Compressor Blade Design

“…Such design variables include those that alter the camber and stagger of the blade. Similar behavior was observed when optimizing a subsonic rotor blade [18]. When modifying the geometry at smaller length scales, the design procedures recommended by Goodhand et al in [4] should be adopted.…”

An adjoint‐assisted multilevel multifidelity method for uncertainty quantification and its application to turbomachinery manufacturing variability

“…The squared exponential covariance has been extensively used in robust optimisation studies involving surface perturbations (see ref. 7,35,25 ). This covariance function is infinitely differentiable, which means that the GP with this covariance function has derivatives of all orders (in the mean squared sense), and is thus very smooth 36 .…”

“…A zero‐mean Gaussian process is imposed along the normal direction to this nominal surface to generate the perturbed surface as, In addition, the random variables and at any two arbitrary points i and j on the surface (see Figure 5) are assumed to have a squared exponential spatial covariance as shown in Equation (42). The squared exponential covariance has been extensively used in robust optimisation studies involving surface perturbations (see References 6,29,39). This covariance function is infinitely differentiable, which means that the GP with this covariance function has derivatives of all orders (in the mean squared sense), and is thus very smooth 40 .…”

“…Determining performance deterioration of turbomachines due to manufacturing variations is important to engine manufacturers when determining trade-offs between manufacturing cost and performance. Many studies have been conducted in the literature with focus on change in performance due to manufacturing variations on blade shapes 4,5,6,7,8,9,10 . The main challenge of modelling the effect of geometric variations on the blade performance is the large number of uncertain parameters that have to be considered.…”

“…We introduce a Gaussian process model for surface variations due to manufacturing process of turbine blades. Unlike previous work in the literature 25,23,5,7 we consider multiple cost functions in a goal-based principal component analysis of the surface variation model. We show that including multiple cost functions increases the the number of PCA modes required to achieve the specified error tolerance in the cost functions.…”