Scenario analysis under worst-case cost minimization

Similar steps are conducted for worst-case cost minimization as in 6.2.1. Figure 20 shows the comparison of individual scenario costs with their cost structures. Very different from the results for expected cost minimization, each of the five scenarios has the same total cost, which means that there is no variation among total scenario costs.

The difference lies on the distribution of cost components within each scenario. While the cost of capital investments (CInv) slightly decreases due to technology advancement in developing DER equipment utilizing renewable energy, as is the same in expected cost minimization, cost of electricity purchase from the national grid and cost of carbon tax increase from scenario 1 to scenario 5. On the other hand, items such as cost of fuel consumption (including direct consumption and DER usage) and cost of operation and maintenance decrease because the shift of traditional power sources to new power sources such as PV and wind farms can reduce the consumption of fuels and the cost in equipment maintenance.

Scenario 1 2 3 4 5 CInv 3,821,129 3,517,859 3,214,283 2,911,013 2,607,437

CEbuyN 3,120,908 3,767,749 4,972,990 6,051,466 7,400,016 CFuel 13,761,584 14,698,551 13,502,594 11,994,906 9,406,361

COM 1,011,117 828,312 635,667 513,519 386,167

CCtax 8,465,881 7,179,963 7,666,898 8,521,528 10,264,906

CSS 8,400 8,400 8,400 8,400 8,400

CSal 188,186 72,454

Cost 30,000,833 30,000,833 30,000,833 30,000,833 30,000,833 Figure 20: Scenario analysis under worst-case cost minimization – Cost structure of each individual scenario (Unit: USD$10)

The distribution of power generation from each power source, sales of electricity to the national grid, and the storage of electricity in each month under worst-case cost

minimization is listed in Appendix 10.

According to the data in Appendix 10, the distribution of toal DER power generation of five scenarios with respect to months can be depicted as shown in Figure 21. The broken-line graph in worst-case cost minimization is significantly different from that in expected cost minimization (Figure 15), with much sharper fluctuations in magnitude.

Figure 21: Scenario analysis under worst-case cost minimization (Taichung Industrial Park) – Comparison of total power generation by months among individual scenarios

In addition, there are dramatic changes among seasons, such that the amount of power generated in summer is apparently higher than that in winter. The distribution in this case indicates that the variation of power generation among scenarios in worst-case cost minimization can be larger than that in the expected cost minimization, although the variation among individual scenario costs is supposed to be minimized.

The composition of power generation can also be presented with respect to different scenarios, as shown in Figure 22. It can be found that when the scenario conditions turn out to be more environmental friendly, such as in scenario 4 and 5, the proportion of power generated by H-PV becomes larger. On the contrary, the proportions taken by HC-GT and HC-GE become less from scenario 3 to scenario 5. In general, the total amount of power generated by DER technologies increases from scenario 1 to scenario 5, with an exception happening in scenario 2.

Figure 22: Scenario analysis under worst-case cost minimization (Taichung Industrial Park) – Comparison of power generation composition by power sources among individual scenarios

Next, the scenario analysis of heat recovery in this microgrid is carried out under worst-case cost minimization, and the outcomes are listed in Appendix 11. Based on these output data, the line graph of total recovered heat vs months for five scenarios can be drawn as shown in Figure 23, while the stacked histogram of recovered heat by different power sources in different scenarios are presented in Figure 24.

It can be observed from Figure 23 that the distribution of total recovered heat in the microgrid along the time horizon differs significantly among five scenarios, unlike the case in expected cost minimization. While the lines of scenario 1 and scenario 2 appear to be flatter along the whole year, the lines of scenario 3, 4, and 5 fluctuate sharply, with much more heat recovered in summer than in winter. This graph also implies that the minimization of variation among scenario costs does not necessarily result in a more uniform distribution in heat recovery.

Figure 23: Scenario analysis under worst-case cost minimization (Taichung Industrial Park) – Comparison of total recovered heat by months among individual scenarios

Figure 24: Scenario analysis under worst-case cost minimization (Taichung Industrial Park) – Comparison of recovered heat composition by power sources among individual scenarios

As for the composition of heat recovery from different sources as shown in Figure 24, it can be realized that more heat is recovered in scenario 3, 4, and 5 as in those

environmentally friendlier conditions, CCHP technologies are used more frequently for energy-saving and green concerns. The increase of the heat recovery proportion taken by H-PV also supports this viewpoint.

Lastly, the comparison of direct fuel condumption among five scenarios is made in Figure 25 under worst-case cost minimization. It can be seen from the diagram that both biomass and natural gas are used in scenario 1 to generate additional heat for the grid, while only natural gas is needed in the other four scenarios. Moreover, the total amount of fuel consumption decreases from scenario 1 to scenario 5, which conforms to the trend that the use of recovered heat increases from scenario 1 to scenario 5, as indicated by Figure 24.

Figure 25: Scenario analysis under expected cost minimization (Taichung Industrial Park) – Comparison of total direct fuel consumption among individual scenarios