Techno-economic analysis of hybrid PV-Battery-diesel system for isolated Dockyard In West Papua

. An isolated dockyard located in West Papua currently relies on two diesel generators (DGs) with a total capacity of 1,100 kW to fulfil its energy demand. However, these DGs operate at low efficiency, resulting in a high levelized cost of energy (LCOE) of Rp9,064 /kWh and generating 496 metric tons CO2 emissions equivalent per year. On the other hand, the dockyard holds significant solar energy potential. There is opportunity to utilise solar photovoltaic (PV) since its LCOE is also decreasing. This study aims to determine the optimal configuration and sizing of a PV-battery-diesel to minimise LCOE, considering CO2 emissions and a maximum capital expenditure (CAPEX) constraint of Rp 16.4 billion. Four topologies are evaluated: DG, PV-DG, PV-battery, and PV-battery-DG topology. The results highlight the PV-battery-DG topology as the lowest LCOE of Rp3,185 /kWh while adhering to the CAPEX constraint. The PV-battery topology is the most effective in reducing emissions by 100%. Both the PV-battery-DG and PV-battery topologies are less sensitive to fuel price volatility but are more influenced by changes in inflation rates. In contrast, the DG and PV-DG topologies are relatively less responsive to inflation changes but significantly responsive to variations in fuel prices.


Introduction
Indonesia, as one of the world's largest archipelagic nations, positions the shipping industry as a priority sector.This industry plays a significant role in supporting the national economy as a primary mode of transportation for goods and commodities.There are 4,419 maritime vessels within Indonesian territory and is supported by a dockyard.There are approximately 240 dockyards and several of the dockyards are located in remote areas with no electric utility [1].Consequently, the dockyards rely on diesel generators (DGs) using high speed diesel (HSD).It leads to negative impacts such as carbon dioxide equivalent (CO2e) emissions, and the high, unstable costs due to fuel price volatility.
The dockyard under study is also located in the remote area, and serves as the repair hub for vessels operating in the eastern part of Indonesia.Currently the dockyard relies on two DGs with capacity of 800 kW and 300 kW to meet its energy demand of 1,431.2kWh/day and a peak load of 108.9 kW.Nonetheless, both DG units operate at low efficiency due to overcapacity.Consequently, the levelized cost of energy (LCOE) is high, priced at Rp9,064 /kWh.Furthermore, the existing system also contributes 496 metric tons CO2e emission per year.A comprehensive reassessment of the dockyard's power generation system is needed to achieve more sustainable power generation, while considering the lowest LCOE, also addressing the CO2e emissions produced and capital expenditure (CAPEX) constraint.
Several studies have been conducted regarding the optimization of PV-battery-diesel systems.Using different methodologies for different locations, hybrid PV-battery-diesel systems have demonstrated lower levelized cost of energy (LCOE) values compared to standalone DG or PV systems [2][3][4][5][6][7][8][9].The systems exhibit the capability to ensure grid stability, with the highest PV penetration reaching up to 90%.Hybrid PV-battery-diesel and PV standalone is less sensitive to the fuel price volatility.Other investigations have delved into control strategies, namely load following (LF) and charging cycle (CC).The outcomes reveal that the LF control strategy is more efficient in emission reduction [2].
This study will analyse four topologies, they are DG, PV-DG, PV-battery, and PVbattery-DG.The aim is to achieve an optimal power system based on the lowest LCOE, the emission reduction, and the CAPEX constraint of Rp16.2 billion.This will be done by conducting a techno-economic analysis of the topology and capacity of PV, battery, and DGs.This paper is arranged as follows: Section 2 describes the existing system and its load profile, while Section 3 outlines the sizing and topologies.In Section 4, the optimization is conducted, while the result and discussion are provided in Section 5. Finally Section 6 concludes the work.

Existing system
Information regarding the generation units can be observed in Table 1.It is shown that the dockyard's electrical system is supplied by 2 DG units with capacities of 800 kW and 300 kW.The initial design allocated the 800 kW DG unit to provide electricity to the dockyard, as depicted in Figure 1.Meanwhile, the 300 kW DG unit was originally intended to supply power to docked ships.Due to significant overcapacity of 800 kW DG, the 300 kW DG unit was eventually used to operate and supply electricity at 50 Hz for loads of 20 kW and 60 kW.To address this situation, the dockyard aims to reevaluate the optimal power system solution for the utility load, thereby discontinuing the use of the 300 kW DG unit with a different frequency.The shipyard has a daily energy consumption (DEC) of 1,431.2kWh/day and an average day load (ADL) of 991.2 kWh/day.The daytime (6:00-17:00) loads contribute approximately 69% of the total daily electricity demand, indicating an opportunity to harness solar energy through the installation of PV systems operating during the daytime.Monthly global horizontal irradiance (GHI) values in the dockyard are shown in Fig. 2. The total GHI over a year is 1,853 kWh/m².The average daily GHI value over the year is approximately 5.08 kWh/m².The highest average daily GHI occurs in October at 5.64 kWh/m², while the lowest average daily GHI takes place in December at 4.23 kWh/m².

System topologies, sizing and optimization
The sizing of each component is performed manually based on the conducted literature review.This is carried out as a reference for the forthcoming simulation results.Thus, the researcher will compare the outcomes of theoretical calculations with HOMER simulation.In this research, four topologies will be analyzed, namely the DG topology, PV-DG topology, PV-Battery topology, and PV-Battery-DG topology.

DG topology
The DG topology is calculated under the assumption, when the efficiency of the DG operates at the optimal point, fuel consumption will be minimised.This, in turn, will lead to reduced fuel costs and potentially result in lower LCOE and lower emissions.Enhancing DG efficiency can be achieved by establishing a minimum operating constraint for the DG, set at minimum 40% of the DG's rated capacity [10][11][12][13].
DGs that fulfil the constraints will be mapped and selected based upon their highest efficiencies corresponding to each load.The efficiency of DGs will be calculated using Equation 1 [14].Where: The value of the mass flow rate is calculated using Equation 2. [9]   =   * (  1000 ) (2) Where: The results of mapping DG capacity that meets operational constraints are presented in Figure 3.Meanwhile, the results of efficiency calculations for DGs that meet the constraints are presented in Figures 4 (a  Afterward, researchers conducted optimization using HOMER software with input parameters as shown in Table 2.These parameters were subsequently employed for all topologies.The simulation results indicate that the optimal capacities for the DG are 24 kW and 109 kW.Therefore, in this case, the researchers employed 2 DG units with respective capacities of 109 kW and 24 kW, which were used for all topologies equipped with DG.

PV-DG topology
The PV-DG topology is analysed under the assumption that PV systems offer a lower LCOE compared to fossil fuels and generate minimal CO2e emissions [15].This topology aims to achieve a reduced LCOE while adhering to a CAPEX constraint of IDR 16.2 billion, all the while producing fewer emissions.The capacity calculation for the PV-DG system follows the methodology used for determining PV capacity in on-grid PV systems.However, the key distinction is that excess energy generated by the PV system cannot be exported to the DG.Thus, the PV system's energy generation should not surpass peak load demand to prevent potential system failures stemming from surplus PV energy in the absence of energy storage.The calculation of PV capacity is performed using Equation 3.
The PVout value, which represents the solar potential in terms of kWh per kWp installed at the research location used in the calculations, is 3.91 kWh/kWp.Meanwhile, the average daily load (ADL) or the daytime electricity consumption stands at 991.2 kWh per day.Based on the calculation results, the system with this topology has a maximum integrated PV capacity of 131 kWp, along with DG capacities of 109 kW and 24 kW.However, the optimization results indicate an optimal PV capacity of 194 kW.This difference arises because the optimization considered economic aspects.

PV-Battery topology
In the previous PV-DG topology, the system harnesses solar energy through the use of PV systems.However, PV utilisation is limited by the absence of an energy storage system.Therefore, the researcher analyses the PV-battery topology under the same assumption that PV exhibits a lower LCOE and results in reduced CO2e emissions.In the PV-battery topology, all energy requirements are supplied by both PV and batteries without the use of a DG system.Before conducting the calculations, there are two primary aspects to consider in this topology: determining the AC coupling or DC coupling system and determining the number of days for autonomy.
In this research, the researcher employed an AC-coupled system due to the relatively high peak load at the research location, which is 108.9 kW.Meanwhile, DC coupling was used for smaller capacities due to inadequate product compatibility.Additionally, AC coupling offers high modularity, making it more convenient for capacity expansion of PV or hybrid systems with other power sources.The researcher also established a 3-day autonomy period because the research location experiences relatively higher rainfall compared to other areas in Indonesia.
After performing calculations under the conditions of an AC-coupled system with 3 days of autonomy, the results yielded PV and battery capacities of 612 kWp and 3,012 kWh, respectively.The PV capacity was increased by 20% to ensure sufficient energy for charging the tripled battery capacity.The battery capacity was enlarged to provide 3 days of autonomy, ensuring the system's operation during night time and extreme conditions, such as adverse weather.Meanwhile, the optimization using HOMER results indicate optimal PV and battery capacities of 881 kWp and 2,587 kWh, respectively.

PV-Battery-DG topology
The PV-battery-DG topology is a combination of the previous topologies with the aim of mitigating their respective limitations to yield a more optimal system.Based on the existing system data in Table 1, the PV-battery-DG combination can offer an improved solution with a lower LCOE and reduced emissions.In this system, PV and batteries take precedence, while the DG is utilised selectively.Hence, in the design of the PV-battery-DG system, careful consideration is required for determining the control mechanism for system operation.In this system, a 1-day autonomy period is established.Unlike the PV-battery topology, where increasing battery capacity is necessary to handle rainy days, in this topology, system reliability is already ensured by the presence of DG.
Based on the calculations conducted, the PV and battery capacities in the PV-battery-DG topology are 510 kWp and 1.004 kWh, respectively.Meanwhile, the optimization results indicate that the optimal capacities for the system are PV 453 kWp and battery 834 kWh.PV and battery will be integrated with 2 DG units with capacities of 109 kW and 24 kW.Nevertheless, PV and battery energy will remain the primary priority due to the application of LF control during operation.
All topologies are technically feasible as they are capable of supplying a daily energy demand of 1431.2 kWh/day and an annual energy of 522.388 kW/year without any unmet load.A summary of the energy produced by each topology can be found in Table 3.The amount of energy from PV supplied to the load is referred to as PV penetration.The DG, PV-DG, PV-battery, and PV-battery-DG topologies have PV penetrations of 0%, 33%, 100%, and 89%, respectively, in sequence.From this data, it can be observed that the hybrid solar-diesel system without batteries can achieve a maximum PV penetration of 33%.Meanwhile, in the hybrid solar-diesel system with batteries or the PV-battery-DG topology, PV penetration can reach up to 89%.To increase the PV penetration from 89% to 100%, or approximately 10% reliability improvement, an increase in battery capacity from 834 kWh to 2,587 kWh, roughly threefold, is required.

Economic analysis
In conducting the economic analysis, the researcher employed LCOE as a parameter.LCOE is a value used to compare the economic viability of one power generation technology with another.It is calculated by summing all the expenses incurred over the project's lifetime and then dividing it by the energy generated to supply the electrical load.The analysis revealed that all topologies have lower LCOE values compared to the base case LCOE of Rp9.064 /kWh when supplemented with two DG units of 800 kW and 360 kW in capacity.The comparison of LCOE for all systems is presented in Fig 5 .Topology 1, consisting of DG units with optimal capacities of 109 kW and 24 kW, has an LCOE value of Rp7.008 /kWh.' Fig. 5. Economic analysis result.
The improvement in DG efficiency can significantly reduce the LCOE of the power generation system by approximately Rp2.056 or roughly 23%.This represents a substantial reduction in LCOE, and when multiplied by the annual electricity consumption of 522,388 kWh/year, it translates to potential operational cost savings of around Rp1 billion per year.Meanwhile, the CAPEX required for the procurement of two new DG units is only Rp 540 million, well below the established CAPEX limit.
Next is Topology 2, which is the PV-DG configuration.The addition of PV is based on literature studies indicating that PV systems have relatively low LCOE values.The results show that adding PV with a 33% penetration of the total annual electricity demand can reduce the LCOE to Rp5,583, which is lower by Rp3,481 or approximately 38% compared to the base case.This implies potential annual savings of Rp1.8 billion if this topology is implemented.The CAPEX required to implement the PV-DG topology is Rp4.22 billion, still well within the CAPEX limit of Rp16.2 billion.
Referring to the LCOE reduction due to the addition of 33% PV in Topology 2 (PV-DG), the researchers then analyzed Topology 3, a system where all electricity is supplied solely by PV and batteries, or the PV-battery topology.The result is an LCOE decrease to Rp5,227/kWh, which is not significantly different from the previous PV-DG topology.The LCOE reduction for the PV-battery topology is Rp3,837/kWh or approximately 42% compared to the base case.The difference in LCOE between the PV-DG and PV-battery topologies is only about Rp311.However, to implement the PV-battery topology, a CAPEX of up to Rp30.2 billion is required, which exceeds the CAPEX limit.This indicates that the PV-battery topology is not recommended due to failing to meet the maximum CAPEX constraint.
Finally, the lowest LCOE is achieved by the PV-battery-DG topology, a hybrid PV system with diesel backup and energy storage in the form of batteries.The PV-battery-DG topology has an LCOE of Rp3,185/kWh, which is 65% lower than the base case or a decrease of approximately Rp5,879/kWh.Implementing this topology could result in savings of around Rp3 billion per year.To implement this topology, a CAPEX of Rp14 billion is required, which is within the CAPEX limit and feasible.The CAPEX for the PV-battery-DG topology with 89% PV penetration is 54% lower than the previous PV-battery topology with 100% PV penetration.This means that to increase PV penetration from 89% to 100%, which adds approximately 10% system reliability, would require an initial investment or CAPEX that is double.

Emission analysis
The calculations of CO2e for all topology schemes have yielded lower CO2e emissions compared to the base case.The emission reductions for each topology are presented in Fig 6 .Sequentially, the emissions reduction for the DG, PV-DG, PV-battery, and PV-battery-DG topologies is 35%, 47%, 100%, and 91%, respectively.The lowest emissions among all systems are observed in the PV-battery topology.This is evident as the PV-battery topology relies on renewable energy from the sun.Meanwhile, the PV-battery-DG topology also yields relatively low emissions, not significantly different from the PV-battery topology.With 89% PV penetration, the PV-battery-DG topology results in CO2 equivalent emissions of 44 tons per year.This represents a reduction of 452 tons per year compared to the base case emissions.Over the 25-year project lifespan, the PV-battery-DG topology with the lowest total LCOE is capable of reducing CO2e emissions by a total of 11,300 tons.

Sensitivity analysis
The sensitivity analysis results illustrate how the LCOE values can change due to variations in input variables, namely inflation rates and the price of diesel fuel (BBM Solar).The sensitivity analysis, involving a 20% change in input variables, is presented in Fig 7 and 8.A negative sign (-) in the sensitivity indicates that the LCOE value at the test point decreases from the initial value.The PV-battery topology with 100% PV penetration proves to be the most sensitive to changes in the inflation rate.With a 20% increase in the inflation rate, the LCOE of the PV-battery topology decreases by approximately 4.5%.This phenomenon occurs because systems with higher PV penetration exhibit long-term investment characteristics.The total costs incurred are more substantial in year zero or the initial investment costs.These initial investment costs are not influenced by inflation since they occur in year zero.In contrast, systems with a higher percentage of fossil energy, such as the DG-only topology, are not sensitive to increased inflation rates.When inflation rises by 20%, the LCOE change for the DG-only topology is only about 0.06%.This is because the total costs in the DG system are dominated by annual fuel and operational costs, which are influenced by inflation rates.Conversely, in the sensitivity analysis of LCOE values to changes in the price of diesel fuel (HSD price), topologies with lower penetration or a more dominant share of fossil energy are more sensitive to these price fluctuations.The DG topology, which relies entirely on BBM Solar for energy (100% penetration), exhibits a sensitivity rate of 19%.With a 20% increase in the price of BBM Solar, the system's LCOE increases by approximately 19%.Meanwhile, the PV-battery topology is not affected by changes in BBM Solar prices as it no longer utilizes DG.The sensitivity rates for the DG-only topology (0% PV penetration), PV-DG topology (33% PV penetration), PV-battery-DG topology (89% PV penetration), and PV-battery topology (100% PV penetration) are 19%, 17%, 4.8%, and 0%, respectively.
The lower the percentage of DG usage, the more resilient the system is to fluctuations in BBM Solar prices.For the PV-battery-DG topology, the sensitivity of LCOE to changes in BBM Solar prices remains relatively low.With a 20% increase in BBM Solar prices, the LCOE only changes by 4.77%.Even if BBM Solar prices are assumed to increase by 100%, the LCOE for the PV-battery-DG topology only increases by 20% to Rp3,820, which is still lower than the unaffected LCOE of Rp5,227 for the PV-battery topology.

Conclusion
After conducting technical, economic, and environmental analyses, the researchers summarized the analysis of the four topologies into a single radar chart in Fig 9 .Three parameters included in the chart are the LCOE, which is the primary priority in determining optimal outcomes, the CAPEX value, which serves as a research constraint, and emissions, which are a consideration in selecting the recommended power generation system.A scale of 5 indicates excellence, while a scale of 1 denotes poor performance.The scales are assigned based on the values of LCOE, CAPEX, and emissions in Figures 5 and 6.The topology with the lowest LCOE value, making it the most favorable, is the PVbattery-DG topology.Furthermore, this topology also has the lowest emissions and CAPEX value below the specified maximum CAPEX limit.The minimum point for the CAPEX parameter is 3.If a topology has a CAPEX value below 3, it means the CAPEX for that topology exceeds the maximum CAPEX limit.Therefore, this topology is highly recommended with a total score of 13.The schematic diagram of the PV-battery-DG topology can be seen in Fig 10.Similarly, the PV-battery topology also exhibits the highest emissions reduction due to all energy coming from renewable sources.The PV-battery topology also has an LCOE value of 4, which is better than the PV-DG and DG-only topologies.However, this topology has a low CAPEX score, indicating a relatively high CAPEX cost that exceeds the maximum CAPEX limit.The total score for the PV-battery topology is 10.
In terms of CAPEX value, the DG topology has the highest score.This is because the CAPEX required for acquiring DG is relatively low compared to PV and battery systems.However, this topology has low LCOE and emissions values.The LCOE for the DG topology is the lowest compared to other topologies.The total score for the DG topology, considering the three parameters, is 9. Another topology with low CAPEX is the PV-DG topology, which has a CAPEX score of 4, indicating relatively low CAPEX requirements due to the absence of an energy storage system.However, solar energy utilization in this system is limited.Therefore, the scores for LCOE and emissions for this system are both 3.The total score for the PV-DG topology is 10, the same as the PV-battery topology but with significantly lower CAPEX.
) and (b).The DGs with the highest efficiency for loads of 108.9 kW, 60 kW, and 20 kW are DG 109 kW and DG 24 kW.