Page 268 - Biofuels for a More Sustainable Future

240 Biofuels for a More Sustainable Future

2.3 Global sensitivity analysis by Latin hypercube design of
experiments
In the computations shown later, parameters were used as the independent
variables in the DOE: plant capacity, interest rate, feedstock (fish oil, waste
cooking oil, or jatropha oil) price, operating rate, biodiesel conversion. The
range of the plant capacity was taken to be 50–100kt/year. The interest rate
was taken to be between 3% and 5%. The cost of one ton each of Jatropha
oil, fish oil, and waste cooking oil was assumed to vary between US$400 and
US$500, US$100 and US$400, and US$0 and US$300, respectively. Bio-
diesel conversion efficiency was taken to range between 90% and 98%. The
operating rate was varied from US$190 to US$400 per ton.
Baseline cost distributions are shown in Fig. 8.5. To generate this figure,
the midpoints of the range of the six selected parameters were chosen. It can
be seen that feedstock cost, the operating cost, and the capital cost constitute
the largest bulk of the costs albeit to varying degrees depending on the feed-
stock. Though technically negative values, the absolute values of salvage
value and by-product credit are included to provide an idea of their relative
magnitudes. It can be seen that these and the maintenance cost are
relatively small.
For the sensitivity analysis of the LCC model, JMP software was used to
determine 50 sampling points for each feedstock via the Latin hypercube
experimental design. These are presented in Tables 8.3–8.5
Table 8.6 presents the results from the sensitivity analysis by DOE of the
LCC of biodiesel from jatropha oil. In this table, the factors that are signif-
icant at a 95% level of confidences are marked with an asterisk. It can be seen
that 9 factors and factor interactions are significant. These include each of the
5 factors: operating rate, feedstock price, capacity, interest rate, and biodiesel
conversion efficiency and 4 interactions: the square of (capacity-75) and the
products (interest rate-0.04) (operating rate-295), (biodiesel conversion
efficiency-0.94) (feedstock price-450), and (feedstock price-450) (inter-
est rate-0.04). The linear effect of each of the 5 factors maybe expected but
the design of experiments approach for sensitivity analysis provides addi-
tional information on the magnitude of each impact in relation to each
other. The factor with the largest positive (unfavorable) impact on the
LCC is the operating rate followed by the feedstock price and the interest
rate. On the other hand, the capacity and the biodiesel conversion efficiency
have negative (favorable) impacts on the life-cycle cost. A large capacity and
a high biodiesel conversion efficiency would result in a lower cost. Similarly,