(417d) Enhanced Power Management for Direct Methanol Fuel Cell/Battery Hybrid System By Improved Cascade Control Structure

Effective control structure for fuel cell hybrid systems is one of the critical milestones in the design phase that needs due attention. In the applications where the load connected to the cell is expected to change its demand in a frequent basis, the device might not be capable in some occasions to satisfy the load requirement. This is attributed to the fact that fuel cells’ electrical outputs are often fluctuating, and their dynamics are relatively slow during some transient periods such as start-up and shutdown.

In view of the fuel cell job of powering the connected loads, its integration in a hybrid system is essential to overcome the aforementioned shortcoming in the applications with changing loads. The literature has a good amount of work treating the hybrid systems consisting of fuel cells and batteries, supercapacitors or other storage devices. There are two main configurations available for hybrid systems (i) the series structure where the battery is the main source of energy to the load, and (ii) the parallel form where both devices feed the connected load. The first structure is considered less challenging since the cell feeds the battery which in turn provides the power to the load. On the other hand, the second configuration requires more considerations in the design and implementation of control strategy. To this end, the literature needs more research works investigating optimal designs of parallel systems to make them more efficient. In this work we provide a proposed parallel configuration structure for direct methanol fuel cell (DMFC)/Li-ion battery hybrid system that maximizes the dependance on DMFC and reduces the use of battery. We address the benefits and traits of the proposed architecture along with selected results.

PARALLEL HYBRID SYTEM CONFIGURATION CONCEPT

The parallel configuration utilizes a power management system (PMS) in a form of a supervisory control strategy. In this system, the electrical signal of the load is sent to PMS and the controller makes the decision whether to deliver the power by DMFC solely or by the combination of fuel cell and battery.

The signature of this work lies in the utilization of an improved cascade control structure that we call a “dual threshold algorithm” installed to regulate the liquid-level of methanol solution tank feeding the oxidation reaction species in DMFC. As discussed briefly in the next part, this control logic extends its service beyond the level management and leads to a stable temperature output of the cell that can be manipulated -intermittently- based on the user specification. The ultimate advantage is the ability to change the power of the DMFC from one state to another as a result of the temperature availability as a manipulated variable. In this fashion, the fuel cell can shift its power to satisfy the changes in load demands. However, this can happen in some periods during the course of operation where PMS will only use the feed of DMFC and this leaves the battery on sleep mode.

IMPORTANCE AND ADVANTAGES

The parallel hybrid configuration requires more investigation and research work to further advance the performance of such system. We argue that with a proper power management system organizing the electrical delivery to the connected load, several advantages could be realized. First, the dependance on the fuel cell could be maximized such that the battery life can be extended. Hence, the expenses associated with battery replacement become lower. Second, the battery can be used as a backup to compensate for the power in the event the fuel cell fails in satisfying load requirement. Thus, the system size is reduced since the battery pack needed would be less. Another aspect of pivotal role is the improvement of system efficiency where there is a good flexibility of optimizing the operational condition for both the fuel cell and battery pack.

CHALLENGES OF IMPLEMENTATION

Figure 1 demonstrates the cascade control structure used to regulate the level in the methanol solution tank of DMFC. The symbols LT, TT respectively designate the level transmitter and temperature transmitter. The level controller (master) is denoted by LC while TC signifies temperature controller (slave). The level set point is denoted by L_sp. The fan is the final control element operated by TC. The output of the master proportional-integral (PI) controller is given by

T^AFC = T_bias + K_p e(t) + K_i ∫ e(θ)dθ (1)

where T^AFC is the temperature of the anode flow channel, T_bias is the bias temperature, K_p and K_i are respectively the proportional and integral gains, while e(t) = L_sp- L(t) is the feedback error. It is evident from this law that the operational temperature of the cell (approximated to be T^AFC) is always prescribed by the master LC. This controller in turn has a single objective of tracking the set point L_sp. Thus, the temperature is continuously manipulated for the sole purpose of level management. The ultimate result is non-ability to operate the cell at the desired output power which is directly affected by temperature.

In view of cascade control mechanism, if the fuel cell is integrated in a hybrid system, the output power of the cell cannot be altered or shifted from one state to another since the temperature is not free for change. Consequently, upon load changes the battery has to be invoked in most occasions which violates the objective we try to achieve of maximizing the dependance on DMFC.

MODIFICATION AND IMPROVEMENT

A modified cascade control architecture detailed in [1] is proposed to improve the electrical performance of the cell. Its main idea is to make the tank operating within an upper and lower threshold enclosing a medium set point. The law of the modified cascade control is given by

T^AFC = T_bias + f(t) [ K_p e(t) + K_i ∫ e(θ)dθ ] (2)

where f(t) is a switching function adopting the binary value of 0 or 1. This modified version works in a fashion that when the level is within the threshold band, the switching function becomes 0 and the output of the controller is reduced to the bias temperature that can be freely selected by the user for any objective, such as changing the output power. On the other hand, if either one of the thresholds is violated, f(t) becomes 1 and the controller is returned to standard cascade so that the level is returned to the medium set point. Our work presented in [1] demonstrates via simulation environment that the cell can operate in periods of significant time where the master and slave controllers are decoupled (f(t)=0) and the power can be shifted successfully from one state to another.

This improvement shows that if the cell is integrated in a hybrid system, the battery usage becomes less frequent and extends its lifetime.

SELECTED RESULTS

This work is based on the DMFC/Li-ion battery hybrid system where we introduce the power management system in a simulation environment using MATLAB and Simulink software suite. We utilize a detailed dynamic modeling of open-cathode DMFC system that is developed by our research group while a widely used Li-ion battery model is imported from Simulink.

The reader should observe that there are several case scenarios for this hybrid system mode, where in some occasions the master and slave controllers are decoupled (f(t)=0), while in other events they are coupled (f(t)=1). Here and for the purpose of brevity, we show one example of the results, and we leave other cases for more detailed manuscripts.

Figure 2 shows the simulation results in the event where the master and slave controllers are decoupled with a switching function of 0.

The upper plot demonstrates the signal of the power requested by the load, the second is the output power from the fuel cell, the third is the battery power (both 2 and 3 as dictated by PMS), while the fourth is the total summation of the fuel cell and battery.

As seen from the figure, the requested power by the load is initially at 47 W and that is totally delivered by the fuel cell. A step change in the load from 47 W to 50 W takes place at 2000 seconds and that is also delivered by the cell while the battery is on sleep mode. The success of the fuel cell to change its output is attributed to the ability to change temperature from one state to another as given by Equation (2) when f(t)=0.

REFERNCES

[1] Zuhair A Al-Yousef, Shyam P Mudiraj, and Oscar D Crisalle. Improved cascade control

structure for water and thermal management in open-cathode direct methanol fuel cells. In

2019 American Control Conference (ACC), pages 5480–5486. IEEE, 2019.