Design of hybrid emergency power supply system based on proton membrane fuel cell and lithium battery

China has a vast territory and various natural disasters occur frequently. The gasoline generators commonly used in emergency power supply equipment for disaster relief and emergency response are bulky, noisy and release harmful gases. Lithium batteries , nickel-metal hydride batteries, lead-acid batteries, etc. have short continuous power supply time and cannot provide charging recovery in emergency situations. This paper proposes an emergency power supply system that uses PEM fuel cells and lithium batteries for co-supply. Continuous power supply can also be guaranteed during the replacement of hydrogen storage containers. The control system uses fuzzy algorithms to dynamically distribute and manage energy according to the lithium battery SOC, the optimal working state of the fuel cell and the load conditions. A prototype for use in emergency situations has been developed. The system has a long continuous power supply time (2 to 3 times that of the current commonly used equipment), no noise, zero emissions, and can achieve good results. It is an ideal emergency power supply equipment for disaster relief and emergency response.

　　1 System composition

　　Fuel cell emergency power supply system.

　　The system consists of a 120W proton membrane fuel cell, a fuel cell controller, a lithium battery and management system, and an energy management unit. The lithium battery indicator is 13.2V/10Ah to ensure the tactical requirement of emergency full power support for 1 hour when the fuel cell fails or the fuel is exhausted and replaced in time. Fuel cell stack indicators: power is 120W, output voltage is 15V~28V. The fuel cell controller mainly controls and monitors the temperature of the stack, input hydrogen and air pressure, flow, and abnormal conditions of the stack, and transmits information to the system controller through the CAN bus. The system controller mainly completes real-time detection of load size, lithium battery SOC and fuel cell stack working conditions, and dynamically manages energy according to fuzzy algorithms, so that each component of the emergency power supply system works in the best state to improve the efficiency of the whole machine and the service life of key components.

　　2 Circuit Design

　　2.1 Charging and battery management circuit

　　The DC voltage is added to the 15th pin of MAX1873 through the isolation diode D5. Q1 is the charging drive signal output switch tube. R4 is the charging current detection resistor, which is used to detect the output current. R2 is the system current detection resistor. R5 and R6 are output charging voltage adjustment resistors.

　　The 15V~28V voltage output by the fuel cell passes through the isolation diode D5 and the total current detection circuit, one way passes through R2 and the DC/DC circuit to the output end, and the other way passes through Q1, inductor L1, D6 and R4 to charge the lithium battery. The voltage on R4 is proportional to the charging current, which is amplified by the voltage error amplifier and converted into a DC component and input into the microprocessor. The microprocessor will output a reverse control voltage from the 14th pin of MAX1873 to reduce the conduction current of Ql. If the current flowing through R4 is too small, the control voltage output by the 14th pin of MAX1873 will increase the current of Ql accordingly, which will make the battery pack have a constant current value. When the current is very small and reaches the minimum charging current or 0, MAX1873 outputs a low-level pulse control signal from the 14th pin to turn off BGl and stop charging the battery. When the control input is at a low level, BG2 is turned on, the charging control pin 6 (ICHG/EN) is at a low level, pin 14 outputs a low level, BG1 is turned off, and charging stops. At this time, the charging current is only 1μA and is in the off state (charging is prohibited).

　　2.2 DC conversion and control circuit

　　The DC/DC conversion circuit uses the XL4012 integrated converter, with an input voltage of 3.6V~36V, a switching frequency of 2800kHz, an output voltage adjustable from 0.8V~28V, a conversion efficiency of up to 95%, a maximum output current of 12A, and a simple peripheral circuit.

　　There are many parameters that need to be tested in the emergency power supply system: output voltage and output current of the fuel cell; charging current, battery voltage and battery SOC of the charging and BMS; output current and output voltage of the output end. Therefore, it is necessary to expand the A/D interface. The system control adopts 89S51CPU, and the A/D adopts TLV2543 chip. The chip has 10 analog voltage inputs and uses a serial interface with the microcontroller, which occupies less port line resources and has a faster conversion speed. The display adopts LCD1602 liquid crystal display. When the backlight is not used, the dynamic current of the liquid crystal is not more than 5mA. It mainly displays the working status of the fuel cell, the SOC and charging and discharging status of the lithium battery, the output voltage, output current information, the efficiency of the whole machine and other power supply information.

　　3 Fuzzy Control Algorithm

　　Keep the fuel cell in the best state and keep the lithium battery state of charge above SOCmin. Use the power share allocated to the fuel cell as a constraint to adjust the output power of the lithium battery. For lithium batteries, when the minimum limit value of the battery SOC (SOCmin) is less than or equal to 30%, the lithium battery must be charged; when the SOC is between 50% and 70%, it can be charged or discharged depending on the load power demand; when the SOC is greater than 90%, it is not charged. The load power Pg and the lithium battery state of charge SOC are used as the input variables of the fuzzy control, and the fuel cell output power Pfc and the lithium battery output power Pb are used as the output variables of the fuzzy controller. The basic domains of fuzzy input variables Pg and SOC are [0, 100]W and [30, 90]%. The input variables are fuzzified, and the fuzzy subsets are {ZO (zero), PS (positive small), PM (positive middle), PB (positive large)}; the domain of fuzzy output variable Pb is [-100, 110]kW, and the fuzzy subsets are also {NB (negative large), NM (negative middle), NS (negative small), ZO (zero), PS (positive small), PM (positive middle), PB (positive large)}; the domain of fuzzy output variable Pfc is [0, 110]kW, and the fuzzy subsets are also {ZO (zero), PS (positive small), PM (positive middle), PB (positive large)}.

　　The fuzzy controller selects the membership function of the input and output fuzzy variables based on the load power Pg and the charge state of the lithium battery as a triangle. The fuzzy control rules are connected by a series of relational words. The most commonly used relational words are if-then, also, or and and. The fuzzy control rules for each output and input are determined. The control quantity given by the fuzzy control algorithm needs to be defuzzified and converted into the basic domain acceptable to the control object. The defuzzification algorithm uses the centroid method.

　　4. Software Design

　　The system control program flow chart is shown in the figure.

　　5 System Simulation

　　A fuzzy controller is established in the Matlab simulation system. The domains of the input variables of the fuzzy control, the target power Pg and the state of charge SOC of the lithium battery, are [-100, 110] W and [30, 90]%, respectively. The domains of the output variables of the fuzzy controller, the fuel cell distribution output power Pfc and the lithium battery distribution output power Pb, are [0, 110] kW and [-100, 110] W, respectively. The lithium battery is 10Ah/13.2V, and the initial state of charge SOC of the battery is 60%. At the same time, the time is 0 to 15 minutes in Matlab/Simulink, and the simulation waveform is shown in the figure.

　　6Prototype testing and evaluation

　　According to the battery SOC and load size, the fuzzy algorithm is used to dynamically allocate and manage the energy of PEM fuel cells and lithium batteries. A prototype has been developed. Actual tests show that the power supply efficiency of the whole machine is above 90%, and the specific power is 120W/500g. When the initial SOC of the lithium battery is 80%, it can continuously supply power for about 16 hours to a metal hydrogen storage tank with a capacity of 600 liters. The continuous working time and maintenance are greatly improved compared with the performance of traditional emergency power supply equipment. It is currently being industrialized and has great promotion value.