Embodied Energy and Off-Grid Lighting, Green Tech, Page 8

Both solar and grid charged LED lamps have fast energy payback periods considering the
amount of avoided kerosene we observed among adopters of LED lighting. The average user of
LED lighting consumed 5.2 MJ of kerosene per night before adopting LED lighting and 2.8 MJ
per night of kerosene after.3 The users in our study uniformly chose to use grid recharging, with
a median daily requirement of 0.5 MJ of primary energy to generate the required electricity
based on the observed frequency of lamp recharging, the Kenya grid mix, and measured
efficiency of the charging system.4 Therefore, a total of 2.3 MJ primary energy use was avoided
each day for the average user in the study. The result is an energy payback time of about one
month for the grid-charged version of the LED lamp. None of the vendors chose to purchase a
solar module for recharging, but, if they had, and kerosene use post-purchase remained
unchanged relative to kerosene use for those who used the AC charged lamps, their payback time
would have been approximately double, i.e., two months.


Figure 3 shows three embodied energy scenarios: the primary energy consumption over two
years for (i) a grid charged LED lamp like the one we deployed, (ii) a corresponding solar
charged LED lamp, and (iii) one month of kerosene offset by the LED lamps (considering that
about 50% of the baseline kerosene was offset, which is cautious but also consistent with our
field observations in the 2008 market test). The figure shows that over a two-year lifetime,5 the
solar charged option has higher primary energy requirements than the grid charged one.

3
The off-grid lighting users we studied were night market vendors in two Kenyan towns: Mai Mahiu and Karagita.
Both towns are relatively small (<20,000) and located in the Rift Valley Province. Before our study, the vendors
relied on various fuel-based lighting technologies to illuminate their nighttime businesses. We surveyed 50 vendors
to establish baseline fuel use trends and carefully measured baseline lighting fuel use for a subset of 23 vendors. We
then offered the opportunity to purchase an LED light with and without a solar charging option to the 23 for whom
we had established a detailed baseline; 14 chose to purchase an improved lighting product. We tracked kerosene
use, user satisfaction, and expenditures for lighting for all 23 vendors over a one-year period). The mean GHG
emissions over the one-year study period for those who did not adopt LED lighting was 130 kg CO2e/vendor-year
from burning approximately 150 mL of kerosene a day. Those who purchased LED lights reduced their year-long
emissions from burning kerosene by approximately 50% to 65 kg CO2e/vendor-year; the mean kerosene
consumption rate for them was 79 mL/day.
4
The Kenya grid had a primary energy heat rate of 5.6 MJ/kWh in 2007 (KNBS, 2008), assuming that the thermal
efficiency of hydroelectricity and geothermal electricity is unity and that the average efficiency of thermal,
cogeneration, and imports is 33%. Based on the measured charging efficiency of the AC charger of 21% and
assumed battery efficiency of 70%, the lamps we offered required 25 Wh of grid electricity for each charging cycle.
The median observed recharging rate for the lamp users was once every three days.
5
Our assumption is that the commercial version of the LED lamp we distributed has a lifetime of about two years,
based on our extensive lab-based testing of off-grid lighting products (Mills and Jacobson 2007) and observations
we made in the field of use patterns and the rigors of actual use. The modified lamps we distributed had shorter
lifetimes in practice due to design flaws in the detachable lamp head and housing we used.


8