Reconstruction of missing daily streamflow data using recurrent neural network

0 Người đánh giá. Xếp hạng trung bình 0

Tác giả: XUÂN HIỀN LÊ

Ngôn ngữ: Vie

Ký hiệu phân loại:

Thông tin xuất bản: Tạp chí Khoa học Kỹ thuật Thủy lợi và Môi trường 2019

Mô tả vật lý:

Bộ sưu tập: Metadata

ID: 407223

Thêm vào giỏ Liên kết toàn văn

Lưu lượng sông là một trong những dữ liệu quan trọng nhất trong thủy văn bởi các dữ liệu này có thể được sử dụng cho các phân tích liên quan tới quản lý tài nguyên nước cũng như dự báo dòng chảy lũ. Việc thiếu dữ liệu dòng chảy có thể dẫn tới các phân tích khoa học không đầy đủ. Để có được những thông tin đáng tin cậy và chính xác hơn thì những dữ liệu bị thiếu này phải được lấp đầy. Mục tiêu của bài báo này là giới thiệu một cách tiệp cận hiệu quả dựa trên mô hình mạng nơ-ron hồi quy để xây dựng lại dữ liệu dòng chảy hàng ngày bị thiếu. Trạm thủy văn Lai Châu được chọn làm trạm mục tiêu cho nghiên cứu điển hình bởi đây là trạm thủy văn nằm ở thượng lưu của lưu vực sông Đà. Kết quả nghiên cứu thể hiện hiệu suất cao của mô hình mạng nơ-ron hồi quy. Với kết quả này, mô hình hoàn toàn có thể được áp dụng cho các trạm thủy văn ở thượng nguồn nơi mà thiếu các dữ liệu về dòng chảy. , Tóm tắt tiếng anh, Streamflow data is one of the most important quantities in hydrology because of these data closelyrelated to water resource management problems as well as flood forecasting problems. The lack ofthese data can lead to inadequate scientific analysis. Therefore, reconstruction of missing data is animportant step to get more reliable and accurate information. The objective of this paper is to introducean effective approach based on the recurrent neural network model to reconstructing missing dailydischarge data. Lai Chau hydrological station, located upstream of the Da River basin, was selected asa case study. The findings of this study demonstrated that the recurrent neural network model yieldsreliable estimates for the problem of missing data. As a result, the RNN model can be applied to otherhydrological stations upstream where the flow data is missing.

1. Đọc trực tuyến

1. Đọc trực tuyến

Tạo bộ sưu tập với mã QR