기계학습을 응용한 유동시스템 Efficacy 향상에 대한 연구

초록

According to the inherent nature of the environment of flow system, the effective flow performance is usually not predictable which, thereby requires the consideration of the flow system efficacy. In this paper, an outlet flow field from the flow system is optimized to provide most desirable efficacy based on CFD analysis. A mixed-flow fan system is used to create flow fields within the surroundings. The efficacy is approximated by the air age of the flow field for the purpose of the optimal goal. This paper focuses on the learning process of the relationship between the system exhaust velocity profiles and the ages of air via the machine learning that extracts the optimal outlet-profile from the system for the highest efficacy. Fluid systems with the same flow rate may show significant performance changes depending on the fore-mentioned exhaust-profiles. We try to find the correlation between velocity profiles and the system efficacy. This study is implemented on the air purifier design to improve the efficacy of the system.