Probing the Financial Sustainability of Eskom’s Open Cycle Gas Turbines (OCGTs) Utilisation (2018–2021)

Abstract

Contributing to achieving sustainability in South Africa’s energy sector, this study probes financial sustainability and its relationship to the environmental sustainability of Eskom. This is because, over the past three financial years (FY2018–2019 to FY2020–2021) of Eskom’s generating plants’ performance, the energy availability factor (EAF) has taken a deep dive, reaching an extremely low generation availability year-end performance of 64.2%, translating to approximately an average of 29,800 MW available generation capacity out of a nominal generation capacity of 46,466 MW in FY2020–2021. Therefore, the study employed a quantitative research methodology, where the relevant financial records were analysed, and the necessary energy calculations made using descriptive analysis in Microsoft Excel. The findings show that the volumes (GWh) produced by the OCGTs during this period far exceed the regulatory approved volumes, thus attracting substantial costs, amounting to ZAR 25.9 bn instead of ZAR 8.9 bn, that could have been spent on the OCGTs if the level of efficiency achieved in FY2016–2017 and FY2017–2018 was maintained. The analysis also revealed that the OCGTs’ long-term financial and environmental sustainability could be achieved through switching from diesel to natural gas, thus resulting in lower fuel costs and lower emissions. Further, potential savings of approximately ZAR 27 bn (excluding capital expenditure) at a 10% load factor can be realised over a ten-year period when the natural gas price is sitting at ZAR 85/GJ (minimum). Finally, in order to attain financial and environmental sustainability, it is recommended that both Eskom’s and the independent power producers’ (IPPs) OCGTs must switch fuel from diesel to natural gas and be run at a 10% load factor, allowing the OCGTs to be run as mid-merit generators.

1. Introduction

Electricity is generated by various power stations and transported through transmissions and distribution powerlines to reach end-users [1]. According to the South African Energy supply report [2], the South African Energy supply is made up of the following energy sources: 65% coal, 11% renewable energy and waste, 18% crude oil, 3% gas, 2% nuclear energy, and 1% geothermal. According to the report, Eskom (Electricity Supply Commission) is the South African government’s utility company that supplies electricity across South Africa and the Southern African Development Community (SADC) region, which generates, transmits, and distributes electricity. The report also indicates that Eskom operates and owns 27 power stations, with approximately 45 GW of total nominal capacity. Eskom’s generating capacity mix consists of coal power stations, nuclear, diesel-powered open cycle gas turbines (OCGTs), pumped storage schemes, hydro-electrics, and wind farms. Eskom owns and operates two diesel-powered open cycle gas turbine (OCGT) peaking power stations, namely, Ankerlig and Gourikwa, with a capacity of 1323 MW and 735 MW, respectively. These two OCGT peaking power stations are intended to be used during peaking periods, morning (05:00–08:00) and evening (17:00–20:00), to supply power to the grid, and also during emergencies [3].

Independent power producers (IPPs) also own and operate two diesel-powered open cycle gas turbine (OCGT) power stations, namely, the Dedisa and Avon peaking power stations, with a capacity of 335 MW and 670 MW, respectively. These IPPs’ OCGT peaking power stations sell electricity to Eskom, and they are also intended to provide power to the grid during peaking periods and also for emergencies [4]. Kim et al. [5] points out that OCGTs are utilised as peaking power plants that run only during peak hours and emergency periods since the capital expenditure cost of OCGTs is the cheapest, but the variable (fuel) cost is the most expensive among conventional plants. The national regulator, National Energy Regulation of South Africa (NERSA), classifies OCGTs as emergency reserves, thus limiting the volumes of GWh generated by these peaking stations due to the high diesel costs [6,7].

Eskom’s coal generating units have experienced the worst chronic breakdowns over the past two years. Consequently, Eskom reverted to frequently rolling out country-wide rotational load shedding as a last resort in order to preserve the integrity of the interconnected power system [8,9]. Additionally, the continued trend of poor power generation spurred the over-utilisation of the emergency reserves, such as diesel-powered OCGTs (including the IPPs’ OCGTs), using them as baseload or mid-merit generators in order to minimize or to avoid country-wide rotational load shedding. The annual load factor for Eskom’s and the IPPs’ OCGTs far surpasses the regulatory approved volumes, thus attracting significant costs between the 2018–2019 to 2020–2021 financial years [5,10,11]. Thus, this study seeks to assess Eskom’s open cycle gas turbine (OCGT) utilisation between 2018 and 2021 to analyse its sustainability dimensions.

Over the past three financial years (FY2018–2019 to FY2020–2021), Eskom’s coal generation has experienced chronic breakdowns at both old and new coal power stations [9,12,13]. Lack of quality maintenance, an ageing coal fleet, prolonged heavy rains, historical plant operation, critical maintenance deferral due to funds, design flaws at Medupi and Kusile, and allegations of sabotage are some of the suggested reasons behind the worsening generating performance [9,13]. The high number of unplanned breakdowns at coal generation plants has necessitated Eskom to continuously over-utilize the emergency reserves, such as their OCGTs (including the IPPs’ OCGTs), using them as baseload or mid-merit generators in order to minimize or avoid country-wide rotational load shedding [9,12,13].

NERSA [6] classifies OCGTs (including IPPs DOE Peaker) as emergency reserves, and they are supposed to be operated during peak periods (morning and evening) at a 1% load factor, approximately 88 h per annum, in order to respond to unforeseen events as well to curtail over-utilisation and inefficiencies [14]. The over-utilisation of OCGTs directly and substantially impacts electricity consumer tariffs and Eskom’s financial performance. Consequently, Eskom spent ZAR 25.9 bn on OCGTs over the past three financial years (FY2018–2019 to FY2020–2021) [8,9,13]. Figure 1 shows a sharp increase in Eskom’s OCGT costs (excluding the IPPs’ OCGTs) from ZAR 320 m in FY2017–2018 to ZAR 3.74 bn in FY2018–2019 (>1100% increase). Eskom will likely face a tough hurdle in recovering these additional OCGT costs, which will worsen their precarious financial position, thus impeding their sustainability goals. It is financially and environmentally unsustainable to use emergency reserves (OCGTs) to compensate for the deterioration of coal generation plants. Short-term decisions on OCGTs will likely have a negative impact on Eskom’s long-term sustainability.

Consequently, the South African energy supply sector contributed 78.1% to gross national GHG emissions in 2015 [16]. The integration of low emission technologies into existing generation plants and electricity planning is essential to achieve climate mitigation objectives. The electricity sector is the largest subsector for mitigation; thus, in line with global trends, the rapid decarbonisation of grid-connected electricity is identified as the key mitigation option worldwide [17,18].

Therefore, this research aims to assess Eskom’s open cycle gas turbine (OCGT) utilisation between 2018 and 2021 via a financial sustainability dimension. The objectives are as follows:

• Evaluate the financial impact (revenue and tariffs) of OCGT (over) utilisation
• Ascertain the role of OCGT over-utilisation on long-term (2022–2030) grid security and Eskom’s sustainability.

2. Methodology

A quantitative approach was used for this study. All data pertaining to the OCGT utilisation for this study was attained through Eskom’s online (publicly available) data [13]. The data portal provides hourly energy produced by each OCGT peaking power station; furthermore, the data portal allows dataset downloads for a period of 5 years. The previous years outside the study period will form the basis for comparison to the current study period.

In evaluating the financial impact of OCGT (over) utilisation on Eskom’s finances, we have used audited Eskom financial reports and integrated reports for the financial years 2018, 2019, 2020, and 2021. The data, stored in Microsoft Excel 2016, were analysed using descriptive statistics, tables, charts, etc.

In attaining the role of OCGT over-utilisation on long-term grid security and sustainability (2022–2030), a comparative analysis by Mutsau (2018) of international natural gas prices against the South African natural gas energy price (2012–2016) was used as input data to conduct a comparative analysis between diesel and naturally powered OCGTs.

Table 1. Descriptive statistics on the average gas energy prices in the EU and South Africa [20].

	Natural Gas Prices in USD/mmBtu
Statistic	European Union	South Africa
Mean	11.54	9.62
Median	12.46	10.19
Maximum	14.09	11.46
Minimum	8.24	7.13
Standard Deviation	2.20	1.48

contains the descriptive analysis of the average natural gas energy prices for the European Union and South Africa extracted from NERSA. The diesel energy prices used in this study are the six-year average wholesale price (2015–2020) [19], taking into account wholesale discounts and tax rebates for electricity generation. Eskom’s and the IPPs’ OCGT diesel costs range from ZAR 3.50/L to ZAR 4.10/L less than wholesale prices [14].

The South African Carbon Tax Act provides information to calculate the future carbon tax savings on OCGT fuel conversion from diesel to gas to drive long-term Eskom operational and financial sustainability [21,22].

The analysis of the financial impact evaluation of the OCGTs was largely quantitative. It was achieved using a spreadsheet (table), using the OCGT costs over the study period to gain a better understanding of how diesel-powered OCGTs affect Eskom’s finances. The spreadsheet analysis of Eskom’s operating cost structure will be developed to better understand the impact of OCGT costs on Eskom’s finances, thereby assisting in unpacking the impact of OCGT costs on Eskom’s revenue, profitability, and tariffs using data from financial years 2015–2016 to 2020–2021, and compared with other previous financial years where the level of operational efficiency in terms of adhering to the regulatory allowed volumes (GWh) was relatively good. The second aspect will be to show how OCGT (over) utilisation impacts tariffs by contributing to the primary energy cost. This was done in a spreadsheet using tables.

According to Notton et al. [23], solar and wind energy sources are naturally time-varying on scales from minutes and hours up to seasons. Consequently, the integration of such intermittent renewable energy systems into the grid presents some new challenges in managing a reliable and safe energy supply; therefore, ‘backup’ energy from other sources, such as OCGTs, will be required to preserve the integrity of the grid, depending on the penetration levels of renewable sources.

The South African electricity supply industry has approximately 5900 MW installed capacity from renewable sources [24]. The Integrated Resource Plan (IRP) indicates an additional capacity of approximately 20,000 MW from solar and wind technologies allocated in the IRP2019, while 10,500 MW of coal generation capacity will be decommissioned by 2030 [16]. The above factors and the deterioration in performance from the Eskom generating fleet will necessitate the diesel-powered OCGTs to be run at higher load factors to fully utilise all supply-side generation plants to meet the demand. The high operational costs from diesel-powered OCGTs will put a serious strain on Eskom’s finances, in accordance with the regulations from NERSA. Lastly, Eskom’s carbon tax exemption comes to an end by 2022; therefore, other fuel sources with low carbon emissions need to be considered in order to drive financial efficiencies.

In ascertaining the role of OCGT over-utilisation on long-term grid security and sustainability, feasibility studies on the conversion of OCGTs to be powered by natural gas at higher load factors were conducted. The issues considered regarding fuel conversion includes the following:

• Reduction in the operational cost of fuel by converting to natural gas.
• Introduction of required flexibility with increasing renewables energy plants in the future [16].
• Reduction in carbon emissions (carbon tax savings that will be realised by switching to natural gas).

In realising the potential financial benefits envisaged in the conversion to natural gas running at a load factor of 10%, the following financial analysis will be conducted:

(i)

Operational costs from diesel at different prices vs. natural gas at different gas prices;

(ii)

Calculate the potential savings of moving from diesel to natural gas;

(iii)

Calculate the carbon tax savings from reducing emissions by switching to natural gas.

Power produced (OCGT) in MWh = Full rating of OCGT Plant (MW) × (365 × 24 × load factor in (%))

Eskom’s OCGTs

Ankerlig: 1323 MW (147 MW × 9 units)

Gourikwa: 735 (147 MW × 5 units)

IPPs’ OCGTs

Avon: 670 MW (167.5 MW × 4 units)

Dedisa: 335 MW (167.5 MW × 2 units)

Annual carbon tax (ZAR/annum) = Annual CO₂ emissions (tons/annum) × carbon tax rate (ZAR/ton CO₂)

Annual CO₂ emissions from Diesel/Natural Gas powered OCGT’s (tons/annum) = annual power generated (MWh/annum) ×1000 × CO₂ emission factor for electricity (kg CO₂/kWh)/1000

The emission factors for diesel and natural gas will be taken from in the IPCC 2006 guidelines for national greenhouse gas inventories, Volume 2: Energy, Chapter 2: Stationary Combustion: “default emission factors for stationary combustion in the energy industries (kg of greenhouse gas per TJ on a Net Calorific Basis)” [25].

Diesel: 74,100 kg CO₂/TJ

Natural gas: 56,100 kg CO₂/TJ

The carbon tax rate per ton of CO₂ equivalent for fuel combustion (electricity generation) has increased from ZAR 120/t CO₂ to ZAR 127/t CO₂ for 2020 [21,22]. Thus, the 2020 rate of ZAR 127/t CO₂ will be used as the carbon tax rate to calculate the annual carbon tax. The carbon tax will be calculated on real rand; thus, the 20-year carbon tax for different fuel excludes inflation.

3. Results and Analysis

3.1. Open Cycle Gas Turbine Utilisation and Cost

Figure 2 shows both Eskom’s and the IPPs’ open cycle gas turbines’ (OCGTs) production in GWh from 2016–2017 to 2020–2021. The OCGT volumes (GWh) produced from 2018–2019 to 2020–2021 were enormously above the NERSA-approved volumes. This happens despite NERSA being acquainted with the challenges confronting generation plant performance, thus allowing Eskom’s OCGT volumes for the first two years of the MYPD4 period (2019–2020 to 2020–2021) to be four and two times the volumes (GWh) produced in the previous financial years—2016–2017 and 2017–2018.

For the financial year 2018–2019, Eskom’s OCGT volumes produced were approximately ten times (from 118 GWh to 1202 GWh) more than in the 2017–2018 financial year, despite the regulator limiting volumes to 105 GWh due to the high operational costs linked with diesel (fuel) [4,7,14]. The same trend was observed for the IPPs’ OCGTs, where the volumes produced for the period precipitously increased to approximately five times the regulatory approved volumes, as shown in Figure 2. For the financial year 2019–2020, Eskom’s and the IPPs’ OCGTs exceeded the regulatory approved volumes by approximately three and eight times, respectively. The same trend was witnessed for the following financial year (2020–2021), with Eskom’s and the IPPs’ OCGT volumes exceeding the regulatory approved volumes by approximately six and eight times, respectively.

Figure 3 shows the OCGT costs (ZAR’m) from FY2016–2017 to FY2020–2021. The increased utilisation of OCGTs (GWh) shown in Figure 2 attracted substantial costs for the period from FY2018–2019 to FY2020–2021, as shown in Figure 3. The expenditure on Eskom’s OCGTs increased from ZAR 328 m (118 GWh generated) in 2017–2018 to ZAR 3768 m (1202 GWh generated) in FY2018–2019, while the IPPs’ OCGT costs increased from ZAR 2291 m (105 GWh generated) in FY2017–2018 to ZAR 4291 m (552 GWh generated). This trend has continued for the first two years of the MYPD4 period (FY219–2020 to FY2020–2021), thus attracting significant OCGT costs, as shown in Figure 3 in ZAR’m.

According to

Table 2. Eskom generation plant technical performance (2016–2021).

FY2016–2020
Generation Technical Performance	2016–2017	2017–2018	2018–2019	2019–2020	2020–2021
Energy Availability Factor (EAF) %	77.3	78.0	71.5	66.6	64.2
Unplanned Capacity Load Factor (UCLF) %	9.9	10.2	17.5	22.9	20.0
Planned Capacity Load Factor (PCLF) %	12.1	10.4	9.5	8.9	12.3
Other Capacity Loss Factor (OCLF) %	0.7	1.4	1.5	1.6	3.5

, these volumes were strongly affected by the decline in Eskom’s generation plant performance. The energy availability factor (EAF) has declined due to a high unplanned capacity loss factor (UCFL), commonly known as unplanned outages [5,10,11]. The key take away from the generation plant performance is that the UCFL has precipitously increased from 10.2% in 2017–2018 to a record high of 22.9% in 2019–2020; consequently, the energy availability factor (EAF) reached extremely low availability levels, up to 66.6%. The unplanned breakdowns for 2019–2020 translate to approximately 10,000 MW out of a possible 45,000 MW generation on average, effectively then not being available to produce electricity.

3.1.1. 2018–2019 OCGT Utilisation

Table 3. 2018–2019 OCGT volume and costs.

FY2018–2019	Volume (GWh)				Cost (ZAR’m)
Open Cycle Gas Turbines (GWh)	Decision	Actual	Variance	% Change	Decision	Actual	Variance	% Change
Eskom’s OCGTs	105	1202	1097	1145%	ZAR 345.00	ZAR 3768.00	ZAR 3423.00	1092%
IPPs’ OCGTs	88	552	464	627%	ZAR 2380.00	ZAR 4291.00	ZAR 1911.00	180%

below shows both Eskom’s and the IPPs’ OCGT volumes produced and costs for 2018–2019 assessed against the regulatory approved volumes and costs (decision) [4]. The regulatory decision kept Eskom’s OCGT use at a 0.5% load factor (105 GWh), which translates to 44 h for the year. This equated to approximately running Eskom’s OCGTs for less than 4 h a month; thus, the OCGT costs are limited to ZAR 345 m. The decision was informed by Eskom’s 2016–2017 and 2017–2018 OCGT production, which was achieved at less than a 0.5% load factor. As per the regulatory decision, the IPPs’ OCGT volumes were kept at a 1% load factor (88 GWh), which translates to 88 h. This then equates to approximately running the IPPs’ OCGTs at the cost of ZAR 2380 m. It must be noted that the IPPs’ OCGT costs include significant capacity payment and energy payment compared to Eskom’s OCGTs [8].

As shown, the actual volumes for Eskom’s OCGTs were substantially above the approved volumes, amounting to 1202 GWh at a cost of ZAR 3768 m, exceeding by 1145% and 1092%, respectively (Table 3). The IPPs OCGT volume (GWh) was also higher, with volumes reaching 552 GWh, exceeding by 627%, while the cost stood at ZAR 4291 m exceeding by 180%, as shown in Table 3. One of the key pieces of information from Eskom indicated that the excessive use of OCGTs was propelled by lower plant availability, which necessitated higher usage to avoid and sometimes limit load shedding in the second half of the year.

3.1.2. 2019–2020 OCGTs Utilisation

Table 4. 2019–2020 OCGT volume and costs.

FY2019–2020	Volume (GWh)				Cost (ZAR’m)
Open Cycle Gas Turbines (GWh)	Decision	Actual	Variance	% Change	Decision	Actual	Variance	% Change
Eskom’s OCGTs	456	1329	873	291%	ZAR 1902	ZAR 4303	ZAR 2401	228%
IPPs’ OCGTs	88	711	623	808%	ZAR 2422	ZAR 4883	ZAR 2461	202%

below shows both Eskom’s and the IPPs’ OCGT volumes produced and costs for 2019–2020 assessed against the regulatory approved volumes and costs (decision). In light of lower Eskom generation plant availability, the regulatory decision revised the volume and associated cost from the previous year and limited the use of Eskom OCGTs to 456 GWh (2.5% load factor) at a cost of ZAR 1902 m. The IPPs’ OCGTs were limited to 88 GWh at a cost of ZAR 2422 m, as depicted in Table 4 [4].

The actual volumes for Eskom’s OCGTs were above the approved volumes, amounting to 1329 GWh at a cost of ZAR 4303 m, exceeding the allowable volumes and cost by 291% and 226%, respectively, as shown in Table 4. The IPPs’ OCGT volume (GWh) was also higher, with volumes reaching 711 GWh, exceeding by a whopping 808%, while cost stood at ZAR 4883 m, exceeding by 202%, as shown in Table 4. It must be noted that both Eskom’s and the IPPs’ OCGTs exceeded the allowed volumes and costs for the two consecutive financial years. This is primarily due to the poor availability of Eskom’s generation plant, which is within management control, resulting in OCGTs being operated with higher load factors than the regulatory approved levels.

3.1.3. 2020–2021 OCGT Utilization

The actual volumes for Eskom’s OCGTs were above NERSA-approved volumes, amounting to 1246 GWh at a cost of ZAR 4075 m, exceeding the allowable volumes and cost by 591% and 429%, respectively, as shown in

Table 5. 2020–2021 OCGT volume and costs.

FY2020–2021	Volume (GWh)				Cost (ZAR’m)
Open Cycle Gas Turbines (GWh)	Decision	Actual	Variance	% Change	Decision	Actual	Variance	% Change
Eskom’s OCGTs	211	1246	1035	591%	ZAR 950	ZAR 4075	ZAR 3125	429%
IPPs’ OCGTs	88	704	616	800%	ZAR 2463	ZAR 4549	ZAR 2086	185%

. The IPPs’ OCGT utilisation in terms of volume (GWh) was higher as well, reaching 704 GWh, exceeding by a whopping 800%, while costs stood at ZAR 4549 m, exceeding by 185%, as shown in Table 5. It must be noted that both Eskom’s and the IPPs’ OCGTs exceeded the NERSA-allowed volumes, and costs for the three consecutive financial years, primarily due to poor availability of Eskom’s generation plants, which is within management control, resulting in OCGTs being operated with higher load factors than the regulatory approved levels.

3.2. OCGTs’ Impact on Revenue and Tariffs

Table 6. 2018–2020 OCGTs’ usage impact on revenue and tariffs. * This means the OCGTs costs that would have been realised at Load Factor < 1%, efficient use of OCGTs during 2016–2017 and 2017–2018 period in line with NERSA approved levels.

Analysis of OCGT Costs	2018–2019		2019–2020		2020–2021		Total
	ZAR’m	c/kWh	ZAR’m	c/kWh	ZAR’m	c/kWh	ZAR’m	c/kWh
Electricity used (GWh)	208,319		205,638		191,852		602,809
Electricity revenue	179,892	86.35	199,468	97.00	204,326	106.50	583,686
Eskom standard average tariff		90.01		106.80		116.15
Operating expenses	150,974	72.47	163,769	79.65	174,175	0.91
Net income (Loss)	(20,729)	−10.00	(20,502)	−10.00	(18,934)	−9.87	(60,165)	−9.98
Total OCGT @ FY17 efficiency level (ZAR’m) *	2851	1.37	2969	1.44	3117	1.62	8937	4.66
Total OCGT costs (ZAR’m)	8059	3.87	9186	4.47	8624	4.50	25,869	13.5
Analysis
Growth in OCGTs	283%		309%		277%		289%
Primary energy costs (ZAR’m)	99,448	47.74	112,119	54.52	115,903	56.72
Primary energy cost as % of revenue	55%		56%		57%
OCGT costs as % of primary energy costs *	2.9%		2.6%		2.8%		8%
OCGT costs as % of primary energy costs	8.1%		8.2%		7.7%		22%

shows the OCGT costs for Eskom and the IPPs and the impact thereof on revenue and tariffs from the 2018/19 to 2020/21 financial years. In 2018/2019, Eskom reported a net loss of ZAR 20.7 bn (10 c/kWh), which was incurred from a standard average tariff set at 90.01 c/kWh, translating to revenue of ZAR 179.9 bn. Additionally, for the second consecutive FY2019/2020, Eskom reported a net loss of ZAR 20.5 bn (10 c/kWh), incurred from an increased standard average tariff set at 106.80 c/kWh, translating to revenue of ZAR 199.5 bn, as shown in Table 6. Further, for the third consecutive FY2020–2021, Eskom posted a net loss of ZAR 18.9 bn from a further increased standard average tariff of 116.15 c/kWh, translating to revenue increase amounting to ZAR 204.3 bn. The losses incurred from 2018–2021 are partly attributed to OCGT over-utilisation, when viewed from the context of the variance between the approved regulatory levels and actuals, as comprehensively discussed in Section 3.1.1 and Section 3.1.3.

The OCGT costs fall under primary energy costs as per the regulatory methodology when assessing the allowable revenue. Further, it encompasses coal costs, OCGTs from IPPs, renewable energy from IPPs, water, and nuclear fuel costs, etc. [4,26]. Primary energy costs are the significant cost item on Eskom income statement, ranging from 55% to 57% of Eskom’s revenue for the study period. The primary energy costs translate to 47.47 c/kWh, 54.52 c/kWh, and 56.72 c/kWh of the set standard average tariff for the FY2018–2019, FY2019–2020, and FY2020–21 financials years, respectively, as shown in Table 6.

During the FY2018/2019 study period, OCGT expenditure (Eskom’s and the IPPs) amounted to ZAR 8.1 bn (8% of the primary energy costs), with Eskom’s OCGTs increasing by a whopping 1092% for the financial year 2017–2018. The total OCGTs incurred for FY2018–2019 translates to 3.87 c/kWh of the set standard average tariff for the respective period, as shown in Table 6. If the level of operational efficiency attained during FY2016–2017 and FY2017–2018 was carried over to FY 2018–2019, taking into account the consumer price index (CPI), the total OCGT costs would have amounted to ZAR 2.85 bn compared to ZAR 8.1 bn incurred due to extensive use of OCGTs.

For the FY2019–2020 study period, OCGT expenditure amounted to R9.2 bn (8% of the primary energy costs), increasing by 283% from the FY2016/17 efficiency level, thus translating to 4.47 c/kWh of the set standard average tariff. This trend continues for FY2020–2021, and the OCGTs’ over-utilisation amounts to ZAR 8.6 bn compared to ZAR 3.1 bn that would have been incurred if FY2016–2017 and FY2017–2018 were carried over to FY2019–2020.

The total OCGT expenditure for the study period amounts to ZAR 25.9 m, translating to 8.4 c/kWh (8.3% of 2019/2020 set standard tariff). The OCGT total expenditure would have amounted to approximately ZAR 8.9 bn (2.83 c/kWh) if level of OCGT utilisation was maintained at the efficiency levels of FY2016–2017 and FY2017–2018, thus making up 3% of the primary energy costs for each respective study period, relative to 8% of primary energy costs realised for each year of the study period (Table 6).

The difference in OCGT expenditure between the regulatory approved expenditure and actuals amounts ZAR 15.4 bn. The unrelenting generation plant challenges continued to ravage the year 2021, with Eskom recording the worst ever load shedding in terms of GWh shed [27]. Under these circumstances, further compounded by the Russia–Ukraine conflict, led to a sharp increase in oil prices. At the end of the FY2022/2023 financial year, OCGT over-utilisation will cumulatively amount to approximately ZAR 45 bn. This will worsen Eskom’s precarious financial position by increasing the operational expenditure, particularly primary energy costs, which negatively impact Eskom’s financial sustainability.

3.3. OCGT Long-Term Sustainability and Grid Security

In this section, we will discuss the potential fuel cost savings that can be realised when OCGTs are powered by natural gas instead of diesel as it is currently. Further, we seek to illustrate the savings at a higher load factor, which will assist with grid security strengthening in light of the energy landscape that is undergoing the transition from fossil dependency to intermittent low carbon generation technologies. Further, this section seeks to elucidate the carbon tax potential savings that could be realised when these OCGTs are powered by natural gas instead of diesel.

3.3.1. Fuel Costs Savings—From Diesel to Natural Gas

Table 7. Eskom’s fuel-switching saving costs analysis for OCGTs.

Eskom’s OCGTs	ZAR/GJ	Total Cost over 10 Years @ 10% Load Factor	10 Year Savings by Using Gas Fuel Instead of Diesel (Excluding Capex)	10 Years % Savings by Using Gas Fuel Instead of Diesel (Excluding Capex)
		ZAR (R)	ZAR (R)
Diesel (5-year average price)	212	R 45,197,264,073.01
Gas	85	R 18,121,544,557.58	R 27,075,719,515.44	60%
Gas	90	R 19,187,517,766.84	R 26,009,746,306.17	58%
Gas	95	R 20,253,490,976.11	R 24,943,773,096.90	55%
Gas	100	R 21,319,464,185.38	R 23,877,799,887.63	53%
Gas	105	R22,385,437,394.65	R 22,811,826,678.36	50%
Gas	110	R 23,451,410,603.92	R 21,745,853,469.09	48%
Gas	115	R 24,517,383,813.19	R 20,679,880,259.82	46%
Gas	125	R26,649,330,231.73	R 18,547,933,841.28	41%
Gas	130	R 27,715,303,441.00	R 17,481,960,632.01	39%
Gas	150	R 31,979,196,278.07	R 13,218,067,794.94	29%
Gas	190	R 40,506,981,952.23	R 4,690,282,120.78	10%
Gas	200	R 42,638,928,370.77	R 2,558,335,702.25	6%

shows potential savings if Eskom’s OCGTs (2058 MW) are run on natural gas instead of diesel over a ten-year period. Further, the OCGTs are modelled to run at a 10% load factor, translating to 876 h per annum. OCGTs with higher load factors greatly assist with flexibility due to increasing grid scale renewable energy plants and lower plant generation availability, thus strengthening grid security. The diesel costs per GJ, using its five-year average South African wholesale diesel price, took into account the wholesale discounts. The natural price used is the four-year statistical (mean, median, minimum, and maximum) gas prices from a study done by Mutsau [22].

Table 7 shows Eskom’s total ten-year cumulative diesel-powered OCGT costs, at ZAR 212/GJ, equalling approximately ZAR 45.2 bn at a 10% load factor.

The potential savings of approximately ZAR 27 bn savings (excluding capital expenditure) at a 10% load factor can be realised over a ten-year period when natural gas price is at R85/GJ (minimum), translating to 60% savings when Eskom’s OCGTs are powered by natural gas, relative to ten-year diesel costs at ZAR 212/GJ; this, confirms the findings from [10,28,29]. At a natural gas price of ZAR 110/GJ, potential savings can approximately reach ZAR 21.8 bn (excluding capital expenditure), translating to 48% savings. At the natural gas price of ZAR 190/GJ, the potential fuel savings decline significantly to ZAR 4.7 bn over a ten-year period, translating to 10%. Thus, potential savings can be realised when the price of natural gas is less than ZAR 210/GJ; as the price of natural gas increases, the potential savings tend to decline.

Table 8. IPPs’ fuel-switching saving costs analysis for OCGTs.

IPPs’ OCGTs	R/GJ	Total Cost over 10 Years @ 10% Load Factor	10 Year Savings by Using Gas Fuel Instead of Diesel (Excluding Capex)	10 Years % Savings by Using Gas Fuel Instead of Diesel (excluding Capex)
		ZAR (R)	ZAR (R)
Diesel (5-year average price)	212	R 24,675,454,790.71
Gas	85	R 9,893,460,647.22	R 14,781,994,143.49	60%
Gas	90	R 10,475,428,920.59	R 14,200,025,870.13	58%
Gas	95	R 11,057,397,193.95	R 13,618,057,596.76	55%
Gas	100	R 11,639,365,467.32	R 12,454,121,050.03	53%
Gas	105	R 12,221,333,740.68	R 12,454,121,050.03	50%
Gas	110	R 12,803,302,014.05	R 11,872,152,776.66	48%
Gas	115	R 13,385,270,287.42	R 11,290,184,503.30	46%
Gas	130	R 15,131,175,107.51	R 9,544,279,683.20	39%
Gas	150	R 17,459,048,200.98	R 7,216,406,589.74	29%
Gas	190	R 22,114,794,387.90	R 2,560,660,402.81	10%
Gas	200	R 23,278,730,934.64	R 1,396,723,856.08	6%

shows potential savings if the IPPs’ OCGTs (1005 MW) are run on natural gas instead of diesel over a ten-year period. The IPPs’ ten-year total cumulative OCGT costs are ZAR 212/GJ, equalling approximately ZAR 24.7 bn at a 10% load factor. The potential fuel savings from IPPs are similar to Eskom’s OCGTs, only varying in generating capacity proportional to the volumes produced. The potential savings of approximately ZAR 14.8 bn (excluding capital expenditure) at a 10% load factor can be realised over a ten-year period when the natural gas price is sitting at ZAR 85/GJ (minimum). This translates to 60% savings when the IPPs’ OCGTs are powered by natural gas, relative to ten-year diesel costs at ZAR 212/GJ. At a natural gas price greater than ZAR 190/GJ, the potential fuel savings decline significantly to less than ZAR 2.6 bn over ten years, translating to less than 10% savings by using natural gas instead of diesel.

Table 7 and Table 8 show the savings that can be realised if the natural gas price sits within a range of ZAR 85/GJ to ZAR 190/GJ. The indicative range is based on a 10% load factor (876 h), and an OCGT diesel price of ZAR 212/GJ. The total (Eskom’s and the IPPs’) savings for the stated natural gas prices ranges between ZAR 42 bn over ten periods at ZAR 85/GJ and ZAR 4 bn at ZAR 85/GJ and ZAR 200/GJ, respectively. At a natural gas price of less than ZAR 150/GJ, Eskom and its customers will realise financial benefit even though the OCGTs will be operating at higher load factors. Consequently, under these attractive conditions (natural gas prices < ZAR 150/GJ), Eskom would be compelled to motivate for the additional natural gas-powered 3000 MW (OCGT/CCGT) allocated in the IRP2019 [19]; and other generating units of existing OCGTs can be used as mid-merit generators, allowing them to be used for longer hours with no negative financial implications for both Eskom and its customers [10,28,29]. Moreover, according to Pauschert [30], installing a gas turbine costs about ZAR 120 million, factoring in inflation and other economic indicators. Furthermore, according to Generac industrial power [31], utilizing natural gas is more sustainable than diesel generators due to the fuel reliability of gas, expansion of demand response, reasonable installation cost, and the total ownership cost of gas is lower. Therefore, all these facts justify that the use of gas turbines is more economical, financially wise, and environmentally sustainable than diesel usage. Hence, investment in gas turbines for gas is a worthwhile investment.

3.3.2. Carbon Tax Costs Savings—From Diesel to Natural Gas

Table 9. OCGTs’ CO₂ emissions at different load factors between diesel and natural has.

Diesel-Powered OCGTs
		Load Factor
		3%	5%	6%	7%	8%	10%
	h/year	262.80	438.00	525.60	613.20	700.80	876.00
	Size of Unit
	MW	MWh
Ankerlig (Eskom)	1323	347,684.4	579,474	695,368.8	811,263.6	927,158.4	1,158,948
Gourikwa (Eskom)	735	193,158	321,930	386,316	450,702	515,088	643,860
Dedisa (IPP)	335	88,038	146,730	176,076	205,422	234,768	293,460
Avon (IPP)	670	176,076	293,460	352,152	410,844	469,536	586,920
CO2 Emissions
Ankerlig (Eskom)	Ton CO2/year	257,634.14	429,390.23	515,268.28	601,146.33	687,024.37	858,780.47
Gourikwa (Eskom)	Ton CO2/Year	143,130.08	238,550.13	286,260.16	333,970.18	381,680.21	477,100.26
Dedisa (IPP)	Ton CO2/year	65,236.16	108,726.93	130,472.32	152,217.70	173,963.09	217,453.88
Avon (IPP)	Ton CO2/year	130,472.32	217,453.86	260,944.63	304,435.40	347,926.18	434,907.72
Natural Gas-Powered OCGTs
Load Factor
		3%	5%	6%	7%	8%	10%
	h/year	262.80	438.00	525.60	613.20	700.80	876.00
	Size of Unit
		MWh
Ankerlig (Eskom)	1323	347,684.4	579,474	695,368.8	811,263.6	927,158.4	1,158,948
Gourikwa (Eskom)	735	193,158	321,930	386,316	450,702	515,088	643,860
Dedisa (IPP)	335	88,038	146,730	176,076	205,422	234,768	293,460
Avon (IPP)	670	176,076	293,460	352,152	410,844	469,536	586,920
CO2 Emissions
Ankerlig (Eskom)	Ton CO2/year	195,050.9484	325,084.914	390,101.8968	455,118.8796	520,135.8624	650,169.828
Gourikwa (Eskom)	Ton CO2/year	108,361.638	180,602.73	216,723.276	252,843.822	288,964.368	361,205.46
Dedisa (IPP)	Ton CO2/Year	493,89.318	82,315.53	98,778.636	115,241.742	131,704.848	164,631.06
Avon (IPP)	Ton CO2/Year	98,778.636	164,631.06	197,557.272	230,483.484	263,409.696	329,262.12

shows the OCGTs’ individual peaking power station CO₂ emissions for different load factors powered by different fuels: diesel and natural gas. Further, it shows how fuel switching from diesel to natural gas translates to CO₂ emissions reductions, consistent with findings from [5,10]. The amount of CO₂ emitted during combustion is a function of the carbon content of the fuel. Diesel and natural gas have different emission factors. The emission factors used for the model were taken from IPCC guidelines for national greenhouse inventories for combustion in energy industries [4]. The total OCGT annual CO₂ emissions reductions at a 10% load factor derived from diesel to natural gas fuel switching results in a 24% reduction, from 1,988,242.31 ton/CO₂ to 1,505,268.47 ton/CO₂.

Figure 4 shows carbon tax for different load factors per each fuel type for one year and also twenty years (in real rand). The carbon tax rate per ton of CO₂ equivalent for fuel combustion (electricity generation) has increased from ZAR 120/t CO₂ to ZAR 127/t CO₂. The one-year and twenty-year carbon tax for OCGTs powered by diesel operating at a 10% load factor amounts to ZAR 252.5 m and ZAR 5.05 bn, respectively. The one-year and twenty-year carbon tax for OCGTs powered by natural gas operating at a 10% load factor amounts to ZAR 191.2 m and ZAR 3.82 bn, respectively. The potential carbon tax savings for a different load that can be realised by switching from diesel to natural gas are shown in Figure 4. The one-year and twenty-year carbon tax savings from OCGTs powered by natural gas operating at a 10% load factor amounts to ZAR 61.3 m and ZAR 1.3 bn, as shown in Figure 5.

4. Conclusions and Recommendations

The utilisation of Eskom’s and IPPs’ open cycle gas turbines between 2019 and 2021 was assessed in this study. The study has found that OCGT usage, volumes produced (GWh), and the associated costs have tremendously increased from previous financial years (2016–2017 and 2017–2018), with Eskom’s and the IPPs’ volumes increasing approximately ten and seven times, respectively. The increased utilisation has attracted substantial costs, totalling ZAR 25.9 bn, for the study period. These findings are consistent with Clark et al. [28] and Silinga et al. [3] about the exorbitant costs of diesel-powered OCGTs. The OCGT volumes (GWh) and associated operational costs far exceed the regulatory approved levels for each financial year under study.

The OCGTs’ over-utilisation was triggered by the sharp decline in EAF for the study period (extremely low availability). The material decline in EAF was largely impacted by a surge in unplanned outages, which reached an alarming state, culminating in country-wide rotational load shedding being implemented to preserve the integrity of the interconnected power system, when there was insufficient generating capacity to meet demands.

The increase in OCGT costs pushed up the primary energy costs, thus contributing 8% to primary energy costs for each financial year under study, translating to 3.87 c/kWh of the set standard average tariff of 86.35 c/kWh for 2018/19, and to 4.47 c/kWh of the set standard average tariff of 106.80 c/kWh for FY2019/2020. For the study period, it is evident that OCGTs were utilised. The NERSA-allowed OCGT costs from FY2018–2019 to FY2020–2021 amounted to ZAR 10.5 bn and the actual costs due to extensive use of the OCGTs amounted to ZAR 25.9 bn. The actual OCGT expenditure far exceeds the regulatory approved costs of ZAR 15.4 bn.

Lastly, the study has found that long-term financial and environmental sustainability can be achieved for OCGTs by switching from diesel to natural gas. Further, potential savings of approximately ZAR 27 bn (excluding capital expenditure) at a 10% load factor can be realised over a ten-year period when the natural gas price is sitting at ZAR 85/GJ (minimum). Clark et al. [28], Fernández [10], and Moniz et al. [29] contend that OCGTs fuelled by natural gas, especially when operated outside peaking periods, have proven to be the most cost-effective option as compared to diesel. Additionally, the total OCGT annual CO₂ emissions reduction at a 10% load factor is 24%; thus, one-year and twenty-year carbon tax savings amounts to ZAR 61.3 m and ZAR 1.3 bn, respectively. Fernández [10] also maintains that natural gas emits less CO₂ per unit than other fossil fuels and with the lowest carbon intensity. Therefore, gas turbines powered by natural gas have lower operational costs and lower emissions than gas turbines powered by diesel. The fuel switching from diesel does not only offer savings in fuel costs and carbon tax but is a viable solution for addressing the environmental challenges through fostering sustainable low carbon development. Powering OCGTs with natural gas will assist South Africa’s efforts to stabilise greenhouse emissions between 2025 and 2035, as envisaged in the nationally determined contributions [32].

In light of these findings, the growing penetration by renewable energy technologies envisaged in the IRP2019 [16], coupled with significant generation capacity from coal generating units to be decommissioned, raises grid security and reliability risks. This situation necessitates OCGTs to be run for a longer period to support the power system under periods of low PV and wind availability [23]. It is recommended that both Eskom’s and the IPPs’ OCGTs switch fuel from diesel to natural gas and run at a 10% load factor, allowing the OCGTs to run as mid-merit generators. Work done on existing OCGTs will provide a foundation for an additional 3000 MW (OCGT/CCGT) allocated in the IRP2019, also to use natural gas.