From driving behavior to energy consumption: A novel method to predict the energy consumption of electric bus

Nan, Sirui; Tu, Ran; Li, Tiezhu; Sun, Jian; Chen, Haibo

doi:10.1016/j.energy.2022.125188

Cited by 32 publications

(7 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Interpretable machine learning has addressed this problem effectively and is applied to interpreting fuel consumption driving behavior. Commonly used algorithms include the Permutation Feature Importance (PFI) (Eslahi, 2022), Partial Dependency Plots (PDP) (Li and Sun, 2021), Local Interpretable Model-Agnostic Explanation (LIME) (Pang and Kong, 2022), and Shapley additive explanations (SHAP) (Nan et al, 2022), etc. The SHAP algorithm has the most extensive application due to its integration of global, interaction, individual interpretation, solid theoretical foundation, and rich functions.…”

Section: Literature Reviewsmentioning

confidence: 99%

High fuel consumption driving behavior identification and causal analysis based on LightGBM and SHAP

Chen,

Liu,

et al. 2024

Preprint

View full text Add to dashboard Cite

Accurate identification of high fuel consumption driving behaviors provides good theoretical support for eco-driving training. To gain a deeper understanding of contributing factors and their impacts on fuel consumption, this study acquired a driving data set based on a driving simulator test and employed the light gradient-boosting machine (LightGBM) algorithm to identify driving behaviors related to high fuel consumption and the SHapley Additive exPlanations (SHAP) algorithm for causal analysis. First, the vehicle kinematics and fuel consumption data were collected and preprocessed. Secondly, the LightGBM algorithm was employed to classify fuel consumption levels, including low, medium, and high. The results show that the LightGBM algorithm performs well in terms of precision, recall, and F1-score. Finally, the SHAP algorithm was used to interpret the influence of contributing factors on high fuel consumption from three perspectives, global interpretation, interaction interpretation, and individual interpretation. The SHAP algorithm can intuitively display the relationship between high fuel consumption and its contributing factors. Specifically, acceleration, speed, roll speed, pitch speed, and engine speed significantly increased the probability of high fuel consumption. This study proposed an efficient combined method for high fuel consumption identification and interpretation, which can reduce the occurrence of high fuel consumption driving behavior, thus achieving the purpose of eco-driving training.

show abstract

Section: Literature Reviewsmentioning

confidence: 99%

High fuel consumption driving behavior identification and causal analysis based on LightGBM and SHAP

Chen,

Liu,

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Hochreiter and Schmidhuber developed the LSTM to address the mentioned shortcomings. LSTM has three gates, an input gate, an output gate, and a forget gate, which can improve its performance in long-term learning tasks [38], [50], [51]. This option is considered critical because this network is prone to overfitting.…”

Section: ) Joint Feature Learning and Time Series Modelingmentioning

confidence: 99%

An Adaptive Regenerative Braking Strategy Design Based on Naturalistic Regeneration Performance for Intelligent Vehicles

Ziadia,

Kelouwani,

Amamou

et al. 2023

IEEE Access

View full text Add to dashboard Cite

The effectiveness of regenerative braking strategies plays an important role in extending the driving range of electric vehicles. Since the driver is still an essential factor in levels 3 and 4 of intelligent electric vehicles, improving user acceptance and adoption of the braking control strategy is crucial. This paper puts forward a new regenerative braking strategy to find a compromise between optimal braking control performance and naturalistic regeneration performance while satisfying the maximum speed preference when driving between two-stop events. Unlike other similar works that only maximize regenerative braking energy while satisfying the physical limits of an electrified powertrain, this paper considers naturalistic regeneration performance. To achieve this, firstly, the power regenerated by three drivers is predicted with a long-horizon (30 seconds), using long-short-term memory networks (LSTM) and non-linear autoregressive exogenous model (NARX). Subsequently, an estimation of the energy recovery maximization rate is performed to give a perception of the naturalistic regeneration performance. As this performance varies, the deceleration planning employs three horizon scales of long, medium, and short, determined by the energy recovery maximization rate. Finally, dynamic programming (DP) is utilized to optimize a deceleration profile. The study utilizes real data of inverter efficiency, transmission efficiency, and motor-to-battery efficiency map. The outcome of this study shows that the proposed regeneration braking strategy is adaptive, improving regeneration efficiency by 39,6% for driver 1, 16% for driver 2, and 26% for driver 3, and forecasting the optimality of some deceleration behaviors.

show abstract

“…In regression analysis, if one independent variable is used, it is called simple regression; if more than one independent variable is used, it is called multivariate regression [74].…”

Section: ) Regression Evaluation Metricsmentioning

confidence: 99%

A Novel Energy Consumption Prediction Model of Electric Buses Using Real-Time Big Data From Route, Environment, and Vehicle Parameters

Ekici,

Akdağ,

Aydin

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Electric vehicles (EVs) have positive impacts on reducing oil dependence and exhaust emissions. However, the range problem of EVs is a factor that raises concerns for individual users and bus operators. For this reason, studies on increasing the range of the electric buses in public transportation is extremely important to ensure optimum operation. In this study, a novel energy consumption model, MTECM (Malatya Trolleybus Energy Consumption Model), is developed using the multi-parameter linear regression method. The real-time big data was collected on the field of Trolleybus vehicles, which have been operated for 8 years in Malatya / Turkey. Firstly, by calculating the correlation of the parameters affecting this model, the parameters that are suitable for the purpose of our study are determined and regression analysis is performed on the original Trolleybus dataset. A total of 75.497.472 data are analyzed for this model. The RMSE (Root Mean Square Error) of MTECM is calculated as 0.29996. The trained model is applied to the 10 busiest routes in Malatya in terms of passenger density. The RMSE value on these routes is calculated between 0.30299 and 0.31421. Based on the results, with lower error rates, the proposed novel model is more efficient than other studies in the literature. In addition, energy consumption can be calculated for any route planned to establish an electric bus operation with MTECM. Therefore, according to the consumption obtained, the correct determination and selection of parameters that significantly affect the investment cost such as route, vehicle length, engine power, and battery capacity can be made.INDEX TERMS Big data analysis, electric vehicle, electric bus, machine learning, regression, trolleybus.

show abstract

From driving behavior to energy consumption: A novel method to predict the energy consumption of electric bus

Cited by 32 publications

References 30 publications

High fuel consumption driving behavior identification and causal analysis based on LightGBM and SHAP

High fuel consumption driving behavior identification and causal analysis based on LightGBM and SHAP

An Adaptive Regenerative Braking Strategy Design Based on Naturalistic Regeneration Performance for Intelligent Vehicles

A Novel Energy Consumption Prediction Model of Electric Buses Using Real-Time Big Data From Route, Environment, and Vehicle Parameters

Contact Info

Product

Resources

About