IEEE Power Electronics Magazine - September 2022 - 53

energy delivered to the output of 1116 Wh during
a day (94% from the extracted PV energy). After
a fault, the performance of the microconverter
depends more strictly on the input voltage and,
consequently, on the operating mode. It features
relatively high efficiency in the buck mode at the
input voltages above 44 V, which corresponds to
the operation during the morning and evening
hours, as could be seen from Figure 4b. During
the peak solar irradiance hours, the MPP voltage
of the PV module drops below 44 V, and the
microconverter starts operating in a boost mode.
In that case, the topology is reconfigured into
the quasi Z-source (qZS) single-switch converter
that suffers from relatively high current stresses.
This results in an efficiency drop of 5%-6% and,
consequently, daily output energy yield dropped
down to 1070 Wh after a fault, which is 90% of
the harvested PV energy. This is an acceptable
performance deterioration considering that the
converter can continue operating after a semiconductor
switch fault.
Post-Fault Operation Issues of ZeroRedundancy
Fault-Tolerant Converters
The efficiency of the zero redundancy FT dc-dc
converters typically deteriorates after a fault, and
the efficiency drop could vary significantly
between the converter operation modes. This
observation imposes an issue related to overloading
of the critical components when attempting to
operate close to the rated power after the faultinduced
topology reconfiguration. If the zero
redundancy FT converter continues delivering the
rated power after the fault, the lifespan of the
healthy components could be shortened as they
will face high thermal stress or even catastrophic
failure, depending on design trade-offs. For example, the
low-cost designs use smaller printed circuit boards and
cheaper semiconductors to meet the cost constraints.
Hence the low-cost implementations of zero redundancy
FT dc-dc converters may require software constraint
(curtailment) of the input power to be introduced in the
control system to ensure safe post-fault operation. On the
other hand, in many applications, including PV, operation
at the maximum power happens rarely and contributes
only a small fraction of the annual energy yield [29].
A simple solution to improve the reliability of FT dc-
1400
1200
1000
800
600
400
200
No Limit 250 W
200 W
150 W
Maximum operating power
(a)
100
20
40
60
80
No Limit 250 W
Maximum operating power
150 W
200 W
(b)
FIG 5 Reliability of the FT PV microconverter before and after the
occurrence of a fault with power curtailment, considering the 60-cell Si
residential PV module on the input side coupled with a dc microgrid of
350 V on the output side. (a) Yearly failure rate. (b) Yearly PV energy
yield predictions.
switches from the MPP tracking to the power derating
mode. However, the microconverter input voltage will be
different from the MPP voltage.
A trade-off between extending the converter reliabildc
converter with zero redundancy and, thus, extend its
lifetime, is to curtail the converter input power at a certain
level during infrequent PV energy production peaks.
In the power curtailment mode, the failure rates of the
critical components will be reduced significantly due to
reduced thermal loading [25]. When the maximum power
of the PV module becomes higher than the predetermined
curtailment power level, the converter control system
ity using power curtailment and associated reduction
in the energy yield of the PV system should be considered.
There are two possible solutions to the power curtailment
problem: operating above and below the MPP
voltage. In practice, the power curtailment to the point
above the MPP voltage is more beneficial as high stepup
converters tend to provide higher efficiency at lower
dc gain [26], [27]. In addition, it is simpler to reach that
point when starting MPP tracking from the open-circuit
voltage of the PV module.
To define the random failure rate of the converter during
its prognosed lifetime, the reliability approach from
the FIDES handbook can be adopted [28]. This methodology
allows for defining how the random failure rate
of components operating under dc stress depends on
the variations in component stress resulting from the
September 2022 z IEEE POWER ELECTRONICS MAGAZINE 53
100 W
Normal
Post-fault
100 W
Normal
Post-fault
PV Energy Yield Prediction [%]
Failure Rate [105 Failure/Year]
IEEE Power Electronics Magazine - September 2022

Table of Contents for the Digital Edition of IEEE Power Electronics Magazine - September 2022
