17TH ERU RESEARCH SYMPOSIUM, 2011: FACULTY OF ENGINEERING, UNIVERSITY OF MORATUWA, SRI LANKA
Feasibility Study on Solar Photovoltaic Office Air Conditioning in Sri Lanka K.A.E.K. Perera, K.A.L. Kollure, W.A.K.S.S. Prasad, K.K.C.K. Perera and P.A.B.A.R. Perera, A.T.D. Perera
Abstract Due to the rapid depletion of fossil fuel resources and global concern on Green House Gas emissions, sustainable energy technologies have become important in Heating, Ventilation and Air Conditioning (HVAC) systems in buildings. As a result, much focus is given on research and application based on solar powered HVAC, in various parts of the world. Such initiatives are of timely importance in Sri Lankan context due to the availability of good solar energy potential. In this work, an existing air conditioning system for an office building of 51m2 in Sri Lanka was taken into study. An economic evaluation was performed using HOMER (Version 2.68 beta) to assess the feasibility of using Solar PV (SPV) to power the air conditioning system. The results show that under present circumstances, the Levelized Cost of Energy (COE) of the optimized photovoltaic stand alone system lies far above that of the grid powered. It further depicts that grid integration would bring down the COE of the SPV system closer to the range of the main grid. Hence in current conditions, grid integration is the way forward for SPV assisted air conditioning in Sri Lanka.
1. Introduction Located near the equator, Sri Lanka (7° 30' N latitude and 81° 30' E longitude) receives abundance of solar radiation throughout the year making it a tropical country with a warm climate. Consequently in urban areas, a significant proportion of building energy consumption is drawn for space cooling. Due to the vast potential in solar resource, a solar assisted solution for space cooling seems most logical compared to other renewable resources. Solar energy is widely used in the tropical and subtropical countries, mostly for standalone applications including rural electrification and refrigeration [1].This mature technology is no stranger to Sri Lanka as it is widely used for rural electrification mainly in the capacities of 50W PV/Battery systems [2]. Yet there exists many technologies in harnessing this solar potential for refrigeration applications such as solar assisted vapour compression, vapour absorption, vapour adsorption, etc. This work attempts to deviate from the norms of electrification and refrigeration towards air conditioning of office spaces. It analyses the economic feasibility of the SPV assisted vapour compression technology for air conditioning in the context of Sri Lanka. For this purpose HOMER (Version 2.68 beta) is used as the design, simulation and optimisation tool. HOMER was chosen due to its capability to model, simulate and optimize both grid connected and standalone energy systems.
2. Methodology In order to evaluate the financial feasibility it is essential to model, simulate and optimize the SPV assisted vapour compression system. The system consists of a Photovoltaic array, a battery bank and a converter. Initially the energy system is mathematically modelled to power it without grid integration (standalone). Subsequently the system is grid integrated where both renewable (solar generated) and conventional (grid) energies were used in combination. Ten different combinations varying the renewable energy component were simulated.
renewable energy potential was taken for this work (Fig. 1). The annual average of solar radiation for this site was 5.7 kWh/m2/d.
Figure 1 - Variation of Monthly solar insolation
2.2 Electrical Load The office under consideration is 51m2 in floor area situated in Hambanthota. It houses an average of 25 employees and works on the weekdays from 9am to 5pm. The air conditioning system consists of 2 split type air conditioners (12,000BTU/h) with an EER of 12 BTU/Wh. The cooling load for the office building was calculated through the Cooling Load Temperature Differential method (CLTD) by its builders based on ASRAE Standards 55-2010. The monthly electric load variation is shown in Fig. 3. It consists of an average daily load of 14.8kWh/d and a peak load of 4.1 kW. 2.3 Simulation of Energy system In the simulation, primary electric load is simulated while taking into consideration the fluctuation of solar irradiation and grid integration in hourly basis throughout the year. Fig. 2 shows the system architecture for a grid integrated system. Based on the simulation life time of the battery bank and level of grid interaction were computed.
2.1 Solar resource Hourly tilted solar irradiation data at Hambantota, (06°07′ N 81°07′ E) a local location blessed with good Figure 2- Grid integrates system architecture
17TH ERU RESEARCH SYMPOSIUM, 2011: FACULTY OF ENGINEERING, UNIVERSITY OF MORATUWA, SRI LANKA Seasonal Profile
5
max daily high
Load (kW)
4 3
mean
2
daily low min
1 0
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Figure 3 – Monthly Load distribution
The components used for the simulation along with their properties are shown in Table 1. 2.4.1 Photovoltaic Panels Mono-crystalline cells were preferred over thin film cells due to their high efficiencies (13% under standard conditions) [1]. A slope of 18˚ was assumed for installation as in a typical roof in Sri Lanka [3]. HOMER sizes the panel according to equation 1. ̅
1 , 1 ̅ , Where, rated capacity of the PV array (kW) Ypv fpv PV derating factor (%), GT solar radiation incident on the PV array in the current time step (kW/m2), GT,STC incident radiation at standard condition (1kw/m2) αp temperature coefficient of power (%/˚C), TC PV cell temperature in current time step (˚C), TC,STC PV cell temperature at standard conditions (25˚C) 2.4.2 Grid Integration The grid electricity was purchased and the excess solar generated electricity was sold back to the grid as a grid integration strategy. Purchasing was done at USD 0.104/kWh [4] whilst the generated was sold at USD 0.19/kWh [5]. Net metering was left out for this analysis. 2.5 Economics and Constrains A lifetime of 25 years was opted for the project with an annual interest rate of 8.8%. The standalone system was granted a 3% capacity shortage without any penalties [2]. 2.6 Optimisation and Costs of Energy HOMER optimises the system taking the net present cost (NPC) as the objective function rather than the cost of energy (COE). According to HOMER, this is because total NPC is more trustworthy than COE. Equation 2 defines COE with regard to this analysis.
Where, Cann,tot Eprim,AC Egrid,sales
Nov
Dec
Ann
,
(2)
!"#$%,&' ( )"#*,+,%+
total annualized cost of the system (USD/yr) AC primary load served (kWh/yr) total grid sales (kWh/yr)
3. Results The results of the standalone system (100% REF) and the grid integrated systems are shown in Table 2. 3.1 Sizing Results It is clearly shown in Fig. 4, that the PV array size increases as the REF increases (grid connected systems) but reduces as we switch from grid connected (90% REF) to a standalone system (100% REF). 8 7
PV Array Size (kW)
2.4 System Components
Oct
6 5 4 3 2 1 0 0
20
40
60
80
100
Renewable Energy Fraction (%)
Figure 4 - PV Array size vs. REF
It is further observed in Fig. 5 that the standalone has a greater percentage of unused excess electricity (30%). Hence it is economical and logical to feed this excess electricity to the grid through net metering or through sell back. Since the price of sell back is higher than the cost of purchase [4], [5] net metering was left out in in this analysis.
Table 1 - System Component details
Characteristic Manufacturer /Model Power Range
PV Modules Sharp/Mono-crystalline 0.0kW – 8kW
Lifetime Purchase Cost Replacement Cost Maintenance Cost
25 years 2500 USD/kW 25 USD/yr
Battery Trojan/T-105 Nominal voltage - 6V; Nominal Capacity – 225Ah (1.35kWh) Lifetime throughput – 845kWh 150 USD/battery 150 USD/battery 1 USD/battery
Converter SMA/Sunny Boy 460W – 5kW 10 years 550 USD/kW 550 USD/kW 50 USD/kW
17TH ERU RESEARCH SYMPOSIUM, 2011: FACULTY OF ENGINEERING, UNIVERSITY OF MORATUWA, SRI LANKA Table 2- Sensitivity results for all cases
Renewable Energy Fraction (%)
COE ($/kWh)
Number of Batteries
PV Size (kW)
Converter Size (kW)
Initial Cost ($)
Total NPC ($)
PV price $2500/kW
PV price $5000/kW
Excess Electricity (As % of (kWh/yr) production)
Stand-alone
5.6
26
3.8
21,990
29,865
0.566
0.87
2,503
30.1
90
7.1
0
4
21,950
18,928
0.351
0.73
393
3.36
80
4.8
0
3
15,050
15,494
0.287
0.542
105
1.18
70
3.8
0
2
11,650
13,585
0.252
0.45
326
4.06
60
3
0
1.6
9,230
11,888
0.22
0.379
238
3.21
50
2.4
0
1.2
7,330
10,733
0.199
0.323
270
3.85
40
1.8
0
1.2
5,680
9,527
0.177
0.274
17.2
0.26
30
1.3
0
0.7
4,005
8,391
0.156
0.227
97.4
1.5
20
0.8
0
0.46
2,478
7,287
0.135
0.185
37.9
0.64
10
0.4
0
0.46
1,378
6,812
0.126
0.154
0
0
0
0
0
0
0 5,636 0.104 0.104 0 grid integrated system being independent of a battery bank and the higher price for sell back of excess electricity to the main grid.
35
Excess Electricity (As a % of Production)
30 25
4. Discussion and Conclusion
20
From this study it is clear that the cost of 100% renewable (standalone) energy at USD 0.56/kWh is yet economically farfetched when compared to the grid energy at USD 0.104/kWh. Nevertheless, a compromise can be reached by grid integrating the system. The cost fraction of the PV array in standalone systems is around 65% but does steadily decrease as the system is grid integrated. Hence it is clear that with the reduction of PV costs, a significant reduction in COE can be expected (Fig. 6). At present, the prices of the PV panels and inverters are on a steady but slow decline [6]. Hence as time goes on, with the reduction in component pricing and increase in grid energy pricing [7], air conditioning systems with solar PV would become a sustainable eco friendly alternative with a better economical feasibility.
15 10 5 0 0
20
40
60
80
100
Renewable Energy Fraction (%)
Figure 5 - Excess Electricity vs. REF
3.2 Cost Results The difficulty with harnessing solar energy for useful work has always been the financial barriers associated. It is shown that (Fig. 6) the COE is at its maximum when the system is designed to operate as a standalone. 0.9
Cost of Energy (USD/kWh)
0.8 0.7
References PV price in 2011 at USD 2500/kW PV price in 2009 at USD 5000/kW Cost of Grid Energy
0.6
[1]
D. S. Kim and C. A. Infante Ferreira, “Solar refrigeration options a state-of-the-art review,” International Journal of Refrigeration, vol. 31, no. 1, pp. 3-15, Jan. 2008.
[2]
T. Givler and P. Lilienthal, “Using HOMER Software, NREL’s Micropower Optimization Model to Explore the Role of Gen-Sets in SmallSolar Power Systems. Case Study: Sri Lanka,” National Renewable Energy Laboratory, Golden, Colorado, May 2005.
[3]
G. T. Still and T. Thomas, “The optimum sizing of gutters for domestic roof water harvesting,” Development Technology Unit, University of Warwick, Dec. 2002.
[4]
“Consultation Paper on Setting Tariffs for the Period 2011-2015,” Public Utilities Commission of Sri Lanka, 2011.
[5]
K. Dissanayake and I. Vithanage, “Solar Energy Development in Sri Lanka,” presented at the SOLAR 2011, Raleigh, North Carolina, United States of America, 2011.
[6]
A. Kolan, F. Wang, and C. Yang, “Primer for Photovoltaic Technology across Texas,” Texas A&M University, College Station, Texas, Dec. 2010.
[7]
“Sustainable Power Sector Support Project,” Asian Development Bank, Sri Lanka, 39415, Nov. 2010.
0.5 0.4 0.3 0.2 0.1 0 0
20
40
60
80
100
Renewable Energy Fraction (%)
Figure 6 - Cost of Energy vs. REF
Furthermore, it is shown that a 36% cost reduction can be made by integrating the system with the main grid (moving from standalone to 90% REF and 10% grid energy). This is a significant cost reduction through a small contribution of grid energy. It is mainly because the
0