Marine Data Acquisition Using Class-Based ML and Hydrophone Diagnostics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems fail to accurately estimate operational parameters of vessels and detect mechanical issues using data from onboard systems effectively, particularly in marine environments.
Innovation Solution
A data acquisition apparatus and method utilizing a vehicle-class based machine learning model to generate estimated equivalents of missing data and detect abnormalities using hydrophone data, integrated with a portable data acquisition unit and remote server for real-time analysis and maintenance alerts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a vehicle-class based machine learning model is used to generate estimated equivalents for missing data, then measurement precision of operational parameters is improved, but device complexity increases
Solution Approach 1:
The machine learning model is trained in advance using historical data from multiple vessels of the same class, creating a pre-trained model that can generate accurate estimates for missing operational parameters. This preliminary training action enables the system to fill data gaps without requiring complex real-time data collection infrastructure for every parameter.
Solution Approach 2:
The patent introduces a data acquisition apparatus as an intermediary component that bridges the gap between available sensor data and required operational parameters. This apparatus uses the machine learning model to transform partial vessel data into complete operational parameter sets, mediating between limited measurements and comprehensive monitoring needs.
2Reliability
If hydrophone data and machine learning models are used for early detection of mechanical issues, then reliability of vessel operation is improved, but loss of time for data processing increases
Solution Approach 1:
The machine learning model is pre-trained on historical hydrophone data and mechanical failure patterns, enabling it to quickly analyze new data and detect anomalies. This preliminary training establishes the detection framework in advance, allowing rapid analysis of incoming hydrophone signals without extensive real-time processing delays.
Solution Approach 2:
The system continuously monitors hydrophone data and compares it against the trained model's expectations, providing feedback when deviations indicate potential mechanical issues. This feedback mechanism enables early detection of problems while maintaining efficient processing through iterative comparison rather than exhaustive analysis.
Data Source
AI summary
A method of estimating an operational parameter of a subject vessel, including receiving at least one vessel data item of a set of vessel data from a subject vessel; associating the vessel data with a class; generating an estimated equivalent for at least one vessel data item using a machine learning model trained using a database of data from vessels of the class of the subject vessel; and providing the estimated equivalent to a user. The vessel data item may be acquired by a dedicated data acquisition unit on a marine vessel. The data acquisition unit may be part of a system for analysing marine data. In another aspect, a method of detecting an abnormality is provided, including receiving hydrophone data from a subject vessel; detecting an abnormality in the hydrophone data; associating the abnormality with an operational system of the subject vessel; and generating a maintenance alert identifying the operational system for further attention.