• 제목/요약/키워드: SILO

검색결과 182건 처리시간 0.025초

원통형 구조물의 발파해체설계에 대한 최신 발파해체 시뮬레이션 기법의 적용 (Application of Advanced Blast Demolition Simulation Method to the Drill and Blast Design for Demolishing Cylindrical Structures)

  • 박훈;석철기;김승곤
    • 화약ㆍ발파
    • /
    • 제26권1호
    • /
    • pp.7-14
    • /
    • 2008
  • 사일로와 같은 대단면 원통형 구조물을 전도시키기 위해서는 구조적 특성 및 사전 취약화를 위한 개구부의 조건을 고려하여 설계해야 한다. 본 연구에서는 원통형 구조물을 전도시키기 위해 구조적으로 동일한 원통형 구조물을 상용해석 프로그램인 3D AEM으로 모델링하여, 개구부의 조건으로는 개구부의 높이, 개구부의 각도, 개구부의 형태를 변수로 설정하여 이들 변수에 따른 원통형 구조물의 전도 붕괴 거동을 모사하고 분석하였다.

Patterns between wall pressures and stresses with grain moisture on cylindrical silo

  • Kibar, Hakan
    • Structural Engineering and Mechanics
    • /
    • 제62권4호
    • /
    • pp.487-496
    • /
    • 2017
  • The focus of this study were to investigate patterns between wall pressures and stresses with grain moisture of soybean and rice varieties widespread cultivated in Turkey in order to determine needed designing parameters for structure analysis in silos at filling and discharge. In this study, the wall pressures and stresses were evaluated as a function of moisture contents in the range of 8-14% and 10-14% d.b. The pressures and von Mises stresses affected as significant by the change of grain moisture content. The main cause of pressure and stress drops is changed in bulk density. Therefore is extremely important bulk density and moisture content of the product at the structural design of the silos. 4 mm wall thickness, were determined to be safe for von Mises stresses in both soybean and rice silos is smaller than 188000 kPa.

DRM-FL: Cross-Silo Federated Learning 접근법의 프라이버시 보호를 위한 분산형 랜덤화 메커니즘 (DRM-FL: A Decentralized and Randomized Mechanism for Privacy Protection in Cross-Silo Federated Learning Approach)

  • 무함마드 필다우스;초느에진랏;마리즈아길랄;이경현
    • 한국정보처리학회:학술대회논문집
    • /
    • 한국정보처리학회 2022년도 춘계학술발표대회
    • /
    • pp.264-267
    • /
    • 2022
  • Recently, federated learning (FL) has increased prominence as a viable approach for enhancing user privacy and data security by allowing collaborative multi-party model learning without exchanging sensitive data. Despite this, most present FL systems still depend on a centralized aggregator to generate a global model by gathering all submitted models from users, which could expose user privacy and the risk of various threats from malicious users. To solve these issues, we suggested a safe FL framework that employs differential privacy to counter membership inference attacks during the collaborative FL model training process and empowers blockchain to replace the centralized aggregator server.