An Automated Materials Optimization Approach for Large, Lightweight, Additively Manufactured Direct Drive Generators with Triply Periodic Minimal Surfaces

Electric generators account for over 98% of all worldwide energy production. Wind energy provided 7.2% of energy in the United States in 2019 and growth of wind energy will be vital in combatting climate change. Advances in additive manufacturing has allowed complex topologies with lattices to behave in ways different or better than its bulk material. In applications such as wind energy where low rotational speeds are present or large off axis forces threaten a gearbox lifespan, the shaft is directly coupled to the generator in a direct drive coupling system. As the United States begins to open its coastlines to offshore wind farms, the number of taller, higher power wind turbines will increase. Major wind turbine manufacturers are creating direct drive systems for these large turbines due to their increase in reliability and fewer moving parts. However, the increase in reliability comes at the cost of mass---specifically structural mass. For example, a 5 MW direct drive generator can weigh 250 tons with structural mass comprising up to 80% of the total. From a materials perspective, such a high structural mass proportion makes direct drive generators a prime application for light weighting. Triply periodic minimal surfaces (TPMS) are surfaces which locally minimize their area. They have been active areas of research due to their high strength to weight ratio and can be used for light weighting. Manufacturing of these TPMS structures is difficult conventionally yet done with ease in additive manufacturing. This work explores developing direct drive generator rotor and stator support structures using an automated optimization scheme and finite element analysis to determine the proper TPMS parameters for maximum stiffness at minimum mass. The TPMS structure is varied in order to determine topology influence on bulk properties. First, the generator support structures are created digitally using implicit modeling for speed of lattice generation. Next, finite element analysis collects the deformations and masses associated with the TPMS structure. Finally, a fully automated optimization routine reads the data and determines the optimal TPMS structure for the rotor and stator. For a 5 MW generator, this approach yields a 34% reduction in structural mass compared to conventional manufacturing techniques. A hybrid additive manufacturing approach was used to create these designs for experimental verification. In this case, a mold was additively manufactured and metal cast using conventional means. This combination enables structures with the complexity of additive manufacturing while maintaining the cost effectiveness of traditional casting. Scaled prototypes of the designs were tested using digital image correlation and compared to their corresponding finite element analysis. Both results agreed within 6.7% for deformations with equivalent loading conditions. This method of optimizing and manufacturing TPMS structures for large direct drive machines holds promise for reducing the mass of direct drive machines. As the generator is in the nacelle at the top of the tower, if it is made lighter, the downstream design can be made lighter and cheaper. This helps reduce their total cost making wind energy a more competitive renewable energy source.