報告書番号

M15004

タイトル（和文）

タービン翼の温度推定における不確かさ評価と簡便な温度推定式の導出

タイトル（英文）

Uncertainty quantification of turbine blade temperature estimation and simplified temperature estimation

概要 （図表や脚注は「報告書全文」に掲載しております）

タービン翼温度推定の信頼性を定量化することを目的に，タービン翼の1次元解析に多項式カオス展開を適用し，モンテカルロ法との比較からその有用性を評価した．ついで，タービン翼の2次元解析に多項式カオス展開を適用し，翼温度分布の信頼範囲を明らかにした．多項式カオス展開の中で導出される代理モデルは数値解析結果を良好に予測し，代理モデルを用いることで，将来運転条件が変化した際の影響を，新たに数値解析を行うことなく迅速に推定できる可能性が示された．

概要 （英文）

Polynomial chaos expansion, which is a computationally efficient method for performing uncertainty analysis on large complicated models, is applied to 1D gas turbine blade heat transfer model and it's availability for uncertainty quantification of blade temperature is evaluated from a comparison with results of Monte Carlo simulation. TBC thickness, gas temperature and blowing ratio are selected as uncertain input parameters. The polynomial method gives good approximations for the probability density function of metal surface temperature with less than 6 model calculations as compared with the Monte Carlo simulation of 10,000 model calculations. The polynomial method is then applied to a 2D turbine blade conjugate heat transfer model, and an uncertainty of blade surface temperature associated to the TBC thickness and the inlet gas temperature is evaluated with 6 model runs. A polynomial approximation (surrogate model) obtained in the polynomial method predicts computational results with great accuracy. An estimation error of blade surface temperature by the surrogate model is less than 0.3 %. It is concluded that the polynomial method is useful for uncertainty analysis of turbine blade temperature.

報告者

キーワード