Đánh giá xác suất phá hoại địa chấn của kết cấu trụ cầu bê tông cốt thép chịu động đất dựa trên mạng nơron nhân tạo=Probabilistic damage assessment of reinforced concrete piers under earthquakes based on artificial neural networks

0 Người đánh giá. Xếp hạng trung bình 0

Tác giả: Hoàng Vĩnh Nguyễn, Duy Hòa Phạm, Hoàng Nam Phan

Ngôn ngữ: vie

Ký hiệu phân loại:

Thông tin xuất bản: Tạp chí Khoa học công nghệ xây dựng (Đại học Xây dựng Hà Nội), 2024

Mô tả vật lý: tr.31-44

Bộ sưu tập: Metadata

ID: 233525

Thêm vào giỏ

Nghiên cứu trình bày một hướng tiếp cận mới dựa trên mô hình mạng nơron nhân tạo (ANN) nhằm xây dựng đồ thị trạng thái phá hủy của trụ cầu bê tông cốt thép (BTCT) chịu động đất. Theo đó, trụ cầu được mô hình hóa sử dụng phần mềm OpenSees xét đến mô hình vật liệu phi tuyến và ảnh hưởng của nền móng. Các phân tích lịch sử thời gian được thực hiện trên một bộ gồm 140 bản ghi gia tốc nền được lựa chọn theo các phổ hiểm họa đều mục tiêu. Dựa trên dữ liệu được thu thập từ các kết quả phân tích, hai mô hình ANN nối tiếp được xây dựng, tối ưu và xác thực
cụ thể, mô hình ANN 1 được sử dụng để thiết lập quan hệ giữa các đại lượng đo lường cường độ dao động nền (IM) và mô hình ANN 2 với dữ liệu đầu vào là các IM được sử dụng để ước lượng đáp ứng kết cấu. Mô phỏng Monte Carlo cuối cùng được áp dụng trên 2 mô hình này để xây dựng đồ thị trạng thái phá hủy của kết cấu ứng với các trạng thái hư hại. Kết quả phân tích trên một trụ cầu thân cột BTCT của cầu Quá Giáng ở thành phố Đà Nẵng cho thấy phương pháp đề xuất thể hiện tính khả thi và tin cậy trong việc xây dựng đồ thị trạng thái phá hủy cho các trạng thái hư hại khác nhau của trụ cầu.This study presents a new approach based on the artificial neural network (ANN) model for constructing fragility curves of reinforced concrete piers subjected to earthquakes. Initially, the pier is modeled using OpenSees software, accounting for the effects of non-linear material behavior and foundation interactions. Subsequently, time-history dynamic analyses are performed on a set of 140 ground motion records selected based on the target uniform hazard spectra for a specific site. Based on the dataset collected from these analyses, two ANN models are built, optimized, and validated. Specifically, ANN model 1 establishes the relationship between the ground motion intensity measure (IM) and pier response, while ANN model 2, using IMs as input data, estimates the pier’s response. Monte Carlo simulations are then applied to the ANN models to construct fragility curves. The analysis results for a reinforced concrete pier of the Qua Giang bridge in Da Nang City demonstrate that the proposed method is both feasible and reliable for constructing fragility curves for different seismic performance levels of the pier.

1. Đọc trực tuyến

1. Đọc trực tuyến

Tạo bộ sưu tập với mã QR