Optimal mesh generation for a blade passage using deep reinforcement learning

A mesh generation method that can generate an optimal mesh for a blade passage at a single attempt is developed using deep reinforcement learning (DRL). Unlike the conventional methods, where meshing parameters must be specified by the user or iteratively optimized from scratch for a newly given geometry, the developed method employs DRL-based multi-condition (MC) optimization to define meshing parameters for various geometries optimally. The method involves the following steps: (1) development of a base algorithm for structured mesh generation of a blade passage; (2) formulation of an MC optimization problem to optimize meshing parameters introduced while developing the base algorithm; and (3) development of a DRL-based mesh generation algorithm by solving the MC optimization problem using DRL. As a result, the developed algorithm is able to successfully generate optimal meshes at a single trial for various blades.

Multi-condition multi-objective optimization using deep reinforcement learning

A multi-condition multi-objective optimization method that can find Pareto front over a defined condition space is developed for the first time using deep reinforcement learning. Unlike the conventional methods which perform optimization at a single condition, the present method learns the correlations between conditions and optimal solutions. The exclusive capability of the developed method is examined in the solutions of a novel modified Kursawe benchmark problem and an airfoil shape optimization problem which include nonlinear characteristics which are difficult to resolve using conventional optimization methods. Pareto front with high resolution over a defined condition space is successfully determined in each problem. Compared with multiple operations of a single-condition optimization method for multiple conditions, the present multi-condition optimization method based on deep reinforcement learning shows a greatly accelerated search of Pareto front by reducing the number of required function evaluations. An analysis of aerodynamics performance of airfoils with optimally designed shapes confirms that multi-condition optimization is indispensable to avoid significant degradation of target performance for varying flow conditions.