• Journal of the Institute of Convergence Signal Processing
• /
• v.23 no.3
• /
• pp.179-184
• /
• 2022

In this paper, we are going to study how we can automatically determine the number of convolutional filters for the optimal model without a search algorithm. This paper proposes HSE Block by connecting SE Block proposed in SENet to a convolutional neural network and connecting a convolutional neural network not learned at the bottom. An experiment was conducted to increase the number of filters by one per 3 epoch using two datasets for the HSEBlock model and to increase the number of filters by the value in the filter. Based on this experiment, the model was constructed with multi-layer HSE Block instead of layer HSE Block, and the experiment was carried out using a dataset that was more difficult to learn than the one used in the previous experiment. The effect of HSE Block was verified by conducting an experiment with the number of HSE Blocks set to 2, 3, 4, and 5 on a dataset that is more difficult to learn than before.

• Journal of the Korean Society of Systems Engineering
• /
• v.19 no.2
• /
• pp.74-91
• /
• 2023

This study provides an in-depth comparison and analysis of various risk assessment models widely used in modern industries, and proposes the most suitable model for risk assessment of offshore wind power in Korea. The assessment models were selected by considering various factors such as the purpose of risk assessment, stakeholder requirements, and characteristics of offshore wind power. We also emphasized the importance of using different risk assessment models in combination in situations of high uncertainty. To systematize the combination of risk assessment models, we used systems engineering which is effective to develop a new system. Systems engineering was used to define the complete, traceable functions from site requirements, and model-based systems engineering was used to manage the design information from requirements to detailed functions in a single model. The developed risk assessment module provide automatic conversion between risk assessment models to enable risk assessment suitable for offshore wind power. The functionality and usability of the offshore wind risk assessment module were verified by the evaluation of three wind power experts.

• Journal of Industrial Technology
• /
• v.19
• /
• pp.253-260
• /
• 1999

Two thermodynamic models were used to predict the partial molar volumes of solutes in supercritical carbon dioxide at infinite dilution: (1) the Peng-Robinson equation of state with various mixing rules including those based on $EOS/G^E$ (2) the Kirkwood Buff fluctuation integral with the hard sphere expansion (HSE) method. The Kirkwood-Buff fluctuation integral method, in which an equation of state for pure component and molecular parameters are required, produced better results especially near the critical point than the Peng-Robinson equation of state with the several mixing rules based an $EOS/G^E$. When the $EOS/G^E$ mixing rules were used, poorer results were obtained compared with the classical mixing rule and Kirkwood-Buff model.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.37 no.1
• /
• pp.1-8
• /
• 2024

This paper employs stochastic processing techniques to analyze explosion risks in plant facilities based on explosion return periods. Release probability is calculated using data from the Health and Safety Executive (HSE), along with annual leakage frequency per plant provided by DNV. Ignition probability, derived from various researchers' findings, is then considered to calculate the explosion return period based on the release quantity. The explosion risk is assessed by examining the volume, radius, and blast load of the vapor cloud, taking into account the calculated explosion return period. The reference distance for the design blast load model is determined by comparing and analyzing the vapor cloud radius according to the return period, historical vapor cloud explosion cases, and blast-resistant design guidelines. Utilizing the multi-energy method, the blast load range corresponding to the explosion return period is presented. The proposed return period serves as a standard for the design blast load model, established through a comparative analysis of vapor cloud explosion cases and blast-resistant design guidelines. The outcomes of this study contribute to the development of a performance-based blast-resistant design framework for plant facilities.

• Journal of Energy Engineering
• /
• v.28 no.4
• /
• pp.29-34
• /
• 2019

In this study, a quantitative risk assessment was carried out for a hydrogen complex station. The complex fueling station to be evaluated was hydrogen-LPG, and the components of each station were analyzed and the risk was evaluated. The final risk is assessed by individual and societal risks, taking into account the impact of damage and the frequency of accidents. As a result of individual risk calculation for the hydrogen-LPG fueling station that is the subject of this study, the hydrogen-LPG type fueling station does not show the unacceptable hazardous area (> 1 × 10E-3) proposed by HSE. The level of individual risk for both the public and the worker is within acceptable limits. In societal risk assessment, the model to be interpreted shows the distribution of risks in an acceptable range(ALARP, As Low As Reasonably Practicable). To ensure improved safety, we recommend regular inspections and checks for high-risk hydrogen reservoirs, dispensers, tube trailer leaks, and LPG vapor recovery lines.