Hybrid power ship energy management method, device, equipment and medium
By combining CNN and DP, high-quality features are extracted and energy management strategies are optimized, the problems of unreasonable energy distribution in traditional energy management systems in complex situations and low accuracy of input data of DP algorithms are solved, and the energy use efficiency and energy management strategies of hybrid ships are improved.
Patent Information
- Application Number
- CN202411693545.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-25
- Publication Date
- 2025-03-14
AI Technical Summary
Traditional ship energy management systems are difficult to cope with complex emergencies, resulting in unreasonable energy distribution, resulting in waste of energy and degradation of ship performance. Dynamic programming algorithms are difficult to implement in high-dimensional continuous states, and the input data is low, resulting in a Bellman dimension disaster.
Combined with convolutional neural network (CNN) and dynamic programming (DP), multi-level features are automatically extracted through CNN, reducing manual feature engineering, and improving the input data quality of DP. Using DP's optimization capabilities, a dynamic programming mathematical model is constructed to calculate and obtain the optimal ship energy management strategy.
It improves the energy use efficiency of hybrid ships, improves energy management strategies, solves the problems of low accuracy of input data of DP algorithm and the Bellman dimension disaster, and achieves more accurate energy prediction and optimization decisions.
Smart Images
Figure CN119623272A_ABST