Optimised Design and Analysis of Solar Water Pumping Systems for Pakistani Conditions

This paper is about the optimized design and analysis of two solar water pumping systems in which one of the systems is designed with a battery bank and other with a cylindrical water tank for a selected site in Pakistan. The design, sizing, cost analysis and steady state analysis of the proposed systems were done in HOMER and dynamic analysis of the designed system with battery bank was performed in MATLAB/Simulink. The simulations performed in HOMER involved proper mapping of the loads which helped to evaluate the PV panel requirement, inverter rating, batteries (in case of battery based solution), modeling of water tank as a deferrable load (in case of solution based of water tank) and detailed cost analysis for a life time of 25 years. To verify the design of the solar water pumping system with battery bank, a simulation in MATLAB/Simulink for study of dynamic behavior of the overall system was performed which involved mathematical modeling of a PV panel, buck-boost converter, inverter, battery bank and motor/pump, a perturb and observe maximum power point tracking algorithm based control system. Analysis was conducted based on the economic results that indicate designed solar water pumping system with water tank would be a cheaper solution as compared to solar water pumping system with a battery bank. This work can be taken as a case study for the understanding and optimized designs of solar water pumping system with battery bank and with cylindrical tank in Pakistani conditions.


Introduction
Pakistan is primarily an agricultural economy which contributes for about 22.2% for overall GDP [1] of the country and employs around 42.3% of the overall country's labour strength [2]. With agriculture being the main stay of economy, an important component of agriculture is water. Pakistan is a country which faces water shortage, with water availability of around 5000 m 3 per capita in 1950s. It has shrunk to around 1000 m 3 per capita currently [2] and by 2025, the water short fall is expected to reach 150.8 Million acre-foot [3]. Bashir et al. [4] stated the reason for shrinking water resources as poor management of water due to availability of less number of small and large dams for water storage which results in wastage of water.
To combat this shortage of water supply for irrigation, water pumps and tube-wells are installed and since the scenario of electricity in country is not great as a short fall of 5000 MW is faced by the country [5]. So, instead of powering the water pumps through electricity especially in remote areas where there is no or limited access of electricity, fossil fuels have been used to power water pumps. Since Pakistan is a country which is not rich in fossil fuels, increased dependence on imported fossil fuel for the purpose of electricity production has increased the circular debt to 1.2 trillion Pakistani rupees during the time span of last five years [6]. Also with the latest governmental policy shift towards renewable and environment friendly alternative energy sources, reliance on fossil fuels is expected to decrease.
In Pakistan, there is a huge potential for renewable which is around 167.7 GW [5]. In renewable energies, potential for photovoltaics in particular is immense in the country. Pakistan is one of the wealthiest countries in term of its solar potential which is up to 100,000 MW and an average solar insolation for the country is about 5.5 kWh/m 2 /day [7].
Keeping in view all the scenarios discussed above, it suggests a direction for implementation of running water pumping systems on renewable sources instead of fossil fuels in Pakistan. The primary reasons are as Pakistan is not a country rich in fossil fuel resources along with its environmental hazards and Pakistan is rich in solar resources which can be tapped to overcome energy requirements for running water pumps.

Literature Review
Choudhary et al. has discussed evolution [8] of solar water pumping system with time. Throughout the world for some time now due to fluctuating prices of fossil fuel [9] and due to their environmental hazards, there has been huge shift towards Solar water pumping systems and a lot of research and development have been going on in the field of designing and optimisation of a Solar water pumping system. Allouhi et al. has discussed in detail about the solar water pumping systems [10] considering its sizing aspects and economic analysis.
There have been many methods used to design and optimise a solar water Over the time many manual computation methods have evolved. They are simple and cheap to implement. One such method [11] uses mathematical formulas to calculate photovoltaic panel requirement by considering a fixed tilt angle along with number of batteries required and cost analysis of an overall system over a fixed life cycle. Discussed [12] also many different methods for PV sizing which are based on the loss of load probability (LLP), also a method is discussed which used simple manual computations and require four coefficients to design a stand-alone PV system throughout Spain. In [13] a method is discussed which does PV system curve sizing using a simple procedure which is based upon observed time series of solar radiation and by incorporating simple geometrical concepts in which curve for sizing is constructed, which incorporates data of daily solar insolation. Computers based method is based on computer simulations but these software are computation intensive. There are two types of analysis which are done using computer based software of a solar water pumping system. One is steady state analysis and other is dynamic analysis of the designed system. In steady state analysis software perform calculations related to sizing of a PV system keeping in view a specific interval of time; they also help to determine the power generated from the source, load mapping and different parameters related to battery and power converter requirements [14]. Many such software for steady state analysis are discussed in [15], few of such software for system sizing are RETScreen [16], PVplanner [17], PVSyst [18], SolarPro [19], SAM (System advisor model) [20], HOMER (Hybrid Optimization Model for Electric Renewables) [21] etc. The other type of analysis of the Solar water pumping system is dynamic analysis which involves analysis and understanding of the subjected system with varying input conditions which normally involve change in irradiance and temperature. Some of the software for dynamic modeling which help to evaluate system with changing conditions of the PV system are TRANSYS (Transient System Simulation Tool) [22], LABVIEW [23], MATLAB [24] etc.
A lot of work is available in literature for conducting steady state analysis of the Solar systems. As we have chosen HOMER for steady state analysis, the work regarding this in the literature is discussed further. Chaichan et al. [25] discussed four different design scenarios for solar water pumping system in Oman using HOMER. Among them three were powered by PV and one system was powered by a diesel generator. After the economical comparison was made cost of energy for diesel generator water pumping system was 0.6092 USD/kWh which was greater than the cost of energy for the water pumping system which was operated by PV source which was USD 0.4743/kWh. Kazim et al. [26] has done modeling of a solar water pumping for a remote area in Sohar, Oman which can be used for irrigation purposes. For the solar water pumping design he used HOMER and the system requirements for a daily load of 2.22 kWh/day came out to be 12 V, 200 Ah four batteries, PV module network of rating of 0.84 kW and DOI: 10.4236/epe.2020.1210032 524 Energy and Power Engineering an inverter of 0.8 kW to serve the design purpose. Cost of the system came out to be 0.309 USD/kWh. In this paper [27] four different systems are designed using HOMER for lightning a street in Salalah Oman which is 10 km long. One system was sourced solely by PV network, second was sourced by wind source only, third system designed was sourced by diesel generator and last was designed using a hybrid approach which used an optimised combination of PV and wind. The hybrid solution of PV and wind proved to be the most cost effective solution which consisted of a wind turbine of 250 kW and 80 kW PV panel network backed by 200 (200 Ah/12V) batteries and power converter of 55 kW size.
As per Kammash et al. [28] Renewable energy center at University of technology is to be lighted using PV Panels network, so they used HOMER for the overall system's sizing. The requirement of the PV Panel network came out to be 8 kW, 20 12 V 200 Ah batteries and a power converter requirement of 4 kW. Initial cost of the overall system came out to be 2000 USD, net cost of the overall system came out to be 32,015 USD and per kWh cost of electricity produced by the system came out to be 0.903 USD. Alkarrami et al. [29] designed a water pumping system for a site in Sirte city Libya, which was sourced by hybrid energy sources. In this he used three different software to design the proposed system namely HOMER Pro, HOMER Beta and iHOGA and as per his analysis HOMER Pro gave the best results. This paper [30] discussed about the design of a PV water pumping system for Bangladeshi conditions which is already in operation in Lalmonirhat, Bangladesh. For PV water pumping system sizing the software used was HOMER. After the analysis the PV Panel requirement came out to be 15.013 kW, battery bank capacity requirement came out to be 2600 Ah and power converter or inverter requirement came out to be 21 kW. The dynamic modeling for the overall system with battery bank, without battery and water tank and a hybrid storage system consisting of small battery bank and a small water tank was done using MATLAB/Simulink. For dynamic modeling a lot of work is available in literature but the scope of this paper is limited to dynamic modeling in MATLAB/Simulink. In [31] a PV operated multi staged centrifugal pump which was driven by an induction motor was discussed and the pumping system was aided with battery bank for excess energy storage purpose as well. For dynamic behavior evaluation of the overall system MATLAB/Simulink was used and at the end results were compared with manufacturer's data and results were differing by a small margin. In this paper [32] MATLAB/Simulink is used to evaluate the dynamic behavior of a Solar water pumping system with brushless DC motor, it has a zeta converter as DC-DC converter which is used for maximum power extraction during its operation.
The simulation results were verified with the manufacturer's data. After simulation, an overall efficiency of 83% was obtained at solar insolation of 1000 W/m 2 and an overall efficiency of 71% was obtained at solar insolation of 400 W/m 2 . A solar water pumping system is modeled in MATLAB/Simulink [33]  der steady or starting state and dynamically changing insolation conditions. Keeping in view the above literature review, while designing and sizing of a solar water pumping system for this work we chose HOMER for steady state analysis as it gave the most consistent and reliable sizing results and was found to be most widely used software. For dynamic analysis of the designed system MATLAB was chosen because as per literature many results of dynamic analysis from MATLAB were compared with manufacture's data and they found simulation results to be in line with manufacturer's data which not only verified the results but also suggest MATLAB to be a good option for dynamic analysis of the designed solar water pumping system.
Since the MATLAB blocks do not allow large set of data for irradiance and temperature to be processed at a rapid speed as they are more complex and reduce computation speed for analysis of large set of input data. To overcome this issue and make our system simpler to increase computation speed, mathematical modeling of different components of the overall solar water pumping system used in our designed system was done in MATLAB namely PV Panel, MPPT, Buck-Boost Converter, Battery Bank, Inverter and water pump using 3 phase induction motor.

Site Selection
A suitable site was selected where a solar water pumping system can be implemented for that purpose an agricultural area known as "Mustafa Research It covers an area of around 239.6 acres of land. The crop cultivated on this land is Rhodes Grass which is grown for commercial purposes and is exported from Pakistan to many countries of the world. The Mustafa Research Farms can be seen in Figure 1, is bordered by blue line in the figure and as this whole area is irrigated by five diesel water pumping systems, the location of the water pumping systems can be seen by circles in blue.

Solar Insolation Details of the Selected Site
The solar insolation details along with clearness index can be seen in Figure 2, which is extracted from the database of National Renewable Energy Laboratory [35] as its database is attached with HOMER. From the figure, it can be seen that the GHI (Global Horizontal Index) varies from 3.6 to 7.26 kWh/m 2 /day and its clearness index varies in the range of 0.601 to 0.69.

Data Collected from Site
The data collected from the selected agricultural site for motor/pump load calculation is as follows and the yearly operation of the motor is summarised in Table 1.    The brand selected for the water pump/motor is Wilo, it has an online tool [36] that was used to evaluate the motor and pump size. The required data (flow rate and total dynamic head the motor) was incorporated in the online tool, the motor rating came out to be of 11 kW which is almost equivalent to 15 hp. The details regarding the motor/pump can be seen in Figure 3. Details led to the selection of Atmos GIGA-N 125/200-11/4 [37]. The further details regarding motor/pump are summarized in Figure 4.

Sizing of the Proposed System Using HOMER
Now, the next step was to map the load details in HOMER as per the motor operation table summarized in Table 1 and the calculated motor/pump load. Then the simulation in HOMER was performed to get the optimum ratings of different design components of the proposed solar water pumping systems. The diagram of the overall proposed solar water pumping system can be seen in Figure   5.
After the simulation was performed in HOMER, the sizing results which came are summarised in Figure 6. The PV Panel network which is composed of Astronergy ASM6612P-320 solar panel [38], its requirement came out to be 73.    [40] was incorporated to serve the purpose.
The overall proposed system can be seen in Figure 7. As the PV panel network requirement came out to be 73.8 kW, so a total of 240 24 V panels will be

Economic Analysis of the Proposed System Using HOMER
The economic analysis of the overall proposed system over its life cycle of 25 cycles was also done using HOMER. The cost summary of the overall proposed system is summarised in Figure 8. It can be seen that the total net present cost of Figure 7. Overall proposed solar water pumping system with battery bank.  the overall system is $273,570, the levelized cost of energy of the system comes out to be $0.48 along with operating cost which is $5692.37 per year. As per cash flow over the life cycle of 25 years for the proposed system which can be seen in Figure 9. The initial investment which is required is $199,981.86, then an annual investment of $4615.22 is required from the 1st year of its operation to the 25th year of its operation. At 21st year a further investment of $99,000 is required to keep the system running. After the end of the cycle of 25 years a salvage value of $76,681.52 is available for the overall system.

Verification of Proposed System with Battery Bank Using Dynamic Modeling
The dynamic modeling of the proposed system was done in MATLAB/Simulink. Mathematical modeling for different design components was done so that speedy simulation with a much overall simplified system can be done. The model of the proposed system can be seen in Figure 10.
The data was simulated for the first seven days of the April, when the motor is in operation to evaluate the dynamic behavior of the proposed solar water pumping system. The solar data which include solar irradiance and temperature for the first seven days of April which makes it a total simulation for 10,080 minutes can be seen in Figure 11. The irradiance varied from 0 to 1021 W/m 2 and temperature varied from 25˚C to 49.5˚C during a week of simulation.
Simulation results are split into two figures (A and B) for the proposed system which can be seen in Figure 12 and Figure 13. It can be seen in Figure 12    The initial SOC (State of charge) of the battery was 80% which declined to 79% after 1st discharge varied between 84% and 100% afterwards throughout the simulation as can be seen in Figure 13. The output voltage was only once 377.5    SOC it no longer takes into consideration the battery's internal resistance hence, there is a slight decrease in overall battery's output voltage. The voltage output of inverter varied from +326.6 V to −326.6 V peak as can be seen in Figure 13. In our system phase to neutral voltage is considered which is equal to ( ) It is important to know that we have considered 4 wire three phase bridge inverter which can supply a voltage from phase to neutral and we have shown only one phase for inverter stage.
The detailed simulation results for the seven day simulation from April month of the motor/pump can be seen in Figure 14. It can be seen from the simulation results in figure that it took about 15 seconds to get to steady state and after that continuous operation of motor is at 1470 rpm which is in accordance with the data sheet, also after the initial surge for 25 sec which is the characteristic for induction motor the current stabilises from 326.6 A to 29.6 A within 15 seconds which is the current for phase to neutral current which is equal to 20.9 A (As per motor data sheet) × 2 = 29.6 A.

Load Determination for the Proposed System
Before proceeding with the design the first step was to determine an optimum size of a tank which could be large enough to keep a storage of water for 1 day. So for this flow rate for 1 day was calculated as can be seen below.  . The details regarding the motor can be seen in Figure 15. The further details regarding the graph can be seen in Figure   16.

Sizing of the Proposed System Using HOMER
After the size of the motor was evaluated, the load was mapped in keeping the operational hours of the motor to be 5 hours. The load of the motor since is associated with a water tank storage so a deferrable load is used in HOMER and as the motor is running for 5 hours, so 55 × 5 = 275 kWh is the storage capacity required. The detailed load mapping in the deferrable load in HOMER can be seen in Figure 17.
The overall proposed system with water tank in HOMER appears as can be seen in the Figure 18. The PV panel used was Astronergy ASM6612P-320 solar [38] for design of PV panel network, the inverter used is SMA American STP60-US-10 [42] which is 60 kW inverter.
After the simulation of the overall proposed solar water pumping system was conducted; the requirement of the PV Panel network came out to be 72.3 kW, the inverter requirement came out to be 59.9 kW, which is fulfilled by 60 kW SMA American STP60-US-10. The overall result is seen in Figure 19.
The overall proposed system with water tank can be seen in Figure 20. To fulfill the requirement of 72.3 kW, 240 panels were used, as the bus voltage is 360 V.

Cost Analysis of the Proposed System with Water Tank Using HOMER
The overall cost summary for the proposed system was performed in HOMER

Sizing of Water Tank
As the total water discharge in one day is 4896 m 3 , so a tank of 5000 m 3 is planned. For this the mathematical details are as follows.   Considering a cylindrical tank of 5000 m 3 with a height of 2 m. The radius of the tank can be calculated using Equation (1).
Here V is for volume in m 3 ; r is radius of the cylinder in meters; h is the height

Conclusions
A detailed sizing and economic analysis of solar system with battery bank was performed using HOMER software to evaluate the steady state analysis of system. It was found out that 73.8 kW of PV panel network, 450 batteries of Trojan SAGM 12 105 and 16.7 kW of inverter are required to realise a system for a motor load of 11 kW. To further study the system, dynamic analysis of the proposed system was done in MATLAB/Simulink. Mathematical modeling of PV panel, battery bank, inverter, buck-boost converter and selected model and pump along with a control system was done to represent the overall system in shape of maximum power point tracking by Perturb and observe algorithm that was used to control the duty cycle. This helped to further validate the designed solar water pumping system in HOMER.
A detailed sizing of another option was also proposed for the selected agricultural site which was supported by a cylindrical water tank instead of a battery bank. For this system, the PV panel network requirement came out to be 72.3 kW, the inverter requirement came out to be 59.9 kW and this all was required for the motor of load 55 kW and the water tank volume requirement came out to be 4896 m 3 .

Economics Comparison of Two Designed Solar Water Pumping Systems
To compare the solar water pumping systems design their economic analysis is considered, which was also performed using HOMER software. It can be seen in the results that overall net present cost of the system with battery bank was $273,570 and the net present cost for the system with water tank came out to be goes in favour of system without battery. This factor is less as compared to battery based system, because construction of tank is relatively cheaper in Pakistan due to availability of cheap labour.
Hence, it can be concluded that solar water pumping system based on water tank is cheaper as compared to solar water pumping system based on battery.