Purpose:
The hardware in-loop experiment is designed to provide verification of the power management shell policy formulation. The experiment aims to verify that the load power can be measured and the corresponding reference power trajectories generated as per the PMS policy and decision epoch constraints.
Procedure:
The experiment was performed using the target hardware specified in Chapters 6. The PMS policy was first simulated using SIMPLORER. Figure 8.9 presents the PMS simulation module. From the previous experiment, the positive battery power limit was found to be 6.75kW. Also from experimental observations, the power delivery bandwidth of the ultracapacitors was set to +/- 10kW. A PMS epoch constraint of 10ms is achieved by setting DT=0.01s and setting the maximum simulation step time to also 0.01s. Both the battery charging power limit and the battery negative slew coefficient (Gnbatt)are set to zero hence disabling the battery regenerative power capability. The battery positive slew coefficient (Gpbatt) is however set to 5kW/second.
CONST Pbmax CONST Pbmin CONST PUCmax
CONST PUCmin
CONST Gbpos CONST Gbneg
CONST DT
LIMIT LIMIT1
LIMIT LIMIT2
MIN MUL1 MIN1
SUM1
UnitDelay GZ2
SUM2
MIN MIN2
CONST CONST1
CONST CONST2
MAX MAX1 SUM3 MUL2
CONST CONST3
MAX MAX2
CONST CONST4
SUM4
UnitDelay
GZ3
Ultracap_REF
Battery_REF Y
t
LoadPower
-5.00k 10.00k
0 5.00k
20.00 30.00 40.00 45.00
2DGraphSel1
LoadP...
-5.00k 10.00k
0 5.00k
20.00 30.00 40.00 45.00
2DGraphSel1
Ultrac...
-5.00k 10.00k
0 5.00k
20.00 30.00 40.00 45.00
2DGraphSel1
Battery...
Pbmax Pbmin PUCmax PUCmin Gpbatt Gnbatt DT
6750 0 10000 -10000 5000 0 0.01 Min Sim.
Step Time 0.001s Max Sim.
Step Time 0.01s
Load Power
Ref Battery Power Ref. Ucap Power
PMS
Implementation of the PMS is as shown in Figure 8.10. Developed in Labview7, the PMS algorithm is executed within the target hardware real-time controller .The PMS deterministic loop is set as 10ms with a Boolean flag to indicate a late finish of the execution. Outputs of the loop are the battery and ultracapacitor feed-forward power references. Acquisition of the measured power is performed using a continuous but interruptible polling subroutine. The measured power is then made available to the PMS via the input variable marked as
‘Numeric’ in Figure 8.10. Experimental data is stored in memory at a rate of 300 samples per test segment.
Figure 8.10 PMS Implementation module
Results:
The simulated and implemented power splits generated for the four test segments by the PMS module are presented in Figure 8.11 through Figure 8.18. In the four simulation plots, the battery reference power trajectories conform to the power limit and rate of power constraints. The ultracapacitors handle the transient power demands as well as the load magnitudes above the specified battery power limit.
Test Segment 1
Power (W)
Test Segment 1
Load Power Battery Power Ultracapacitor Power
Time (s)
Figure 8.11 Power management in Segment 1- (Simulation)
Figure 8.12 Power management in Segment 1-(Implementation)
Test Segment 2
Power (W)
Test Segment 2
Load Power Battery Power Ultracapacitor Power
Time (s)
Figure 8.13 Power management in Segment 2 – (Simulation)
Figure 8.14 Power management in Segment 2- (Implementation)
Test Segment 3
Power (W)
Test Segment 3
Load Power Battery Power Ultracapacitor Power
Time (s)
Figure 8.15 Power management in Segment 3- (Simulation)
Figure 8.16 Power management in Segment 3- (Implementation)
Test Segment 4
Power (W)
Test Segment 4
Load Power Battery Power Ultracapacitor Power
Time (s)
Figure 8.17 Power management in Segment 4- (Simulation)
Figure 8.18 Power management in Segment 4- (Implementation)
Discussion:
Results of the PMS implementation match the simulated ones. For all four-test segments, the PMS loop proved deterministic by generating the reference power splits within the PMS epoch. No ‘Finished-Late’ flags were reported in the experiment. The experiment verifies that the concept of the PMS generating reference power trajectories by measuring load variation on the DC bus is achievable. At this point, it is timely to reiterate a claim in Chapter 1, which states that, “power split decisions are made using only power fluctuations at the DC Bus rather than the conventional methods of monitoring the throttle input (driver input). This leads to the ability of including propulsion as well as non-propulsion loads in the implementation framework”.
Since the previous experiments validate the battery model, an offline method to test the effect of introducing the PMS is accomplished by subjecting the vehicle model to the battery reference power generated by the PMS. With this, a comparison of battery current and voltage with and without the PMS can be made. Figure 8.19 presents simulated comparisons of battery currents while Figure 8.20 compares the voltage profiles. In all four battery current profiles, the introduction of the PMS is manifested as a reduction in current ramp rate. The 5kW/s positive slew rate coefficient (Gpbatt) effectively limits the battery current rise to approximately 100A/s. The clamping of the battery current at 150A is most apparent in Segment 3 of Figure 8.19.
The voltage waveforms (Figure 8.20) show that with the PMS activated, the battery voltage operates within the specified bounds. Voltage excursions do not exceed the predetermined 45V lower level. Although not apparent from the experiment, operating the battery system in under such conditions is electrically and thermally preferred [62], and does result in an extension of the battery life cycle.
Battery Current (A)
Time (s) Test Segment 1
Without PMS With PMS
Battery Current (A)
Time (s) Test Segment 2
Without PMS With PMS
Battery Current (A)
Time (s) Test Segment 3
Without PMS With PMS
Battery Current (A)
Time (s)
Test Segment 4 Without PMS With PMS
Figure 8.19 Battery Current – with and without PMS
Battery Voltage (V)
Test Segment 1 Without PMS With PMS
Time (s)
Battery Voltage (V)
Test Segment 2 Without PMS With PMS
Time (s)
Battery Voltage (V)
Test Segment 3 Without PMS With PMS
Time (s)
Battery Voltage (V)
Test Segment 4 Without PMS With PMS
Time (s)
Figure 8.21 provides a comparison of the battery energy expenditure with and without power management. The dotted line marks the energy expenditure without a PMS. As seen, the gain in terms of energy savings at the end of the 4 tests is marginal. In this case, only 3.4%. Although the PMS policy limits the peak power delivery of the battery through the use of ultracapacitors, the total energy to propel the vehicle still comes from the battery pack. In addition, to transfer the energy from the battery to the ultracapacitor results in power conversion losses. Therefore, a large energy saving is not to be expected at the end of the short test cycles. As discussed, mitigating high power peaks from the battery system does however result in a long-term gain of extending the battery life as well as the long-term energy storage efficiency.
Time (sec) Time (sec)
Battery Energy (J) Battery Energy (J)
Without PMS WithPMS
Figure 8.21 Battery energy expenditure after four tests segments