3. Future technologies#

The base of future technologies made available to the model aligns with the MBIE generation stack used for EDGS 2024. This lists many future consented, planned or generic plants, and includes detailed estimates of costs (investment, connection, and maintenance) as well as other important parameters like fuel efficiency, relevant substation, or earliest possible commissioning year.

For the Steady scenario, we use the MBIE Reference generation stack. For the Shift scenario, we use the Innovation generation stack, which contains more potential new generation plants at lower capital costs.

We make the following adjustments and additions to the MBIE generation stack:

Future costs in the MBIE generation stack[1] are assumed to be static. We allow these to decline over time, based on the United States’ National Renewable Energy Laboratory (NREL)[2] assumptions of future cost declines for solar, wind, and geothermal technologies.
We add data for distributed solar generation, offshore wind, and diesel peakers.
We assume all utility solar plants in the MBIE generation stack include single-axis tracking capabilities. This is in alignment with existing assumptions on utility-scale solar, such as work by Alan Miller[3].
Some plants have had fixed install dates updated to reflect recent developments.

3.1. Plant fixed install dates#

New plants in MBIE’s generation stack may have fixed install dates, where the model is forced to build a plant during a specific year. EECA has updated some fixed commissioning dates since the public generation stack was first compiled. We also implement a partial year method for these, where plants installed later in the year are only available for a fraction of the commissioning year. This method is important for short-term or modelled historical generation investment, where we have more detailed information and can be more precise about build timings. The entire list of fixed plants, and the month we assume they become available, is listed in REFERENCE ME

Where possible, the listed months consider full generation availability, not initial synchronisation.

Table 76 Plant fixed install dates and months#
PlantName	Year	Month
Tauhara Stage 1[4]	2024	November
Te Huka Unit 3[5]	2024	December
Lauriston solar farm[6]	2025	February
Twin Rivers solar farm[7]	2025	December
Harapaki wind farm[8]	2024	July
Kawerau (TOPP2)[9]	2026	March
Kaiwaikawe[10]	2026	September

This method aligns with the TIMES-NZ treatment of seasons. For exammple, if a plant’s first generation is in December, the model allows it to only be available for part of summer in that year. Where a plant has a fixed install date, but we are unsure of the month, we default to assuming it is first available for generation in July, or half of that year.

3.2. Learning curves#

Data from the NREL ATB database[11] was used to produce learning curves for certain solar, wind, and geothermal technologies, where the costs of technologies decrease over time assuming technological advancement. The cost data from NREL is projected from 2022 to 2050 across a range of scenarios and technology types.

NREL projections consider three different scenarios: Conservative, Moderate, and Advanced. The NREL scenarios are defined as follows[12]:

Conservative: Small changes in technology with decreases in public and private research and development investments.
Moderate: Innovation in the market is more widespread with current levels of investment in public and private research and development.
Advanced: Innovation in the market is widespread with an increase of investment in public and private research and development.

For the Steady scenario we apply the Moderate projections, and for Shift we use Advanced.

Learning curves were applied to reduce the future capital cost (CAPEX) and fixed operating and maintenance costs (FOM) to plants that met the following criteria:

Only solar, wind, and geothermal plants have had reduced future costs applied.
The plant status must be less advanced than “Fully Consented” or “Under Construction”.
Plants must have an earliest or fixed commissioning year no earlier than 2030, or no information on commissioning year.

All remaining plants have their CAPEX and FOM remain fixed across the model horizon. The NREL technology categories used for each plant type are as follows:

Table 77 Technology types and NREL category matching#
Technology Type	NREL technology used
Utility solar (tracking)	Tracking PV
Onshore wind	Wind Turbine Technology 1
Offshore wind (fixed)	Offshore Wind Fixed-Bottom
Offshore wind (floating)	Offshore Wind Floating
Geothermal	Geothermal - Hydro / Flash

To apply the learning curves, the percentage indices of the NREL CAPEX and FOM data was found. The percentage indices were found using \(PI_i=(Expense_i)/(Expense_{2023})\), where \(PI\) is the percentage index and \(i\) is the year.

The base year was set to 2023 to match the base year of the MBIE data. For the forecasted CAPEX of the future technology, the percentage indices were applied to the capital cost, which does not include the cost of connection. The cost of connection was divided by the capacity of the plant and then added as a constant to the projected CAPEX. For the FOMs the percentage indices were able to be applied without extra addition of other costs.

3.3. Offshore wind#

Offshore wind costs use the NREL data, converted to 2023 NZD. Note that NREL CAPEX costs for offshore wind[13] include the connection costs. The distances to land and water depth are different for the different classes of offshore wind technology.

The potential future capacity of offshore wind was estimated in PWC’s National Impacts Report[14] on the offshore wind industry in New Zealand. This report lists offshore wind opportunities that were in early development at the time of publication. The proposed wind farms and sizes for the selected regions are added to TIMES-NZ, using NREL cost data. While some of these offshore wind opportunities may never go ahead, these act as an upper limit on offshore wind potential in the model.

The earliest commissioning year for offshore wind farms was set to 2035, as all potential developers project at least a 10-year project implementation time.

3.4. Distributed rooftop solar#

We do not model distributed solar uptake in TIMES, and instead provide exogenous assumptions on rooftop solar uptake.

This is because TIMES uses a system cost perspective when selecting optimal technologies. This often means that it will not invest in distributed solar, preferring the economies of scale of utility-scale solar. This is true even when considering the efficiency and grid maintenance benefits of off-grid generation[15].

Because it would not be realistic to assume no distributed solar installations, we instead use external forecasts of distributed solar installation from MBIE’s EDGS scenarios[16]. We use the Reference projections for our Steady scenario, and the Innovation projections for Shift. This means that TIMES-NZ distributed solar uptake rates are not the product of any other model properties; they are hardcoded assumptions.

3.5. Diesel peakers#

We allow the model to also build diesel peakers, as this could prove a useful option for the sector given declining gas availability. Diesel OCGT peakers do not exist in the MBIE generation stack, so have set their parameters equivalent to existing natural gas OCGT peakers. However, the heat rates were adjusted to 11,000 GJ/GWh, implying a fuel efficiency of 32.7%, which is in line with existing assumptions on the operation of the Whirinaki diesel plant.

3.6. New fuels in existing assets#

We add additional fuel options to some existing plants to incorporate potential future renewable fuels. All gas-fired electricity generation can also use biomethane directly. Similarly, the Huntly Rankines have the option to use black pellets, if these are produced via torrefaction facility. See documentation on biofuel assumptions for more details on the production and distribution of these fuels.