System performance under worst-case cost minimization

With all the conditions held equal as those in 6.1.1, the microgrid model for Taichung Industrial Park is then optimized as to minimizing the worst-case cost of the same five scenarios (i.e. the highest scenario cost). The resulting planning of the capacity design for each DER power source is listed in Table 6, with the estimated cost of scenario 3 to be $300,008,330 for a time span of one year. Although the expected cost of all five scenarios is now higher than that in expected cost minimization, the variations between individual scenario costs are significantly suppressed. And in most cases, the variations can be reduce to zero (like in the current case), which means that the five scenarios may have the same cost.

Table 6: Worst-case cost minimization – DER sources capacity design (Unit: 10 kW) DER Source H-ICES H-ICEL FC HC-FC HC-GT HC-GE PV H-PV WT Capacity 300 2,000 0 294 1,781 2,000 294 6,000 6,000 Resulting total cost of scenario 3: $300,008,330

Comparing the capacity design in Table 6 to that in Table 5, it can be found that the two sets of capacity allocation are very similar. The only difference lies on the capacity of HC-GT, which is 17,810 kW under worst-case cost minimization but only 9,440 kW under expected cost minimization. The difference in capacity of HC-GT contributes to part of the increase in the overall expected cost of the current setting, if compared with the setting in expected cost minimization. However, the increase of the overall expected

cost is not only caused by the increased fixed cost in capacity enlargement, but also some other factors, for instance, the less capacity utility rate. On the contrary, the increased system capacity may be helpful in accommodating the possible sudden rise in demand once in a while. The reserved room in capacity for emergency usage may sometimes prevent the local community from suffering a huge loss caused by severe grid failure due to overload, which may also effectively lower the worst-case cost and reduce the variations among different scenario costs.

Appendix 6 shows the predicted distribution of power generation by different DER sources in the microgrid under worst-case cost minimization. Except for fuel cells (FC) which is assigned zero capacity, all the eight power sources installed are recommended to operate continiously from January to December under the demand and conditions in scenario 3. The diagrammatic distribution of power generation among different DER sources and the overall electricity usage is presented in Figure 12. It can be seen from the chart that the internal DER power generation in this scheme does not supply the microgrid with much more amount of electricity in summer than in winter, unlike the scheme shown in Figure 10.

Instead, it can be observed from Figure 12 that a great amount of power generated in summer is sent to storage and the shortage in supply is covered by electricity purchase from the main grid. This phenomenon looks abnormal but is possible to happen, as the microgrid is subjected to too much variation in demand and supply and too many uncertainties in system components. The program would simply take all the input conditions into account and determine which scheme to form, based on economic equivalent concerns. It should be noted that all non-economic factors, such as

environmental impacts, are converted to be evaluated on the same monetary units with the other economic factors, as set forth as one of the assumptions in the current model.

Another feature of the DER power generation distribution, as shown in Figure 12, is that the amount of electricity generated from H-PV varies significantly from month to month.

Unlike typical internal combustion engines, photovoltaic equipment allows more flexibility in accommodating the sharp changes in throughput from time to time, which makes the PV system capable of fitting in many energy-saving schemes or operation plans and become favorable in robust optimization design.

Figure 12: Distribution of power generation and usage under worst-case cost minimization – Microgrid for Taichung Industrial Park

As carried out in the previous section, the system performance of trigeneration (CHP and CCHP) under worst-case minimization is also predicted and listed in Appendix 7.

The waste heat collected from power generation can be recovered to support the heating and cooling demand of local community. Some recovered heat is used for heating directly, while the other can be used for cooling through the function of absorption chillers and the rest be reused in generating electricity.

The distribution of heat recovery by different DER power sources under worst-case cost minimization in scenario 3 is illustrated in Figure 13. It can be identified that the scheme of heat recovery here differs significantly from that in Figure 11. In the current case, a large portion of electricity demand in summer is scheduled to be satisfied with power purchase from the national grid or energy outflow from the storage, rather than by DER

generation. As a result, not too much heat can be recovered for other use. In addition, it can be seen that the recovered heat from H-PV is significantly less than in the case of expected cost minimization.

Figure 13: Distribution of recovered heat under worst-case cost minimization – Microgrid for Taichung Industrial Park