MESSAI.iO

AI summary

65% confidence

A machine learning-based energy management system using LSTM network is developed for fuel cell hybrid buses, outperforming a rule-based strategy in charge-sustaining mode conditions and fuel economy.

Generated by MESSAI extraction pipeline · review against source PDF

Generic MECRepresentative model
Click to animate flow
loading 3D model…

Representative MEC — matched on the paper’s system type only, not its reactor or geometry.

Extraction

Reported parameters

No values extracted from this paper yet.

Open in lab for full controls, parameter editing, and template overlays.

Open in lab →

What they did

System
MEC

What worked

No outcome metrics extracted yet.

Abstract

In this work, a machine learning-based energy management system is developed using a long short-term memory (LSTM) network for fuel cell hybrid buses. The neural network implicitly learns the complex relationship between various factors and the optimal power control from massive data. The selection of the neural network inputs is inspired by the adaptive Pontryagin’s minimum principle (APMP) strategy. Since an estimated value of the global average fuel cell power is required in the machine learning-based energy management strategy (EMS), some global features of driving cycles are extracted and then applied in a feedforward neural network to predict the average fuel cell power appropriately. The effectiveness of the machine learning-based energy management, with the integration of the mechanism of estimating the average fuel cell power based on the forward neural network, is tested under two different driving cycles from the training environment, with comparisons to a commercially used rule-based strategy. Based on the simulation results, the learning-based strategy outperforms the rule-based strategy regarding the charge-sustaining mode conditions and fuel economy. Moreover, compared to the best offline hydrogen consumption, the machine learning-based strategy consumed 0.58% and 0.36% more than the best offline results for both driving cycles. In contrast, the rule-based strategy consumed 1.80% and 0.96% more than optimal offline results for the two driving cycles, respectively. Finally, simulations under battery and fuel cell aging conditions show that the fuel economy of the machine learning-based strategy experiences no performance degradation under components aging compared to offline strategies.

Key findings

  • The LSTM network implicitly learns the complex relationship between various factors and optimal power control from massive data.
  • The selection of neural network inputs is inspired by the adaptive Pontryagin’s minimum principle (APMP) strategy.
  • The machine learning-based energy management strategy outperforms the rule-based strategy in terms of fuel economy.

Keywords

Artificial neural networkEnergy managementOffline learningArtificial intelligenceFuel efficiencyFeedforward neural network

Identifiers

DOI
10.3390/vehicles4040072
Journal
Vehicles
Year
2022