3.1 OSeMOSYS
Energy modelling has emerged as an extremely useful tool for policy makers. This is due to open-source datasets, accompanied with many different open source tools, making modelling more accessible globally. This study is conducted using the Open Source Energy Modelling System (OSeMOSYS). OSeMOSYS is a fully developed systems optimisation model that can assess long-term energy planning. Unlike other established energy systems models, it likely requires less time commitment to build and operate, and is potentially easier to learn (Howells et al. 2011). It was designed to provide an open tool that can be used by anyone, increasing their human capacity in terms of understanding energy models (Strachan and Pye 2009; as cited by Howells et al. 2011). OSeMOSYS is designed to represent energy carriers and technologies within an energy system. Technologies are defined by costs, (capital, fixed and variable), operational life and capacity factors (Howells et al. 2011). Whilst OSeMOSYS is used to produce the results, data are inputted via the Climate Compatible Growth (CCG) clicSAND interface, which provides an organised open source data file. With results being produced through clicSAND or via OSeMOSYS Cloud. For this study the Philippines starter data kit, provided and produced by CCG, was used (Allington et al. 2021).
3.2 Temporal Structure
The modelling period is from 2015 to 2070. Within the model, each year is split into eight timeslices: four seasons, each with two 12-hour dayparts. Each day is from 0600–1800 and night is 1800–0600 hours. Season 1 is December to February, season 2 March to May, season 3 June to August and season 4 September to November. The eight timeslices represent each season, such as summer day and summer night.
3.3 Model structure
In terms of the model structure there are ten commodities and accompanying technologies: coal, oil, gas, biomass, solar, wind, hydropower, geothermal, concentrated solar power and nuclear. As for nuclear, despite a closed plant reopening in June (Hilotin 2022), it is not included in this study due to being a least-cost focused study as the construction costs are far greater than other options. Despite there currently being no concentrated solar power (CSP) plants in the Philippines, it is included as a generation option. Solar power is represented by three types of generation: Solar PV utility, utility scale with 2-hour storage and solar PV with storage. Wind is also represented by three types of generations: onshore, offshore and onshore with storage. Hydropower is also split into three: large (>100MW), medium (10–100MW) and small (<10MW). Natural gas is also split into two technologies: a combined cycle gas (turbine) plant and simple cycle gas (turbine) plant. Then finally, there are two other technologies: transmission and distribution. These represent the improvements to the grid to make it compatible with increasing capacity. Moreover, there are many parameters that define the technologies in the model: capacity factor, availability factor, capacity to activity unit, operational life, residual capacity, capital costs, fixed costs and variable costs. Capacity factor represents the capacity available per timeslice as a fraction of total installed capacity, from 0.1 to 1, whereby a higher value signifies more availability. As well as this operational life represents how long each power plant will last before being retired. A reserve margin of 15% is used and a discount rate of 0.1, as they are the conventional values always used. Finally, Figure 1 shows the reference energy system (RES) for the Philippines, showing the linkages between commodities and consumption.
In terms of capacity, the data within the SAND were updated to represent current figures. Residual capacity, the currently installed capacity that the model has no choice to include when it decides what to invest in, was updated in two ways. Firstly, the Department of Energy Power Statistics (2021) were used to update current installed capacity, with installed capacity for 2021 shown in Table 2. Secondly, the data were further updated to better represent future project installation and their retirements (Table 3).
3.4 Costs
Costs within the model are split into three: capital, fixed and variable. Capital costs are the amount of investment needed per unit of capacity. The Renewable Power Generation Costs report (IRENA 2022) was used to update the model values, whereby table 4 shows the cost of renewable plants per kW in 2021. Crucially, the average capital cost of CSP plants in 2021 was $9091 per kW; however, from 2010 to 2020 it had previously dropped from $9492 to $4746. The large rise in costs was due to only one CSP plant opening globally in 2021 (IRENA 2022), which had higher than usual costs. Therefore, due to the trend of falling costs of CSP, the 2020 value is used as the baseline for projecting future costs. As well as this, IRENA (2022) also found that solar PV has dropped around 7.5% in costs each year since 2010 making it the cheapest renewable energy source, whereas others have increased such as large scale hydropower at almost 6% a year since 2010. Furthermore, using the percentage change per year in capital costs for renewables, future cost projections up until 2030 were made and these values were used for the rest of the modelling period.
Fixed costs relate to the operational and maintenance costs for a technology per unit of capacity. Fixed costs of renewables are falling and projected to continue to do so, making them more competitive. This is the case even for the most expensive renewables, which as of 2021 are wind and concentrated solar power. Solar is the cheapest of all renewables due to the low maintenance needs of plants. In terms of fossil fuels, oil is the cheapest, and coal costs are similar to that of renewables.
Then finally, variable costs are the fuel costs. These are the extraction and importation of fuels used in power plants, which relate to coal, gas, oil and biomass. Importing is more expensive than extracting fuels, and oil is the most expensive in both cases, whilst coal is the cheapest fuel to import and biomass is the cheapest fuel to extract. Also, it is important to note that most renewables have no variable costs; there is no fuel cost, except for biomass which is still small.
3.5 Energy Demand
The energy demand statistics come from the Department of Energy Power Statistics (2021). The total energy demand is split into residential, commercial and industrial. Figure 2 shows the growth of the total energy demand from 2015 to 2070. Furthermore, projections of energy demand beyond the data available were based on the 5.97% average annual increase of demand from 2003 to 2021 in the Philippines. Finally, the 2019 demand figures are used as the baseline to project future demand from 2022. This is because 2020 and 2021 were COVID-19 years, so using 2019 gives a more realistic figure for demand without the impact of COVID-19.
3.6 Modelled Scenarios
There were six scenarios modelled.
- Least Cost (LC): no constraints applied to find cost optimal solution, except no nuclear; this was also used as the baseline.
- Coal Phase Out (CPO): gradual phase out of coal power plants and production by 2050.
- Philippines Energy Plan (PEP): forcing the model to make renewables generate 35% of total demand by 2030 and 50% by 2040, which are the goals directly from their policy.
- Combined Energy Plan (CEP): combination of the CPO and PEP scenarios.
- Nationally Determined Contribution (NDC): modelling the NDC of 75% emission reductions by 2030.
- Net Zero (NZ): net zero emissions by 2050.