中国电－电混合动力燃料电池公交车燃料电池系统衰减现象分析

Fuel cell system degradation analysis of aChinese plug－in hybrid fuel cell city bus

Jianqiu Li

Zunyan Hu

Liangfei Xu

Minggao Ouyang

Chuan Fang

Junming Hu

Siliang Cheng

Hong Po

Wenbin Zhang

Hongliang Jiang

Abstract

Fuel cell hybrid electric vehicles areattracting increasing attention and research． However， the durability of fuelcell systems （FCS） is the main bottleneck in commercial application． A plug－infuel cell city bus was developed by our group， and it completed several monthsof demonstration operation． A two－stage flowchart was proposed to clarify theFCS degradation analysis． The average cell voltage decline rate isapproximately 346 μV／h at a current density of 120 A， under which the initialvoltage is about 0．7 V． The estimated lifetime in the third stage reachesthe design target． The final electrochemically active surface areas areapproximately 80％ of the initial value． A detailed degradation analysis offuel cell voltage uniformity is presented， and a five－region degradationotherness analysis is proposed for performance uniformity analysis． Thedifference of equivalent resistance has been verified to be the main factor indegradation otherness of the regions and in system performance degradation．

该文章主要是统计分析，总结规律，不是针对性的电堆和单节电池内部材料失效的分析。

Fig． 2 e Fuel cell powertrain system．

两个电堆一个在上一个在下。

Table 1 e Parameters of the fuel cell  city bus．

Fig． 3 e Flowchart of fuel cell  performance degradation analysis．

Fig． 6 e Average power demand during the  demonstration operation．

Fig． 7 e Relationship between battery SoC  and FCS output current．

从图8来看电流和SOC并不是典型值。

To prolong the lifetime of the FCS， a  soft－run control strategy is proposed to guarantee the quasi－steady state  operation of the fuel cell stack， and the FCS is designed as a range  extender． Therefore， the fuel cell stack is limited to provide a stable  output current， which is nearly equivalent to the average power demand．

这个想法贯穿于增程式燃料电池系统设计中。

Fig． 8 e Histograms of FCS output  current．

大比例运行在170A，0．62A／cm2。

Fig． 9 e Daily hydrogen consumption per  100 km．

使用了Plugin模式，耗氢中等，8kg／100km，有几个接近于0的表明几乎在用电池。

Fig． 10 e Voltage degradation process  under 120 A output current．

Fig． 11 e Stage division of the  demonstration operation．

0．7V／cell＊120A＊135cell＊2＝22．7kW，不知道哪里搞错了。

Fig． 12 e Voltage degradation process  under 170 A and 0 A output current．

Table 2 e Decline rates of three stages  under different output currents．

Fig． 13 e Remaining ECA percentage of  fuel cell stack．

The sealing of the FCS of the city bus  fuel cell system is worse than that in the bench test because of a jolt  during operation． Therefore， air can leak into the anode after FCS shutdown．  Without the nitrogen purging before operation， oxygen in the anode can lead  to serious carbon support corrosion during startup．

Fig． 14 e Voltage range degradation  process under 170 A and 120 A output current． （a） （b） daily average voltage  range， （c） （d） voltage rangeetime curve．

The voltage range of the FCS is defined  as the difference of maximum and minimum monolithic voltage 单节电池的极差

Fig． 15 e Voltage mean square error  degradation process under 170 A and 120 A output current． （a） （b） daily  average voltage mean square error， （c） （d） voltage mean square error－time  curve．

Fig． 16 e Degradation process of fuel  cell voltage non－uniformity．

Fig． 17 e Reversible and irreversible  degradation of fuel cell voltage non－uniformity．

Arrow A is the irreversible increase part  of voltage uniformity degradation， arrow C is the initial reversible part， and  arrow B is the last reversible part．

这里没看明白，四个箭头区，三个区域。

The reversible part mainly depends on the  water and reaction gas pressure distribution difference of every fuel cell．

The irreversible part is mainly caused by  a change in the parameters．

Fuel cell A has a higher average voltage，  a better voltage mean square error， but a worse voltage range．

The only difference between the two fuel  cell stacks is position． Fuel cell A is located in the upper position， and  the two fuel cell stacks use a public intake manifold， which is assumed to  supply more reaction gas to fuel cell A．

Fig． 18 e Region division of fuel cell  stack．

Every 27 fuel cell is defined as one  region， and an average value in this region is used to analyze the  degradation process．

Table 3 e Fitting results of the  polarization curves of five regions．

这里的r指的是电池内阻。

Table 3 shows that the deviation of the  equivalent resistance is the most obvious and that the constant part  difference has a weak effect on performance uniformity． By contrast， the  regularity is not apparent for fuel cell stack B because the average values  in regions 1－4 are close and only the last region is worst．

The constant part shows a slight  decrease trend along the flow channel． This trend is consistent with the  reaction gas pressure drop along the flow channel， which will affect the Vreversible． The deviation of ohmic polarization in the different regions is significant．

Fig． 21 e Evolution process of voltage  chromatogram in fuel cell A under 170 A．

The figure shows that the voltage in the  middle part of fuel cell A is lower than that of the two sides． This result，  as indicated by Refs．， reflects that the cells at the center are overheated  and dry during operation．

In the course of the operation， the lower  regions expanded from the middle part to the two sides． The water balance in  fuel cell A is unstable． Therefore， the temperature and humidity management  may be a problem for the system， thereby requiring an improved structure  design or control algorithm．

Fig． 22 e Evolution process of loading  curve parameter fitting results．

An analysis by an engineer indicates that  the increased resistance may be due to three reasons： aging of the humidifier，  corrosion of the membrane or other components， and decrease of the fuel cell  stack preload． The aging of the humidifier in particular is detected through  a voltage response speed experiment

Fig． 23 e Parameter changing process of  region otherness

The constant parts of regions four and  five are always lower than those of the other regions， thereby proving that  the voltage otherness is due to the lower reaction gas in the rear．

Conclusion

This paper analyzed the degradation  process of an FCS equipped in a plug－in fuel cell city bus． A two－stage  analysis method is proposed to obtain an overall understanding of the performance  degradation process． Several useful conclusions are obtained on the basis of  the statistical characteristics of daily operational data and fuel cell  voltage．

The power demand of a  fuel cell city bus is nearly a stable value with a fixed route． The FCS uses  a soft－run strategy based on feedbacks of battery SoC and the predicted power

requirement．

The average cell  voltage decline rate in three stages is analyzed． Throughout the entire  demonstration operation， the average cell voltage decline rate at an average  current

density of 0．435 A／cm2 is approximately  346 uV／h． The estimated lifetime in the third stage reaches the design target．  However， the sealing and leak of the gas supply system in the second stage  caused a tenfold acceleration in the decline of FCS．

An estimated ECA degradation rate based on the everyday average  open circuit voltage is obtained． This rate is fast in the first two－thirds  of the demonstration， and the final ECA

is approximately 80％ of the initial  value．

The degradation  process of fuel cell voltage uniformity is discussed． The voltage uniformity  worsens with performance degradation． The voltage mean square error shows

that the voltage uniformity of fuel cell  A is better， while the voltage range shows that the reliability of fuel cell  B is better．

极差和标准方差的规律不一致。

The voltage uniformity is divided into  the irreversible increase part， which is mainly caused by parameter change，  and the reversible increase part， which mainly depends on the water and  reaction gas pressure distribution difference of every fuel cell． Both parts  are essential to voltage uniformity．

A five－region  degradation otherness analysis and simplified polarization curve fitting are  proposed for performance degradation analysis． The voltage of the middle part  is lower than that of the other regions． During operation， the lower regions  expand from the middle part to the two sides presumably because of humidity  and temperature management．

 The constant part voltage drop of parameter  fitting results along the flow channel has been observed． The increase of ohmic  polarization is the main factor in voltage decline and

region otherness． The deviation of the  equivalent resistance for different regions is more than 7％． Three potential reasons  for the increased resistance are proposed．

In future research， a next－generation FCS  will be designed and optimized to prolong the service life of the FCS． An improved  water management system and reduction of carbon support corrosion are key  technologies for the next generation system． Moreover， a mathematical model  will be developed on the basis of additional demonstration operation

data．