Optimal Sizing of Decentralized Photovoltaic Generation and Energy Storage Units for Malaysia Residential Household Using Iterative Method

World's fuel sources are decreasing, and global warming phenomena cause the necessity of urgent search for alternative energy sources. Photovoltaic generating system has a high potential, since it is clean, environmental friendly and secure energy sources. This paper presents an optimal sizing of decentralized photovoltaic system and electrical energy storage for a residential household using iterative method. The cost of energy, payback period, degree of autonomy and degree of own-consumption are defined as optimization parameters. A case study is conducted by employing Kuala Lumpur meteorological data, typical load profile from rural area in Malaysia, decentralized photovoltaic generation unit and electrical storage and it is analyzed in hourly basis. An iterative method is used with photovoltaic array variable from 0.1kW to 4.0kW and storage system variable from 50Ah to 400Ah was performed to determine the optimal design for the proposed system.


Introduction
World's electricity production is severely depending on coal, oil and natural gas as energy supply.However, fuel sources are decreasing, and global warming phenomena cause the necessity of urgent search for alternative energy sources.The use of renewable energy (RE) reduces the dependency on fossil fuels, and it is proven that RE has great potential and can be utilized to fulfill world energy demand [1].
Among other alternative energy available, photovoltaic (PV) system is the most promising RE in Malaysia.In the future, it is expected more installation of decentralized PV generation units combined with electrical storage in Malaysia's distribution network.
Grid tied decentralized PV generation system has numbers of advantages; modularity system, reduce dependency on conventional energy supply, environmental friendly [2] and the excess energy can be sold to utility [3], which entitled for bill reduction.Even though its potentials are proven, the coincidence of local load requirement with PV generation needs to be studied.
This paper introduces the new power management for interaction between electrical household demand, decentralized PV generation and electrical storage unit.This paper identifies the optimum configuration for PV array and storage capacity required to supply a specific load demand over a year.The optimization of system's parameters were performed under criteria of favoured cost of energy (COE), payback period (PP), degree of autonomy (DA) and degree of own consumption (DO).

System configurations
Figure 1 shows a schematic diagram for the proposed system.Referring to the figure, PV generator is used as a decentralized power source operating in grid parallel operation mode with public network.Electrical storage unit acts as the main backup power supply.Excess energy from PV plant can be sold to utility, and utility also act as backup energy when electrical storage reaches maximum depth of discharge (DOD) The residential loads considered in the case study are only resistive loads.Figure 2 shows a typical hourly load profile for a residential house in Malaysia [4], [5] with the total daily load demand of 2.215 kWh.It is assumed that the energy requirement remains the same in each day for a year period.

System design
In this paper, the proposed power management strategies algorithm is summarized in Figure 3.The simulation is performed by iterating the PV array variable (0.1kW -4.0kW) and also the storage system variable (50Ah -400Ah).The priority of the management strategies is to optimize all PV energy output to supply demand.When PV output is higher than load demand, it is used to charge the battery and sell the surplus power to public network.However, when PV energy generation is not enough to supply the household load, battery will be discharge.If battery reaches lowest DOD, the additional energy for the load is supply or bought from utility.

PV array modeling
PV output size must able to fulfill load demand with extra energy to cover system losses.In the analysis, Kuala Lumpur's hourly climate data is used.The hourly energy generated by PV array, E pv is calculated using Equation (1) [3], [6].PSH is peak sun hour over the period of interest (one hour), P mp_stc is PV's maximum power output in standard test condition, STC, N pv is number of PV module in the array, K temp is temperature de-rating factor K wire is cable efficiency, K inv is inverter efficiency, K mm is module mismatch, and K dirt is dirt derating factor.

Battery modeling
Estimation of battery state of charge (SOC), E bat analysis is shown as Equation ( 2) [4], [7] (charging/discharging and self discharge efficiency is ignored).I bat is battery capacity in Ah, and V bat is battery's nominal voltage.Minimum SOC, E SOC.min is 50% of its full capacity.

Inverter size selection
Inverter is expected to deliver maximum AC load in household.Hence, inverter's power rating for battery system, P inv_bat is selected using Equation ( 3), where P ACload is total power from AC load demand and 1.25 is set as oversized factor [2].
Inverter's power rating for PV array, P inv_pv is selected using Equation 4: where SF inv_pv is inverter-to-PV sizing factor (0.80) [2], and N pv is quantity of PV modules in the system.

System optimization criteria
In order to select an optimum combination, four analyses were used to determine the ideal parameters for PV plant and electrical storage unit.To find optimum configuration with the lowest investment possible, it is inevitable to use economic analysis.Therefore, Cost of Energy (COE) and Payback period (PP) is used.Meanwhile, Degree of Autonomy (DA) and Degree of own-much energy generated from PV plant are used in residential household.

Cost of energy
COE (RM/kWh) is defined as the average cost per kWh of electrical energy produced by the system when a lifetime, investment cost, replacing, operation & maintenance, and capital cost is considered [8].Based on technical datasheet, the batteries need to be replaced for every 5 years and inverter need to be replaced every 10 years [5].Lower COE is desirable and it is calculated by dividing producing electricity annualized life cycle cost, LCC 1year with E PV [9].The market price for system components is summarized as in

Payback period
PP is used to determine the time it would take to recover or break even on initial investment cost and ongoing O&M cost [11].E sell is energy fed back to utility during surplus energy from PV and E buy is energy bought from utility during lack of PV output.Meanwhile, feed in tariff, FIT is different for each country, and for Malaysia, it is RM 1.35/kWh in 2015 [10].Each energy unit bought from utility is provided by Tenaga Nasional Berhad, TNB [12].Shorter payback period are more desirable.
where E bat_start is initial battery's energy and E bat_end is battery's energy at the end of the period.Higher DA is desirable for optimum configuration.

Degree of own-consumption
Meanwhile, DO shows how much energy is provided from decentralized PV plant that is directly used by the residential household compared to the provided power by PV generator.It is calculated as Equation 10[13].However, contrast with DA, optimum design tends to have lower DO.4, it shows the results from COE analysis as the optimization criteria.It can be seen that when battery capacity ranged from 50Ah to 100Ah is paired with PV capacity from 3kW to 4kW, the system obtained the lowest COE of RM 0.57/kWh.Figure 5 is obtained after payback period analysis was performed; it is observed that for all combination of battery capacity at the PV ranged from 0.3kW to 0.9kW gives the payback period more than 20 years.Hence, it can be concluded that any configuration within the specified range of PV capacity is not feasible.This is due to insufficient PV output, which causes the total benefit obtained from selling energy to utility is inadequate to break-even on LCC system.However, for all pairs of battery capacity at the PV ranged from 3.45kW to 4kW gives PP lower than 8 years.
Hence, the lowest PP and DO values obtained are 7.333 years and 0.1436p.urespectively at the combination of 4kW PV array and 50Ah battery capacity.Meanwhile, the highest DA obtained using the configuration of 4kW PV capacity and 100Ah is 0.9998p.u.

Conclusions
The optimal sizing of decentralized photovoltaic generation and energy storage units for Malaysia residential household has been presented.From the results obtained, the most favorable design is either a system of PV capacity 4kW and 50Ah or 100Ah battery capacity, which is obtained from the lowest DO or the highest DA value.However, based on the observation, it is wiser to select the best combination parameters using DA analysis.It is because DA indicates which design has high degree of energy independence from local utility with lowest COE and low PP.The selected parameter pair is able to use most energy from PV generator directly in residential household, with low energy need to be bought from utility and has small changes between initial and ending battery's capacity.
This has proven that it is possible to effectively use RE sources and storage that allows the system to be independent from utility, and gain profit from solar FIT.

Figure 1 .Figure 2 .
Figure 1.System configurations for a residential load in grid parallel operation mode with PV array, electrical storage and utility.

Figure 4 , 5 ,
Figure 4,5, 6 and 7  show the analyses results based on the optimization criteria as stated in equations 7, 8, 9 and 10.Referring to Figure4, it shows the results from COE analysis as the optimization criteria.It can be seen that when battery capacity ranged from 50Ah to 100Ah is paired with PV capacity from 3kW to 4kW, the system obtained the lowest COE of RM 0.57/kWh.Figure5is obtained after payback period analysis was performed; it is observed that for all combination of battery capacity at the PV ranged from 0.3kW to 0.9kW gives the payback period more than 20 years.Hence, it can be concluded that any configuration within the specified range of PV capacity is not feasible.This is due to insufficient PV output, which causes the total benefit obtained from selling energy to utility is inadequate to break-even on LCC system.However, for all pairs of battery capacity at the PV ranged from 3.45kW to 4kW gives PP lower than 8 years.Hence, the lowest PP and DO values obtained are 7.333 years and 0.1436p.urespectively at the combination of 4kW PV array and 50Ah battery capacity.Meanwhile, the highest DA obtained using the configuration of 4kW PV capacity and 100Ah is 0.9998p.u.

Figure 5 .
Figure 5. Results of PP analysis.

Figure 6 .
Figure 6.Results of DA analysis.

Figure 7 .
Figure 7. Results of DO analysis.