HYIMFO: Hybrid method for optimizing fiber orientation angles in laminated piezocomposite actuators

2021 
Abstract Numerous engineering applications employ composite materials due to the advantages that the combination of their components can provide. Fiber-reinforced composites are a particular group of composites which allows designing optimized structures by tailoring the orientation of the fiber. For this reason, new additive manufacturing technologies have been developed nowadays. These tools allow the manufacturing of fiber-reinforced structures designed by using optimization algorithms. Consequently, there are several works in recent literature determining the optimized fiber orientation in composite structures, where the methods based on gradient stand out due to their better efficiency. There are two main approaches to determine the fiber orientation by using gradient-based methods, Continuous Fiber Angle Optimization (CFAO), where the angle is the design variable and commonly present local minima problems, and Discrete Fiber Angle Interpolation Models , which consider candidate angles  a priori and by using weighting functions choose one of them as the optimized angle. This work proposes a method named HYMFO, which uses a combination of both approaches. To verify the efficiency of the method, two design problems of Laminated Piezocomposite Actuators (LAPA) are solved by using several mesh discretizations with the discrete methods DMO , HPDMO, BCP, and NDFO, the continuous method SPIMFO and the newly proposed hybrid method . Therefore, the method is not evaluated with the traditional problem of maximizing the stiffness in structural design. A new fiber orthogonality regularization is presented, facilitating the appearance of continuous paths in fiber distribution by imposing a specified maximum angle between adjacent fibers. A comparative analysis based on the performance of the structure and the computational cost is carried out to determine the most appropriate method for each situation.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    73
    References
    0
    Citations
    NaN
    KQI
    []