Scholarlink

<p>This study investigates the long-term dynamic response of the Yusufeli Arch Dam, Turkey’s tallest and one of the world’s highest double-curvature concrete arch dams, through the integration of Structural Health Monitoring (SHM) techniques and Artificial Neural Networks (ANNs). High-sensitivity accelerometers installed on the dam body continuously recorded the dam’s ambient vibration response throughout the reservoir impoundment phase. Simultaneously, environmental parameters including reservoir water level and air temperature were collected. For the SHM procedure, data were collected between 22 November 2022 and 15 March 2024. The analysis focused on the first six natural frequencies, along with the corresponding temperature and reservoir water level recorded on the same dates. In total, approximately 537 paired datasets of frequency (between 1.657 and 5.322&#xa0;Hz for all modes), temperature (minimum −&#xa0;15.42&#xa0;°C, maximum 27.82&#xa0;°C), and water level (between 0 and 192.85&#xa0;m), spanning the transition from empty to full reservoir conditions, were used in this study. These inputs formed the basis of a feed-forward backpropagation ANN model developed to predict natural frequency variations across the first six vibration modes. The trained model demonstrated high prediction accuracy and was used to generate frequency response maps across a range of environmental conditions. These maps provide valuable insight into the influence of hydrostatic and thermal effects on the dam’s dynamic characteristics. The proposed approach offers a powerful decision-support framework for dam safety assessment, enabling early anomaly detection and improved maintenance planning under real-world operating conditions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Long-term frequency prediction of Yusufeli arch dam during reservoir impoundment under varying air temperatures using artificial neural networks

  • 期刊论文

  • 发表日期：

    2026年7月21日

查看全文
  • Ebru Kalkan Okur,
  • Fatih Yesevi Okur,
  • Ahmet Can Altunişik

摘要

This study investigates the long-term dynamic response of the Yusufeli Arch Dam, Turkey’s tallest and one of the world’s highest double-curvature concrete arch dams, through the integration of Structural Health Monitoring (SHM) techniques and Artificial Neural Networks (ANNs). High-sensitivity accelerometers installed on the dam body continuously recorded the dam’s ambient vibration response throughout the reservoir impoundment phase. Simultaneously, environmental parameters including reservoir water level and air temperature were collected. For the SHM procedure, data were collected between 22 November 2022 and 15 March 2024. The analysis focused on the first six natural frequencies, along with the corresponding temperature and reservoir water level recorded on the same dates. In total, approximately 537 paired datasets of frequency (between 1.657 and 5.322 Hz for all modes), temperature (minimum − 15.42 °C, maximum 27.82 °C), and water level (between 0 and 192.85 m), spanning the transition from empty to full reservoir conditions, were used in this study. These inputs formed the basis of a feed-forward backpropagation ANN model developed to predict natural frequency variations across the first six vibration modes. The trained model demonstrated high prediction accuracy and was used to generate frequency response maps across a range of environmental conditions. These maps provide valuable insight into the influence of hydrostatic and thermal effects on the dam’s dynamic characteristics. The proposed approach offers a powerful decision-support framework for dam safety assessment, enabling early anomaly detection and improved maintenance planning under real-world operating conditions.