• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.99-99
• /
• 2023

기후변화는 물 관리의 가장 큰 리스크 요인이므로 물 관리 계획을 수립하는 과정에서 기후변화의 영향을 고려하는 것이 필수적이다. 기후변화에 대한 수자원 예측 관련 연구가 이루어지고 있으나, 대부분의 연구에는 수문학적 모델링이나 시뮬레이션이 동반되는데, 이 과정에는 시간과 비용이 많이 들어가며, 지역이나 연구목적에 따른 정밀한 매개변수의 보정은 전문지식이 필요하기 때문에 현업에서 연구결과를 의사결정에 활용하기에는 한계가 있다고 볼 수 있다. 이에 따라 수문학적 모델링의 입력 및 출력 결과를 딥러닝의 학습자료로 하여 수문모델을 사용하지 않아도 효율적으로 결과를 도출할 수 있는 딥러닝 기반 Surrogate 모형에 대한 연구가 이루어지고 있으나 수자원 분야에 접목된 사례는 부재한 실정이다. 따라서 이 연구를 통해 국내 유역을 대상으로 Surrogate 모형을 구축한 뒤, 그 성능을 평가하고자 한다. 이를 위한 Surrogate 모형 구축 과정은 다음과 같다. 충주댐 유역을 대상으로 과거 20년간의 강우 및 기온 자료를 수집한 뒤, 이 자료를 바탕으로 기후변화의 영향을 고려한 3,162개의 시나리오를 생성한다. 그 후 장기유출모형 IHACRES에 생성된 시나리오를 입력자료로 하여 유입량 결과를 도출하고, 이 결과를 Python코드 기반의 딥러닝 학습자료로 하여 최적 예측 결과를 도출해내는 Surrogate 모형을 생성한 뒤 기존 장기유출모형과의 성능을 비교하고자 한다. 이와 같은 Surrogate 모형은 추가적인 데이터와 매개변수의 보정 과정이 없어도 장기유출모형과 같은 결과를 짧은 시간내에 상당히 정확하게 모사할 수 있어 시간과 비용을 줄일 수 있으며, 비전문가도 쉽게 사용할 수 있다는 장점을 가진다.

• Proceedings of the KIPE Conference
• /
• 2019.11a
• /
• pp.206-207
• /
• 2019

리튬이온 배터리의 상태를 모니터링 하는 방법에 있어서, 대표적으로 배터리의 충전 상태(SOC)와 배터리의 건강 상태(SOH)를 추정하여 상태 지표로 사용된다. 본 연구에서는 리튬 이온 배터리의 상태 지표를 위한 용량 정보의 추정을 데이터 기반의 근사 모델을 이용하여 수행하였다. 다양한 근사 모델링 방법을 적용하여 추정되는 용량 정보를 비교하고, 모델링 방법에 따른 용량 추정 성능을 확인하였다. 또한, 이를 바탕으로 리튬이온 배터리의 용량을 예측하고 예측 성능을 분석하였다. 본 연구를 통하여 근사모델을 이용하는 경우, 리튬이온 배터리의 용량 추정은 물론 예측을 수행하는 방법으로서의 활용 가능성을 확인하였으며, 또한 제안하는 방법을 이용하여 보유하고 있는 모니터링 데이터를 활용하여 리튬이온 배터리의 성능을 평가하는데 있어 효과적으로 활용될 수 있을 것으로 판단된다.

• Korean Chemical Engineering Research
• /
• v.61 no.4
• /
• pp.542-549
• /
• 2023

Since the growing interest in surrogate modeling, there has been continuous research aimed at simulating nonlinear chemical processes using data-driven machine learning. However, the opaque nature of machine learning models, which limits their interpretability, poses a challenge for their practical application in industry. Therefore, this study aims to analyze chemical processes using Explainable Artificial Intelligence (XAI), a concept that improves interpretability while ensuring model accuracy. While conventional sensitivity analysis of chemical processes has been limited to calculating and ranking the sensitivity indices of variables, we propose a methodology that utilizes XAI to not only perform global and local sensitivity analysis, but also examine the interactions among variables to gain physical insights from the data. For the ammonia synthesis process, which is the target process of the case study, we set the temperature of the preheater leading to the first reactor and the split ratio of the cold shot to the three reactors as process variables. By integrating Matlab and Aspen Plus, we obtained data on ammonia production and the maximum temperatures of the three reactors while systematically varying the process variables. We then trained tree-based models and performed sensitivity analysis using the SHAP technique, one of the XAI methods, on the most accurate model. The global sensitivity analysis showed that the preheater temperature had the greatest effect, and the local sensitivity analysis provided insights for defining the ranges of process variables to improve productivity and prevent overheating. By constructing alternative models for chemical processes and using XAI for sensitivity analysis, this work contributes to providing both quantitative and qualitative feedback for process optimization.

• Transactions of the Korean hydrogen and new energy society
• /
• v.22 no.3
• /
• pp.357-362
• /
• 2011

Complete mixture preparation of reactants prior to catalytic reforming is an enormously important step for successful operation of a fuel reformer. Incomplete mixing between fuel and reforming agents such as air and steam can cause temperature overshoot and deposit formation which can lead the failure of operation. For that purpose it is required to apply computational models describing coupled kinetics and transport phenomena in the mixing region, which are computationally expensive. Therefore, it is advantageous to analyze the gas-phase reaction kinetics prior to application of the coupled model. This study suggests one of the important design constraints, the required residence time in the mixing chamber to avoid substantial gas-phase reactions which can lead serious deposit formation on the downstream catalyst. The reactivity of various gaseous and liquid fuels were compared, then liquid fuels are far more reactive than gaseous fuels. n-Octane was used as a surrogate among the various hydrocarbons, which is one of the traditional liquid fuel surrogates. The conversion was slighted effected by reactants composition described by O/C and S/C. Finally, threshold residence times in the mixing region of a hydrocarbon reformer were studied and the mixing chamber is required to be designed to make complete mixture of reactants by tens of milliseconds at the temperature lower than $400^{\circ}C$.

• Proceedings of the Korean Society of Propulsion Engineers Conference
• /
• 2007.04a
• /
• pp.143-147
• /
• 2007

Kerosene and diesel are compounded fuels with various types of hydrocarbon elements and difficult to model the chemical kinetics. This study focuses on the prediction of the non-equilibrium reaction of fuel-rich combustion with detailed kinetics developed by Dagaut using PSR(perfectly stirred reactor) assumption. In Dagaut's surrogate model for kerosene and diesel, chemical kinetics consists of 2352 reaction steps with 298 chemical species. Also, Frenklach's soot model was implemented along with detailed kinetics to calculate the gas properties of fuel rich combustion efflux.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.49 no.9
• /
• pp.709-719
• /
• 2021

Recent progress in analysis and flight simulation methods enables wider use of a virtual certification and reduces number of certification flight tests. Aerodynamic database (AeroDB) is one of the most important components for the flight simulation. It is composed of aerodynamic coefficients at a range of flight conditions and control deflections. This paper proposes and efficient method for construction of AeroDB that combines Gaussian Process based Variable Fidelity Modeling with adaptive sampling algorithm. A case study of virtual certification of a F-16 fighter is presented. Four AeroDB were constructed using different number and distribution of high-fidelity data points. The constructed database is then used to simulate gliding, short pitch, and roll response. Compliance with certification regulations is then checked. The case study demonstrates that the proposed method can significantly reduce number of high-fidelity data points while maintaining high accuracy of the simulation.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.34 no.7
• /
• pp.89-96
• /
• 2006

The combustion temperature in gas generator should be kept below around 1,000K to avoid any possible thermal damages to turbine blade by adopting either fuel rich or oxidizer rich combustion. Thus, non-equilibrium chemical reaction dominates in the gas generator. Meanwhile, Kerosene is a compounded fuel mixed with various types of hydrocarbon elements and difficult to model the chemical kinetics. This study focus to model the non-equilibrium chemical reaction of kerosene/LOX with detailed kinetics developed by Dagaut using PSR(Perfectly stirred reactor) assumption. Also, droplet evaporation time is taken into account by calculating for the residence time of droplet and by decoupling reaction temperature from the reactor temperature. In Dagaut’s surrogate model for kerosene, chemical kinetics of kerosene consists of 1592 reaction steps with 207 chemical species. The comparison of calculation results with experimental data could provide very reliable and accurate numbers in the prediction of combustion gas temperature, species fraction and other gas properties.