Dynamic model of a high power PEM fuel cell system on the basis of artificial neural networks

Author(s):  
A. U. Chavez ◽  
S. M. Duron ◽  
L. G. Arriaga ◽  
R. Munoz
Author(s):  
Fengxiang Chen ◽  
Jieran Jiao ◽  
Yang Yu ◽  
Yuan Gao ◽  
Sichuan Xu

With the advantages of high power density, rapid startup, low operating temperature and no emission of pollutants, proton exchange membrane (PEM) fuel cell is considered to be the most promising candidate for the next generation power source of Clean Energy Automotive. PEM fuel cell operation necessitates thermal management to satisfy the requirements of safe and efficient operation by keeping the temperature within a certain range independent of varying load conditions. As for a high power PEM fuel cell system (eg. 80kw) without the external gas to gas humidifier, the temperature of the stack inlet coolant had better track to a time-varying curve produced by the working condition, which introduce the temperature difference between the cathode inlet and outlet, and thus it improves the relative humidity of the inlet air of the cathode. Compared to the traditional stack outlet coolant temperature regulation problem, the new plant is a two inputs and two outputs system, furthermore, the stack inlet coolant temperature control is a tracking problem which is different to the outlet coolant temperature regulation (regulation problem). Considering that the PEM fuel cell without the external humidifier is a promising scheme which has been adopted by the Mirai fuel cell vehicle [1], we actively aim to control both the inlet and outlet coolant temperature as desired simultaneously. In this paper, a two inputs and two outputs decouple control scheme is developed to achieve our aim. Firstly, based on the energy conservation and continuity equation, we establish a dynamic thermal model for the cooling system consisted of a water circulation pump and a radiator coupled to a fan, integrated with the fuel cell stack. Secondly, the static coupling characteristics of the control variable is analyzed according the relative gain matrix method. Then two specific control strategies are designed. One is based on frequency domain pure PID control technique. Considering the coupling phenomenon between two control channels, another technique is based on decouple theory feed-forward decouple control technique. Both of them try to regulate the outlet and inlet coolant temperature through tuning mass flow rate of water circulation pump and duty ratio of radiator. Finally, all the control strategies are demonstrated on the platform of Matlab / Simulink. The results show that both of them can control the stack inlet and outlet coolant temperature simultaneously, but the second strategy has much better performance than the first.


2021 ◽  
Vol 7 ◽  
pp. 3199-3209
Author(s):  
Junlong Zheng ◽  
Yujie Xie ◽  
Xiaoping Huang ◽  
Zhongxing Wei ◽  
Bahman Taheri

Electronics ◽  
2021 ◽  
Vol 10 (8) ◽  
pp. 928
Author(s):  
Ferenc Hegedüs ◽  
Péter Gáspár ◽  
Tamás Bécsi

Nonlinear optimization-based motion planning algorithms have been successfully used for dynamically feasible trajectory planning of road vehicles. However, the main drawback of these methods is their significant computational effort and thus high runtime, which makes real-time application a complex problem. Addressing this field, this paper proposes an algorithm for fast simulation of road vehicle motion based on artificial neural networks that can be used in optimization-based trajectory planners. The neural networks are trained with supervised learning techniques to predict the future state of the vehicle based on its current state and driving inputs. Learning data is provided for a wide variety of randomly generated driving scenarios by simulation of a dynamic vehicle model. The realistic random driving maneuvers are created on the basis of piecewise linear travel velocity and road curvature profiles that are used for the planning of public roads. The trained neural networks are then used in a feedback loop with several variables being calculated by additional numerical integration to provide all the outputs of the original dynamic model. The presented model can be capable of short-term vehicle motion simulation with sufficient precision while having a considerably faster runtime than the original dynamic model.


Sign in / Sign up

Export Citation Format

Share Document