*2.3. Plant Costs*

Plant cost results are presented in Figure 1. These results are a high-resolution LCoE, refined through co-optimized debt-finance and taxation variables. Note in Figure 1 that Incumbent Coal plant has a generalized average total cost of \$64/MWh comprising fuel (\$30/MWh), Operations & Maintenance (O&M) of (\$8.71/MWh), debt (\$4.17/MWh), taxation (\$5.44/MWh) and equity (\$15.74/MWh). The OCGT cost structure focuses on the *"carrying cost"* of capacity (at \$14/MWh), and has a marginal running cost of \$123/MWh (including variable O&M). Finally, results in Figure 1 are based on static capacity factors, but in the NEMESYS model, plant costs arise on a dynamic basis with capacity factors determined by the dispatch necessary to meet final demand.

#### **3. Model Results and Discussion**

#### *3.1. Overview of Power System*

The NEMESYS Model has been populated with the plant cost results from Section 2.3, and half-hour load data (using Queensland power system final demand from 2016). From this, multiple scenarios are simulated. A long run (own-price) demand elasticity of −0.30 is applied to all variation cases [49–51]. To keep modeling results tractable, the power system is modeled as a single, non-interconnected gross-pool energy-only market. Recall also that the level of government-initiated CfDs are exogenously determined and designed to achieve a certain VRE market share. The base scenario is calibrated with 0% VRE plant (i.e., the power system commences as a thermal system with zero renewable plant), and variation scenarios span up to 40% VRE market share.

Consistent with Equation (6), the objective of the power system model is to minimize resource costs and maximize consumer welfare whilst meeting a reliability constraint of no more than 0.002% unserved energy (i.e., the NEM's long-standing reliability criteria). To assist interpretation of subsequent results, critical outputs for the two bookend scenarios (i.e., 0% and 40% VRE market share) are presented in Table 2.


**Table 2.** Overview of key model results.

\* Dispatch-weighted price, excludes any implicit or explicit value of CO2.

Note from Table 2 that the single-region power system has an initial final Energy Demand of 54,717 GWh per annum with peak or Maximum Demand of 9118 MW. Given demand elasticity of −0.30 and the variation in wholesale prices with 40% VRE market share, energy demand and maximum demand rise to 56,386 GWh and 9393 MW, respectively. The opening plant stock is dominated by 6720 MW of coal plant, and in order to meet the reliability constraint (given plant outages) a reserve plant margin of ~11% is necessary. In order to meet a 40% VRE market share, about 3800 MW of wind and 2700 MW of solar PV capacity is added to the plant stock, and given optimal conditions, 2520 MW of coal plant retires. To meet reliability constraints, 800 MW of CCGT plant and 750 MW of OCGT plant is added—albeit operating at relatively low annual capacity factors.

Note also from Table 2 that the power system commences with a dispatch-weighted spot market price of \$82.53/MWh and a system average cost of \$78.74/MWh. With VRE market share of 40%, total system cost increases to \$82.37/MWh whereas the underlying power system price falls to \$65.84/MWh—components of this gap being underwritten by government-initiated CfDs with an imputed CO2 value in the range of \$25–\$35/t. Note that each technology earns a di fferent dispatch-weighted price according to their production profile. In any scenario, OCGT plant earns the highest average spot price (i.e., they increase output at times of high spot prices, and turn o ff in low spot price periods). Wind plant on the other hand earns a lower average spot price than thermal plant (i.e., coal, CCGT, and OCGT plant) since the stochastic production profile of Australian wind generators has a slight o ff-peak bias, and in addition, as more wind plant is added to the power system it has an impressing e ffect on wind's *earned price* [6,24,35]. It is worth noting that the dispatch-weighted

price of Solar PV plant falls below wind once the technology reaches 7% market share due to the relatively tight correlation amongs<sup>t</sup> all solar PV plant output.

Modeled power system CO2 emissions fall from 53.4 Mt pa to 32.0 Mt pa between the 0% and 40% VRE market share scenarios. Unserved energy in both scenarios is ~0.001% of total load, and thus the plant stock in both scenarios meet the reliability criteria.
