IEEE Electrification Magazine - June 2016 - 12

subsidized for the first 50 or 100 units. However, even half
of ₹500 a month is not affordable for many of these households. It is possible that homes in the lowest-income group
may manage with less power and use one tubelight, one
bulb, and one fan instead of two. That will help, but the
quality of life would suffer. Many of the households would
still not be able to afford their power bills. At the same
time, slightly better-off homes would like to add refrigerators, mixers, and computers, adding to the power bill and
making it more unaffordable.
On the other hand, DISCOMs lose money even when
they supply power at ₹5 per unit. The cost of power from
plants using oil and gas (most of which is imported) is quite
high in India. However, India can produce power from coal
at a cost of ₹2 to ₹3 per unit (T. Buckley). Therefore, even
though coal is a pollutant, power production using coal has
been increasing rapidly in India. Even then, DISCOMs cannot break even when they supply power to homes. First,
they have to take into account transmission and distribution (T&D) losses (for rural homes it varies from 40% to
70%). Second, coal power takes time to ramp up and to
ramp down. Therefore, one cannot size coal plants for peak
loads. One needs other power sources with faster responses. These are usually oil/gas-based plants, where the cost of
generation is higher. This increases the cost of power for
DISCOMs. With regard to this, once the costs of meter reading, billing, collections (for large numbers of homes, each
paying small amounts), and the overhead costs are added,
DISCOMs start losing money. When state governments
push DISCOMs to supply power at subsidized rates (for
example, lower tariffs for the first 50 or 100 units), DISCOMs
lose even more. They have no incentive to continue to supply power or expand connections to homes not on the grid,
as they know that these homes can afford (and pay) even
less. Hence, at peak hours, they find one reason or another
to carry out load shedding. One retired chief engineer of an
Indian DISCOM remarked, "We are happy when there is
load shedding as we lose less money." This sums up the
reality faced in India.

Can Rooftop Solar Panels Address These Issues?
Recently, rooftop solar panels have been touted as an
alternative source for power generation. A 500-W solar
panel in most parts of India could generate most of the
power required. As there would be no T&D losses, the
solution looks promising. At an installed cost of ₹50 per
Wp, the rooftop solar photovoltaic (PV) would amount to
a little over ₹3 per unit, assuming a depreciation over 20
years and an interest rate of 7%. (In India, the commercial interest rate varies from 13% to 16% today. Homes
may be able to put their savings in fixed deposits and
earn about 7%.) This would be attractive. However, solar
power is available only during the daytime, and even
then, it fluctuates. On the other hand, DISCOMs face
peak demand both in the daytime and in the evening.
Hence, they are likely to resort to load shedding mostly

I E E E E l e c t r i f i c ati o n M agaz ine / j un E 2016

during these times. Thus, a rooftop solar installation
would require a battery, which increases costs considerably by almost four times, and as a result, solar power no
longer remains attractive.
Furthermore, a solar PV produces dc power that needs
to be converted to alternating current (ac) and synchronized to the ac grid. When 10-kW of dc solar power is converted and synchronized to ac, the conversion loss could be
as low as 3%. However, when 250-500-W solar- dc power is
used, these losses could be as high as 15% as long as the
converter cost is a small percentage of the solar-panel cost
(P. Kaur et al.). The problem gets further compounded as
input and output power of a battery is only dc. Alternating
current power needs conversion to dc before it charges the
battery, and the battery output needs conversion to ac
before it drives the load. Each of these conversions is also
likely to have a 15% loss. In addition, there is battery loss (as
high as 8%-10% for low-cost lead-acid batteries) and over
half the solar power is lost before it reaches the load.
The approach looks more absurd when one examines
the load to be driven. Some 62% of India's home load is
composed of ceiling fans and lighting (Global Buildings Performance Network). With the advent of brushless dc (BLdc)
motors, a dc-powered ceiling fan consumes only 40% of the
power consumed by conventional ac-powered inductionmotor-based ceiling fans (Global Buildings Performance
Network). If one uses an ac-powered ceiling fan, another
converter with about a 15% loss will be required. Similarly,
conventional ac-powered compact fluorescent lamp lighting is being replaced by light-emitting diode (LED) lighting.
LEDs use dc power and are best powered by dc. Electronics
(such as TVs, cell phones, and computers) are increasingly
being used in homes, and all electronics need dc power.
Taking solar power through multiple conversions to power
the ac home grid and then converting it to dc to power each
of these devices is indeed ridiculous. The stage is set for dc
microgrids for homes powered by rooftop solar panels having batteries and connected to the incoming ac power
through a converter as shown in Figure 2. This is the solar-dc
microgrid for a home.
One of the key challenges in designing such a
microgrid is to keep the losses low. The problem is not as
straightforward as it appears. The solar-PV panel's voltage
[at a maximum power point (MPP)] would vary during the
day. The battery's voltage would vary depending on its
state of charge. The grid power would be converted at an
independent voltage, and the load would be expected to
operate at some fixed voltage. If one uses dc-dc converters
to connect these units together, the losses may not be very
different from that of an ac home grid. The design would
therefore require some smart power electronics such that
the solar PV operates at its MPP and the battery is charged
and discharged optimally while driving the load with minimal losses. As discussed in the subsequent section, this
has been achieved such that for power in the range of 100-
500 W, solar-PV power aided by a battery drives the

Table of Contents for the Digital Edition of IEEE Electrification Magazine - June 2016