Absorbed energy capacity, and dynamics of improved perovskite solar cells: Introducing SVM-PSO-GA algorithm to predict vibrational information.

Using perovskite solar cells due to their lightweight, cost-effective, and adaptable nature, is regarded as one of the most encouraging renewable energy options. So, improving the stability of this semiconductor device is very important. One of the suggestions for improving the dynamic stability of this structure is choosing laminated composite layers with various angle ply as the metal layer. So, in this work, for the first time, absorbed energy capacity, and dynamics of improved perovskite solar cells via both mathematical modeling and a hybrid machine learning method is presented. Laminated composite layers are placed in the last layer of the solar cell so that it can improve the energy absorption and dynamics of this kind of semiconductor device. So, with this assumption, in the current work, absorbed energy capacity dynamic deflection and natural frequency sensitivity analysis of solar doubly curved cells with respect to geometrical and physical parameters is carried out. High-order shear deformation theory, which displays the parabolic transverse shear stresses throughout the thickness by satisfying the free surface requirements, is employed in this work to simulate the displacement field. The governing equations with varied boundary conditions are generated and solved, accordingly, using the virtual work principle and exact solution. After obtaining the dataset using mathematical simulation, a hybrid machine learning method using coupled support vector machine (SVM), genetic algorithm (GA), and Particle swarm optimization (PSO) is used to test, train, and validate the outputs. After that, the SVM-PSO-GA algorithm is used to validate the current work's results. Lastly, some detailed recommendations for enhancing the stability of solar cells are provided. [ABSTRACT FROM AUTHOR]