At present, the following key issues exist in the simulation research of auxiliary power supply systems:
(1) In the overall actual operation of the auxiliary power supply system, the instantaneous values of the output voltages are different due to the different starting times of the two inverters, resulting in a circulation between the two converters. This affects the normal operation of the converter and even causes the converter to be paralyzed. Therefore, it is necessary to adopt a certain control strategy to minimize the circulation between the two inverters.
(2) In the auxiliary power supply system, the charging mode of the battery pack is based on the constant current/constant voltage control strategy, and it is quite complicated. In order to simulate the real working conditions of the whole system, the output voltage and current of the charger need to be constantly changed according to the State of Charge (SOC) of the battery pack. Therefore, it is necessary to study the control strategy of charging the battery pack.
(3) The entire auxiliary power supply system with a large number of components is in a large scale, and the operation of the system is controlled by certain logic. Therefore, research on the control unit of the system is required.
In view of the above key problems, the topological structure and control principle of the key modules of the auxiliary power supply system are studied. The following research is made on the auxiliary converter parallel system, the battery pack charge management strategy and the overall control logic of the system.
3.1. The Stable Operation of Auxiliary Converter Parallel System
A single auxiliary converter converts the 3000 V DC voltage of the traction converter into stable three-phase 380 V/50 Hz AC voltages, which can quickly and stably respond to load abrupt conditions. The basic topology is shown in
Figure 3.
The auxiliary converter control adopts the SVPWM (Space Vector Pulse Width Modulation) control method, and the SVPWM modulation pulses drive the auxiliary converter to output a stable voltage.
Figure 4 shows the feedforward dual-loop control block diagram with the decoupling of output voltage and current. The inverter output voltage
uo, the output current
io, and the bridge arm current
iL are transformed into a synchronous rotation dq coordinate system. The inverter output voltage d-axis component
uod tracks the d-axis component reference value
uodref through the PI (Proportion Integral) controller. PI controller output value is superimposed on the output current d-axis component
iod and feedforward output voltage q-axis compensation component
-wcuoq as the d-axis component reference value
iLd*, for the bridge arm current
iLd. The bridge arm current
iLd tracks the d-axis component reference value
iLd* through the PI regulator. The PI controller output value is superimposed on the output voltage d-axis component
uod and the feedforward bridge arm current q-axis compensation component
-wLiLq as the d-axis component of bridge arm voltage
ud. Similarly, the q-axis component of the bridge arm voltage
uq can be obtained. Convert the d and q components of the bridge arm voltage into three-phase modulated waves
um (m = a,b,c) under three-phase stationary coordinates.
um generates a 6-way switch drive signal through the modulation module. This control method with a strong robustness can stabilize the output voltage of the auxiliary power supply system, ensuring the system will quickly return to a steady state when the load is switched.
The auxiliary converter uses parallel operation in actual work. The basic simplified model of two auxiliary converters in parallel is shown in
Figure 5, where
E1∠φ1 and
are respectively the no-load output voltage and the output current of the auxiliary converter 1.
r1 + jX1 is the output impedance of the auxiliary converter 1, equivalent to the sum of the output impedance of the auxiliary converter 1 and the line impedance value to the AC bus. The parameters of the two auxiliary converters are the same.
ZL is the load and
V∠0 is the load voltage.
shown in
Figure 5 is system current circulation.
In terms of the auxiliary converter parallel system, the parameters of the two auxiliary converters are the same, which is
r1 + jX1 = r2 + jX2 = Z, so the following formula (1) can be derived:
where
,
and
are vector form of two output currents and the system loop current, respectively.
It can be seen from the above equation that the system circulating current value is proportional to the no-load output voltage difference of the two auxiliary converters, and inversely proportional to the equivalent output impedance of the single inverter.
Larger circulations can affect the normal operation of the converter and even cause paralysis of the system. The common way to reduce the circulation is the droop control method introducing virtual impedance [
21]. The structure diagram is shown in
Figure 6. It mainly includes power measurement module, droop characteristic control module, virtual impedance module and current and voltage double closed loop control module.
According to the instantaneous power calculation method, when the coordinate system is a two-phase dq rotating coordinate system, the calculation formula of the instantaneous active power
p and the reactive power
q are:
The inverter output active power p and reactive power q obtained by Formula (2) have an AC component and a relatively high frequency. Therefore, a low-pass filter is added to obtain the average active power P and the average reactive power Q. As shown in Formula (3), ωc is the cutoff frequency of the low pass filter. P and Q are passed into the droop feature module as the final power input semaphore.
The droop characteristic control module is mainly realized by the absorption of active power and reactive power, which are triggered by the difference of the amplitude and phase of the output voltage of each single module in the parallel system. In the parallel system, the output resistive components of the auxiliary converter is small, and the difference between the single output voltage and the bus voltage is small, so it supposes:
X1 = X2 = X, r1 = r2 ≈ 0, sinφi = φi, cosφi = 1 (i = 1,2). Then the output active power
P1 and reactive power
Q1 of the auxiliary converter 1 are:
where
U1,
V and
φ1 are output voltage amplitude of the auxiliary converter 1, common bus voltage amplitude and the difference between the phase angle of the voltage at the outlet of the converter 1 and the phase angle of the bus voltage, respectively.
The output active and reactive power
P2 and
Q2 of the auxiliary converter 2 shares the resemblance:
It can be seen from the above formulas that the change of the output active power is mainly affected by the phase of the output voltage, and the output reactive power is mainly affected by the change of the output voltage amplitude. The basic governing equation for the drooping characteristic can be obtained as follows:
where
f and
U are frequency and amplitude reference values of the auxiliary converter output voltage,
f0 and
U0 are the frequency and amplitude of the output voltage when the auxiliary converter is unloaded,
KP and
Kq are droop parameters of active and reactive power, respectively.
Figure 7 shows the drooping characteristic curve of the parallel system. The droop control can achieve the active power and reactive power balance between modules.
According to the
P and
Q calculated by the power measurement module, the output voltage amplitude
u0* by the inverter can be obtained by the droop control. Since the output side of the inverter is equipped with a filter inductor under actual working conditions, the output impedance of the inverter is biased when the filter inductance is large, that is, the equivalent output impedance is biased at the power frequency. Therefore, the virtual impedance
Zvir is designed to be inductive. After passing through the virtual impedance module, a new voltage reference value
u*oref is obtained. Considering the influence of the virtual impedance, the calculation formula of
u*oref in the dq coordinate system is shown in the Formula (7).
where
u*odref and
u*oqref are the dq components of the output voltage reference value
u*oref after passing through the virtual impedance module,
uod* and
uoq* are the dq components of the output voltage reference value
uo* after the droop control module,
ω and
Lvir are the angular frequency in the dq coordinate axis and the virtual inductance, respectively.
Finally, the reference voltage
u*oref is generated by the feedforward dual-loop decoupling control of the voltage and current, which is shown in
Figure 4 to generate the trigger
um of the inverter, and the
um drive SVPWM control generates a trigger pulse to control the output voltage of the inverter.
In the auxiliary converter parallel system, the droop control method with virtual impedance is used to reduce the system circulating current value, achieve the balance between active power and reactive power, and improve the quality of the output voltage of the parallel system.
3.2. Battery Pack Constant Current Constant Voltage Charging Design
As the auxiliary stage of the auxiliary converter, the charger mainly supplies a sTable 110V DC power to the battery pack and DC load. The basic topology of the China standard EMU charger is shown in
Figure 8.
The charger also uses the current and voltage double closed loop control strategy. The output current and voltage values are compared with the reference value and then controlled by PI method. The output signal is sent to the IGBT pulse control device to regulate the stable output of the voltage and current.
Using the lithium battery as an example for the China standard EMU auxiliary power supply system battery pack, there are various polarization phenomena during the charging process. The ideal state of battery charging is to reach the goal of short charging time and small battery damage [
22]. Increasing the charging current can shorten the charging time, however, it also causes irreparable damage to the battery. Considering three factors of charging time, battery capacity and life, American scientist Joseph A. Mas has proposed a rapid battery charging curve through a large number of experimental studies, which is shown in
Figure 9.
The battery pack charging diagram shown in
Figure 10 can be obtained by considering the battery pack charging requirements, the fast charging curve and the actual charging conditions of the battery pack. The charging current is expressed by the charging magnification C. There are two stages in charging the battery pack. The first stage is the five-step constant current stage, the battery pack is first charged quickly with the maximum current (0.5 C) that can be accepted. When the battery pack voltage rises to the set voltage
V1, the charging current is reduced to 0.25 C. When the charging current decreases, the battery voltage decreases by a certain magnitude, and then continues to rise at a lower constant charging current until it reaches
V1 again, and this process is repeated. The second stage is the constant voltage phase, the purpose is to keep the battery pack voltage at the set voltage
V1 until the end of charging.
The battery pack energy management system (BMS) sends the maximum allowable charging current to the charger in real time, and the charger controls the output voltage and current according to the battery state. The basic parameters of the Chinese standard EMU battery pack are shown in
Table 1. According to above analysis, a battery pack charging flow chart can be obtained, as shown in
Figure 11.