運用類神經網路建置後設模型並離線最佳化海底渦輪機模擬系統參數

Offline optimization of marine turbine parameters based on Neural network

陳奕廷(I-Ting Chen)

指導教授 ： 洪一薰

國立臺灣大學/工學院/工業工程學研究所/碩士(2020年)

https://doi.org/10.6342/NTU202002532

摘要

隨著環保意識抬頭，綠色能源逐漸興起，其中海流能被視為極具潛力的選項，然而發展海流能需要大量的物理實驗數據輔助，藉由電腦模擬物理實驗又缺乏效率，因此本研究期望以類神經網路架設後設模型模擬超級電腦模擬系統，已大幅增加研究效率。本研究主要分為三個部分，首先對原資料進行篩選和整理，以剔除失準資料並將資料調整成適合後續訓練的型態，接著以電腦模擬系統為目標架設後設模型，最後利用粒子團演算法和基因演算法以近似物理實驗為目標找出最佳參數，除此之外以類神經網路的後設模型為基礎，成功將後設模型轉換為非線性模型的型式，求解後證實非線性模型有機會找出優於萬用啟發式演算法的參數解。此研究架構能夠在提高效率又不失精準度的情況下，找出模擬系統的參數解。

關鍵字

非線性模型 ； 類神經網路 ； 後設模型 ； 粒子團演算法 ； 基因演算法 ； 離線最佳化

並列摘要

With the rise of environmental awareness, green energy is gradually emerging, especially the ocean current energy, which is regarded as a very promising option. However, the development of the ocean current energy requires a large amount of physical experiment data, and the computer simulation of physical experiments is inefficient. This paper utilizes the neural network to develop our surrogate model describing the response surface between the design parameters of the simulation and experimental quantities of interests, and the efficiency has been greatly increased. This study is mainly divided into three parts. First, remove the inaccurate data and adjust the data to a form that is suitable for training. Then, construct the surrogate model with the goal of approximating simulation system. Last, find the optimal solution of the simulation with the goal of approximating physical experiment system by using the particle swarm algorithm, the genetic algorithm and the nonlinear model, and finally confirm that the nonlinear model has the opportunity to find a better solution than the metaheuristics.

並列關鍵字

nonlinear model ； neural network ； surrogate model ； particle swarm algorithm ； genetic algorithm ； offline optimization

參考文獻

Ashby, S. F., Bosl, W. J., Falgout, R. D., Smith, S. G., Tompson, A. F. B., Williams, T. J. (1999). A numerical simulation of groundwater flow and contaminant transport on the CRAY T3D and C90 Supercomputers. The International Journal of High Performance Computing Applications, 13(1), 80-93.

Chau, K. W. (2007). Application of a PSO-based neural network in analysis of outcomes of construction claims. Automation in construction, 16(5), 642-646.

Chow, T. T., Zhang, G. Q., Lin, Z., Song, C. L. (2002). Global optimization of absorption chiller system by genetic algorithm and neural network. Energy and buildings, 34(1), 103-109.

Cui, Y., Olsen, K. B., Jordan, T. H., Lee, K., Zhou, J., Small, P., ... Levesque, J. (2010, November). Scalable earthquake simulation on petascale supercomputers. In SC'10: Proceedings of the 2010 ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis (pp. 1-20). IEEE.

Das, G., Pattnaik, P. K., Padhy, S. K. (2014). Artificial neural network trained by particle swarm optimization for non-linear channel equalization. Expert Systems with Applications, 41(7), 3491-3496.