English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

The magnitude versus offset (MVO) curve, a type of frequency domain marine controlled-source electromagnetic data, is the most common way to identify electromagnetic anomalies in oil and gas reservoirs. However, in actual exploration, it can be difficult to identify the boundary of the high resistance anomaly when there are response signals of multiple emission frequencies. Also, the noise would reduce the accuracy of manually detecting electromagnetic anomalies. The robustness of the bidirectional long short-term memory (LSTM) network is relatively strong, and the LSTM neural network would get the most out of the sequence information of the data for feature extraction purposes and to achieve automatic classification and identification. Therefore, this paper proposes a method of using bidirectional LSTM to solve the problem of anomaly identification in marine controlled-source electromagnetic data. The LSTM unit was applied to establish anomaly identification models of single-layer LSTM, two-layer LSTM, and bidirectional LSTM, respectively. In this paper, theoretical data were calculated by a one-dimensional uniform layered medium model, and the synthetic noise data were constructed by adding random noise with different signal-to-noise ratios. The three types of models were trained, verified, and tested, respectively, to compare the accuracy of electromagnetic anomaly identification. According to the comparison, a conclusion can be drawn that the bidirectional LSTM model suggests the best manifestation of learning the characteristics of the sample. Its electromagnetic anomaly identification accuracy reached 100% in the theoretical dataset and 79.58% in the synthetic noise dataset.

A high resistivity anomaly identification method for marine controlled-source electromagnetic data