Objectives Based on the discrete-module-beam (DMB) method, this study aims to address the connector stiffness optimization problem of modular box-pontoon-type offshore floating photovoltaic (OFPV) platforms. The floating foundation of such OFPV systems consists of multiple modular units connected by connectors. The connector stiffness design significantly impacts the platform’s structural safety and reliability under complex ocean environments.
Methods This research employs comprehensive methods. First, the DMB hydroelastic analysis method is introduced. The connector stiffness matrix form is given, with stiffness values for different degrees of freedom defined, and the numerical modeling method for hydroelastic response briefly outlined. This method allows for efficient calculation of the structure's response to environmental loads. Second, two genetic-algorithm-based approaches are presented. The linear-weighted genetic algorithm converts the multi-objective optimization problem into a single-objective one by assigning weights to different objectives. The NSGA-II (non-dominated sorting genetic algorithm II) is used as a multi-objective optimization algorithm, which can identify a set of Pareto-optimal solutions instead of a single one. Three encoding techniques for stiffness, namely real encoding, exponent encoding, and scientific notation encoding, are elaborated. Each encoding method has its own crossover and mutation operators. For example, real encoding directly operates on stiffness values, while exponent encoding and scientific notation encoding have unique operation mechanisms. The performance of these encoding methods is compared through population initialization and individual distribution analysis across different evolutionary generations. In addition, the concept of equivalent zero stiffness and equivalent infinite stiffness is introduced to reduce the solution space. This enhances the efficiency of the optimization process.
Results The results show that the NSGA-II algorithm can obtain the Pareto front with the objectives of minimizing the maximum structural shear force and minimizing the maximum structural bending moment. The Pareto front can be regarded as a set of optimal solutions corresponding to different weight settings obtained by the linear-weighted method. Analysis of population initialization and individual distribution reveals that the scientific notation encoding performs better in terms of search efficiency within the solution space. It can explore a wider range of stiffness values, including both low and high magnitudes, compared to the other two encoding methods.
Conclusions In conclusion, the developed optimization theory model is effective in performing multi-objective optimization on the connector stiffness of modular box-pontoon-type OFPV platforms. The scientific notation encoding provides a more efficient way to search for optimal solutions. However, it should be noted that the OFPV system is complex, and future research can focus on considering additional objectives and combinations to further optimize the design. This research provides a valuable reference for the design and optimization of floating-type photovoltaic platforms in ocean engineering.
DownLoad: