In Table 2, under “Summer (July 2018),” the “Average daily renewable energy generation” is listed as 14,079.17 kWh. However, in Table 3 (page 18), the same value appears as “Renewable energy generation (kWh)” for summer. But in Table 3, the “ONO with ESS” summer value is also listed as 14,079.17 kWh—which is identical to the total generation, not the utilized amount. This seems misleading. Shouldn’t “renewable energy generation” be the same in both tables, and “utilization” be a percentage of that? The way it’s presented conflates generation with consumption.
In Section 5.1.2, the authors claim that improving prediction accuracy of renewable energy has negligible impact on performance (only ~0.5% improvement in winter). However, they only test errors up to 100% reduction (perfect prediction). Given the intermittency of solar power, couldn’t large overpredictions lead to insufficient grid power being planned, risking bus charging shortfalls? The study seems to assume the grid can always compensate, but what if grid capacity is constrained during peak hours?
The delay model uses a truncated normal distribution based on “minimum driving time,” “time-dependent extra driving,” and “anticipated extra delay.” But in Figure 8, the red line (“anticipated extra delay”) is shown at roughly +2 minutes, yet the distribution appears skewed right. If delays are skewed, a normal distribution may underestimate the frequency of large delays. Did the authors test other distributions (e.g., log-normal) to see if reliability results change significantly?
Table 3 shows that adding an ESS improves renewable energy utilization by 7.9% in winter but only 1.4% in summer. This makes sense intuitively (more solar in summer), but the absolute increase in renewable usage is higher in summer (191 kWh vs. 118 kWh in winter). Why is the percentage improvement so much lower in summer? Is it because the baseline utilization is already high, or because the ESS is frequently full and can’t store excess? The explanation is lacking.