Rail Vehicle Fleet Energy Optimization via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Optimizing the overall electrical energy balance of a rail vehicle fleet is challenging due to the need for each vehicle to adhere to individual target criteria while minimizing energy consumption and maximizing energy generation, which requires complex situation-dependent rules that are difficult to implement effectively.
Innovation Solution
A method utilizing computer units trained by machine learning to select actions for the state influencing systems of rail vehicles, taking into account vehicle-related, location-related, and route-related state parameters, to optimize the overall electrical energy balance of the fleet, thereby enabling 'swarm intelligence' for energy-efficient operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If complex situation-dependent rules are implemented to optimize energy balance, then energy optimization capability is improved, but system complexity and implementation difficulty increase
Solution Approach 1:
Each rail vehicle's computer unit independently selects actions for its state influencing system based on machine learning training, without requiring complex centralized control rules. The system serves itself by using learned patterns to autonomously optimize energy balance decisions at the vehicle level.
Solution Approach 2:
The patent replaces complex mechanical rule-based control systems with a machine learning-based intelligent system. The computer units use trained machine learning models to substitute for explicit situation-dependent rules, enabling energy optimization through learned behavior rather than programmed logic.
2Productivity
If machine learning is used to select actions for state influencing systems, then energy optimization is improved, but computational requirements and training complexity increase
Solution Approach 1:
The system divides the fleet into individual rail vehicles, each with its own computer unit running machine learning models. This segmentation allows distributed computation and training, reducing the computational burden on any single system while maintaining fleet-wide optimization capabilities.
Data Source
AI summary
A method for energy-optimized operation of a rail vehicle fleet. The fleet includes n rail vehicles, each with a state-influencing system for influencing a vehicle state to generate and/or consume electrical energy and a computer unit trained by machine learning. For every i of 1 to n during operation of the rail vehicle fleet an action to be applied to the state-influencing system of the i-th rail vehicle is selected by the computer unit of the i-th rail vehicle while taking into account at least one target criterion for the i-th rail vehicle and according to vehicle, location, and/or route-related status parameters. The action, when applied to the state-influencing system of the i-th rail vehicle, contributes to the optimization of an electrical total energy balance of the state-influencing systems of the rail vehicle fleet, and the selected action is applied to the state-influencing system of the i-th rail vehicle.