Optimisation of the growth conditions for manufacturing of single-wall carbon nanotubes (SWCNTs) was challenging until recent demonstration of “High-throughput screening and machine learning for the efficient growth of high-quality single-wall carbon nanotubes” by Ji et al. 2021 in Nano Research (DOI:10.1007/s12274-021-3387-y). The high-throughput screening of growth conditions was conducted by depositing patterned cobalt (Co) nanoparticles on a marked silicon wafer catalysts and varying the temperature, reduction time, carbon precursor, and growth time during chemical vapor deposition. The quality (crystallinity) of the SWCNTs was characterised by the G/D peak intensity (IG/ID) measured by Raman spectroscopy. 1664 samples were used to train and validate machine learning models for prediction of the quality resulting from a particular combination of growth parameters. Here, we expand the work and train an artificial neural network with improved prediction accuracy. The quality depends on the growth parameters in a non-linear fashion with multiple local optima, explaining why it is so hard to optimise the growth conditions without machine learning. With AI growth conditions for high-quality SWCNTs were identified.

In the Laser Powder Bed Fusion (L-PBF) process, the interactions between the metal vapor dynamics, induced gas flow dynamics, and the dynamics of the powder particles play an essential role in dictating the printability, stability, and quality of the built process. Complex phenomena such as the powder entrainment, spattering, and the powder bed denudation have significant effects on the roughness of the printed surface, porosity of the built part, and even the success of the printing process. However, very few efforts have been reported on studying those factors and their influences. In this study, a comprehensive three-dimensional (3D) CFD-DEM coupled simulation is proposed to study the induced particle dynamics under the effects of the metal vapor spouting during the process of the single scan track. The model proposed in this work has the capability to consider non-conventional process variables such as the chamber pressure or the gravitational force to study their effects on spatter and denudation suppressions. In addition, the 3D model provides the ability to study physical phenomena that are very difficult to measure via experimental approach such as the induced gas flow velocity surrounding the metal vapor hole and the velocity and temperature of the entrained particles. Beside the capability to simulate the powder entrainment, spatter and denudation formation, to our best knowledge, the highly comprehensive model proposed in this study is the first 3D model that has a more realistic initial condition of the powder bed and the metal vapor compared to other 3D model proposed in literature, and the first model that can mimic the powder absorption of the melt-pool.

Relying on experimental trial and error for finding optimal parameters that can produce fully dense IN718 components is expensive and time consuming. The present study built up a simulation package including micro-scale simulation, part scale simulation to guide the choice of processing parameters to produce high density component in SLM process. The suggested parameters from simulation model are used to fabricate the 3D part in SLM process and the density of fabricated part is measured by using both Archimedes and microscopy method. It is observed that the density of fabricated part can reach 99.9 % by using just one experiment. Thus the time and cost can reduce significantly.

鑒於台灣現在農村社會勞動力人口減少加上農民普遍高齡化，會讓水果採收的良率降低且增加錯過最佳成熟時機點的機率，此次專題目的是先以番茄作發想而達到提供基礎影像辨識以邁向水果採收自動化。此程式為Python的架構，引進opencv作為顏色辨識的基礎，第一步先以HSV區隔出顏色，接著以圓形辨識圈出番茄位置，但非實際番茄大小，再以輪廓程式將圈出番茄的圓圈改以方框框出、最後將先前所有圓形圈出的番茄進行顏色比例的換算，並顯示在輸出圖片上的方框上方。目前是以五張圖片作為程式撰寫的基準圖，而目前能抓出成熟番茄位置的成功率依序為“100%”、“50%”、“100%”、“50%”、“100%”；非成熟番茄的成功率則依序為“0%”、“50%”、“75%”、“16.67%”、“0%”，但在顏色比例上，因目前程式的算法撰寫上尚未完成，所以在同一張照片中，成熟及非成熟的番茄皆各顯示一種數字而已。總結來說，以上皆為未加入機械學習的方式來找出番茄位置及顏色比例，因此單純靠顏色分別、輪廓辨識來算，結果是達不到預期的，需加入機械學習才能增加精準度，但我們認為在此架構上，如果為辨識位置固定、位置皆非重疊的物品上，尚有一定的辨識成功率，且程式撰寫上也較為簡易；不過我們也認為將如果此程式撰寫完整後，可連接上機械手臂，能進一步實驗透過程式自動化採收水果的步驟，其成功的話是必有極大的實用性的。