Offshore towed-streamer positioning based on variance component estimation. (15th June 2023)

• Record Type: Journal Article
• Authors: Yu, Wenkun; Duan, Chufeng; Kuang, Cuilin; Zhang, Haonan; Wu, Peida; Dai, Wujiao
• Is Part Of: Ocean engineering. Volume 278(2023)
• Publication Date: 2023-06-15
• Subjects: Towed-streamer positioning -- Variance component estimation -- Offshore seismic survey -- Polynomial fitting; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2023.114355
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: In offshore towed-streamer seismic surveys, fusing observations from diverse navigation sensors to position streamers and hydrophones accurately is essential for high-quality seismic imaging and analysis. However, streamer positioning accuracy can be limited when nominal sensor precisions adopted in observation weighting become unrealistic in complex marine environments. This study used variance component estimation (VCE) to refine the weight determination of heteroscedastic observations involved in streamer positioning. First, a polynomial fitting model for streamer positioning was established. Subsequently, a simplified VCE method was presented in which the observations were classified based on both sensor type and distribution geometry. Finally, a six-streamer survey dataset acquired from the South China Sea was used to investigate the performance of the proposed method. A comparative analysis was performed between conventional empirical weighting methods and VCE methods implemented in various observation classification schemes. The experimental results verified the superiority of the stochastic model refinement in fully exploiting observations from multiple navigation sensors in offshore streamer positioning. Compared with the empirical weighting, VCE method reduced the mean standard deviation of the hydrophone positions by about 0.5 m. Highlights: The shape of streamer is modeled as a polynomial curve. The analytical towed-streamer positioning model is formulated by linearizing different types of navigation observations. The variance component estimation is performed based on posteriori residuals classified by both sensor types and geometry. The superiority of variance component estimation over empirical weighting is verified in streamer positioning experiment.