Use of Artificial Intelligence in BIST Analysis as a Decision Support System for Fault Diagnosis in Multibeam Systems

The reliability of multibeam sensors is essential for the quality of hydrographic surveys conducted by the Brazilian Navy. However, intermittent failures recorded in the Built-In Self-Test (BIST) of the EM710 echosounder hindered the issuance of conclusive technical reports due to the large volume and complexity of the records. To address this challenge, a decision support system based on generative artificial intelligence was developed, capable of automating the creation of PowerShell and MATLAB scripts for reading and graphical visualization of BIST data.

The approach consisted of structuring the log files, extracting relevant technical parameters, and generating comparative graphical representations. This solution made it possible to identify subtle failures, progressive degradation trends, and hidden correlations between variables, such as the relationship between increased transmission channel temperature and impedance failures. Visual analysis reduced the time required for test interpretation from several hours to just a few minutes, while also expanding the accessibility of information to analysts without specialized electronics expertise.

The practical application was carried out with data from the Hydrographic Survey Vessel Taurus, in a real operational scenario. The results demonstrated significant gains in response time, diagnostic accuracy, and communication of technical findings. Furthermore, the methodology proved to be adaptable to other onboard systems that also generate extensive and complex records, reinforcing its relevance as a decision support tool in critical environments.

It is concluded that the use of generative artificial intelligence in processing BIST logs represents an effective alternative to increase the reliability of multibeam sensor analysis. The integration between technical expertise and analytical automation proved decisive in reducing diagnostic uncertainties, supporting predictive maintenance, and enhancing the safety of hydrographic surveys.