Concrete Crack Detection with Dense Block and SE Attention Mode

HUANGRUI Sungkyunkwan University 2023 국내석사

In order to enhance the accuracy and robustness of the current crack detection algorithm, we propose an improved algorithm based on FC-Densenet(Fully Convolutional Densenet), named the Densely Connected Network with Attention Mode(DCAN), for pixel-level crack detection. We have incorporated SE attention modules at five positions within the FC-Densenet framework. Additionally, we have collected a private dataset specifically focusing on tiny cracks, which closely resembles real-world crack scenarios. To validate the effectiveness of our method, we conducted a series of experiments on three publicly available crack datasets as well as our private dataset. Compared to the baseline neural network, our proposed approach demonstrates superior performance across six evaluation metrics. We observed that as the dataset size increases, the advantages of our method become more pronounced in terms of the mIoU and F1 metrics. For instance, on the Crack500 and Cross datasets, our method achieved F1 scores of 83.24 and 80.40, respectively. Additionally, the mIoU scores reached 71.11 and 68.80 on these two datasets, respectively.

Development of Smart Exsolution Perovskite Catalysts for Water Gas Shift Reaction

Huang, Rui 포항공과대학교 일반대학원 2022 국내박사

Perovskite exsolution catalysts have many advantages pertaining to the Cr-free High Temperature Water Gas Shift (HT-WGS) reaction. For example, exsolved NPs are distributed uniformly on parent perovskite oxide surfaces. Therefore, they provide homogeneous and stable active sites for HT-WGS, avoiding serious aggregation with increases in temperature or operating time. Besides, reduction treatment leaves abundant oxygen vacancies on these surfaces. Such vacancies can boost CO adsorption and oxidation, thereby increasing HT-WGS activity. Furthermore, strong adhesion of exsolved particles on support surfaces increases catalyst coking resistance and thermal stability during HT-WGS. Despite these recent developments in perovskite exsolution catalysts for methane reforming, the use of such catalysts for HT-WGS has seldom been reported. In this thesis paper, LaFeO3 perovskite oxides are selected as parent supports, while Ni serves as a perovskite B-site dopant. Firstly, the feasibility of applying exsolution to HT-WGS (Chapter 2) was investigated. The potential of such exsolved catalysts was confirmed, as both their activity and stability were proven to be better than those of commercial catalysts. Therefore, a series of strategies were designed to improve exsolution in catalysts. For the second step, the effect of reduction duration on catalytic activity was studied (Chapter 3). As reduction duration increased, the sizes of exsolved NPs did not change significantly, while primary perovskite oxide structures were maintained. However, more Ni particles were nucleated, thus more NPs were formed on oxide surfaces. H2-TPR results predicted that oxygen-consuming exsolution processes had already been completed within 5 h. After that, Ni NPs nucleation and growth mainly occurred under reduction. Based on O2-TPD, CO-TPD, and CO-TPR analyses, exsolved metals verifiably and importantly function as active sites for CO adsorption and oxygen delivery, namely by dissociating H2O. To further develop exsolved metal activity, their catalytic mechanisms need to be characterized. Such considerations require the confirmation of catalyst active sites during the water gas shift reaction. In Chapter 4, active sites are quickly qualified in examples of fundamental supported noble metal catalyst systems. The mechanistic origins of high WGS activity, caused by metal surface interaction (MSI), were investigated using density functional theory (DFT) calculations. MSI was found to be essential for high WGS activity. With the understanding obtained from Chapter 4, WGS reaction mechanisms on exsolved catalysts were further discussed in Chapter 5. By annealing Pr0.4Sr0.6CoxFe0.9-xNb0.1O3-δ (PSCxFN) solid precursors in hydrogen gas environments at high temperatures, perovskite oxides with exsolved metal nanoparticle catalysts were synthesized. Fe or Fe/Co alloy nanoparticles exsolved to the surface after reduction, and oxygen vacancy concentrations in oxide matrices increased with Co doping level at B-sites. HT-WGS activity showed a volcano-shaped dependence with Co doping level in PSCxFN. DFT calculations revealed interfaces between exsolved metal nanoparticles and oxide matrices were active sites for water dissociation reaction steps, as well as that reaction barriers decreased with Co doping level. These results provide critical insights into how metal-oxide interfaces impact the catalytic activity of exsolution materials. Such a mechanistic understanding can guide the design of exsolution materials for other energy and environmental devices. 페로브스카이트 용출 촉매는 Cr이 없는 고온 수성가스 전환반응 (HT-WGS) 촉매에 대해 많은 이점을 가질 수 있다. 예를 들어, 용출된 나노입자는 페로브스카이트 산화물 지지체의 표면에 균일하게 분포되어 온도 또는 반응 시간이 증가함에 따라 심각한 응집 현상 없이 균일하고 안정적인 활성체를 제공한다. 뿐만 아니라, 환원 처리는 표면에 풍부한 산소 결함을 남긴다. 이로 인한CO 흡착 및 산화를 촉진하여 HT-WGS의 활성을 높일 수 있다. 또한, 용출 입자가 지지체 표면에 강력하게 결합되어 HT-WGS의 코킹 저항 효과와 열적 안정성이 증가한다. 그럼에도 불구하고 최근 메탄 개질을 관련된 페로브스카이트 용출 촉매의 개발은 활발하지만 HT-WGS 촉매에 대해서는 보고된 바가 적다. 이 논문에서는 LaFeO3 페로브스카이트 산화물은 지지체로 선택되고 니켈은 B 사이트의 도펀트로 선택된다. 먼저 2장에서 HT-WGS에 대한 용출 촉매가 적용할 가능성을 조사하였다. 상용 촉매보다 활성과 안정성이 모두 우수한 것으로 입증돼 잠재력이 큰 것으로 확인됐다. 따라서 exsolution 촉매 능력을 향상시키기 위해 일련의 전략이 설계되었다. 3장에서 촉매 활성에 대한 환원 처리 시간의 영향에 대해서 연구했다. 환원 시간이 증가함에 따라 용출된 나노입자의 크기는 현전히 변하지 않았으며 페로브스카이트 산화물 구조는 유지되었다. 그러나 더 많은 Ni 입자핵이 생성되어 표면에 더 많은 입자가 형성되다. H2-TPR 결과는 산소 소모하면서 입자의 용출 과정은 5시간 이전에 이미 완료되었으며, 그 이후 시간에는 환원 상태에서 Ni 입자의 핵 형성 및 성장하는 과정으로 추측된다. O2-TPD, CO-TPD 및 CO-TPR 분석에 바탕하여, 용출 금속은 물의 해리, 일산화탄소의 흡착 및 산소의 전달을 위한 활성체로서 중요한 기능을 수행하는 것으로 입증되었다. 이러서 촉매 활성을 더 발전시키려면 촉매의 반응 메커니즘을 고려해야 한다. 이를 위해 WGS의 활성체를 확인해야 한다. 4장에서는 활성체를 신속하게 확인하기 위해 널리 사용된 귀금속/지지체 촉매 시스템을 예로 들어 설명했다. 금속 표면 상호 작용(MSI)로 인한 높은 WGS 활성의 기원은 밀도 범함수 이론 (DFT) 계산을 사용하여 조사되었다. MSI는 높은 WGS 활동에 필수적인 것으로 밝혀졌다. 4장에서 얻은 기본적 이해를 바탕으로 용출 촉매에 대한 WGS 반응 메커니즘은 5장에서 더 논의되었다. 고온의 수소 가스 환경에서 Pr0.4Sr0.6CoxFe0.9-xNb0.1O3-δ (PSCxFN) 고체 전구체를 열처리하여 금속 나노입자 촉매가 용해된다. Fe 또는 Fe/Co 합금 나노입자는 표면으로 용출되었고, 산화물 매트릭스의 산소 공공 농도는 B-site에서 Co 도핑 수준에 따라 증가하였다. HT-WGS 활성은 PSCxFN에서 Co 도핑 수준에 대한 화산 모양 의존성을 보여주었다. DFT 계산은 용출된 금속 나노입자와 산화물 매트릭스 사이의 계면이 물 해리 반응 단계의 활성체로 확인되었으며 반응 장벽은 Co 도핑 수준에 따라 감소함을 보여주었다. 용출 물질의 촉매 활성을 결정하는 데 있어 금속 산화물 계면의 역할에 대한 중요성을 검증하였다. 이러한 기계론적 이해는 다른 에너지 및 환경 장치용 용출 촉매의 설계에 참고할 수 있다.

Vibrations of piezoelectric crystal plates with thickness-graded material properties

Huang, Rui Princeton University 2001 해외박사(DDOD)

Approximate two-dimensional equations for piezoelectric crystal plates with thickness-graded material properties are deduced from the three-dimensional theory of linear piezoelectricity by using Mindlin's general procedure of series expansion. Closed form solutions are obtained from the approximate equations for vibrations of piezoelectric crystal plates with different thickness-graded material properties. First, the vibrations of asymmetric bimorph plates of piezoelectric ceramics are studied by using the first-order two-dimensional equations from power series expansions, which were obtained in a previous study. Two correction factors are used in the first-order equations to adjust the cut-off frequencies of thickness-shear and thickness-stretch modes. To eliminate the needs for corrections factors, a new system of two-dimensional equations is deduced by using the new series expansions of mechanical displacements and electric potential, which were introduced in a previous study for homogeneous plates. First and second-order equations are extracted and they are reduced for homogeneous plates with uniform material properties. The first-order equations for homogeneous plates are employed to study the thickness-shear vibrations of AT-cut quartz plates in contact with a viscous liquid layer. The second-order equations are employed in the study of extensional vibrations of piezoceramic disks. The mechanical effects of electrodes on thickness-shear, flexural, and face-shear vibrations of AT-cut quartz plates are studied by using the first-order equations, in which both the mass density and the elastic stiffness of the electrodes are taken into account. Computational results from the two-dimensional equations are compared with the exact solutions or experimental data, whichever applicable. Agreements are close within certain frequency ranges.

Nonlinear Model Predictive Control and Dynamic Real Time Optimization for Large-scale Processes

Huang, Rui Carnegie Mellon University 2010 해외박사(DDOD)

This dissertation addresses some of the theoretical and practical issues in optimized operations in the process industry. The current state-of-art is to decompose the optimization into the so-called two-layered structure, including real time optimization (RTO) and advanced control. Due to model discrepancy and inconsistent time scales in different layers, this structure may render suboptimal solutions. Therefore, the dynamic real time optimization (D-RTO) or economically-oriented nonlinear model predictive control (NMPC) that directly optimizes the economic performance based on first-principle dynamic models of processes has become an emerging technology. However, the integration of the first-principle dynamic models is likely to introduce large scale optimization problems, which need to be solved online. The associated computational delay may be cumbersome for the online applications. We first derive a first-principle dynamic model for an industrial air separation unit (ASU). The recently developed advanced step method is used to solve both set-point tracking and economically-oriented NMPC online. It shows that set-point tracking NMPC based on the first-principle model has superior performance against that with linear data-driven model. In addition, the economically-oriented NMPC generates around 6% cost reduction compared to set-point tracking NMPC. Moreover the advanced step method reduces the online computational delay by two orders of magnitude. Then we deal with a realistic set-point tracking control scenario that requires achieving offset-free behavior in the presence of plant-model mismatch. Moreover, a state estimator is used to reconstruct the plant states from outputs. We propose two formulations using NMPC and moving horizon estimation (MHE) and we show both approaches are offset-free at steady state. Moreover, the analysis can be extended to NMPC coupled with other nonlinear observers. This strategy is implemented on the ASU process. After that, we study the robust stability of output-feedback NMPC in the presence of plant-model mismatch. The Extended Kalman Filter (EKF), which is a widely-used technology in industry is chosen as the state estimator. First we analyze the stability of the estimation error and a separation-principle-like result indicates that the stability result is the same as the closed-loop case. We further study the impact of this estimation error on the robust stability of the NMPC. Finally, nominal stability is analyzed for the D-RTO, i.e. economically-oriented NMPC, for cyclic processes. Moreover, two economically-oriented NMPC formulations with guaranteed nominal stability are proposed. They ensure the system converges to the optimal cyclic steady state.

Characteristics analysis and optimization of the synchronous permanent magnet planar motor with Halbach magnet arrays

Huang Rui 창원대학교 2007 국내석사