燃料电池电堆低化学计量比单节电池局部水淹氢燃料饥饿对相邻节电池的影响：电流汇聚再分布和碳腐蚀[失效分析其三]

A cell interaction phenomenon in amulti-cell stack under one cell suffering fuel starvation

Zunyan Hua

Liangfei Xu

Jianqiu Lia

Junming Hu

Xin Xu

Xiaoli Du

Weihua Sun

Minggao Ouyang

Abstract

Cell interaction is the main factorresulting in a shorter lifespan for multi-cell stacks than for a single fuelcell stack. To explain this mechanism, we propose a cell interaction phenomenonin which one cell experiences fuel starvation. A specific voltagedistribution along the straight channel direction has been observed with aninnovative multipoint monitoring method. This phenomenon also can be used forfuel starvation diagnosis. This study proposes an ingenious simplifiedtwo-chamber model to analyze the current and voltage redistributionmechanism under fuel starvation. The reliability of this model has beenvalidated in this paper. The model shows that current convergence resultedby fuel starvation in one cell can lead to a concomitant local currentconvergence in nearby normal cells. Based on our calculation, >85% ofreaction current concentrates in the anode inlet region of the fuel-starvedcell, resulting in a 70% current convergence in the two adjacent normal cells.A serious corrosion of the fuel-starved cell is observed in the post-mortemstudy. The faulty cell presents a 5° contact angle decrease and a 28% ECSAloss. Scanning electron microscopy and Transmission electron microscopy resultsshow that the decline of anode outlet regions’ cathode catalyst layers aremore serious. Some optimal strategies have been proposed to solve thisproblem.

While for fuel starvation, a remarkable  voltage difference can be observed in the bipolar plate. Additionally, fuel  starvation may result in serious carbon corrosion.

The four-cell stack was assembled by  Shanghai Shen-li High Tech Co. Ltd.

Fig. 1. Experiment equipment and test  method.

Table 1 Basic information of fuel cell  stack and operation condition.

Fig. 2. Explanation of anode flooding and  current redistribution phenomenon.

Table 2 Mean contact angle results

Fig. 3. SEM and TEM result of fresh and  disabled samples.

Fig. 4. Modeling analysis of cell  interaction.

Fig. 5. Two-chamber fuel cell model  validation.

Table 3 Current and voltage results of  the three normal cells.

Fig. 6. Current distribution in the  four-cell stack under 50 kPa anode pressure.

Fig. 7. Anode pressure sensitivity  analysis.

Conclusion

In this paper, we analyzed a cell interaction  phenomenon in a multicell stack with one cell experiencing fuel starvation.

When using an innovative two-voltage  sampling equipment to monitor voltages along the anode channel, we can see  that a fuel starvation condition in a multi-cell stack where one cell  experiences fuel starvation shows specific voltage and current  characteristics. Specifically, the voltages of all cells in the anode  inlet region decrease. However, in the anode outlet region, only the  fuel-starved cell has lower voltage, and the other three cells have even  higher voltage than in their inlet regions. Such a phenomenon is useful  for diagnosing fuel starvation.

We proposed a cell interaction model to  explain the current and voltage redistribution phenomenon. We used a  two-chamber model to indirectly obtain current distribution within the fuel  cell stack. The estimated cross-plate resistance is consistent with the  theoretical value, which validates that the estimated current distribution in  this paper is credible.

Because of current convergence in the  fuel-starved cell, reaction currents concentrate in the anode inlet regions  of all four cells. About 85% of the current of the fuel-starved cell  concentrates in the hydrogen inlet part. Under the influence of a faulty  cell, current convergences of 73%, 68%, and 56% are observed in cells 1, 3,  and 4, respectively.

When far from the fuel-starved cell, the  current difference in the two chambers decreases. An increase in anode  pressure reduces the hydrogen flow rate, making anode starvation more  serious.

We observed serious corrosion of the  fuel-starved cell in our postmortem study. Compared with the other three  cells, only the fuelstarved cell shows a 28% ECSA loss. The contact angle  test shows that cathode GDL corrosion in the anode outlet regions of the  fuel-starved

cell is worse than in the other cells: a  5° contact angle decrease is observed in this test, and a 3° decrease is  observed in the anode inlet region. SEM results of the MEA show that the  cathode catalyst layers of the fuel-starved cell are much thinner in the  anode outlet region. Additionally, we also observed platinum particle  growth and aggregation in the TEM test. TEM results also show that platinum  particle degradation is more pronounced closer to the anode outlet.

Cell interaction in a multi-cell stack is  an interesting phenomenon. Also, the fuel starvation has specific voltage and  current characteristics that can be used to diagnose fuel starvation and help  explain why fuel cell stacks have shorter lifespans than single cells.  Although a twochamber model is credible in this work, it is still too simple  for an overall explanation about all the reactions. Future work is needed to develop  a more accurate 1D kinetic model to analyze the fuel cell stack evolution  process and propose an optimal control strategy to totally solve this  problem.