Session: 26-01 Probabilistic and Machine Learning Methods Development and Applications

Paper Number: 121684

121684 - Bayesian Inference for the Calibration of Progressive Damage Model of Dovetail Specimens From Laminated Composite Fan Blade 

    Compared with the traditional titanium fan blade, laminated composite fan blades are characterized by high specific strength and stiffness, good designability, better flutter resistance and outstanding durability. It is the preferred alternative for the fan stage of most advanced high bypass ratio turbofan engines in the world. Under service operating conditions, the dovetail part bears a significant centrifugal load, which is one of the key points of the multi-level "building block" approach of compliance demonstration for special airworthiness regulation. From the perspective of structure features, there are a large number of layers and ply drop-offs inside due to the rapid thickness transition in blade root. Moreover, the competing failure modes of composite materials further lead to high discrepancy and difficulty in predicting the ultimate strength of the dovetail element.

    Progressive damage is associated with the damage phenomenon that occurs in composite materials. The respective progressive damage analysis (PDA) could effectively simulate the process of laminated composite material from damage initiation to the ultimate failure, and is widely used to predict the ultimate strength of composite aerospace structures. Therefore, the damage initiation failure criterion and theoretical stiffness degradation model based on Continuum Damage Mechanics (CDM) directly determine the orthotropic stiffness attenuation status during the progressive damage process of composite structures. The parameters incorporated into the instant stiffness reduction via material property degradation or gradually softening behavior based on dissipated energy is extremely important for accurately predicting the bearing capacity of the specimen.

       In this work, the actual dovetail structure of a laminated composite fan blade is simplified in order to conduct radial tensile strength tests, which simulates the interacting centrifugal force. Load-displacement curves of these non-standard specimens are combined with finite element models with embedded progressive damage analysis, where Hashin failure criterion is applied to capture intralaminar damage initiation for instant material property degradation and continuum damage evolution respectively. Posterior probability distributions of stiffness degradation parameters were adjusted based on Bayesian inference. After making comparisons between the calibrated model representing the as-manufactured damage properties and experimental data, the inverse results indicate that the two failure criteria have a relatively small impact on the identification results of the critical tensile load, while they have a significant impact on the identification results of the initial damage load; Bayesian inference could not only consider the interference of modeling errors and measurement noise, but need a relatively scarce amount of experimental data to conduct a inverse analysis, then obtain more accurate important prior probability distributions. The proposed methodology may lay solid foundation for the subsequent forward uncertainty quantification of critical quality parameters of composite fan blade.

Presenting Author: Xu Tang School of Mechanical Engineering, Shanghai Jiao Tong University

Presenting Author Biography: The authors have expertise in laminate design, high-fidelity structural FE simulation and airworthiness certification under special regulations of advanced turbofan engine woven/laminated composite fan blade design features, e.g. static strength, vibration characteristics, low/high cycle fatigue, bird impact induced dynamic response, aeromechanics, etc.

Authors:

Xu Tang School of Mechanical Engineering, Shanghai Jiao Tong University
Yong Chen a) School of Mechanical Engineering, Shanghai Jiao Tong University; b) Engineering Research Center of Gas Turbine and Civil Aero Engine