超合金の高温低サイクル疲労におけるき裂分布の統計情報に基づく複合損傷シミュレーション解析

抄録

Actual components used under severe conditions often suffer from the complicated material damage affected by multiple influencing factors. “Multifactor damage simulation analysis” was proposed to solve such complex damage problems. In this paper, a discrete cluster model of material and the associated damage state matrix [D_N] are introduced to solve a high temperature low cycle fatigue damage problem of Co-base superalloy FSX414.
The damage state at Nth cycle is determined by the damage state at (N-1)th cycle expressed as follows.
[D_N]=[Λ][D_N-1]
As this equation is similar to the discrete dynamical system and operator matrix [Λ] is non-linear, the chaotic behavior, of damage evolution may occur. Instead of solving the matrix equation directly, simulation was conducted using 2-dimensional model of 2500 clusters at 5×5mm area for the dendrite and grain boundary structure of FSX414. Damage of crack initiation and growth was calculated deterministically and the randomness was only introduced at the initial condition of the material.
High temperature low cycle fatigue test was conducted to follow up the damage process at the total strain range of 1% and at the temperature of 1123K. Surface crack morphologies were investigated by the replication technique at 20% and 40% of the failure life.
The trends of crack numbers, maximum crack length and mean crack length against fatigue cycles were obtained by the simulation. These trends were complicated like chaos due to the interaction of cracks and material structure. The crack length distribution by the simulations agreed well with the experimental results and better agreement was obtained by using intermittent inspection informations. These results suggested that the method would be effective for damage prediction of actual components based on inspection informations.