Since a hybrid vehicle is powered by two power sources, the engine and the battery, its design is much more complex than that of a traditional vehicle powered by a single power source (engine). At present, the method of experience and trial calculation is mainly used, referring to the selection mode of traditional cars, determining the power level of each component and the range of related parameters through some empirical formulas, then selecting parts that meet the conditions, and then verifying the selected parameters. Or adjustment, the parts selected in this way have many limitations. Although they can meet the performance of the car, the matching between the parts and each other is not optimal, and the process requires repeated verification and adjustment many times.

　　Foreign research mainly focuses on energy management and distribution strategies, and there are few documents specifically discussing matching optimization. Among them, the ADV IOR software has a matching function. This software mainly adjusts and calculates existing component parameters, and is targeted at specific operating conditions and vehicle types. This paper takes the series fuel cell hybrid vehicle as the research object. Through the analysis and transformation of the research object, the power source selection and matching problem of the series hybrid vehicle is transformed into a mathematical optimization problem, and different energy management strategies are used to solve the problem. Considering the performance of parts and components, use optimization algorithms to solve them, and conduct analysis and research based on the results obtained.

　　1 Introduction to optimized configuration

　　The configuration of a series fuel cell hybrid vehicle is shown in Figure 1. The entire vehicle solution consists of a fuel cell engine, DC/DC, power battery, motor and transmission. The solid lines in Figure 1 are mechanical connections and the dotted lines are electrical connections. The fuel cell engine generates electricity, and after DC/DC conversion, it is electrically coupled with the battery to provide energy to the motor, which drives the transmission to drive the entire vehicle.

　　Figure 2 is a schematic diagram of the energy input and output of components. In order to reduce optimization parameters and processing needs, the fuel cell engine and DC/DC are considered as a whole, represented by PS in the figure. The control unit is responsible for the energy management strategy, determining the working status and output power of the PS and battery stack, and the working status of the motor.

　　For the fuel cell vehicle studied, its vehicle parameters are shown in Table 1, and its power performance indicators are shown in Table 2.

　　2 Transformation and processing of optimization problems

　　The above optimization problem can be described as: for the specified vehicle structural parameters, dynamic indicators and selected working conditions, select powertrain components with different power levels and characteristics so that the vehicle can fuel the vehicle while meeting the dynamic indicators. The best economy. That is, under the constraints of dynamic indicators, find the optimal components and their performance parameters to achieve the best fuel economy of the vehicle.

　　2.1 Determination of optimization function

　　The objective function is cycle fuel economy, which consists of two parts: hydrogen consumption of the fuel cell engine and hydrogen consumption converted from changes in battery SOC, namely

　　In the formula, b is the engine fuel consumption rate, U and I are the voltage and current of the battery respectively, Peng is the output power of the fuel cell engine, Qfuel is the low calorific value of the fuel, Xeng, Xbat and Xmot are the power levels of the engine, battery and motor respectively. T is the total time of cycle conditions.

　　Since all the mechanical energy of the series hybrid vehicle is output by the motor, the selection of the motor is based on the premise that it satisfies the dynamic index of the entire vehicle. Based on this, the maximum power of the motor is obtained, and the distribution of the operating point in the high-efficiency area of the motor is determined by the corresponding The cyclic working conditions are converted, and finally the motor power that meets the requirements, as well as performance indicators such as efficiency maps are obtained. In this way, the parameters of the optimization problem can be reduced, and the above optimization objective function can be transformed into a selection problem of battery and fuel cell power levels. The total power demand can be determined by equation (2)

　　In the formula, Pneed is the total power demand of the vehicle, m is the mass of the vehicle, CD and A are the air resistance coefficient and windward area of the vehicle respectively, v is the vehicle speed, and D is the rotation mass conversion coefficient.

　　At this time, the engine power demand and battery power demand are

　　In the formula, V (SOC, Pneed) is the power distribution coefficient, which is determined by the control strategy.

　　2.2 Transformation of constraints

　　The constraints are power indicators, that is, the power components can meet the requirements of the maximum vehicle speed, maximum climbing angle and acceleration time, as shown in Table 2, which are converted into corresponding constraints on the power of the fuel cell engine and battery, that is,

　　In the formula, Pneed (vmax) is the power demand at the highest speed, Pi (v) is the power demand at vehicle speed v and slope i, Geff and Gmot are the mechanical transmission efficiency and motor efficiency respectively, Tmot is the output torque of the motor, i0 , ig are the main reducer and transmission speed ratios respectively, r is the wheel radius, Pfc_max and Pbat_max are the maximum output power of the fuel cell and battery respectively.

　　For the maximum vehicle speed, the power required to reach the maximum vehicle speed is obtained from equation (2), that is

　　The power requirement of the vehicle can be obtained from the maximum gradeability requirement as

　　The constraint of the acceleration time is the acceleration phase. The motor torque Tmot(t) and the corresponding vehicle speed v(t) at each moment should satisfy

　　In the formula, fw and ff are the wind resistance and rolling resistance respectively at the current vehicle speed, and vset is the terminal speed given in the acceleration time indicator.

　　2.3 Control strategy

　　For different control strategies, it means that the vehicle has different performance and requirements for the corresponding parts. The article gives two typical energy management strategies used by series hybrid vehicles, namely switching control strategy and power following control strategy. Control Strategy. By comparing and analyzing the optimization results of different control strategies and control parameters, the results of component selection and optimization suitable for different design requirements are obtained through comprehensive consideration and analysis.

　　(1) Switching control strategy The engine switch is determined by the upper and lower limits of the battery SOC. When the engine is working, it works at the optimal fuel economy point. This control strategy mainly relies on the battery to follow and respond to the power demand of the vehicle.

　　(2) Power following control strategy The engine starts and stops based on power demand and battery SOC. The engine works on the optimal fuel economy curve and tries to maintain the battery SOC at the set value to ensure driving range.

　　The control strategy is mainly used to allocate power according to working conditions, but it is difficult to express it in mathematical expressions. This paper uses model modeling to implement the control strategy.

　　2.4 Introduction to optimization algorithms

　　For the above optimization problems, engineering optimization algorithms can be used to deal with them, and engineering methods can be used to deal with the convergence and boundary conditions in the optimization problems, and finally obtain better optimization results.

　　For the optimization problem described above, two optimization algorithms, sequential quadratic programming and partition matrix method, are used in this paper to optimize and solve the problem. The mathematical model obtained by converting the above engineering problem and the control strategy represented by the modeling are substituted into the two optimization algorithms, the cut-off conditions are set, and then the following results are obtained through iterative calculations. The two optimization algorithms are as follows.

　　(1) Sequential quadratic optimization algorithm (SQP)

　　For a given initial value, a quadratic programming sub-problem is constructed. By solving the quadratic programming sub-problem, the search direction and step size are obtained, and the solution is iteratively optimized. The advantage of this algorithm is that it is mature and very effective for smoothing problems, but it requires the optimization function to be differentiable, and at the same time it obtains a local optimal solution.

　　(2) Division matrix method (DIRECT)

　　For the optimization interval, first normalize and calculate the function value of the center point, set the number of iterations, determine the optimal matrix set, select a rectangle in the set to divide and calculate its center point value, until the number of iterations is reached or the optimization interval iteration is completed. This algorithm does not require the optimization function to be differentiable, can obtain global optimization, does not require the selection of initial values, and does not require control parameters during the optimization process. However, this algorithm does not have a convergence criterion to judge whether the optimization has converged, so it can only optimize a small number of variables.

　　3 Optimization result analysis

　　Based on typical bus urban working conditions, the average output power based on the working conditions can be obtained. Converted to the fuel cell engine, the power is 42kW. That is, the optimal operating point of the fuel cell engine is located near 42kW. From this comparison, the power of the fuel cell can be obtained The level is around 52kW. In order to meet the requirements of working conditions, the instantaneous power the battery needs to provide reaches 150kW. For a large-capacity battery of 80A#h, the voltage level required for 3C discharge reaches 625V, which requires 52 cells and 10 cells. Only a battery module composed of batteries can meet the requirements, and the results obtained cannot be used.

　　In view of the above optimization problem, a model of the complete vehicle, components and control strategy was established, and two optimization algorithms were used for optimization simulation. The final results are shown in Table 3.

　　3.1 Comparative analysis of control strategies

　　The switch-type control strategy is generally used in vehicles with poor dynamic response of fuel cell engines, batteries as the main power source, and large hybrids. Therefore, relatively large batteries are generally used. For this control strategy, when the battery SOC is between 013 and 018, it means that the battery needs to provide power in a wide range. Under this control method, the fuel cell rarely participates in the work, and the battery provides power under most working conditions.

　　When the battery SOC is between 014 and 016, due to the small SOC control range, that is, the battery's working range is relatively small, the number of battery modules finally optimized is smaller than the control strategy for the battery SOC between 013 and 018. When the battery SOC is between 013 and 018, it is mainly powered by the battery. The battery capacity is large, the charge and discharge current is small, the efficiency is high, and the power consumption is replenished at the highest efficiency point of the fuel cell, so the economy is good; while the battery SOC is Between 014 and 016, the battery capacity is small, the charge and discharge current is large, and the efficiency is low. When the engine starts and stops, due to the limitation of the power change rate, more fuel is consumed, so the economy is better than the battery SOC between 013 and 018. difference between.

　　The power following control strategy is suitable for vehicles where the fuel cell engine has good dynamic response, uses the fuel cell engine as the main power source, and is a moderate or mild hybrid vehicle. This control strategy requires the fuel cell to provide most of the driving energy. As can be seen from Figures 3 and 4, the fuel cell provides most of the energy. The battery is used to make up for the insufficient power demand of the fuel cell. Its power output is basically 20kW. below, and the battery SOC is basically maintained around 015.

　　Due to the dynamic changes of the fuel cell engine, its operating point is not the optimal economic point. However, because the fuel cell has high efficiency in a wide range, the fuel economy obtained by the power following control strategy is better than that of the switching control strategy. Battery SOC between 014 and 016 has better economic efficiency, but it is worse than battery SOC between 013 and 018.

　　Since the average power demand under working conditions is about 50kW, in order to meet the requirements of dynamic indicators, the required power is about 200kW. Although the two different control strategies lead to different main power output components based on working conditions, due to the limitations of dynamic indicators , the final result is that the power level of the components is higher.

　　3.2 Comparative analysis of different optimization algorithms

　　Comparing the two optimization algorithms, we can see that for the switching control strategy, in the battery SOC range of 013~018, since most of the working conditions are mainly operated by the battery, the optimization results obtained are the same, while when the battery SOC is In the interval of 014~016, the optimization results obtained by the two algorithms are quite different, but the fuel economy indicators are basically the same. This is mainly because the end conditions used by the two optimization algorithms are different, resulting in different final optimal results. See Figure 5. ,Figure 6.

　　It can be seen from Figures 5 and 6 that during the entire cycle, the fuel cell engine is initially turned off, and most of the driving energy is provided by the battery. When using a smaller battery (see Figure 5), the fuel cell engine mainly starts when the battery SOC is lower than the set value or the power demand is large in the last period of the working condition. The fuel cell works for a long time. At this stage, the fuel cell is used to provide power and Charging the battery is beneficial to improving system efficiency, but the discharge current of the battery is large, which is detrimental to the life of the battery, and the efficiency of the battery is low. When a larger battery is used (see Figure 6), the discharge current of the battery is reduced, which is beneficial to the life of the battery and the efficiency of charging and discharging. The fuel cell is mainly used to maintain the SOC value of the battery.

　　4 Conclusion

　　(1) Different energy management strategies correspond to different selection requirements for power components. When selecting, the impact of energy management strategies on component performance needs to be fully considered.

　　(2) For the same control strategy, there may be multiple sets of optimization solutions, so it is necessary to consider which combination of components to use based on the actual component performance.

　　(3) The SQP algorithm requires a suitable initial value to be given in advance, but the DIRECT algorithm does not. However, since the DIRECT algorithm does not have a convergence criterion, the optimization solution time is longer than that of the SQP algorithm, but both algorithms can solve the problem.

Read recommendations:

601248 300mAh 3.7V

3.7V 18650 lifepo4 battery.What is the process of 18650 lithium battery cells?

2200mah 18650 battery.University and Toyota Research Institute develop AI algorithm to predict batte

703048 battery sales

AG3 battery