Commercial EV Energy Prediction Using Driving Position Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional commercial electric vehicle energy consumption prediction methods are limited by the difficulty in accurately acquiring vehicle characteristic parameters, leading to poor prediction accuracy when applied to real-world conditions, as they are typically based on simulated laboratory environments.
Innovation Solution
A method using machine learning algorithms to predict energy consumption by acquiring discharge duration and driving position characteristic data, which are then substituted into an energy consumption prediction model, with model training based on historical driving data and battery capacity to improve accuracy and match actual operating conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional energy consumption prediction methods based on simulated laboratory environments and vehicle dynamics models are used, then the method can be applied with available data, but the prediction accuracy deteriorates due to difficulty in accurately acquiring vehicle characteristic parameters in real-world conditions
Solution Approach 1:
The patent replaces the conventional vehicle dynamics model-based prediction method with a machine learning model. Instead of relying on physical parameters like vehicle mass, frontal area, and rolling resistance coefficients that are difficult to measure accurately, the system uses discharge duration data and driving position characteristic data as inputs to a trained neural network model, substituting mechanical measurement requirements with data-driven computational prediction.
Solution Approach 2:
The patent creates a virtual copy of the vehicle's energy consumption behavior through machine learning training. By training the model on historical discharge duration data, driving position data, and corresponding energy consumption values, the system learns to replicate the vehicle's actual energy consumption patterns without needing to physically measure all vehicle characteristic parameters in real-time.
2Measurement precision
If machine learning-based prediction models using discharge duration and driving position data are used, then prediction accuracy improves by reflecting actual driving environments, but the device complexity increases due to model training and data processing requirements
Solution Approach 1:
The patent performs model training in advance using historical data before actual prediction is needed. The training phase processes discharge duration data, driving position characteristic data, and energy consumption data to establish the predictive model. Once trained, the model can perform rapid predictions without requiring complex real-time processing, separating the computationally intensive training phase from the simpler inference phase.
Solution Approach 2:
The system uses the vehicle's own historical operating data (discharge duration, driving position, energy consumption) to train and improve its prediction capability. The model learns from the vehicle's actual performance patterns, making the system self-improving without requiring external calibration or additional sensing infrastructure.
Data Source
AI summary
The present application relates to an electric vehicle energy consumption prediction method and apparatus, a computer device, a computer-readable storage medium, and a computer program product. The method includes: acquiring discharge duration data of an electric vehicle; acquiring driving position characteristic data of the electric vehicle; and inputting the discharge duration data and the driving position characteristic data into an energy consumption prediction model to obtain energy consumption prediction data of the electric vehicle. The energy consumption prediction model is obtained based on a machine learning algorithm.