Marine Vessel Neural Network Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current anomaly detection methods in marine vessel systems are inflexible and unreliable, particularly for propulsion systems with multiple operationally interconnected elements, as they rely on static threshold settings which are cumbersome and complex to manage, failing to adapt to the dynamic changes in equipment condition over time.
Innovation Solution
A method involving sensors that measure vibration data, processing this data using Fast Fourier Transform (FFT) to convert it into frequency-domain spectra, normalizing it, and training a convolutional autoencoder neural network to detect anomalies in real-time, allowing for dynamic threshold adjustments and improved anomaly detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If static threshold settings are used for anomaly detection, then the system is simple to implement, but the system becomes inflexible and unreliable when equipment conditions change over time
Solution Approach 1:
The patent implements dynamic anomaly detection by replacing static thresholds with a neural network model that continuously adapts to changing equipment conditions. The system processes vibration data through FFT transformation and uses a trained neural network to dynamically adjust detection parameters, enabling the system to respond to evolving equipment states while maintaining implementation feasibility through automated learning.
Solution Approach 2:
The system transforms the detection approach by changing from fixed threshold parameters to dynamic parameters generated by neural network processing. The neural network learns optimal detection parameters from training data and automatically adjusts them during operation, allowing the system to adapt to different equipment conditions without manual reconfiguration.
2Reliability
If dynamic anomaly detection is implemented to adapt to changing conditions, then the system becomes more reliable and accurate, but the system complexity increases
Solution Approach 1:
The patent replaces traditional mechanical threshold-setting mechanisms with a neural network-based intelligent system. Instead of manually configuring and adjusting threshold parameters, the system uses machine learning models that automatically learn from data, substituting complex manual configuration processes with automated computational approaches that improve reliability while managing complexity through software-based solutions.
Solution Approach 2:
The neural network system performs self-training and self-adjustment by learning from historical vibration data and automatically optimizing detection parameters. The system serves itself by continuously improving its detection capabilities without requiring external intervention for parameter tuning, thereby enhancing reliability while containing complexity within the automated learning framework.
3Use of energy by moving object
If traditional threshold-based monitoring is used, then the system requires minimal processing, but it fails to detect anomalies in complex propulsion systems with multiple interconnected elements
Solution Approach 1:
The patent applies segmentation by dividing the complex vibration signal into frequency components through FFT transformation. This breaks down the complex propulsion system monitoring task into manageable frequency bands, allowing the neural network to analyze specific frequency ranges independently and detect anomalies more effectively while optimizing processing resource allocation across different signal components.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides a more accurate and reliable system for detecting anomalies in marine vessel propulsion systems, enabling early fault detection and improving operational efficiency by adapting to changing conditions without the need for complex threshold settings.
Implementation Method 1
generating a Fast Fourier Transform (FFT) on the time-domain reference sensor data to provide a plurality of reference spectra files in frequency-domain
Data Source
AI summary
A computer implemented method and apparatus for a marine vessel data system, the method comprising: receiving data from at least one sensor configured to measure vibration and operationally arranged to the marine vessel to provide time-domain reference sensor data; maintaining the time-domain reference sensor data within a data storage system; generating a Fast Fourier Transform (FFT) on the time-domain reference sensor data to provide a plurality of reference spectra files in frequency-domain, wherein each reference spectra file comprises spectra data defined by amplitude information and frequency information, and each spectra file is associated with condition information determined based on collection of the time-domain reference sensor data; normalizing each reference spectra file by converting the frequency information to order information using the condition information to provide normalized reference spectra files; and training a convolutional autoencoder type of neural network using the normalized reference spectra files.